The Stock Exchange of Hong Kong Limited Main Board (Chapter 18C – Specialist Technology Company) · Filed 2025-12-30 · Full English Translation
Knowledge Atlas Technology (ZhipuAI, stock code 2513) is a Beijing-based artificial intelligence company developing large language models and AI solutions for enterprise customers in China.
Revenue has grown at a compound annual rate exceeding 130% from 2022 to 2024, rising from roughly $792K to $1.7M to $4.3M across those years (figures in 万元 units: 574, 1,245, and 3,124 respectively). First-half 2025 revenue reached approximately $2.6M. Gross margins have consistently exceeded 50% across all reported periods. R&D spending is heavy, totaling approximately $22M in the first half of 2025 alone, nearly double the $11.8M spent in H1 2024. The company serves over 8,000 institutional customers and has empowered roughly 80 million devices.
ZhipuAI is listing on the Hong Kong Stock Exchange Main Board under the Specialist Technology (Chapter 18C) framework, offering 37.4 million H shares at HK$116.20 each, implying gross proceeds of approximately $557M. Prior to IPO, the company raised the equivalent of roughly $1.15B across previous funding rounds. Notable backers include Legend Capital, Qiming Venture Partners, Meituan (4.27% stake), and Ant Group. Proceeds are intended to fund continued R&D investment and business expansion, though a detailed allocation was not disclosed in the extracted filing.
The three most significant risks are intense competition and rapid technological change in AI that could render current models obsolete, heavy dependence on third-party computing infrastructure, and evolving Chinese regulatory requirements governing AI that could impose material new compliance burdens.
| Period | Revenue | Net Profit | Gross Margin |
|---|---|---|---|
| 2022 | $8M | $-20M | 54.6% |
| 2023 | $17M | $-109M | 64.6% |
| 2024 | $43M | $-408M | 56.3% |
| 2024H1 | $6M | $-170M | 48.9% |
| 2025H1 | $26M | $-325M | 50.0% |
| Date | Total Assets | Total Liabilities | Equity |
|---|---|---|---|
| 2022-12-31 | $50M | $75M | $-25M |
| 2023-12-31 | $395M | $530M | $-136M |
| 2024-12-31 | $604M | $1.1B | $-546M |
| 2025-06-30 | $704M | $1.6B | $-848M |
| Project | Amount (USD) | Focus |
|---|---|---|
| Strengthening research and development capabilities in general-purpose large AI models | $365M | Continuously strengthen R&D capabilities in general-purpose large AI models, accounting for approximately 70% of net proceeds (HK$2,921.4 million) |
| Optimization of MaaS platform | $52M | Continuously optimize the MaaS platform by offering the latest foundation models and training/inference tools and infrastructures, accounting for approximately 10% of net proceeds (HK$417.3 million) |
| Development of business partner network and strategic investments | $52M | Development of business partner network as well as strategic investments, accounting for approximately 10% of net proceeds (HK$417.3 million) |
| Working capital and other general corporate purposes | $52M | Working capital and other general corporate purposes, accounting for approximately 10% of net proceeds (HK$417.3 million) |
北京智譜華章科技股份有限公司 (A joint stock company established in the People's Republic of China with limited liability)
Sole Sponsor, Sponsor-Overall Coordinator, Overall Coordinator, Joint Global Coordinator, Joint Bookrunner and Joint Lead Manager
IMPORTANT IMPORTANT: If you are in any doubt about any of the contents of this prospectus, you should seek independent professional advice.
Knowledge Atlas Technology Joint Stock Company Limited 北京智譜華章科技股份有限公司 (A joint stock company established in the People's Republic of China with limited liability)
| Item | Details | |---|---| | Number of Offer Shares under the Global Offering | 37,419,500 H Shares (subject to the Over-allotment Option) | | Number of Hong Kong Offer Shares | 1,871,000 H Shares (subject to reallocation) | | Number of International Offer Shares | 35,548,500 H Shares (subject to reallocation and the Over-allotment Option) | | Offer Price | HK$116.20 per H Share, plus brokerage of 1.0%, SFC transaction levy of 0.0027%, Stock Exchange trading fee of 0.00565% and AFRC transaction levy of 0.00015% (payable in full on application in Hong Kong dollars and subject to refund) | | Nominal value | RMB0.10 per H Share | | Stock code | 2513 |
Sole Sponsor, Sponsor-Overall Coordinator, Overall Coordinator, Joint Global Coordinator, Joint Bookrunner and Joint Lead Manager
Hong Kong Exchanges and Clearing Limited, The Stock Exchange of Hong Kong Limited and Hong Kong Securities Clearing Company Limited take no responsibility for the contents of this prospectus, make no representation as to its accuracy or completeness, and expressly disclaim any liability whatsoever for any loss howsoever arising from or in reliance upon the whole or any part of the contents of this prospectus.
A copy of this prospectus, having attached thereto the documents specified in "Appendix VII—Documents Delivered to the Registrar of Companies and Documents on Display—A. Documents Delivered to the Registrar of Companies" to this prospectus, has been registered by the Registrar of Companies in Hong Kong as required by section 342C of the Companies (Winding Up and Miscellaneous Provisions) Ordinance (Chapter 32 of the Laws of Hong Kong). The Securities and Futures Commission and the Registrar of Companies in Hong Kong take no responsibility for the contents of this prospectus or any other document referred to above.
The Offer Price will be HK$116.20 per Offer Share, unless otherwise announced. Applicants for Hong Kong Offer Shares may be required to pay, on application (subject to application channels), the Offer Price of HK$116.20 for each Hong Kong Offer Share together with a brokerage fee of 1.0%, a SFC transaction levy of 0.0027%, AFRC transaction levy of 0.00015% and a Hong Kong Stock Exchange trading fee of 0.00565%.
The Sponsor-Overall Coordinator (for and on behalf of the Underwriters) may, with the consent of our Company, reduce the number of Offer Shares and/or the Offer Price below that stated in this prospectus at any time on or prior to the morning of the last date for lodging applications under the Hong Kong Public Offering. In such a case, notices of the reduction in the number of Hong Kong Offer Shares and/or the Offer Price will be published on the websites of the Stock Exchange at www.hkexnews.hk and our Company at www.zhipuai.cn as soon as practicable but in any event not later than the morning of the day which is the latest day for lodging applications under the Hong Kong Public Offering. For further information, see the sections headed "Structure of the Global Offering" and "How to Apply for Hong Kong Offer Shares" in this prospectus.
Prior to making an investment decision, prospective investors should carefully consider all of the information set out in this prospectus, and in particular, the risk factors set out in the section headed "Risk Factors."
We are incorporated, and substantially all of our businesses are located, in the PRC. Potential investors should be aware of the differences in the legal, economic and financial systems between the PRC and Hong Kong and that there are different risk factors relating to investment in PRC-incorporated businesses. Potential investors should also be aware that the regulatory framework in the PRC is different from the regulatory framework in Hong Kong and should take into consideration the different market nature of the H Shares. Such differences and risk factors are set out in "Risk Factors," "Appendix IV—Summary of Principal Legal and Regulatory Provisions" and "Appendix V—Summary of Articles of Association" to this Prospectus.
The obligations of the Hong Kong Underwriters under the Hong Kong Underwriting Agreement to subscribe for, and to procure applicants for the subscription for, the Hong Kong Offer Shares are subject to termination by the Sponsor-Overall Coordinator (for and on behalf of the Hong Kong Underwriters) if certain grounds arise prior to 8:00 a.m. on the day that trading in the H Shares commences on the Stock Exchange. Such grounds are set out in "Underwriting—Underwriting Arrangements and Expenses—The Hong Kong Public Offering—Grounds for Termination" in this prospectus.
The Offer Shares have not been, and will not be, registered under the U.S. Securities Act or any state securities laws in the United States and may not be offered, sold, pledged or otherwise transferred within the United States or to, or for the account or benefit of, any U.S. Investors. The Offer Shares will be offered and sold outside the United States to persons that are not, and are not acting for the account or benefit of, U.S. Investors in offshore transactions in reliance on Regulation S. There has not been and will not be any public offering of the H Shares in the United States.
Our Company is a Specialist Technology Company (as defined in Chapter 18C of the Listing Rules). The securities of Specialist Technology Companies carry high investment risks including risks of share price volatility and inflated valuation due to the difficulty in valuing such companies. Investors should fully understand the investment risks of a Specialist Technology Company and the risks disclosed by our Company before making their investment decisions.
We have adopted a fully electronic application process for the Hong Kong Public Offering. We will not provide printed copies of this prospectus to the public in relation to the Hong Kong Public Offering.
This prospectus is available at the website of the Hong Kong Stock Exchange at www.hkexnews.hk and our website at www.zhipuai.cn. If you require a printed copy of this prospectus, you may download and print from the websites above.
IMPORTANT NOTICE TO INVESTORS OF HONG KONG OFFER SHARES FULLY ELECTRONIC APPLICATION PROCESS We have adopted a fully electronic application process for the Hong Kong Public Offering and below are the procedures for application. We will not provide printed copies of this prospectus to the public in relation to the Hong Kong Public Offering. This prospectus is available at the website of the Stock Exchange at www.hkexnews.hk under the "HKEXnews > New Listings > New Listing Information" section, and our website at www.zhipuai.cn. If you require a printed copy of this prospectus, you may download and print from the website addresses above. To apply for the Hong Kong Offer Shares, you may: (1) apply online through the HK eIPO White Form service through the designated website www.hkeipo.hk; or (2) apply electronically through the HKSCC EIPO channel and cause HKSCC Nominees to apply on your behalf by instructing your broker or custodian who is a HKSCC Participant to give electronic application instructions via HKSCC's FINI system to apply for the Hong Kong Offer Shares on your behalf.
We will not provide any physical channels to accept any application for the Hong Kong Offer Shares by the public. The contents of the electronic version of this prospectus are identical to the printed document as registered with the Registrar of Companies in Hong Kong pursuant to Section 342C of the Companies (Winding Up and Miscellaneous Provisions) Ordinance. If you are an intermediary, broker or agent, please remind your customers, clients or principals, as applicable, that this prospectus is available online at the website addresses above. Please refer to "How to Apply for Hong Kong Offer Shares" for further details of the procedures through which you can apply for the Hong Kong Offer Shares electronically. Your application through the HK eIPO White Form service or the HKSCC EIPO channel must be made for a minimum of 100 Hong Kong Offer Shares and in multiples of that number of Hong Kong Offer Shares as set out in the table below. If you are applying through the HK eIPO White Form service, you may refer to the table below for the amount payable for the number of H Shares you have selected. You must pay the respective amount payable on application in full upon application for Hong Kong Offer Shares.
If you are applying through the HKSCC EIPO channel, you are required to pre-fund your application based on the amount specified by your broker or custodian, as determined based on the applicable laws and regulations in Hong Kong.
| No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | |---|---|---|---|---|---|---|---| | 100 | 11,737.19 | 2,500 | 293,429.69 | 30,000 | 3,521,156.31 | 600,000 | 70,423,126.20 | | 200 | 23,474.37 | 3,000 | 352,115.63 | 40,000 | 4,694,875.08 | 700,000 | 82,160,313.90 | | 300 | 35,211.56 | 3,500 | 410,801.57 | 50,000 | 5,868,593.86 | 800,000 | 93,897,501.60 | | 400 | 46,948.75 | 4,000 | 469,487.51 | 60,000 | 7,042,312.62 | 935,500(1) | 109,801,390.94 | | 500 | 58,685.94 | 4,500 | 528,173.44 | 70,000 | 8,216,031.39 | | | | 600 | 70,423.12 | 5,000 | 586,859.39 | 80,000 | 9,389,750.15 | | | | 700 | 82,160.32 | 6,000 | 704,231.26 | 90,000 | 10,563,468.94 | | | | 800 | 93,897.50 | 7,000 | 821,603.14 | 100,000 | 11,737,187.70 | | | | 900 | 105,634.69 | 8,000 | 938,975.01 | 200,000 | 23,474,375.40 | | | | 1,000 | 117,371.88 | 9,000 | 1,056,346.90 | 300,000 | 35,211,563.10 | | | | 1,500 | 176,057.82 | 10,000 | 1,173,718.76 | 400,000 | 46,948,750.80 | | | | 2,000 | 234,743.75 | 20,000 | 2,347,437.55 | 500,000 | 58,685,938.50 | | |
(1) Maximum number of Hong Kong Offer Shares you may apply for and this is 50% of the Hong Kong Offer Shares initially offered. (2) The amount payable is inclusive of brokerage, SFC transaction levy, the Stock Exchange trading fee and AFRC transaction levy. If your application is successful, brokerage will be paid to the Exchange Participants (as defined in the Listing Rules) or to the HK eIPO White Form Service Provider (for applications made through the application channel of the HK eIPO White Form service) while the SFC transaction levy, the Stock Exchange trading fee and the AFRC transaction levy will be paid to the SFC, the Stock Exchange and the AFRC, respectively.
No application for any other number of Hong Kong Offer Shares will be considered and any such application is liable to be rejected.
If there is any change in the following expected timetable of the Hong Kong Public Offering, we will issue an announcement to be published on the websites of the Stock Exchange at www.hkexnews.hk and our Company at www.zhipuai.cn.
| Event | Date and Time | |---|---| | Hong Kong Public Offering commences | 9:00 a.m., Tuesday, December 30, 2025 | | Latest time for completing applications under the HK eIPO White Form service through the designated website www.hkeipo.hk(2) | 11:30 a.m. on Monday, January 5, 2026 | | Application lists of the Hong Kong Public Offering open(3) | 11:45 a.m. on Monday, January 5, 2026 | | Latest time to (a) complete payments of HK eIPO White Form applications by effecting internet banking transfer(s) or PPS payment transfer(s) and (b) give electronic application instructions to HKSCC(4) | 12:00 noon on Monday, January 5, 2026 |
• 配发结果专页，位于指定配发结果网站 www.tricor.com.hk/ipo/result 或 www.hkeipo.hk/IPOResult . . . . . . . . . . . . . . . . . . . . . . . . . . . .
• 通过拨打 +852 3691 8488 配发结果电话查询热线（服务时间为上午9时至下午6时）. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
就全部或部分成功的申请，H股股票将予寄发或存入中央结算系统 . . . . . . . . . .
就全部或部分成功的申请（如适用）或全部或部分不成功的申请，HK eIPO白表电子自动退款指示／退款支票将予寄发 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
H股在联交所主板开始买卖 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
If you are instructing your broker or custodian who is a HKSCC Participant and will submit an electronic application instruction on your behalf through HKSCC's FINI system in accordance with your instruction, you are advised to contact your broker or custodian for the earliest and latest time for giving such instructions, as this may vary by broker or custodian.
Application lists of the Hong Kong Public Offering close(3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Announcement of the level of applications in the Hong Kong Public Offering; the level of indications of interest in the International Offering; and the basis of allocation of the Hong Kong Offer Shares to be published on the websites of our Company at www.zhipuai.cn(5) and the Stock Exchange at www.hkexnews.hk . . . . . . . . . . . . . . .
The results of allocations in the Hong Kong Public Offering (with successful applicants' identification document numbers, where appropriate) to be made available through a variety of channels, including:
• in the announcement to be posted on the websites of our Company at www.zhipuai.cn(6) and the Stock Exchange at www.hkexnews.hk, respectively . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
• the "Allotment Results" page on the designated results of allocations website at www.tricor.com.hk/ipo/result or www.hkeipo.hk/IPOResult . . . . . . . . . . . . . . . . . . . . . . . . . . . .
• from the allocation results telephone enquiry line by calling +852 3691 8488 between 9:00 a.m. and 6:00 p.m. . . . . . . . . . . .
from 11:00 p.m. on Wednesday, January 7, 2026 to 12:00 midnight on Tuesday, January 13, 2026 from Thursday, January 8, 2026 to Tuesday, January 13, 2026 (except Saturday, Sunday and Hong Kong public holidays)
H Share certificates in respect of wholly or partially successful applications to be despatched or deposited into CCASS . . . . . . .
HK eIPO White Form e-Auto Refund payment instructions/refund checks in respect of wholly or partially successful applications (if applicable) or wholly or partially unsuccessful applications to be despatched . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dealings in the H Shares on the Main Board of the Stock Exchange to commence at . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Notes: (1) All dates and times refer to Hong Kong local dates and times, except as otherwise stated. (2) You will not be permitted to submit your application under the HK eIPO White Form service through the designated website at www.hkeipo.hk after 11:30 a.m. on the last day for submitting applications. If you have already submitted your application and obtained an application reference number from the designated website prior to 11:30 a.m., you will be permitted to continue the application process (by completing payment of the application monies) until 12:00 noon on the last day for submitting applications, when the application lists close. (3) If there is/are a "black" rainstorm warning signal, a tropical cyclone warning signal number 8 or above and/or Extreme Conditions in force in Hong Kong at any time between 9:00 a.m. and 12:00 noon on Monday, January 5, 2026, the application lists will not open or close on that day. See the section headed "How to Apply for Hong Kong Offer Shares—E. Severe Weather Arrangements" for further details. (4) Applicants who apply for Hong Kong Offer Shares by giving electronic application instructions to HKSCC via HKSCC's FINI system should refer to the section headed "How to Apply for Hong Kong Offer Shares—A. Applications for Hong Kong Offer Shares—2. Application Channels". (5) None of the websites or any of the information contained on the websites forms part of this prospectus. (6) The H Share certificates will only become valid evidence of title at 8:00 a.m. on the Listing Date provided that the Global Offering has become unconditional in all respects. Investors who trade the H Shares on the basis of publicly available allocation details prior to the receipt of H Share certificates or prior to the H Share certificates becoming valid evidence of title do so entirely at their own risk. (7) HK eIPO White Form e-Auto Refund payment instructions/refund checks will be issued in respect of wholly or partially unsuccessful applications pursuant to the Hong Kong Public Offering and in respect of wholly or partially successful applicants. Part of the applicant's identification document number, or, if the application is made by joint applicants, part of the identification document number of the first-named applicant, provided by the applicant(s) may be printed on the refund check, if any. Such
data would also be transferred to a third party for refund purposes. Banks may require verification of an applicant's identification document number before encashment of the refund check. Inaccurate completion of an applicant's identification document number may invalidate or delay encashment of the refund check.
(8) Applicants who have applied on the HK eIPO White Form service for 500,000 or more Hong Kong Offer Shares may collect H Share certificates in person from our H Share Registrar, Tricor Investor Services Limited, at 17/F, Far East Finance Centre, 16 Harcourt Road, Hong Kong from 9:00 a.m. to 1:00 p.m. on Thursday, January 8, 2026 or any other places or date as notified by us as the date of despatch/collection of H Share certificates/HK eIPO White Form e-Auto Refund payment instructions/refund checks. Applicants being individuals who are eligible for personal collection may not authorize any other person to collect on their behalf. Individuals must produce evidence of identity acceptable to our H Share Registrar at the time of collection.
Applicants who have applied for Hong Kong Offer Shares through HKSCC EIPO channel should refer to "How to Apply for Hong Kong Offer Shares—D. Despatch/Collection of H Share Certificates and Refund of Application Monies" for details.
Applicants who have applied through the HK eIPO White Form service and paid their application monies through single bank accounts may have refund monies (if any) despatched to the bank account in the form of HK eIPO White Form e-Auto Refund payment instructions. Applicants who have applied through the HK eIPO White Form service and paid their application monies through multiple bank accounts may have refund monies (if any) despatched to the address as specified in their application instructions in the form of refund checks in favor of the applicant (or, in the case of joint applications, the first-named applicant) by ordinary post at their own risk.
The H Share certificates and/or refund checks for applicants who have applied for less than 500,000 Hong Kong Offer Shares and any uncollected H Share certificates will be despatched by ordinary post, at the applicants' risk, to the addresses specified in the relevant applications.
Further information is set out in the section headed "How to Apply for Hong Kong Offer Shares—D. Despatch/Collection of H Share Certificates and Refund of Application Monies."
The above expected timetable is a summary only. For details of the structure of the Global Offering, including its conditions, and the procedures for applications for Hong Kong Offer Shares, see "Structure of the Global Offering" and "How to Apply for Hong Kong Offer Shares," respectively.
If the Global Offering does not become unconditional or is terminated in accordance with its terms, the Global Offering will not proceed. In such a case, our Company will make an announcement as soon as practicable thereafter.
This prospectus is issued by our Company solely in connection with the Hong Kong Public Offering and does not constitute an offer to sell or a solicitation of an offer to buy any security other than the Hong Kong Offer Shares offered by this prospectus pursuant to the Hong Kong Public Offering. This prospectus may not be used for the purpose of, and does not constitute, an offer or a solicitation of an offer to subscribe for or buy any security in any other jurisdiction or in any other circumstances. No action has been taken to permit a public offering of the Offer Shares or the distribution of this prospectus in any jurisdiction other than Hong Kong. The distribution of this prospectus and the offering and sale of the Offer Shares in other jurisdictions are subject to restrictions and may not be made except as permitted under the applicable securities laws of such jurisdictions pursuant to registration with or authorization by the relevant securities regulatory authorities or an exemption therefrom.
You should rely only on the information contained in this prospectus to make your investment decision. We have not authorized anyone to provide you with information that is different from what is contained in this prospectus. Any information or representation not made in this prospectus must not be relied on by you as having been authorized by us, the Sole Sponsor, Sponsor-Overall Coordinator, the Overall Coordinators, the Joint Global Coordinators, the Joint Bookrunners, the Joint Lead Managers, the Capital Market Intermediaries, any of the Underwriters, any of our or their respective directors, officers or representatives, or any other person or party involved in the Global Offering. Information contained on our website, located at www.zhipuai.cn, does not form part of this prospectus.
| Section | Page | |---|---| | EXPECTED TIMETABLE | iv | | CONTENTS | vii | | SUMMARY | 1 | | DEFINITIONS AND ACRONYMS | 24 | | GLOSSARY OF TECHNICAL TERMS | 36 | | FORWARD-LOOKING STATEMENTS | 39 |
RISK FACTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
WAIVERS FROM STRICT COMPLIANCE WITH THE REQUIREMENTS UNDER THE LISTING RULES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
INFORMATION ABOUT THIS PROSPECTUS AND THE GLOBAL OFFERING . . . . . . . . .
DIRECTORS, SUPERVISOR AND PARTIES INVOLVED IN THE GLOBAL OFFERING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CORPORATE INFORMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
INDUSTRY OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
REGULATORY OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HISTORY, DEVELOPMENT AND CORPORATE STRUCTURE . . . . . . . . . . . . . . . . . . . . . . .
BUSINESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DIRECTORS, SUPERVISOR AND SENIOR MANAGEMENT . . . . . . . . . . . . . . . . . . . . . . . . .
RELATIONSHIP WITH OUR CONTROLLING SHAREHOLDERS . . . . . . . . . . . . . . . . . . . . .
SUBSTANTIAL SHAREHOLDERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FINANCIAL INFORMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CORNERSTONE INVESTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FUTURE PLANS AND USE OF PROCEEDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNDERWRITING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
STRUCTURE OF THE GLOBAL OFFERING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HOW TO APPLY FOR HONG KONG OFFER SHARES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ACCOUNTANTS' REPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
At the assistant level, we provide ready-to-use products powered by our large models. This level is intended to serve individual users, who access these products mainly through our MaaS platform directly.
Acceleration: At the acceleration level, we provide API services (including standard API and customized API through fine-tuning), agent development tools and various knowledge-based and tool-based services that allow institutional customers and developers to build their own applications and products using our model and capabilities.
Adaptability: At the adaptability level, we provide dedicated cloud, private cloud and on-premises deployment options along with services to institutional customers with specific computing resource or data security requirements that cannot use API services directly.
SUMMARY As a commercialization platform, our MaaS platform provides the foundation and infrastructure to translate our technology into products and services. As a product development platform, our MaaS platform provides institutional customers with tooling and infrastructure to develop their own AI applications and models.
Revenue from individual users: Revenue from individual users is generated through product subscriptions, under which individual users pay a subscription fee to access our AI products directly.
Revenue from institutional customers: Revenue from institutional customers is generated through three approaches: (1) access to API services (for the Acceleration level), (2) providing customized deployment services (for the Adaptability level), and (3) customized model development services.
SUMMARY INDUSTRY: A LARGE AND FAST-GROWING MARKET China's AI market is characterized by very rapid development. The LLM market in China is expected to expand rapidly from RMB6.5 billion (approximately US$0.9 billion) in 2024 to RMB188.6 billion (approximately US$25.9 billion) in 2030, at a CAGR of approximately 76.3%.
(Source: Frost & Sullivan) COMPETITIVE STRENGTHS We believe our competitive strengths are: 1.
We have established a unique technology advantage through developing and iterating our GLMbased model family with strong fundamental research capabilities. Our models have consistently achieved top-tier benchmark performance, demonstrating leadership in capabilities including coding, reasoning, long-context understanding, and knowledge and instruction following.
2.
We are the pioneer of MaaS (Model-as-a-Service) in China. Our MaaS platform has evolved into a comprehensive, full-stack, one-stop platform for diverse AI applications in enterprise settings.
3.
Our MaaS platform drives an expansive ecosystem with 8,000+ institutional customers and approximately 80 million device empowerments, spanning diverse industries such as technology, healthcare, finance, automotive, education, retail and more.
4.
Our team features extensive AI research expertise, particularly in pre-training and alignment of large models. Our co-founder and CEO Dr. Jie Tang (唐杰) has 20+ years of AI research experience, having produced 1,000+ publications and received many prestigious awards, which enabled us to build unique frontier-AI research capabilities.
5.
We were founded in 2019, before the era of ChatGPT, with an early-mover advantage in building, training and commercializing large models in China. We have accumulated significant technological advantages in the model development and commercial deployment.
SUMMARY OUR STRATEGIES Our primary strategies are: 1.
We are committed to continuously investing in model research and development. We seek to achieve industry-leading model capabilities by consistently advancing in core model research areas, including but not limited to reasoning, planning and multi-step problem-solving.
2.
We will actively recruit new customers across different industries and verticals. We will also increase ARPU (average revenue per user) by expanding the application scope of models across diverse scenarios within existing customers.
3.
We will continue to improve the capability of and invest in our AI products to attract more individual users through our MaaS platform. We will scale our individual user base and drive recurring revenue through product subscriptions.
4.
We will continue to promote open-source and foster an ecosystem with academic institutions, industry partners and customers. We will invest resources in enriching the third-party application ecosystem around our platform to create a comprehensive and sustainable ecosystem.
5.
We will explore new international markets for our models and products and establish a local customer base as we believe the GLM framework has significant advantages in international markets.
We have been pursuing AGI from the beginning. Our AGI-oriented model exploration roadmap demonstrates our progressive technology advancements and our continuous drive to achieve AGI. Superhuman performance in one or more tasks, learning to do one or more tasks without examples
Superhuman performance in one or more tasks, learning to do one or more tasks without examples. Our GLM-4 represents this stage.
One AI for all tasks, learning to do new tasks with minimal instruction. Our AutoGLM represents this stage.
One AI connected to all tools and taking real-world actions. Our CogAgent represents this stage.
Multimodal perception, understanding and generation across all sensory inputs.
Our models' capabilities and product offerings are continuously expanding across the four stages of our AGI-oriented model exploration roadmap.
(RMB in thousands, unless stated otherwise) Revenue 57,400 124,512 312,440 44,910 190,933 Cost of revenue (25,960) (60,437) (150,940) (20,558) (92,505) Gross profit 31,440 64,075 161,500 24,352 98,428 Gross profit margin (%) 54.8% 51.5% 51.7% 54.2% 51.6% Research and development expenses (57,027) (192,093) (283,434) (102,768) (117,023) Selling and distribution expenses (14,133) (26,879) (69,009) (21,282) (43,225) Administrative expenses (32,393) (65,697) (54,009) (27,386) (27,073) Other income and losses — net 11,459 11,289 23,978 4,483 14,523 Operating loss (100,654) (209,305) (220,974) (122,601) (74,370) Finance income — net 5,199 18,561 19,774 10,217 12,225 Loss before income tax (95,455) (190,744) (201,200) (112,384) (62,145) Income tax expense (67) (177) (3,126) (109) (3,034) Net loss (95,522) (190,921) (204,326) (112,493) (65,179) Adjusted net loss(1) (34,428) (97,254) (118,344) (64,823) (47,393) Total assets 627,695 1,086,011 1,048,831 1,107,025 1,222,388 Total liabilities 281,498 404,026 440,019 438,022 479,756
SUMMARY Notes: (1) Adjusted net loss is a non-GAAP measure. It represents the net loss for the year/period adjusted to exclude share-based compensation expenses and fair value changes of the Group's financial instruments. For details, please refer to the section headed "Financial Information — Non-GAAP Measures" in this prospectus.
RECENT DEVELOPMENT Since June 30, 2025, we have achieved a number of new developments. Our GLM-4.6 achieved the No. 1 rank in coding among all models globally on the CodeArena evaluation platform in November 2025, surpassing models from leading global AI companies including OpenAI, Anthropic and Google. We also announced the release of GLM-4.1V-Thinking, which we believe demonstrates state-of-the-art performance in vision reasoning tasks; GLM-Z1-Air, a highly efficient reasoning model designed for deployment on mobile devices and edge computing scenarios with only 3 billion parameters; and GLM-Z1RumineX, a long thinking model with exceptional mathematical reasoning capabilities.
We also made significant progress in our commercialization efforts. Our revenue for the third quarter of 2025 was RMB105 million, and our revenue from individual users grew to approximately 48% of our total revenue for the first nine months ended September 30, 2025. Our institutional customer base grew to over 8,000 as of June 30, 2025 while institutional customers that paid for our products exceeded 600 as of September 30, 2025.
We have continued to improve our financial performance. Our gross profit margin remained stable at approximately 55% for the third quarter of 2025. Our quarterly operating loss was reduced to RMB32 million in the third quarter of 2025.
Benchmark #1 in Chinese language comprehension and generation Benchmark #2 globally in agentic capability
Benchmark #1 globally in agentic capability (lightweight) Benchmark #1 in Chinese language comprehension and generation (lightweight)
Reflection and Rumination Models Reflection and rumination models are designed to imitate human deep thinking. Reflection models perform iterative self-evaluation to thoroughly analyze problems and verify responses, similar to how humans reflect on their thoughts. Rumination models extend this by connecting and seeking information from both internal knowledge and external sources, enabling a higher level of analysis—like how humans ruminate and deliberate on difficult questions.
GLM-Z1 is our flagship reflection model, featuring advanced chain-of-thought reasoning capabilities. According to Frost & Sullivan, GLM-Z1 achieves Benchmark #1 in Chinese language mathematical reasoning. Through self-reflection, self-examination and reinforcement learning from human feedback, GLM-Z1 is able to question its own solutions and evaluate whether they meet the standard of correct answers.
GLM-Z1-Rumination is our flagship rumination model. GLM-Z1-Rumination is capable of integrating information from multiple sources and conducting in-depth research, leveraging internal knowledge and online search results while thinking. It provides comprehensive and coherent responses covering both micro-level specifics and macro-level analysis to high-complexity questions. Similar to how humans solve difficult problems, GLM-Z1-Rumination can formulate a structured approach, execute it systematically and
综合模型产品组合。我们建立了全面的先进人工智能模型产品组合，在语言、多模态、智能体及代 码能力方面展现出业界领先的表现。凭借我们广泛而强大的产品储备，客户和开发者可始终找到最适合 其特定需求的解决方案。
可扩展应用。我们的模型和智能体旨在无缝运行于各类硬件、应用场景和业务工作流中。这些 模型和智能体能够处理复杂任务、支持人工智能原生、多模态及全方位对话，并执行深度推理。例 如，我们的模型和智能体可协助机构客户简化业务工作流、大规模处理和分析运营数据，并为决策 提供支持。此外，我们的MaaS平台提供智能体工作台，涵盖多种智能体模板及基于场景的解决方 案。通过该智能体工作台，客户可借助精简的模型微调、增量模型训练及提示词工程，快速定制智 能体。
易于基础设施适配。我们与算力基础设施合作伙伴合作，共同设计先进的算力基础设施，使我们 的MaaS平台能够提供集成的计算、网络、训练通信及推理加速能力。该合作还使我们的模型具有广 泛的适配性，支持从15亿到2,300亿参数的模型规模，并可在云端及各类芯片上实现大规模实时部署。 特别是，这种适配性使我们的模型能够在手机、个人电脑及智能汽车等大众消费设备上运行，从而 惠及大量终端消费者。
本地化部署方面，我们的模型托管于客户自有的基础设施内。这种方式使组织能够利用其专有或 敏感数据，将人工智能模型定制化于其特定领域。本地化部署在性能优化和基础设施配置方面提供 更强的管控能力，适用于复杂或高度专业化的应用场景。
云端部署方面，我们的模型托管于可扩展且可靠的云基础设施上。这种方式适合追求敏捷性和易 于实施的企业。借助云端，客户无需投资昂贵的本地基础设施，即可快速、高性价比地部署人工智 能解决方案。
对于本地化部署，我们在大型模型及相关服务交付至客户指定地点并经客户验收时确认收入。对于 云端部署，我们在合同期内确认收入。具体而言，对于订阅制合同，我们通常在合同期内按比例确认收 入；对于用量制合同，我们根据客户在服务提供时的资源使用量确认收入。有关我们收入确认政策的详 细信息，请参阅"财务信息——重要会计政策及估计——重要会计政策信息——收入确认"。
要使机器像人类一样思考，我们必须赋予人工智能三种核心人类能力：深度思考、认知和工具使 用。我们据此开发了人工智能模型，并将其归纳为三个相应类别：反思与沉思模型、多模态模型及智能 体模型。我们还开发了代码模型，可自主生成代码并提升编程效率。上述四个类别均基于我们的GLM系 列基础模型。基础模型和反思与沉思模型属于语言模型的更广泛类别。
认知 多模态模型 CogView（图像生成） GLM-4.5V（视觉理解与推理） 基础模型 CogVideoX（视频生成） GLM-Realtime（实时视频通话） GLM-4.5 GLM-4-Voice（端到端语音模型）
工具使用 人工智能智能体 AutoGLM（"从对话到行动"—— 用于自主完成任务的智能体） AutoGLM-Rumination（"边思考边工 作"——具备深度思考能力的自主任务 完成智能体） CoCo（企业智能体）
基础模型是经过预训练的大语言模型，是开发各类专业模型的基础。GLM-4.5是我们的旗舰基础模 型，于发布时开源。通过多阶段训练以及包含微调和强化学习的全面后训练，GLM-4.5在智能体、推理 和代码任务方面表现卓越。GLM-4.5亦支持多模态扩展和大上下文处理，能够理解高层次提示词并自主 生成实用解决方案。GLM-4.5的模型规模为3,550亿参数，我们还开发了GLM-4.5-Air，这是一款参数量 为1,060亿的轻量化版本。
反思与沉思模型旨在模拟人类的深度思考。反思模型通过迭代自我评估来深入分析问题并验证回 答，类似于人类对自身思维的反思。沉思模型在此基础上进一步延伸，通过连接并检索内部知识及外部 来源的信息，实现更高层次的分析——如同人类对难题的沉思与推敲。
GLM-Z1是我们的旗舰反思模型，具备先进的思维链推理能力。根据弗若斯特沙利文（Frost & Sullivan），GLM-Z1在中文数学推理基准测试中排名第一。通过自我反思、自我审查及人类反馈强化学 习，GLM-Z1能够质疑自身的解答并评估其是否达到正确答案的标准。
GLM-Z1-Rumination是我们的旗舰沉思模型。GLM-Z1-Rumination能够整合多方来源的信息，在思 考过程中充分利用内部知识和在线搜索结果，开展深度研究。它能够对高复杂度问题提供全面、连贯的 回答，兼顾微观细节和宏观分析。与人类解决难题的方式类似，GLM-Z1-Rumination可以制定结构化的
基准测试。根据2025年7月对十二项行业标准基准测试¹的综合评估，GLM-4.5全球排名第三、中国排名第一，并在全球开源模型中排名第一。在上述十二项基准测试中，GLM-4.5取得综合得分63.2，而同行业模型的得分区间为46.3至65.0。
全球排行榜。2025年9月，GLM-4.5在Chatbot Arena和WebDev Arena上均跻身全球第五，上述排行榜为业界公认的全球评测平台，分别对大模型的综合能力及编程能力进行排名。
Token消耗量。自GLM-4.5发布至2025年12月初，本公司在OpenRouter（一家为各类大模型提供API接入服务的全球领先平台）上的token消耗量持续位居全球前十、中国企业前三。这一持续表现印证了GLM-4.5的强劲竞争力和市场认可度，彰显其在效率、可扩展性及实际应用场景中的先进性。
热度排名。GLM-4.5发布后仅48小时内，便在全球最大开源模型平台Hugging Face的热度榜上跻升全球第一。
¹ 基准测试是衡量大型语言模型（LLM）在多项任务上能力的结构化标准化评估。本公司用于评估GLM-4.5的十二项行业标准基准测试涵盖三大类别：（i）智能体基准测试，包括TAU-Bench、BFCL V3和BrowseComp。GLM-4.5在上述智能体基准测试中取得综合得分58.1，而同行业模型的得分区间为45.0至61.1；（ii）推理基准测试，包括MMLU-Pro、AIME 24、MATH-500、SciCode、GPQA、HLE及LCB（2407-2501）。GLM-4.5在上述推理基准测试中取得综合得分68.8，而同行业模型的得分区间为63.5至74.2；（iii）编程基准测试，包括SWE-Bench Verified和Terminal-Bench。GLM-4.5在上述编程基准测试中取得综合得分50.9，而同行业模型的得分区间为36.7至55.5。
幻觉率。根据检索增强生成（RAG）大模型幻觉排行榜，2025年9月，GLM-4.5的幻觉率位居全球第二低、中国最低。该基准测试通过评估大模型在面对刻意误导性问题时生成不实答案（即"幻觉"）的频率，对各大模型进行综合评级。
2025年9月，本公司发布GLM-4.6，作为基础模型的进一步升级版本，主要在编程能力方面实现显著增强。2025年11月，GLM-4.6在CodeArena上荣登全球第一，该平台为最新一代业界公认的全球评测平台，专门用于评估模型的编程能力。
深思与沉思模型在生成答案前会额外投入时间进行"深度思考"，使其在处理复杂推理任务时表现更为出色。本公司在基础模型之上，构建了深思模型（GLM-Z1）和沉思模型（GLM-Z1-Rumination）。
GLM-Z1是一款深思模型，专为解决具有确定性的问题而设计，旨在提供更为精准、准确的解答。该模型以基础模型为基础，通过扩展强化学习以及针对数学、编程和逻辑等任务的进一步训练而研发。
GLM-Z1-Rumination专为解决具有不确定性的问题而设计，尤其适用于需要迭代收集和处理外部信息的开放式探索性问题。与GLM-Z1相比，GLM-Z1-Rumination具备更深层次、更长时间的思考能力，并能够借助工具解决更具开放性和复杂性的问题。
多模态模型能够处理和整合来自文本、图像、音频和视频等多种模态的信息。本公司已开发多款服务于不同功能的多模态模型，包括CogView（图像生成）、GLM-4.5V（视觉理解与推理）、CogVideoX（视频生成）、GLM-Realtime（实时视频通话）及GLM-4-Voice（语音模型）。
AI智能体融合了推理、规划和工具调用能力，能够在无需人工持续介入的情况下自主执行多步骤任务。
本公司的基础智能体模型为AutoGLM。AutoGLM代表了本公司AI宇宙演进历程中的重大跨越——从"对话"迈向"行动"，弥合了对话式AI与现实世界任务执行之间的鸿沟。AutoGLM专为通过图形用户界面（GUI）自主控制数字设备而设计，是量身打造的基础智能体，能够将类人推理转化为具体行动。AutoGLM在AgentBench上取得了最优（SOTA）性能，AgentBench是斯坦福大学2024年AI指数报告所认可的智能体AI基准测试。
2025年8月，本公司发布AutoGLM升级版（又称"AutoGLM 2.0"），该版本以当时最新的基础模型GLM-4.5及视觉理解与推理模型GLM-4.5V为底层驱动。升级版AutoGLM使其能够在更广泛的移动应用程序和网站上模拟人类操作行为。AutoGLM可在云端自主完成所请求的任务，无需占用用户的手机或电脑，令用户得以在任务执行过程中不间断地继续使用自己的设备。
AutoGLM Rumination是AutoGLM的高级版本，是一款自主AI智能体，专为探索开放性问题并基于其发现采取行动而设计。AutoGLM Rumination具备"边思考边工作"的特点——它利用GLM-Z1-Rumination模型提供的卓越推理能力，同时融合了AutoGLM的交互操作功能。AutoGLM Rumination能够处理涉及深度推理、迭代研究并产生可执行成果的复杂任务。
我们还开发了CoCo，这是一款精密的企业级AI智能体，旨在跨企业环境提供智能自动化服务。
CodeGeeX是一款强大的编程模型，旨在提升编程效率并优化工作流程。它使开发者能够根据自然语言描述自动生成代码，或补全未完成的代码行或代码块，从而显著提高生产力。
截至最后实际可行日期，上述精选模型和智能体均已实现商业化。
• 拥有丰富研究与行业经验的管理及顾问团队。
详见"业务——我们的优势"。
• 吸引并留住顶尖人才。
详见"业务——我们的战略"。
我们的行业顾问弗若斯特沙利文（Frost & Sullivan）确认，且董事认为，我们属于专业技术行业中的可接受行业，即上市规则第18C章所定义的新一代信息技术项下的人工智能领域。
| 专业技术产品 | 可接受行业 | 可接受行业——子类别 | 功能及适用性分析 | |---|---|---|---| | 我们的每一款模型及智能体（详见上文"——我们的模型与智能体"） | 新一代信息技术——人工智能 | AI驱动的算法编程：图像识别、视听学习、自然语言处理、机器学习及深度学习 | 大模型是一套复杂的算法体系，通过迭代调整数十亿参数，从海量数据中识别并利用规律，以提升准确性。此类算法使我们的模型能够实现图像识别、视听学习、自然语言处理、机器学习及深度学习等功能。例如，我们的多模态模型能够处理和整合来自文本、图像、音频及视频等多种模态的信息。 | | | | AI解决方案：面向不同垂直行业设计和提供AI解决方案 | 我们的模型构成面向广泛行业的AI解决方案的基础，涵盖科技与互联网、公共服务及传统企业部门（如金融服务、制造业和能源行业）。例如，在科技与互联网领域，我们的大模型帮助客户大规模审阅文件并分析运营数据；在公共服务领域，我们为城市公共交通管理开发了大模型，实现对公交客流量的精准监测及到站时间的实时预测。更多展示我们AI解决方案在不同垂直行业应用的案例，详见"业务——我们的商业应用案例"。 |
基于以下分析以及董事和弗若斯特沙利文的意见，独家保荐人认为，本招股说明书所述本集团提供的模型和智能体属于专业技术行业中的可接受行业，即上市规则第18C章所定义的新一代信息技术项下的人工智能领域。截至最后实际可行日期，我们的专业技术产品无需取得任何监管批准。
作为大模型公司，我们在大语言模型（LLM）市场运营，该市场是AI市场的细分领域。LLM的发展近年来取得了重大飞跃，尤其是自2022年以来。目前最前沿的研究主要在美国和中国开展，两国均涌现出多家领先企业。
个人消费者方面，该市场预计到2030年将增长至1,011亿元人民币，2024年至2030年的复合年增长率为63.5%。以机构客户为主要增长动力，中国企业级大语言模型市场预计到2030年将达到904亿元人民币，2024年至2030年的复合年增长率为63.7%。
中国大语言模型市场的参与者可分为独立提供商和非独立提供商。独立提供商的特点是从运营初期便围绕大语言模型技术、产品和商业模式原生构建；而非独立提供商通常是涉足人工智能领域的科技巨头。独立提供商与非独立提供商所面临的竞争态势存在很大差异。例如，非独立提供商凭借其既有的多元化业务线积累了庞大的用户基础，有助于推广其大语言模型产品。然而另一方面，如果某科技巨头旗下有与企业客户自身业务直接竞争的业务线，企业客户可能不愿选择该科技巨头提供的大语言模型产品。此外，某些行业的企业客户对于落入或被视为落入特定科技巨头势力范围较为敏感，更倾向于采用"纯粹"提供商的人工智能解决方案。
| 排名 | 公司 | 类型 | 收入（十亿元人民币） | 市场份额 | |------|------|------|------------------|----------| | 1 | 公司A | 非独立 | 0.44 | 9.4% | | 2 | 本公司 | 独立 | 0.31 | 6.6% | | 3 | 公司B | 非独立 | 0.30 | 6.4% | | 4 | 公司C | 非独立 | 0.29 | 6.1% | | 5 | 公司D | 非独立 | 0.22 | 4.7% |
注： 1) 公司A成立于1999年，为在深圳证券交易所上市的上市公司。
2) 公司B成立于1999年，为同时在香港联合交易所及纽约证券交易所上市的上市公司。
3) 公司C成立于2014年，为在香港联合交易所上市的上市公司。
4) 公司D成立于2000年，为同时在香港联合交易所及纳斯达克上市的上市公司。
我们的核心是一家由数据科学家和工程师组成的公司，研发贯穿我们工作的每个方面。在日常工作中，我们专注于提升基础模型的智能水平；为越来越多的行业垂直领域及其他使用场景改进并开发实用且具有成本效益的人工智能智能体；并与业务合作伙伴协作设计和改进计算基础设施，使我们的模型即服务（MaaS）平台能够提供全面的能力。我们通过首要的人才资源，以及强大的技术基础设施和严格的研发流程来实现上述目标。
截至2025年6月30日，我们拥有657名研发团队成员，具备自然语言处理、复杂系统中的高级决策以及多模态语义分析等相关领域的背景和经验。详见"业务——研究与开发——人才"。
我们的知识产权对于支撑我们成功的创新至关重要。我们通过专利、版权、商标、域名、商业秘密、保密协议及其他措施的组合来保护我们的知识产权。截至最后实际可行日期，我们在中国拥有86项已注册专利，其中84项为发明专利，以及232项在中国的专利申请。截至同一日期，我们在中国拥有160项版权、314项商标及59个域名。
我们已设计并采取了全面措施以保护我们的知识产权。我们与员工、特定顾问及咨询师签订包含保密条款、竞业禁止条款及知识产权归属条款的劳动协议。他们确认，其在受雇于我们期间所开发的与我们相关的知识产权，包括我们内部开发的内容，均为我们的财产。在业绩记录期间及直至最后实际可行日期，我们未曾面临任何涉及侵犯第三方商标、许可证及其他知识产权的重大争议或索赔。
作为一家大模型公司，我们在大语言模型市场中运营，该市场是人工智能市场的一个细分领域。大语言模型市场竞争激烈。根据弗若斯特沙利文的数据，大语言模型供应商的竞争因素包括：(i) 技术能力，如自主研发的大语言模型预训练框架以及模型定制化与优化能力；(ii) 灵活的商业模式和交付策略；(iii) 生态系统构建能力；以及 (iv) 具有深厚技术背景和丰富经验的人才。我们与国内外独立及非独立大语言模型供应商均存在竞争关系。未来我们也可能面临新进入者带来的竞争加剧。对我们而言重要的主要竞争因素包括：我们服务产品的范围、性能和安全性、用户体验、研发能力及人才储备。有关我们所处行业竞争格局的更多详情，请参阅"行业概览"及"业务——竞争"。
在业绩记录期间，我们实现了强劲的收入增长。我们的收入从2022年的5,740万元人民币增长至2023年的1.245亿元人民币，并进一步增长至2024年的3.124亿元人民币，复合年增长率超过130%。我们的收入也从截至2024年6月30日止六个月的4,490万元人民币大幅增长至截至2025年6月30日止六个月的1.909亿元人民币。
2024 and the six months ended June 30, 2024 and 2025, respectively. Our losses during the Track Record Period were primarily due to our significant investments in research and development. Our R&D expenses increased from RMB84.4 million in 2022 to RMB528.9 million in 2023 and further to RMB2,195.4 million in 2024. Our R&D expenses increased by 85.6% from RMB859.2 million in the six months ended June 30, 2024 to RMB1,594.7 million in the six months ended June 30, 2025.
尽管我们的净亏损绝对金额在业绩记录期间有所增加，我们预计通过增加收入和提升运营效率来扭转净亏损局面。
收入增长是实现盈利的关键。我们已建立强大的人工智能模型和智能体产品组合，以赋能广泛的行业，应对各行业的独特挑战并优化工作流程。借助大语言模型市场的巨大市场潜力以及我们的技术领先优势，我们处于有利地位，能够根据新兴市场机遇升级我们的人工智能模型和智能体，并持续实现收入增长。我们收入的增长将逐步覆盖相关成本和费用，从而总体上减少我们的净亏损。我们计划通过以下方式增加收入：（i）利用中国大语言模型市场的增长潜力；（ii）持续推动人工智能模型和智能体的迭代与升级；（iii）拓宽客户覆盖范围；以及（iv）将我们的解决方案扩展至新的行业领域。
提升运营效率也是实现盈利的重要因素。我们计划优化研发效率，提高运营管理效能。自成立以来，我们在研发及我们的MaaS平台上进行了大量投入。凭借深厚的学术根基作为技术领先地位的基石，我们持续推进技术进步和基础模型的迭代，提升盈利能力。我们还采取了措施控制一般及行政费用，提高运营效率。我们简化了工作流程并优化了资源配置，以确保以具有成本效益的方式满足运营需求。此外，我们对销售及营销活动采取了聚焦的方式，以进一步提高相关费用的使用效率。通过将资源和精力集中于基于云的部署并利用数据驱动的洞察，我们优化了销售和营销策略。
详情请参阅"业务——业务可持续性"。综合考虑我们的历史增长、市场机遇以及我们的盈利计划，我们认为我们拥有可持续的商业模式。
我们的客户包括企业、公共部门实体及个人用户。在2022年、2023年、2024年及截至2025年6月30日止六个月各年度／期间内，来自我们五大客户的收入分别占我们总收入的55.4%、61.5%、45.5%及40.0%。在2022年、2023年、2024年及截至2025年6月30日止六个月各年度／期间内，来自我们最大客户的收入分别占我们总收入的15.4%、14.7%、19.0%及11.0%。
我们的供应商主要包括：（i）计算资源提供商，如计算硬件和计算服务提供商；（ii）硬件设备供应商，包括服务器、存储设备和网络设备供应商；（iii）研发支持服务提供商，如数据清洗和大模型评估服务提供商；以及（iv）营销服务提供商。在2022年、2023年、2024年及截至2025年6月30日止六个月各年度／期间内，向我们五大供应商的采购额分别占我们总采购额的54.5%、
53.6%、47.3%及50.2%。在2022年、2023年、2024年及截至2025年6月30日止六个月各年度／期间内，向我们最大供应商的采购额分别占我们总采购额的33.1%、16.4%、15.6%及13.4%。
于最后实际可行日期，控股股东——北京联排、刘博士、唐博士、李博士、徐博士、张博士、惠惠及智灯，凭借彼等之间订立的经修订一致行动人协议，合共持有约33.03%的股份权益。详情请参阅"历史、发展及公司架构——一致行动人安排及我们的控股股东"。
全球发售完成后，我们的控股股东集团将合共持有约30.22%的股份（假设超额配股权未获行使）。因此，于上市后，彼等将继续作为上市规则所定义的我们的控股股东集团。有关我们控股股东的更多详情，请参阅"与我们控股股东的关系"。
我们已向联交所申请批准（i）根据未上市股份转换而转换的H股，以及（ii）根据全球发售（包括根据行使超额配股权可能发行的任何H股）发行的H股的上市及买卖许可，其依据为（除其他事项外）我们根据上市规则第18C.03条作为商业公司（定义见上市规则）满足相关规定，参考我们于上市时的预期市值，按每股发售股份116.20港元的发售价计算，该市值超过60亿港元。
我们完成了八轮Pre-IPO投资，合共募集资金逾人民币83.60亿元（RMB8,360 million）。有关Pre-IPO投资及Pre-IPO投资者的更多详情，请参阅"历史、发展及公司架构——Pre-IPO投资"。
Our operations and the Global Offering involve certain risks and uncertainties, including (i) risks related to our research and development, (ii) risks related to our commercialization, (iii) risks related to our operations, (iv) risks related to our intellectual property, (v) risks related to our financial condition and need for additional capital, (vi) risks related to where we conduct business, and (vii) risks related to the Global Offering, which are set out in the section headed "Risk Factors" in this prospectus. You should read that section in its entirety carefully before you decide to invest in the Offer Shares. Some of the major risks we face include, but are not limited to:
• The AI industry is characterized by constant changes. If we are not able to upgrade, enhance or innovate our technologies and services, our business, results of operations, financial condition and prospects could be adversely affected.
• We have made and expect to continue to make substantial investments in R&D. If we cannot continuously invest in our R&D activities while achieving technological innovation, our business, results of operations, financial condition and prospects may be materially and adversely affected.
• The development of AGI is still at an early stage and there are substantial uncertainties in the future realization of AGI.
• We rely on third parties to provide computing resources to us, and any disruption of their services or fluctuation of prices could adversely affect our business, results of operations and financial condition.
• Our commercial success depends on the performance of our models. Any failure in research and development efforts to offer high-quality models and solutions could harm our business, results of operations, financial condition and prospects.
• The AI industry is subject to evolving and extensive regulation in China. Future laws and regulations may impose additional requirements and other obligations that could materially and adversely affect our business, results of operations, financial condition and prospects.
• We may not be able to obtain or maintain adequate intellectual property rights protection for our business, or the scope of such intellectual property rights protection may not be sufficiently broad.
• We may not be able to compete effectively against current or future competitors.
• The size of our addressable market and the demand for our solutions may not increase as rapidly as we anticipate due to a variety of factors, which would materially and adversely affect our business, results of operations, financial condition and prospects.
• Any failure of our MaaS platform to perform as required could harm our business, results of operations, financial condition and prospects.
• We have a limited track record in the commercialization of our business.
The following tables set forth the summary of key financial information during the Track Record Period, extracted from the Accountants' Report as set out in Appendix I to this document. The key financial information set forth below should be read together with, and is qualified in its entirety by reference to, our financial statements in this document, including the related notes. Our consolidated financial information was prepared in accordance with the IFRS Accounting Standards.
The following table summarizes our results of operations and as percentage of our total revenue for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | | | | | |---|---|---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | Revenue | 57,409 | 100.0 | 124,538 | 100.0 | 312,414 | 100.0 | 44,909 | 100.0 | 190,877 | 100.0 | | Cost of revenue | (26,049) | (45.4) | (44,056) | (35.4) | (136,525) | (43.7) | (22,950) | (51.1) | (95,453) | (50.0) | | Gross profit | 31,360 | 54.6 | 80,482 | 64.6 | 175,889 | 56.3 | 21,959 | 48.9 | 95,424 | 50.0 | | Other income | 1,784 | 3.1 | 9,965 | 8.0 | 19,281 | 6.2 | 4,174 | 9.3 | 4,614 | 2.4 | | Selling and marketing expenses | (15,139) | (26.4) | (101,198) | (81.3) | (387,475) | (124.0) | (144,194) | (321.1) | (208,570) | (109.3) | | General and administration expenses | (32,316) | (56.3) | (66,302) | (53.2) | (133,603) | (42.8) | (51,452) | (114.6) | (185,165) | (97.0) | | Research and development expenses | (84,377) | (147.0) | (528,884) | (424.7) | (859,217) | (1,913.2) | (1,594,661) | (835.4) | — | — | | Impairment losses on financial assets | (98,719) | (172.0) | (625,723) | (502.4) | (2,538,352) | (812.5) | (1,029,493) | (2,292.4) | (1,899,225) | (995.0) | | Loss from operations | (5,694) | (9.9) | (26,332) | (21.1) | (38,321) | (12.3) | (12,212) | (27.2) | (53,270) | (27.9) | | Finance costs | (31) | (0.1) | (19,786) | (15.9) | (17,008) | (5.4) | (763) | (1.7) | — | — | | Share of profits less losses of associates | — | — | (453) | (0.4) | 21,254 | 6.8 | 324 | 0.7 | 14,147 | 7.4 | | Changes in fair value of financial instruments measured at fair value through profit or loss ("FVPL") | 5,972 | 10.4 | 26,022 | 20.9 | 66,271 | 21.2 | 7,004 | 15.6 | 9,791 | 5.1 | | Changes in the carrying amounts of financial instruments issued to investors | (468,859) | (150.1) | (201,174) | (448.0) | (429,295) | (224.9) | — | — | — | — | | | (45,209) | (78.7) | (161,471) | (129.7) | — | — | — | — | — | — | | Loss before taxation | (143,650) | (250.2) | (787,957) | (632.7) | (2,958,007) | (946.8) | (1,235,551) | (2,751.2) | (2,357,852) | (1,235.3) | | Income tax | — | — | — | — | — | — | — | — | — | — | | Loss for the year | (143,650) | (250.2) | (787,957) | (632.7) | (2,958,007) | (946.8) | (1,235,551) | (2,751.2) | (2,357,852) | (1,235.3) | | Loss attributable to: | | | | | | | | | | | | Equity holders of the Company | (143,374) | (249.7) | (787,960) | (632.7) | (2,956,491) | (946.3) | (1,235,551) | (2,751.2) | (2,351,173) | (1,231.8) | | Non-controlling interests | (276) | (0.5) | 3 | 0.0 | (1,516) | (0.5) | — | — | (6,679) | (3.5) |
To supplement our consolidated financial statements that are presented in accordance with IFRS Accounting Standards, we also use adjusted loss for the year (a non-IFRS measure), as an additional financial measure which is not required by or presented in accordance with IFRS Accounting Standards. We believe that this non-IFRS measure facilitates comparisons of operating performance from period to period by eliminating potential impact of certain items. We believe that this measure provides useful information to investors and others in understanding and evaluating our consolidated financial statements in the same manner as they help our management. However, our presentation of adjusted loss for the year (a non-IFRS measure) may not be comparable to similar measures presented by other companies. The use of such non-IFRS measure has limitations as an analytical tool, and you should not consider them in isolation from, or as substitute for analysis of, our consolidated financial statements or financial condition as reported under IFRS Accounting Standards. We define adjusted loss for the year/period (a non-IFRS measure) as loss for the year/period adjusted for adding back equity-settled share-based compensation expenses, changes in the carrying amount of financial instruments issued to investors, and listing expenses.
| | Year Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Net loss for the year/period | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Add: | | | | | | | — Equity-settled share-based compensation expenses(1) | 1,024 | 5,502 | 23,579 | 4,217 | 158,852 | | — Changes in the carrying amount of financial instruments issued to investors(2) | 45,209 | 161,471 | 468,859 | 201,174 | 429,295 | | — Listing expense(3) | — | — | — | — | 17,731 | | Adjusted net loss for the year/period (non-IFRS measure) | (97,417) | (620,984) | (2,465,569) | (1,030,160) | (1,751,974) |
During the Track Record Period, we derived our revenue from providing large model services through our MaaS platform. The following table sets forth a breakdown of revenue by segment both in absolute amount and as a percentage of our total revenue for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | |---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | | | (RMB in thousands, except for percentages) | | | | (Unaudited) | | | On-premise deployment | 54,815 | 95.5% | 112,614 | 90.4% | 263,930 | 84.5% | | Cloud-based deployment | 2,594 | 4.5% | 11,924 | 9.6% | 48,484 | 15.5% | | **Total** | **57,409** | **100.0%** | **124,538** | **100.0%** | **312,414** | **100.0%** |
| | Six Months Ended June 30, | | |---|---|---| | | 2024 | | 2025 | | | On-premise deployment | 26,806 | 59.7% | 161,777 | 84.8% | | Cloud-based deployment | 18,103 | 40.3% | 29,100 | 15.2% | | **Total** | **44,909** | **100.0%** | **190,877** | **100.0%** |
Starting from 2024, our large model on-premise deployment services have generated revenue from overseas customers, primarily from customers in Southeast Asia. The following table sets forth a breakdown of revenue derived from on-premise deployment by geographical location of our customers both in absolute amount and as a percentage of our total revenue derived from on-premise deployment for the years/periods indicated.
| | Years Ended December 31, | | | | | | Six Months Ended June 30, | | | | |---|---|---|---|---|---|---|---|---|---|---| | On-premise deployment | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | (RMB in thousands, except for percentages) | | Mainland China | 54,815 | 100.0% | 112,614 | 100.0% | 262,479 | 99.5% | 26,806 | 100.0% | 142,990 | 88.4% | | Southeast Asia(1) | — | — | — | — | 1,080 | 0.4% | — | — | 17,927 | 11.1% | | Hong Kong | — | — | — | — | 371 | 0.1% | — | — | — | — | | Others(2) | — | — | — | — | — | — | — | — | 860 | 0.5% | | **Total** | **54,815** | **100.0%** | **112,614** | **100.0%** | **263,930** | **100.0%** | **26,806** | **100.0%** | **161,777** | **100.0%** |
Notes: (1) Includes Malaysia and Singapore. (2) Includes the United States.
During the Track Record Period, our cost of revenue consisted of (i) payroll cost, mainly representing the wages and benefits of service personnel involved in the deployment and maintenance of our services; (ii) computing service fee paid to providers of computing power necessary for providing our services; (iii) provision for warranty; (iv) depreciation and amortization; (v) technical service and consultation fees as we outsourced to third-party service providers certain standard labor intensive tasks to save costs, such as data annotation; (vi) tax and surcharges; and (vii) others.
The following table sets forth a breakdown of our cost of revenue by nature both in absolute amount and as a percentage of our total cost of revenue for the years/periods indicated.
| | Year Ended December 31, | | | | | | Six Months Ended June 30, | | | | |---|---|---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | Payroll cost | 16,434 | 63.1% | 15,375 | 34.9% | 74,262 | 54.4% | 10,832 | 47.2% | 37,628 | 39.4% | | Computing service fee | — | — | 11,996 | 27.2% | 30,173 | 22.1% | 8,451 | 36.8% | 35,845 | 37.6% | | Provision for warranty | 2,101 | 8.1% | 4,522 | 10.3% | 12,171 | 8.9% | 1,672 | 7.3% | 6,670 | 7.0% | | Depreciation and amortization | 1,912 | 7.3% | 6,436 | 14.6% | 6,769 | 5.0% | 549 | 2.4% | 2,074 | 2.2% | | Technical service and consultation fees | 4,553 | 17.5% | 623 | 1.4% | 9,126 | 6.7% | 356 | 1.6% | 6,165 | 6.5% | | Taxes and surcharges | 57 | 0.2% | 1,713 | 3.9% | 541 | 0.4% | 361 | 1.6% | 1,109 | 1.2% | | Others | 992 | 3.8% | 3,391 | 7.7% | 3,483 | 2.5% | 729 | 3.1% | 5,962 | 6.1% | | **Total** | **26,049** | **100.0%** | **44,056** | **100.0%** | **136,525** | **100.0%** | **22,950** | **100.0%** | **95,453** | **100.0%** |
The following table sets forth a breakdown of our cost of revenue by segment in absolute amount and as a percentage of our total cost of revenue for the years/periods indicated:
| | Year Ended December 31, | | | | | | Six Months Ended June 30, | | | | |---|---|---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | On-premise deployment | 25,429 | 97.6% | 35,833 | 81.3% | 89,674 | 65.7% | 10,030 | 43.7% | 66,237 | 69.4% | | Cloud-based deployment | 620 | 2.4% | 8,223 | 18.7% | 46,851 | 34.3% | 12,920 | 56.3% | 29,216 | 30.6% | | **Total** | **26,049** | **100.0%** | **44,056** | **100.0%** | **136,525** | **100.0%** | **22,950** | **100.0%** | **95,453** | **100.0%** |
The following table sets forth a breakdown of our gross profit and gross profit margins, for the years/periods indicated.
| | Year Ended December 31, | | | | | | Six Months Ended June 30, | | | | |---|---|---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | On-premise deployment | 29,386 | 53.6% | 76,781 | 68.2% | 174,256 | 66.0% | 16,776 | 62.6% | 95,540 | 59.1% | | Cloud-based deployment | 1,974 | 76.1% | 3,701 | 31.0% | 1,633 | 3.4% | 5,183 | 28.6% | (116) | (0.4%) | | **Total** | **31,360** | **54.6%** | **80,482** | **64.6%** | **175,889** | **56.3%** | **21,959** | **48.9%** | **95,424** | **50.0%** |
We recorded accumulated losses during the Track Record Period mainly because, despite achieving substantial revenue growth, we strategically made significant investments in research and development to support the development of advanced models and the ongoing improvement of our foundation models, and investments in sales and marketing activities to increase our customer base.
Our net loss increased from RMB143.7 million in 2022, to RMB788.0 million in 2023, and further increased to RMB2,958.0 million in 2024. Our net loss increased from RMB1,235.6 million for the six months ended June 30, 2024 to RMB2,357.9 million for the six months ended June 30, 2025. Such overall upward trend in our net loss was primarily due to (i) the significant increase in research and development expenses as we expanded our R&D team and procured substantially more computing services from third parties and related computing hardware, generally in line with our business expansion; and (ii) the increase in selling and marketing expenses as we expanded our sales and marketing team and made more advertising investment in order to swiftly take advantage of emerging market opportunities.
The following table sets forth a summary of our consolidated statement of financial position as of the dates indicated.
| | As of December 31, | | | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | Total non-current assets | 62,445 | 996,675 | 1,359,902 | 1,360,356 | | Total current assets | 299,531 | 1,863,478 | 3,015,867 | 3,740,355 | | Total current liabilities | (531,855) | (3,608,888) | (7,838,055) | (10,829,216) | | Net current liabilities | (232,324) | (1,745,410) | (4,822,188) | (7,088,861) | | Total assets less current liabilities | (169,879) | (748,735) | (3,462,286) | (5,728,505) | | Total non-current liabilities | (10,309) | (233,858) | (492,859) | (422,336) | | Net liabilities | (180,188) | (982,593) | (3,955,145) | (6,150,841) |
Our net liabilities increased from RMB180.2 million as of December 31, 2022 to RMB982.6 million as of December 31, 2023, primarily attributable to total comprehensive loss of RMB788.0 million mainly driven by the research and development expenses we incurred, and capital contributions from equity holders of RMB18.8 million, partially offset by the recognition of the equity settled share-based transactions of RMB5.5 million. Our net liabilities further increased from RMB982.6 million as of December 31, 2023 to RMB3,955.1 million as of December 31, 2024, primarily attributable to total comprehensive loss of RMB2,958.1 million mainly driven by the research and development expenses and capital contributions from equity holders of RMB41.0 million, partially offset by the recognition of the equity settled share-based transactions of RMB23.6 million. Our net liabilities further increased to RMB6,150.8 million as of June 30, 2025, primarily attributable to total comprehensive loss of RMB2,357.4 million mainly driven by the research and development expenses, partially offset by the recognition of the equity settled share-based transactions of RMB161.7 million.
Our net current liabilities increased from RMB232.3 million as of December 31, 2022 to RMB1,745.4 million as of December 31, 2023, primarily attributed to an increase of RMB2,721.9 million in financial instruments issued to investors, partially offset by an increase of RMB1,030.4 million in cash and cash equivalents.
Our net current liabilities increased from RMB1,745.4 million as of December 31, 2023 to RMB4,822.2 million as of December 31, 2024, primarily attributed to an increase of RMB3,497.1 million in financial instruments issued to investors, partially offset by an increase of RMB1,019.8 million in cash and cash equivalents.
Our net current liabilities increased from RMB4,822.2 million as of December 31, 2024 to RMB7,088.9 million as of June 30, 2025, primarily attributed to an increase of RMB2,887.8 million in financial instruments issued to investors. For more information, see "Financial Information—Discussion of Selected Items of Consolidated Statements of Financial Position."
During the Track Record Period, we recognized the financial instruments issued to Pre-IPO Investors as financial liabilities. The financial instruments issued to investors will be re-designated from liabilities to equity as a result of the termination of all special rights of the Pre-IPO Investors upon the Global Offering. For more details, see Note 26 of the Accountants' Report set forth in Appendix I to this prospectus. We expect to turn our net liabilities position into net asset position upon such re-designation.
The following table sets forth our cash flows for the years/periods indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Net cash used in operating activities | (68,246) | (648,017) | (2,244,919) | (994,678) | (1,327,150) | | Net cash generated from/(used in) investing activities | 32,921 | (784,965) | (48,559) | (138,775) | (556,373) | | Net cash generated from financing activities | 191,196 | 2,463,043 | 3,312,073 | 816,341 | 2,165,110 | | Net increase/(decrease) in cash and cash equivalents | 155,871 | 1,030,061 | 1,018,595 | (317,112) | 281,587 | | Cash and cash equivalents at the beginning of the year/period | 63,057 | 218,928 | 1,249,175 | 1,249,175 | 2,268,164 | | Effect of exchange rate changes | — | 186 | 394 | 3 | 2,219 | | Cash and cash equivalents at the end of the year/period | 218,928 | 1,249,175 | 2,268,164 | 932,066 | 2,551,970 |
我们的现金消耗率是指以下各项的月平均金额：(i) 经营活动使用的净现金，(ii) 购置物业及设备使用的净现金，(iii) 购买/处置财富管理产品使用/产生的净现金，以及 (iv) 租赁付款。我们的历史现金消耗率在截至2022年、2023年及2024年12月31日止各年度以及截至2025年6月30日止六个月分别为人民币300万元、人民币1.059亿元、人民币1.945亿元及人民币3.273亿元，主要反映我们在研发活动及业务运营方面的投入。于追踪记录期间，我们记录了购置物业及设备的支出，主要由于我们采购及租赁计算硬件及办公室，以配合业务扩张，用于运营活动及研发活动。
截至2025年10月31日，我们的现金及现金等价物、以公允价值计量且其变动计入损益的短期投资，以及可动用的已承诺银行信贷额度合计为人民币89.431亿元。我们估计，假设超额配股权未获行使，并以每股发售股份HK$116.20的发售价为基准，在扣除全球发售相关的承销佣金及其他估计发售费用后，我们将获得约人民币37.853亿元（HK$41.734亿元）的全球发售所得款项净额。
假设未来平均现金消耗率将为人民币3.273亿元，与截至2025年6月30日止六个月的现金消耗率水平相近，并基于以下假设：(i) 我们的员工人数不会显著增加，尤其是研发部门；(ii) 我们预计不会有大规模资本投入；以及 (iii) 我们预计不会大量购置固定资产——我们估计，截至2025年10月31日的现金及现金等价物、以公允价值计量且其变动计入损益的短期投资，以及可动用的已承诺银行信贷额度，将足以维持我们27.3个月的财务可行性；若计入上市所得款项净额的10%（即拨作营运资金及其他一般企业用途的部分），则可维持28.5个月；若同时计入全部上市所得款项净额，则可维持38.9个月。我们将继续密切监控经营活动的现金流量，并通过多种方式维持财务可行性，包括但不限于银行信贷额度及外部融资。
截至2025年6月30日止六个月，我们经营活动使用的净现金为人民币13.272亿元，主要归因于税前亏损人民币23.579亿元，主要由以下项目抵销：(i) 向投资者发行的金融工具账面金额变动人民币4.293亿元，以及 (ii) 贸易及其他应付款项增加人民币2.383亿元。
截至2024年12月31日止年度，我们经营活动使用的净现金为人民币22.449亿元，主要归因于税前亏损人民币29.580亿元，主要经以下调整：(i) 贸易及其他应收款项增加人民币4.150亿元，以及 (ii) 以公允价值计量且其变动计入损益的金融资产公允价值变动人民币6,630万元。上述影响部分由以下项目抵销：(i) 向投资者发行的金融工具账面金额变动人民币4.689亿元，(ii) 物业及设备折旧人民币2.703亿元，以及 (iii) 贸易及其他应付款项增加人民币4.160亿元。
截至2023年12月31日止年度，我们经营活动使用的净现金为人民币6.480亿元，主要归因于税前亏损人民币7.880亿元，主要经贸易及其他应收款项增加人民币2.695亿元调整。上述影响部分由以下项目抵销：(i) 贸易及其他应付款项增加人民币1.432亿元，(ii) 向投资者发行的金融工具账面金额变动人民币1.615亿元，以及 (iii) 物业及设备折旧人民币6,380万元。
截至2022年12月31日止年度，我们经营活动使用的净现金为人民币6,820万元，主要归因于税前亏损人民币1.437亿元，主要经贸易及其他应收款项增加人民币1,960万元调整。上述影响部分由以下项目抵销：(i) 向投资者发行的金融工具账面金额变动人民币4,520万元，(ii) 贸易及其他应付款项增加人民币1,290万元，(iii) 物业及设备折旧人民币1,660万元，以及 (iv) 合同负债增加人民币1,310万元。
| | 截至12月31日止年度 | | | 截至2025年6月30日止六个月 | |---|---|---|---|---| | | 2022年 | 2023年 | 2024年 | | | 收入增长率 | — | 7.1% | 27.6% | 20.4% | | 流动比率(1) | 1.2 | 1.5 | 3.3 | 0.6 | | 速动比率(2) | 0.5 | 0.4 | 0.3 | 0.5 | | 资产负债率(3) | 0.5 | 0.4 | 0.3 | 12.0% |
(1) 流动比率按所示日期的流动资产除以流动负债计算。
(2) 速动比率按所示日期的流动资产减存货后除以流动负债计算。
(3) 资产负债率按计息银行及其他借款总额与租赁负债之和除以该期末权益总额再乘以100%计算。
| | 以每股发售股份HK$116.20的固定发售价为基准 | |---|---| | 本公司股份市值(1) | HK$511.55亿元 | | 未经审核备考经调整每股有形资产净值(2) | HK$17.85 |
注： (1) 市值的计算基于全球发售完成后（假设超额配股权未获行使）预期已发行的440,230,190股H股。 (2) 每股H股归属于本公司股东的未经审计备考经调整综合有形资产净值，乃经参考「附录二——未经审计备考财务资料」所述调整后得出，并基于已发行股份数目为440,230,190股（即截至2025年6月30日计及股份细分后已发行及流通的402,810,690股股份，加上根据全球发售将予发行的37,419,500股H股），且未计及可能因行使超额配股权或股份激励计划而发行的任何股份。
Assuming that the Over-allotment Option is not exercised, after deducting the underwriting commissions and other estimated offering expenses payable by us in connection with the Global Offering, and based on the Offer Price of HK$116.20 per Share, we estimate that we will receive net proceeds of approximately HK$4,173.4 million from the Global Offering. We intend to use the proceeds from the Global Offering for the purposes and in the amounts set forth below:
• Approximately 70.0% (or HK$2,921.4 million) will be used to continuously strengthen our research and development capabilities in general-purpose large AI models;
• Approximately 10.0% (or HK$417.3 million) will be used to continuously optimize our MaaS platform by offering the latest foundation models and training/inference tools and infrastructures;
• Approximately 10.0% (or HK$417.3 million) will be used for the development of our business partner network, as well as for strategic investments; and
• Approximately 10.0% (or HK$417.3 million) will be used for working capital and other general corporate purposes.
We did not declare or pay any dividend during the Track Record Period. We do not currently have a formal dividend policy or a fixed dividend payout ratio. We currently intend to retain all available funds and earnings, if any, to fund the development and expansion of our business and we do not anticipate paying any cash dividends in the foreseeable future. Investors should not purchase our ordinary shares with the expectation of receiving cash dividends. Any future determination to pay dividends will be made at the discretion of our Directors and may be based on a number of factors, including our future operations and earnings, capital requirements and surplus, general financial condition, contractual restrictions and other factors that our Directors may deem relevant. Regulations in the PRC currently permit payment of dividends of a PRC company only out of accumulated distributable after-tax profits less any recovery of accumulated losses and appropriations to statutory and other reserves that we are required to make, as determined in accordance with its articles of association and the accounting standards and regulations in China. As advised – 21 –
根据我们的中国法律顾问的意见，考虑到上述情况，由于我们存在累计亏损，我们在某一年度可能没有足够的可分配利润向股东派发股息，或即使我们实现盈利，我们也只能在以下条件满足后方可从可分配利润中宣派或支付股息：(i) 累计亏损已由税后利润弥补，以及 (ii) 已按照相关法律、法规及本公司章程文件提取足额的法定储备及其他储备。鉴于本招股说明书所披露的累计亏损情况，我们在可预见的未来不太可能有资格从利润中支付股息。
上市费用是指与上市及全球发售相关的专业费用、包销佣金及其他费用。我们于2022年、2023年、2024年及截至2025年6月30日止六个月分别录得上市费用人民币零元、零元、零元及人民币1,770万元。我们的上市费用估计约为人民币15,850万元（港币17,470万元）（含包销佣金），占全球发售所得款项总额的4.0%（按每股发售价港币116.20元及超额配售权未获行使计算）。其中，约人民币12,140万元（港币13,380万元）直接归属于股份的发行，将于上市完成时计入权益；约人民币3,710万元（港币4,090万元）将计入我们的综合损益表。我们预计将产生的上市费用包括：约人民币11,830万元（港币13,040万元）的包销相关费用及开支，以及约人民币4,020万元（港币4,430万元）的非包销相关费用及开支，后者包括法律顾问及申报会计师的费用及开支约人民币2,180万元（港币2,400万元），以及其他费用及开支约人民币1,840万元（港币2,030万元）。
2025年1月16日，本公司及其九家子公司被列入由美国工业和安全局（BIS）管理的实体清单。此次列入限制了我们在未获BIS许可的情况下购买或以其他方式获取受《出口管理条例》（EAR）约束的特定货物、软件及技术的能力。然而，由于我们在业绩记录期间并未依赖任何受EAR约束的项目，该事件对我们的业务及财务表现未产生任何重大不利影响，且我们的董事认为（假设EAR限制范围或被列入实体清单的主体范围不会扩大），该事件在近期亦不会对我们的业务及财务表现产生任何重大不利影响。鉴于上述情况，我们的国际制裁法律顾问建议，本公司被列入实体清单一事不会对本集团的业务及财务状况造成任何重大不利影响，亦不会在将来造成此类影响。详情请参阅"业务——美国出口管制法律及法规"。
2025年7月及8月，我们发布了GLM-4.5、GLM-4.5V及AutoGLM的更新版本（亦称"AutoGLM 2.0"）。GLM-4.5是我们的旗舰基础模型。通过多阶段训练以及结合微调与强化学习的全面后训练，GLM-4.5在智能体任务、推理任务及代码编写任务方面均取得了优异表现。GLM-4.5V是我们的基础视觉语言模型（VLM），专为通用视觉理解与推理而设计，能够自主完成各类高度复杂的视觉理解与推理任务。AutoGLM更新版本由GLM-4.5及GLM-4.5V驱动，能够在更广泛的移动应用程序及网站上模拟人类操作。该模型可在云端自主完成用户指定的任务，无需占用用户的手机或电脑，用户可持续正常使用自己的设备而不受干扰。2025年9月，我们发布了GLM-4.6，这是我们基础模型的进一步更新版本，主要特点在于增强了代码编写能力。详情请参阅"业务——我们的模型"及"业务——我们的AI智能体"。
SUMMARY For the nine months ended September 30, 2025, we had over 12,000 institutional customers, representing a substantial increase compared to the six months ended June 30, 2025. In addition, our average daily token consumption volume was 4.2 trillion in November 2025.
Based on our unaudited management accounts, we estimate our revenue in the three months ended September 30, 2025 to increase by over 60% compared to the same period in 2024. We expect to record a significant increase in net loss for 2025, as we expect to incur substantial amount of research and development expenses in this year and be affected by changes in the carrying amounts of financial instruments issued to investors.
The Directors confirm that, up to the date of this prospectus, there has been no material adverse change in our financial or trading position since June 30, 2025, and there is no event since June 30, 2025 which would materially affect the information shown in the Accountants' Report in Appendix I to this prospectus.
The outbreak of the COVID-19 pandemic in 2020 disrupted normal life and daily routines worldwide, prompting governments to implement a range of restrictive measures to curb the outbreak. In particular, the PRC government imposed stringent measures to combat the spread of the pandemic, including travel restrictions, mandatory suspensions of business operations, quarantine requirements, remote and alternative working arrangements, limits on social and public gatherings and lockdowns of cities or regions. In May 2023, the World Health Organization declared that the COVID-19 pandemic is no longer a global health emergency, and the PRC government began to ease restrictions and quarantine measures in China as the pandemic came under control. We are an AI company primarily focused on the development of AI models and agents. Our business operations, including research and development activities, were not materially and adversely affected by the COVID-19 pandemic during the Track Record Period. Our revenue increased significantly by 116.9%, from RMB57.4 million in 2022 to RMB124.5 million in 2023. Based on the foregoing, our Directors are of the view that the COVID-19 pandemic did not, and is not expected to, have any material adverse impact on our business, financial condition or results of operations during the Track Record Period and up to the Latest Practicable Date.
We submitted a filing to the CSRC for application of the Listing and the Global Offering on July 3, 2025. The CSRC confirmed our completion of filing on December 15, 2025.
In this prospectus, unless the context otherwise requires, the following terms shall have the meanings set out below. Certain other terms are explained in "Glossary of Technical Terms" of this prospectus.
"Accountants' Report" | the accountants' report for the Track Record Period prepared by KPMG, the text of which is set out in Appendix I to this prospectus;
"Amended Concert Party Agreement" | the concert party agreement entered into by Beijing Lianpai (北京鏈湃科技發展中心(有限合夥)), Dr. Liu, Dr. Tang, Dr. Li, Dr. Xu, Dr. Zhang, Huihui and Zhideng dated April 5, 2023;
"Articles of Association" or "Articles" | the articles of association of our Company adopted on June 28, 2025, which shall become effective as of the date on which the H Shares are listed on the Stock Exchange, as amended from time to time, a summary of which is set out in "Appendix V—Summary of Articles of Association" to this prospectus;
"Beijing Knowledge Future" | Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), a limited liability company established under the laws of the PRC on March 15, 2024, and a wholly owned subsidiary of our Company;
"Beijing Knowledge Haiying" | Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), a limited liability company established under the laws of the PRC on November 6, 2024, and a wholly owned subsidiary of our Company;
"Beijing Knowledge Huixing" | Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), a limited liability company established under the laws of the PRC on October 29, 2024, and a non-wholly owned subsidiary of our Company;
"Beijing Knowledge Linghang" | Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), a limited liability company established under the laws of the PRC on May 22, 2024, and a wholly owned subsidiary of our Company;
"Beijing Knowledge Qingyan" | Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), a limited liability company established under the laws of the PRC on December 5, 2023, and a wholly owned subsidiary of our Company;
"Beijing Knowledge Qingying" | Beijing Knowledge Qingying Culture Media Co., Ltd. (北京智譜清影科技文化傳媒有限公司), a limited liability company established under the laws of the PRC on March 13, 2025, and a wholly owned subsidiary of our Company;
"Beijing Knowledge Xingyao" | Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), a limited liability company established under the laws of the PRC on September 24, 2024, and a wholly owned subsidiary of our Company;
"Beijing Lingxin Intelligent" | Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), a limited liability company established under the laws of the PRC on November 19, 2021, and a wholly owned subsidiary of our Company;
"Beijing Lianpai" | Beijing Lianpai Technology Development Center (Limited Partnership) (北京鏈湃科技發展中心(有限合夥)), formerly known as Beijing Kaiaigeer Technology Development Center (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), a limited partnership established in the PRC on May 10, 2019, and one of the Controlling Shareholders;
成都智谱华章科技有限公司（Chengdu Knowledge Atlas Technology Co., Ltd.），一家于2024年12月27日依据中华人民共和国法律设立的有限责任公司，为本公司的全资子公司；
北京智谱华章科技股份有限公司（Knowledge Atlas Technology Joint Stock Company Limited），一家于2019年6月11日依据中华人民共和国法律设立的有限责任公司，并于2025年3月26日转制为股份有限公司；
具有《上市规则》赋予该词的含义，除非上下文另有要求，统称北京联湃、刘德兵博士、唐杰博士、李涓子博士、许斌博士、张鹏博士、汇汇及智登，"控股股东"指其中任何一方，详情载于本招股说明书"与控股股东的关系"一节；
于全球发售完成后，现有股东持有的合计178,282,205股非上市股份（于股份分拆后）转换为H股。上述非上市股份转换为H股事宜已于2025年7月3日向中国证监会备案，并已向联交所上市委员会申请H股上市。中国证监会于2025年12月15日出具备案通知；
The Guide for New Listing Applicants, as published by the Stock Exchange on November 29, 2023 and effective on January 1, 2024, as amended or supplemented or otherwise modified from time to time;
Shares of our Company for which an application has been made for listing and permission to trade on the Stock Exchange;
Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), a limited liability company established under the laws of the PRC on September 20, 2023, and a wholly owned subsidiary of our Company;
the application for Hong Kong Offer Shares to be issued in the applicant's own name by submitting applications online through the designated website at www.hkeipo.hk;
the HK eIPO White Form service provider designated by our Company as specified on the designated website at www.hkeipo.hk; – 27 –
the application for the Hong Kong Offer Shares to be issued in the name of HKSCC Nominees and deposited directly into CCASS to be credited to your or a designated HKSCC Participant's stock account through causing HKSCC Nominees to apply on your behalf, including by instructing your broker or custodian who is a HKSCC Participant to give electronic application instructions via HKSCC's FINI system to apply for the Hong Kong Offer Shares on your behalf;
the Operational Procedures of HKSCC in relation to CCASS, containing the practices, procedures and administrative requirements relating to operations and functions of CCASS, from time to time in force;
a participant admitted to participate in CCASS as a direct clearing participant, a general clearing participant or a custodian participant;
the 1,871,000 new H Shares initially being offered by our Company for subscription at the Offer Price pursuant to the Hong Kong Public Offering, subject to reallocation as described in "Structure of the Global Offering" of this prospectus;
the offer for subscription of the Hong Kong Offer Shares to the public in Hong Kong at the Offer Price (plus brokerage of 1.0%, SFC transaction levy of 0.0027%, Stock Exchange trading fee of 0.00565% and AFRC transaction levy of 0.00015%), subject to and in accordance with the terms and conditions set out in this prospectus;
the underwriters of the Hong Kong Public Offering whose names are set out in "Underwriting—Hong Kong Underwriters" of this prospectus;
the underwriting agreement dated December 29, 2025 relating to the Hong Kong Public Offering entered into by our Company, Dr. Liu, the Sponsor-Overall Coordinator and the Hong Kong Underwriters;
Huangshi Knowledge Atlas Technology Co., Ltd. (黄石智譜華章科技有限公司), a limited liability company established under the laws of the PRC on November 11, 2025, and a wholly owned subsidiary of our Company;
Zhuhai Hengqin Huihui Enterprise Management Partnership (Limited Partnership) (珠海橫琴慧惠企業管理合夥企業（有限合夥）), formerly known as Ningbo Huihui Enterprise Management – 28 –
Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業（有限合夥）), a limited partnership established in the PRC on June 23, 2021, one of our Employee Ownership Platforms and one of the Controlling Shareholders;
individuals or company(ies), who or which, to the best of our Directors' knowledge, information and belief, having made all reasonable enquiries, is not a connected person of our Company within the meaning of the Listing Rules;
the 35,548,500 H Shares initially being offered for subscription under the International Offering, together with, where relevant, any additional H Shares which may be issued by our Company pursuant to the exercise of the Over-allotment Option, and subject to reallocation as described in "Structure of the Global Offering" of this prospectus;
the offer of the International Offer Shares at the Offer Price outside the United States to persons that are not, and are not acting for the account or benefit of, U.S. Investors in offshore transactions in reliance on Regulation S and subject to the terms and conditions of the International Underwriting Agreement, as further described in "Structure of the Global Offering" of this prospectus;
all applicable laws and regulation to economic sanctions, export controls, trade embargoes and wider prohibitions and restrictions on international trade and investment related activities, including those adopted administered and enforced by the U.S. Government, the EU and its member states, the UK, UN or Government of Australia and other competent government authorities;
King & Wood Mallesons, our legal advisers as to International Sanctions laws in connection with the Listing;
the group of international underwriters expected to enter into the International Underwriting Agreement relating to the International Offering;
the international underwriting agreement relating to the International Offering to be entered into by our Company, Dr. Liu, the Sponsor-Overall Coordinator and the International Underwriters on or about January 6, 2026;
the period comprising the three years ended December 31, 2024 and the six months ended June 30, 2025;
Wuhan Knowledge Atlas Technology Co., Ltd. (武汉智谱华章科技有限公司), a limited liability company established under the laws of the PRC;
Beijing Zhipu Huazhang Technology Co., Ltd. (北京智谱华章科技有限公司), a joint stock limited company incorporated under the laws of the PRC;
Zhipu AI Technology (Hong Kong) Limited (智谱人工智能科技（香港）有限公司), a company incorporated under the laws of Hong Kong;
Jincheng Yaoda Technology Limited, a company incorporated under the laws of the British Virgin Islands on January 6, 2022, and a wholly owned subsidiary of our Company;
JINGSHENG HENGXING TECHNOLOGY PTE. LTD., a company incorporated under the laws of Singapore with limited liability on November 23, 2023, and an indirect wholly owned subsidiary of our Company;
the joint bookrunners as named in "Directors, Supervisor and Parties involved in the Global Offering";
the joint global coordinators as named in "Directors, Supervisor and Parties involved in the Global Offering";
the joint lead managers as named in "Directors, Supervisor and Parties involved in the Global Offering";
December 22, 2025, being the latest practicable date for the purpose of ascertaining certain information contained in this prospectus prior to its publication;
the date, expected to be on or about Thursday, January 8, 2026, on which dealings in our H Shares first commence on the Main Board;
the Rules Governing the Listing of Securities on The Stock Exchange of Hong Kong Limited, as amended or supplemented or otherwise modified from time to time;
the stock exchange (excluding the option market) operated by the Stock Exchange which is independent from and operated in parallel with the GEM of the Stock Exchange;
Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), a limited liability company established under the laws of the PRC on April 19, 2013, and a wholly owned subsidiary of our Company;
the offer price per Offer Share (exclusive of brokerage fee of 1.0%, SFC transaction levy of 0.0027%, Stock Exchange trading fee of 0.00565% and AFRC transaction levy of 0.00015%) of HK$116.20 at which Hong Kong Offer Shares are to be subscribed in the manner further described in "Structure of the Global Offering" in this prospectus;
China International Capital Corporation Hong Kong Securities Limited, Huatai Financial Holdings (Hong Kong) Limited, BOCOM International Securities Limited, Guotai Junan Securities (Hong Kong) Limited and China Merchants Securities (HK) Co., Limited;
the option expected to be granted by our Company to the International Underwriters, exercisable by the Sponsor-Overall Coordinators (for and on behalf of the International Underwriters) pursuant to the International Underwriting Agreement, pursuant to
which our Company may be required to allot and issue up to an aggregate of 5,612,900 additional H Shares at the Offer Price to, among other things, cover over-allocations in the International Placing, if any, further details of which are described in "Structure of the Global Offering";
the pre-IPO investment(s) in our Company, details of which are set out in "History, Development and Corporate Structure—Pre-IPO Investments" in this prospectus;
Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), a limited liability company established under the laws of the PRC on May 14, 2024, and a wholly owned subsidiary of our Company;
ordinary share(s) in the share capital of our Company with a nominal value of RMB0.10 each upon the completion of the Share Subdivision; before the completion of the Share Subdivision, ordinary share(s) in the share capital of our Company with a nominal value of RMB1.00 each;
the sub-division of the Shares by the Company where the Company subdivided its Shares from one Share of RMB1.00 each into ten Shares of RMB0.10 each, which will become effective immediately prior to the Listing;
Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), a limited liability company established under the laws of the PRC on July 21, 2021, and a wholly owned subsidiary of our Company;
Shenzhen Knowledge Lingxin Intelligent Technology Co., Ltd. (深圳智譜聆心智能科技有限公司), a limited liability company established under the laws of the PRC on August 6, 2025, and a wholly owned subsidiary of our Company;
(i) 美国人士；或 (ii) 《美国联邦法规》第31篇第850.229条所定义的美国人士，即实施第14105号行政命令并于2025年1月生效的最终规则，包括任何美国公民、合法永久居民、依据美国或美国任何司法管辖区法律组建的实体（包括境外分支机构），或任何在美国境内的人士；
香港祥泰瑞丰科技有限公司（Hong Kong Xiangtai Ruifeng Technology Limited），于2023年9月21日依据香港法律注册成立的有限责任公司，为本公司的间接全资附属公司；
珠海横琴智登企业管理合伙企业（有限合伙）（Zhuhai Hengqin Zhideng Enterprise Management Partnership (Limited Partnership)），前称宁波智登企业管理合伙企业（有限合伙）（Ningbo Zhideng Enterprise Management Partnership (Limited Partnership)），于2021年6月23日在中国设立的有限合伙企业，为本公司员工持股平台之一及控股股东之一；
浙江智谱新篇科技有限公司（Zhejiang Knowledge Xinpian Technology Co., Ltd.），于2025年2月24日依据中国法律设立的有限责任公司，为本公司的全资附属公司；
珠海智谱未来科技有限公司（Zhuhai Knowledge Future Technology Co., Ltd.），于2024年12月18日依据中国法律设立的有限责任公司，为本公司的全资附属公司；及
珠海智谱领航科技有限公司（Zhuhai Knowledge Linghang Technology Co., Ltd.），于2024年12月16日依据中国法律设立的有限责任公司，为本公司的全资附属公司。
增值税（VAT）。
为便于查阅，本招股说明书以中英文两种语言列载中国法律法规、政府机构、机构、自然人或其他实体（包括本公司若干附属公司）的名称，如有任何不一致，以中文版本为准。公司名称及其他术语的英文译名仅供识别之用。
DEFINITIONS AND ACRONYMS Certain amounts and percentage figures included in this prospectus were subjected to rounding adjustments. Accordingly, figures shown as totals in certain tables may not be arithmetic aggregation of the figures preceding them. For the purpose of this prospectus, references to "provinces" of China include provinces, municipalities under direct administration of the central government and provincial-level autonomous regions.
GLOSSARY OF TECHNICAL TERMS In this prospectus, unless the context otherwise requires, explanations and definitions of certain terms used in this prospectus in connection with our Group and our business shall have the meanings set out below. The terms and their meanings may not correspond to standard industry meaning or usage of these terms. "AGI"
一种具有方向和大小的多维数学对象，用于表示数据特征。
本招股说明书载有前瞻性陈述。凡非属历史事实的陈述，包括有关我们对未来的意图、信念、预期或预测的陈述，均为前瞻性陈述。
本招股说明书载有若干与本公司及我们附属公司相关的前瞻性陈述及信息，该等陈述及信息以我们管理层的判断以及管理层所作的假设及管理层目前可获得的信息为基础。在本招股说明书中，当"目标"、"预期"、"相信"、"可能"、"预计"、"展望未来"、"拟"、"或许"、"应当"、"计划"、"预测"、"寻求"、"应该"、"将会"、"希望"等词语及其否定形式和其他类似表述与本集团或我们的管理层相关联时，均旨在识别前瞻性陈述。该等陈述反映了我们管理层就未来事件、运营、流动资金及资本资源的当前看法，其中部分内容可能无法实现或可能发生变化。该等陈述须受若干风险、不确定因素及假设的影响，包括本招股说明书所述的其他风险因素。敬请注意，依赖任何前瞻性陈述涉及已知及未知的风险与不确定性。可能影响前瞻性陈述准确性的风险与不确定因素包括但不限于以下各项：
在符合适用法律、规则及法规要求的前提下，我们不承担任何义务，亦不承诺对本招股说明书中的前瞻性陈述进行更新或以其他方式修订，无论是否因新信息、未来事件或其他原因所致。由于上述及其他风险、不确定因素及假设，本招股说明书中所讨论的前瞻性事件及情况可能不会按照我们预期的方式发生，或根本不会发生。因此，您不应过分依赖任何前瞻性信息。本招股说明书中所有前瞻性陈述均受本节警示性陈述的限定。
在本招股说明书中，有关我们或董事意图的陈述或提述均以本招股说明书日期为准。该等信息可能随未来发展情况而有所变化。
You should carefully consider all of the information in this prospectus, including the risks and uncertainties described below, before making an investment in our H Shares. The following is a description of what we consider to be our material risks. Any of the following risks could have a material adverse effect on our business, results of operations, financial condition and prospects. In any such case, the market price of our H Shares could decline, and you may lose all or part of your investment.
These factors are contingencies that may or may not occur, and we are not in a position to express a view on the likelihood of any such contingency occurring. The information given is as of the Latest Practicable Date unless otherwise stated, will not be updated after the date hereof, and is subject to the cautionary statements in the section headed "Forward-Looking Statements" in this prospectus.
We believe there are certain risks and uncertainties involved in our operations, some of which are beyond our control. We have categorized these risks and uncertainties into: (i) risks related to our research and development, (ii) risks related to our commercialization, (iii) risks related to our operations, (iv) risks related to our intellectual property, (v) risks related to our financial condition and need for additional capital, (vi) risks related to the jurisdictions where we conduct business, and (vii) risks related to the Global Offering.
Additional risks and uncertainties that are presently not known to us or not expressed or implied below or that we currently deem immaterial could also harm our business, results of operations, financial condition and prospects. You should consider our business and prospects in light of the challenges we face, including those discussed in this section.
The AI industry is characterized by constant changes. If we are not able to upgrade, enhance or innovate our technologies and services, our business, results of operations, financial condition and prospects could be adversely affected.
The AI industry in general, and LLM market we focus on in particular, are characterized by constant changes. We expect that new technologies and modes of applications will continue to emerge and evolve. Rapid and significant technological breakthroughs continue to shape the AI industry, including developments of large language models, multimodal integration and AI agents. Thus, our future business, results of operations, financial condition and competitive position depend on our ability to develop and introduce new and enhanced models that incorporate and integrate the latest technological advancements and our ability to commercialize our technology in ways that satisfy evolving customer demands. Our peers, especially non-independent providers who already have an established portfolio of products and services before entering the LLM market, may have or have invested more resources in developing new technologies or have more experience in commercializing their LLM offerings. We may encounter significant unexpected technological challenges, or delays in developing new and enhanced models, which require us to invest significant resources in R&D and also require that we:
• design innovative, accurate and efficiency-enhanced features and functions that differentiate our services from those of our competitors;
• respond effectively to technological changes and new product and solution announcements by our competitors; and
adjust to changing customer preferences, market conditions and regulatory landscape quickly and cost-effectively.
Considering the rapid advancement of technology, we may be unable to upgrade our technologies promptly, efficiently or cost-effectively, or at all. In addition, the emergence of new or alternative technologies could replace or reduce the demand for our models and solutions, render our technologies obsolete or unattractive. As a result, our business, results of operations, financial condition and prospects may be materially and adversely affected.
We have made and expect to continue to make substantial investments in R&D. If we cannot continuously invest in our R&D activities while achieving technological innovation, our business, results of operations, financial condition and prospects may be materially and adversely affected.
R&D is crucial to our business and operations. However, our AI technology, or the AI industry as a whole, is still at a development stage and is susceptible to significant uncertainties. We have been investing heavily in our R&D efforts with a focus on elevating the intelligence of our foundation model. Our research and development expenses were RMB84.4 million, RMB528.9 million, RMB2,195.4 million, RMB859.2 million and RMB1,594.7 million in 2022, 2023 and 2024 and the six months ended June 2024 and 2025, respectively, representing 147.0%, 424.7%, 702.7%, 1,913.2% and 835.4% of our total revenue in the respective periods. The AI industry is subject to rapid technological changes and is evolving quickly in terms of technological innovation. We must invest substantial resources in R&D to advance our technology, expand our offerings and ensure that our models and solutions remain innovative and competitive. As a result, we may continue to incur significant R&D expenses in the future. However, R&D activities are inherently uncertain. We cannot guarantee that we will be able to continue invest significantly in R&D activities or our R&D efforts, especially our investments in the development of AGI, will yield anticipated benefits or recognition. Through the five stages of LLM, we have developed large models and agents across the first three stages. However, we cannot guarantee that we will be able to reach the self-perception stage and consciousness stage. Even if we succeed in our R&D efforts and generate the results we expect, we may still encounter practical difficulties in commercializing our development results. New technologies or new approaches to known technologies could render our existing technologies and solutions that we are developing or expect to develop in the future unattractive, expensive or even obsolete, thereby limiting our ability to recover related development costs, which could result in a decline in our revenues, profitability and market share.
The development of AGI is still at an early stage and there are substantial uncertainties in the future realization of AGI.
It is widely recognized within the industry that continued advances in AI research will ultimately result in the realization of AGI, which will significantly expand the AI market and generate substantial benefits for industry participants. We share this belief and, since our inception, have been dedicated to advancing our model performance and investing heavily in our R&D activities in pursuit of AGI. However, we cannot assure you that we, or any of our competitors, will ultimately achieve AGI. We believe the industry as a whole is still at an early stage of development, and further fundamental breakthroughs may prove exceptionally difficult or unattainable regardless of collective or individual effort. If AGI is not technically achievable, or if meaningful progress toward AGI slows or stalls, the commercial benefits we achieved may not justify the time, capital and resources we have invested. As a result, our business, results of operations, financial condition and prospects could be adversely affected.
We are exposed to risks relating to our R&D team and our senior management.
As a leader in the development of general-purpose AI in China, we are fundamentally a company comprised of data scientists and engineers. Our success depends on the continued service of our technology leadership and R&D team and senior management. According to Frost & Sullivan, the LLMs market in
China has a high demand for talent, especially for experts with deep technical backgrounds and extensive experience. Leading industry players have attracted top-tier talent and built strong technical teams, while new entrants face intense competition in acquiring skilled professionals. Our core scientists have significant industry experience, and their knowledge and relationships would be difficult to replace. See "Business—Research and Development—Talent." If one or more of our core scientists were unable or unwilling to continue to contribute their services to us, we may not be able to replace them in a timely manner, or at all. In addition, the status and reputation of our core scientists and senior management in the AI industry and the academic community could directly or indirectly affect us. As a result, our business may be severely disrupted, and our financial condition and results of operations may be materially and adversely affected.
In addition, we rely on our R&D team to support our rigorous R&D process. Our R&D team consists of members with background and experience in fields such as natural language processing, advanced decision-making in complex systems and multimodal semantic analysis. The competition for these highly skilled and qualified employees in our industry is increasingly intense. To help attract, retain and motivate key individuals, employee incentives such as share incentive schemes have been, and will continue to be, an important part of their compensation. Our employee hiring and retention also depend on our ability to build and maintain a diverse and inclusive working environment and be recognized as an attractive employer and desirable workplace. If our share-based payment expenses or other compensation programs and workplace culture cease to be viewed as competitive, our ability to attract, retain and motivate key individuals would be weakened, which would in turn materially and adversely affect our business, financial condition and prospects.
We rely on third parties to provide computing resources to us, and any disruption of their services or fluctuation of prices could adversely affect our business, results of operations and financial condition.
We are highly dependent on high-performance and cost-effective computing resources, such as computing hardware and computing services, for R&D activities and business operation. Computing power forms the foundation for training and inference of AI models and is essential for scaling and commercializing AI technologies. As models become increasingly complex and model parameters grow exponentially, the demand for computational resources continues to rise. In 2022, 2023, 2024 and the six months ended June 30, 2024 and 2025, our computing service fees amounted to RMB14.6 million, RMB311.7 million, RMB1,552.8 million, RMB603.2 million and RMB1,145.1 million, accounting for 17.3%, 58.9%, 70.7%, 70.2% and 71.8% of our research and development expenses, respectively. Such resources may be subject to price fluctuations, shortages or changes in the terms of service, including the discontinuation of access to core technologies. Any such events could increase our operating costs and expenses, limit the scale of the deployment of our solutions, or otherwise adversely affect our business. We may not be able to obtain adequate replacements on a timely basis, or at all, and we may be forced to purchase computing power at premium prices, which could have a material adverse effect on our business, financial condition and results of operations.
In addition, any limitation on the capacity of our third-party suppliers could impede our ability to deliver solutions and conduct R&D activities in a timely manner, which could adversely affect our business, financial condition and results of operations. Any incident affecting our computing power suppliers' infrastructure that may be caused by cyber-attacks, natural disasters, fire, flood, severe storm, earthquake, power loss, telecommunications failures, terrorist or other attacks and other similar events beyond our control could negatively affect our services. In addition, the continued operations of computing power suppliers depend on highly specialized and regulated global supply chains. Therefore, geopolitical conflicts and changes of the relevant laws and regulations, including export controls, economic sanctions and trade tariffs, may affect the price and services provided by our computing power suppliers, which could negatively affect our business, financial condition and results of operations. A prolonged service disruption affecting our services for any of the foregoing reasons would adversely impact our ability to serve our customers, damage our reputation, expose us to liability, cause us to lose customers or harm our business. We may also incur significant costs for using alternative resources or taking other actions in preparation for, or in reaction to, events that damage the third-party services we use.
Our commercial success depends on the performance of our models. Any failure in research and development efforts to offer high-quality models and solutions could harm our business, results of operations, financial condition and prospects.
Our models are complex and deployed across a wide range of application scenarios. The performance of our models depend on various factors, including our legitimate usage of open-source software, quantity and quality of data used, adequacy of training and others. If the data used are inaccurate, incomplete, unrepresentative or biased, or if our algorithms are flawed, incorrectly designed or calibrated, our models may generate inaccurate, sub-optimal, unfair or otherwise inappropriate outputs, including in relation to matching results. This may adversely affect user experience and outcomes, and may give rise to complaints, negative publicity or increased scrutiny from regulators and other stakeholders. In addition, the correct and proper configuration of our models and the proper provision of implementation, analytical and maintenance services may influence customer experience. Any defects, errors, instability, security vulnerabilities or unintended bias in our models, or any failure to deploy, configure or support our models properly, may result in contract terminations or non-renewals, reduced customer payments, negative publicity or legal claims against us.
In addition, inability to meet specific customer demands may result in customer dissatisfaction or reputational damage, which could materially harm our business. As our business and customer base expand, it is essential to maintain efficient R&D efforts to meet customer demands on a larger scale. However, there is no assurance that we will be able to recruit or retain sufficient qualified personnel with experience in developing and deploying our models.
We may not be able to compete effectively against current or future competitors.
While the LLM market is still at an early stage of development, it is already highly competitive and is expected to become increasingly competitive. We currently face intense competition from our competitors and may face even greater competition in the future. As the development of large models becomes increasingly resource-intensive, both in China and globally, we expect the industry to consolidate rapidly, with development efforts concentrated among a small group of leading players. Our competitors may have better financial, R&D or marketing resources, stronger brand recognition, better support from upstream and downstream business partners, stronger ability to seek business partners to secure better pricing on computing resource service, more advanced models in terms of scope and performance, or the capability to expand customer bases more quickly than we do. As a result, our competitors may be able to respond more quickly and effectively to new or changing opportunities, technologies, standards or customer requirements than us and may have the ability to initiate or withstand significant regulatory changes and industry evolvement. Market competition may also result in continued pricing pressures, which may lead to price reductions in certain of our solutions, which may, in turn, materially and adversely affect our profitability and market share.
In addition, new competitors or alliances may emerge with larger market share, broader customer bases, more widely adopted proprietary technologies, greater marketing expertise, stronger financial resources and larger sales forces than us. In light of these factors, even if our models are more advanced, effective and cost-efficient than those of our competitors, current or potential customers may accept competitors' models in lieu of ours. If we are unable to successfully compete in the market, our business, results of operations, financial condition and prospects may be materially and adversely affected.
The size of our addressable market and the demand for our solutions may not increase as rapidly as we anticipate due to a variety of factors, which would materially and adversely affect our business, results of operations, financial condition and prospects.
We are pursuing opportunities in a market that is undergoing rapid changes, including technological and regulatory changes, and it is difficult to predict the timing and scale of such opportunities. If the LLM – 43 –
market experience a shift in customer or prospective customer demand, our solutions may not compete as effectively, or at all, and they may not be deployed by customers. Given the evolving nature of the LLM market, it is difficult to predict customer demand for our solutions or the future market growth. The addressable market for our solutions may be smaller than we have estimated, our future growth opportunities and sales growth may be smaller than we estimate, and our future business, results of operations and financial condition may be materially and adversely affected. Even if the LLM market grows substantially, there is no guarantee that demand for our solutions will correlate with that growth if we fail to effectively pursue such opportunities. There is also no guarantee that our business will be successful simply because of the future or trends of the addressable market of our solutions. If demand does not develop or if we cannot accurately forecast customer demand, our future business, results of operations and financial condition would be materially and adversely affected.
Any failure of our MaaS platform to perform as required could harm our business, results of operations, financial condition and prospects.
Our success depends on the quality, reliability and efficiency of our MaaS platform. Any deficiencies could lead to customer dissatisfaction, loss of business opportunities or harm to our reputation. In particular, the successful performance of our MaaS platform depends, among other things, on:
• our ability to keep abreast of technology and industry trends and continue to advance our platform.
There can be no assurance that we will successfully address any of these or other factors that may affect the market acceptance of our platform. If we are unsuccessful in achieving and maintaining market acceptance of our MaaS platform, our business, financial condition, results of operations and prospects could be adversely affected.
If we fail to retain existing customers, attract new customers or increase the spending by existing customers, our business, results of operations, financial condition and prospects may be materially and adversely affected.
We have been expanding our customer base to cover various industries. Our abilities to retain existing customers, attract new customers, as well as increase the spending by existing customers depend on a number of factors, including our ability to offer effective solutions that address the evolving needs of our customers at competitive prices, the strength of our technologies and the effectiveness of our sales and marketing efforts. If we fail to retain existing customers or attract new customers, we may not be able to grow our revenue as quickly as we anticipate, or at all. As our customer base grows and diversifies into other sectors, we may be unable to provide customers with solutions that meet the specific demand of such customers, and we may be unable to provide quality customer support, which could result in customer
dissatisfaction, decreased overall demand for our solutions and loss of expected revenue. In addition, our inability to meet customer expectations may damage our reputation and could consequently limit our ability to retain existing customers and attract new customers, which would materially and adversely affect our business, results of operations, financial condition and prospects.
We have a limited track record in the commercialization of our business.
We have a limited track record in launching, sales and marketing and commercialization of our large models. Our ability to successfully commercialize our business may involve inherent risks, longer timelines and higher costs as the consumer market for large models in China is still at its early stage of development. We may incur additional marketing expenses and resources on enhancing the market acceptance of our technologies and solutions. In addition, the success of our sales and marketing efforts depends on our ability to attract, motivate and retain qualified and professional employees in our commercialization team who have, among other things, adequate AI knowledge to communicate effectively with AI professionals, sufficient experience in sales and marketing in the AI industry, and extensive industry resources. Furthermore, competition for experienced sales and marketing personnel is intense. If we are unable to attract, motivate and retain a sufficient number of qualified sales and marketing personnel to support our business, our commercialization of our business may be adversely affected.
Due to our limited track record in commercializing our business, we cannot guarantee that our efforts to promote customer adoption of our models will succeed, that our sales results will meet our forecast, that third parties will deploy and operate our models effectively and meet overall user experience, or that we will be able to fully maintain quality control over our models, which may materially and adversely affect the mass commercialization of our business, and, in turn, materially and adversely affect our business, results of operations, financial condition and prospects.
If our attempt to expand our offerings or business model is unsuccessful, our business, results of operations, financial condition and prospects may be materially and adversely affected.
We provide innovative AI-powered solutions designed to address diverse needs of our customers across different industries. However, expanding our offering categories into new industry verticals, or expanding our business model to involve new services, involves additional risks and challenges. Our lack of familiarity with new verticals or industries may make it more difficult for us to keep pace with evolving customer demands and preferences. In addition, there may be one or more existing market leaders in any vertical or industry that we decide to expand into. Such companies may be able to compete more effectively than us by leveraging their experience in doing business in that market as well as their deeper industry insight and greater brand recognition among customers. We may be required to develop new supply-chain relationships and capabilities. We will need to comply with laws and regulations specific to these industries. Expansion into any new industry sectors and development of new solutions may place significant strain on our management and resources and incur substantial R&D and other costs and expenses before generating any revenues. Failure to expand successfully could have a material adverse effect on our business, results of operations, financial condition and prospects.
If we are unable to ensure compatibility of our models with a variety of hardware, platforms and applications developed by others, including our partners, we may become less competitive and our business, results of operations, financial condition and prospects may be harmed.
Our models may be integrated with a variety of hardware, platforms and applications, and we need to modify and enhance our models to adapt to changes in hardware and software technologies in a timely and cost-effective manner. As of June 30, 2025, our models were compatible with over 40 major global chip platforms. Compatibility of our models with hardware, platforms and software developed by others is critical to the performance of our solutions. Failure to ensure compatibility of our models may negatively affect our competitive edge, and our business, results of operations, financial condition and prospects would be harmed.
RISK FACTORS RISKS RELATED TO OUR OPERATIONS The AI industry is subject to evolving and extensive regulation in China. Future laws and regulations may impose additional requirements and other obligations that could materially and adversely affect our business, results of operations, financial condition and prospects.
The AI industry in China is evolving and we may experience strengthened regulatory environment along with rapid industry evolution. Government authorities in China may continue to issue new laws, regulations and rules governing the AI industry. For example, on July 10, 2023, the CAC and six other ministries jointly issued the Interim Measures for the Administrative Measures on Generative Artificial Intelligence Services (《生成式人工智能服務管理暫行辦法》) (the "Measures on Generative AI Services"), which imposes compliance requirements for providers of generative AI services. The Measures on Generative AI Services require generative AI service providers to take effective measures to enhance the accuracy and reliability of the content created by generative AI. Generative AI service providers shall (i) assume the responsibilities of content producers and perform network information security obligations; (ii) assume the responsibilities of processors of personal information to protect personal information; and (iii) process training data such as conducting pre-training optimization in accordance with applicable laws and regulations. In addition, the providers of generative AI services with public opinion attributes or the capacity for social mobilization shall apply for security assessment and complete the filing formalities of algorithms in accordance with the Provisions on the Administration of Algorithm Recommendation for Internet Information Services (《互聯網信息服務算法推薦管理規定》). Moreover, the Administrative Provisions on Deep Synthesis in Internet-based Information Services (《互聯網信息服務深度合成管理規定》), jointly promulgated by the CAC and two other ministries on November 25, 2022, which became effective on January 10, 2023, impose certain compliance obligations upon service providers using deep synthesis technology to provide internet-based information services. See "Regulatory Overview—Regulations on Information Industry—Regulations on the Application of Artificial Intelligence Technologies." The interpretation and implementation of existing measures are evolving and the PRC regulatory agencies, including the CAC, may further adopt new laws, regulations, rules, or detailed implementation and interpretation related to the above-mentioned measures, which may negatively affect us. As such, we cannot assure you that our compliance measures are, and will be, always considered sufficient under applicable laws and regulations. If we are unable to comply with the then applicable laws and regulations, such actual and alleged failure could subject us to significant legal, financial and operational consequences.
We are subject to complex and evolving laws, regulations and governmental policies regarding cybersecurity, data security and personal information. Actual or alleged failure to comply with privacy and data protection laws, regulations and governmental policies could damage our reputation, deter current and potential customers from using our solutions and could subject us to significant legal, financial and operational consequences.
In recent years, cybersecurity, data protection and personal information protection have become an increasing regulatory focus of government authorities across the world. The PRC government has enacted a series of laws, regulations and governmental policies relating to cybersecurity, data protection and personal information protection in the past few years. For instance, on November 7, 2016, the Standing Committee of the National People's Congress promulgated the Cybersecurity Law of the People's Republic of China (《中華人民共和國網絡安全法》), effective since June 1, 2017. The Cybersecurity Law created the first national-level data protection framework for "network operators," which may potentially include all organizations in China that provide services over the internet or through other types of information network. On June 10, 2021, the Standing Committee of the National People's Congress promulgated the Data Security Law of the People's Republic of China (《中華人民共和國數據安全法》), effective since September 1, 2021. The Data Security Law sets out a number of obligations on data security and privacy undertaken by entities and individuals engaged in data-related activities. On August 20, 2021, the Standing Committee of the National People's Congress enacted the Personal Information Protection Law of the – 46 –
People's Republic of China (《中華人民共和國個人信息保護法》), which became effective on November 1, 2021. This law establishes principles and specific rules for personal information processing throughout its entire lifecycle, and imposes obligations on personal information processors. On September 24, 2024, the State Council promulgated the Regulation on Network Data Security Management (《網絡數據安全管理條例》), which came into effect on January 1, 2025 and further provides rules on network data security. See "Regulatory Overview—Regulations Relating to Internet Information Security and Privacy Protection—Regulations Relating to Cybersecurity and Internet Information Security."
The above regulatory developments relevant to cybersecurity, data protection and personal information protection could generally impact the data collection, use, storage and other data processing activities conducted by the enterprises in technology industry, including us. Failure or alleged failure to comply with these laws and regulations may expose us to potential legal liability, harm our reputation and brand and, consequently, our business. In addition, the laws and regulations regarding cybersecurity, data protection and personal information protection in China are generally complex and evolving, with uncertainty as to the interpretation and application thereof, which may lead to uncertainty about the scope of our responsibility in this regard. We may be unable to safeguard the data of our users due to factors beyond our control. We may source training data from third-party vendors, public websites, public datasets or other publicly accessible sources during the operation of our business. There is no guarantee as to the effectiveness of the measures we have taken to urge or supervise such third-party vendors or public data controllers to abide by applicable cybersecurity, data protection and personal information protection laws and regulations. If any of these entities fails, or is deemed to have failed, to obtain such data in a reasonable and lawful manner, or to comply with applicable cybersecurity and data privacy and protection laws and regulations, it could have a material adverse effect on our service as well as our reputation. Moreover, we cannot guarantee that we can effectively filter, desensitize, anonymize data obtained from third-party vendors and publicly accessible sources before further processing to comply with relevant laws and regulations due to the inherent limitations of current data processing technologies. As such, we cannot assure you that our cybersecurity, data protection and personal information protection measures are, and will be, always considered sufficient under applicable laws and regulations. Additionally, the effectiveness of our protection measures is also subject to system failure, interruption, inadequacy, security breaches or cyber attacks. If we are unable to comply with the then-applicable laws and regulations, or to address any cybersecurity, data protection and personal information protection concerns, such actual or alleged failure could damage our reputation, deter current and potential users from using our solutions and subject us to significant legal, financial and operational consequences.
In addition, on December 28, 2021, the CAC, the NDRC, the MIIT and several other administrations jointly promulgated the Measures for Cybersecurity Review (《網絡安全審查辦法》, the "CAC Measures"), effective on February 15, 2022, which provides that entities meeting certain standards shall be subject to a cybersecurity review. See "Regulatory Overview—Regulations Relating to Internet Information Security and Privacy Protection—Regulations Relating to Cybersecurity and Internet Information Security."
Although we are not obligated to apply for a cybersecurity review pursuant to the CAC Measures with respect to our proposed Global Offering, regulatory authorities may initiate cybersecurity review if they determine that our solutions or data processing activities have or could have influence on the national security. However, the CAC Measures provide no further explanation or interpretation for the criteria on determining the risks that "have or could have an influence on national security." In addition, since the interpretation and implementation of these laws and regulations with respect to cybersecurity review keep evolving, therefore, we cannot assure you that there will not be any additional regulatory requirements regarding cybersecurity review relating to new laws and regulations.
In accordance with the Regulations on Protection of Critical Information Infrastructure (《關鍵信息基礎設施安全保護條例》), which were promulgated by the State Council on July 30, 2021 and became effective on September 1, 2021, a critical information infrastructure (the "CII") refers to important network facilities or information systems in important industries or fields such as public communication and information service, energy, communications, water conservation, finance, public services, e-government
affairs and national defense science, which may endanger national security, people's livelihood and public interest in case of damage, function loss or data leakage. In addition, competent authorities and administration departments of each important industry and field shall be responsible for formulating determination rules and determining the critical information infrastructure operator (the "CIIO") in the respective critical industry or field. As of the Latest Practicable Date, we had not received any notification from relevant regulatory authorities regarding our identification as a critical information infrastructure operator. We cannot guarantee that authorities will not classify our MaaS platform as CII in the future. If deemed a CIIO in the future, where we procure network products or services and such procurement affects or may affect national security, we would be required to apply for a cybersecurity review with the Cybersecurity Review Office. In addition, if the relevant regulatory authorities determine that the network products or services used by us or our data processing activities affect or may affect national security, we may be subject to cybersecurity review. Any such review process could result in increased compliance costs, delays to our business operations or product deployment, and may have a material adverse effect on our business, financial condition and results of operations.
Further, the Measures on Security Assessment of Cross-border Data Transfer (《数据出境安全评估办法》), promulgated on July 7, 2022 and effective on September 1, 2022, and the Provisions on Promoting and Regulating Cross-border Data Flows (《促进和规范数据跨境流动规定》), promulgated and effective on March 22, 2024, provide that the transfer of personal information and important data by data handler meeting certain volume thresholds or other standards as provided therein shall apply for security assessment, file with a standard contract for cross-border data transfer or obtain a personal information protection certification. As our business continues to grow, there may be circumstances where we engage in such cross-border data transfers. In such cases, in order to satisfy the legal and regulatory requirements, we may need to comply with the foregoing requirements as well as any other limitations under PRC laws then applicable. Complying with these laws and requirements could cause us to incur substantial expenses or require us to alter or change our practices in ways that could harm our business.
We anticipate that new laws and regulations concerning cybersecurity, data protection and personal information protection will continue to be proposed and adopted. The impact of such future laws and regulations on our business cannot yet be fully determined. New laws and regulations, amendments to or re-interpretations of existing laws and regulations may require us to incur additional costs and restrict our business operations. In addition to the possibility of fines, lawsuits, regulatory investigations, public censure, other claims and penalties, significant costs for remediation and damage to our reputation, we could be materially and adversely affected if legislation or regulations are expanded to require changes in our data processing practices and policies or if the applicable legislation or regulations are interpreted or implemented in ways that negatively impact our business, results of operations, financial condition and prospects. Any inability to adequately address cybersecurity, data protection or personal information protection concerns, even if unfounded, or to comply with applicable laws, regulations, standards and other obligations relating to cybersecurity, data protection or personal information protection could require significant resources and efforts, which may have a material effect on our business, results of operations, financial condition and prospects.
We are subject to risks associated with international trade policies, geopolitics and trade protection measures, and our business, financial condition and results of operations may be materially and adversely affected.
We are subject to the risks associated with international trade policies, geopolitics and trade protection measures, and our business, financial condition and results of operations could be adversely affected.
In recent years, complexities in international relations, such as the geopolitical tensions between China and the U.S., have presented new challenges and ongoing uncertainties. For example, in April 2025, the U.S. government announced substantial new tariffs affecting a wide range of products and jurisdictions and has indicated an intention to continue developing new trade policies. In response, certain other governments
announced or implemented retaliatory tariffs and other protectionist measures. While China and the U.S. made announcement on a joint statement to substantially move down the tariff levels in May 2025, there is no assurance as to how the U.S.-China trade tensions might develop or whether there will be any changes to the scope and extent of goods that are or will be being subject to tariffs or new trade policies introduced by the two countries. These circumstances could reduce levels of international trade, investment, technological exchange, and other economic activities. They might also lead to changes in political and economic relations between countries, sanctions, export controls, and other geopolitical issues. These developments have created a dynamic and unpredictable trade landscape, which may adversely impact our business, financial condition, results of operations and prospects.
Trade disputes, tariffs and other political tensions between the U.S. and other countries may also exacerbate unfavorable macroeconomic conditions including inflationary pressures, foreign exchange volatility, financial market instability and economic recessions or downturns. Any prolonged economic downturn or escalation in trade tensions could materially and adversely affect our business, financial condition, results of operations and prospects.
We may be subject to the risks associated with export controls, economic sanctions and investment restrictions, and our business, financial condition and results of operations may be materially and adversely affected.
Geopolitics and international trade tensions have led to certain countries and organizations utilizing economic sanctions, export controls, investment restrictions and other restrictive measures targeting high-technology solutions, including in the AI industry. These policies have introduced uncertainties to global investment activities, increased compliance costs and limited access to critical resources necessary to R&D activities and operations for companies operating in affected industries.
On January 16, 2025, we and nine of our subsidiaries ("Listed Entities") were added to the entity list (the "Entity List") administered by the U.S. Department of Commerce, Bureau of Industry and Security ("BIS"). The addition of Listed Entities to the Entity List (the "Entity List Addition") restricts our ability to purchase or otherwise access goods, software and technology (collectively, the "Items") that are subject to the Export Administration Regulations (the "EAR") without a license from the BIS. Items subject to the EAR include, among other things, U.S.-origin items, as well as certain items of non-U.S.-origin that contain more than a de minimis portion of U.S.-origin controlled content, and non-U.S.-origin items that are the direct product of certain U.S. origin controlled software or technology. For further information, see "Regulatory Overview—U.S. Export Control laws and Regulations." In order to address the EAR-related risks after the Entity List Addition, we have put in place a series of export control compliance measures. However, there can be no assurance that our export control compliance measures could be strictly followed and implemented, or that the implementation of such export control compliance measures would be sufficient for us to address concerns under the EAR.
Additionally, there can be no assurance that the U.S. or other governments or organizations will not impose additional sanctions or export controls on us and/or our subsidiaries. Any such event could materially and adversely affect our business, financial condition and results of operations. Given recent calls by the U.S. government to enhance control over cloud computing leases, we cannot assure you that the U.S. government will not impose additional restrictions on the provision of cloud computing services for AI training to PRC enterprises. Such measures could be implemented through the enactment of new legislation by the U.S. Congress or amendments to the EAR. In addition, on April 9, 2025, the Remote Access Security Act, which seeks to address perceived gaps in current export control laws that have permitted companies affiliated with the PRC Communist Party to obtain access to restricted U.S. technology via cloud services, was reintroduced. The bill received unanimous bipartisan support in the U.S. House Foreign Affairs Committee. If enacted into federal law, such legislation could adversely affect our business, financial condition and results of operations.
Furthermore, as our business is closely interrelated with our customers and suppliers, any imposition of economic sanctions or export control, that impact our customers or suppliers could materially and adversely affect our business, financial condition and results of operations. Moreover, we may be subject to review and enforcement under domestic and foreign laws that govern foreign investment and acquisitions. In both U.S. and non-U.S. jurisdictions, these regulatory requirements may apply different requirements based on the nature of the company and the profiles of the investors involved. As a result, investments by particular investors may need to be filed with local regulators or could even be prohibited under certain circumstances, which limits our ability to engage in strategic transactions that might otherwise be beneficial to us and our investors. These laws are also regularly changed and updated. For example, recently issued U.S. government regulations, such as the final rule (the "Final Rule") implementing Executive Order 14105 which became effective in January 2025, restricts direct and indirect investment by U.S. persons (as defined under the Final Rule) into companies with specified connections to China that use specific technologies of concern. Notably, on February 21, 2025, the U.S. government issued the "America First Investment Policy" proposing to further expand the set of technologies of concern. These rules are aimed at exerting greater U.S. government oversight over U.S. direct and indirect investments involving China in certain sectors, and may introduce new hurdles and uncertainties for cross-border collaborations, investments and funding opportunities of China-based issuers, including us. As advised by our International Sanctions Counsel, we are deemed a covered foreign person as our business activities are notifiable under the Final Rule. However, due to the addition of our Company to the Entity List, pursuant to the Final Rule, U.S. persons (as defined under the Final Rule) are prohibited to acquire any equity interest in our Company that is not yet publicly traded, except through valid exemptions, or unless such investments were completed prior to the effective date of the Final Rule. Accordingly, as advised by our International Sanctions Counsel, U.S. persons (as defined under the Final Rule) are prohibited to invest in our Company in the Global Offering. Accordingly, we have structured the Global Offering so that the Offer Shares will be offered and sold outside the United States to persons that are not, and are not acting for the account or benefit of, U.S. Investors, which term includes U.S. persons as defined under the Final Rule as well as Regulation S, in offshore transactions. The investment restrictions could limit our ability to raise funds, in particular, from U.S. Investors, and the liquidity and market price of our publicly traded securities could be adversely affected due to a lack of participation from these investors. As advised by our International Sanctions Counsel, once our Company's H Shares are issued and become publicly traded, U.S. persons are exempted under the publicly traded securities exception regardless of whether the company engages in covered activities. Hence, the investments in our Group's publicly traded H Shares after Listing will not be subject to any reporting requirement or prohibition under the Final Rule. Continuing changes in both U.S. and non-U.S. jurisdictions to foreign investment laws and rules could adversely affect our strategic initiatives, financial performance and growth prospects.
We have a limited operating history, which may make it difficult to evaluate our current business and predict our future performance.
We have a limited operating history. Since our inception in 2019, we have been dedicated to developing high-performance, safe and reliable general-purpose intelligent systems. Our operations provide a limited basis for you to assess our ability to successfully market and commercialize our solutions. Certain of our solutions are still at various stages of development. Consequently, predictions about our future success or viability may not be as accurate as they could be if we had a longer operating history. We will encounter risks and difficulties frequently experienced by early-stage companies in rapidly evolving fields. If we do not address these risks and difficulties successfully, we may not be successful in our future business and operations.
As a result of our limited operating history, particularly in light of the rapidly evolving and competitive nature of our industries, it may be difficult to evaluate our current business or reliably predict our future performance based on our historical performance. Our historical results may not provide a meaningful basis for evaluating our business, results of operations, financial condition or prospects. We may encounter unforeseen expenses, difficulties, complications, delays and other known and unknown factors, and may not be able to achieve promising results in future periods. If we cannot address these risks and overcome these difficulties successfully, our business, results of operations, financial condition and prospects will suffer.
We may not be successful in implementing our business plans and strategies effectively, or at all, which could materially and adversely affect our business, results of operations, financial condition and prospects.
Our business plans and strategies are based on our assumptions of future events, which may entail certain risks and uncertainties inherent to our business. These assumptions may not be correct, which could affect the commercial viability of our business plans and strategies. As such, we cannot guarantee that our business plans and strategies will be implemented successfully as scheduled or at all.
If we fail to implement our business plans and strategies effectively and efficiently, we may be unable to expand our operations, manage our growth, seize market opportunities as expected or remain competitive. Furthermore, even if we implement our business plans and strategies effectively and efficiently, there may be other unexpected events or factors beyond our control that may prevent us from achieving desirable and profitable results, such as changes in local laws and regulations and governmental policies, the availability of skilled professionals and changes in customer demand. Moreover, our business plans and strategies may divert the attention of our senior management, increase our operating costs, and increase our cash outflows for operating and investing activities. Accordingly, if our business plans and strategies cannot be successfully implemented, or if they do not yield ideal results, we may have significant difficulties in recovering our costs and therefore experience a material adverse impact on our business, financial condition and prospects.
Our business is dependent on the strengths and market acceptance of our brand. If we fail to maintain and enhance our brand, or if we incur excessive expenses in this effort, our business, results of operations, financial condition and prospects may be materially and adversely affected.
Our business and financial performance depends on the strength and market acceptance of our brand. To enhance the market acceptance of our brand, we participate in online and offline events, such as industry conferences, product launch events and developer forums from time to time to showcase client success stories and developer breakthroughs and to deepen industry connections. However, these events may substantially increase our marketing expenses and we cannot assure you that these activities will be successful or that we will be able to achieve the promotional effect we expect.
If we are unable to maintain our reputation, enhance our brand recognition or promote our solutions, or if we incur excessive expenses in this effort, our business, results of operations, financial condition and prospects may be materially and adversely affected.
Any flaws or misuse of AI technologies, whether actual or perceived, intended or inadvertent, committed by us or by other third parties, could have a material adverse effect on our reputation, business, results of operations, financial condition and prospects.
AI technologies are in the process of rapid development and continue to evolve. Similar to many innovations, AI technologies present risks and challenges, such as potential misuse by third parties for inappropriate purposes or biased applications that may breach public confidence or violate applicable laws and regulations in China and other jurisdictions. Flaws or misuse of AI technologies may also lead to litigation or other proceedings initiated by certain individuals claiming for infringement of legitimate rights such as privacy or personality rights. Such misuse could affect customer perception, public opinions, views of policymakers and regulators and result in decreased adoption of AI technologies.
In addition, we cannot assure you that any measures we have taken or may take to prevent the misuse of our technologies will always be effective, or that our technologies will not be misused or applied in a way that is inconsistent with our intention or public expectation. Any inappropriate, abusive or premature usage of AI technologies, whether actual or perceived, whether intended or inadvertent and whether by us or by third parties, may dissuade prospective customers from adopting AI solutions, may impair the general
acceptance of AI solutions by the society, may attract negative publicity and adversely impact our reputation and may even violate applicable laws and regulations in China and subject us to legal or administrative proceedings and/or other organizations and heightened scrutiny by the regulators. Each of the foregoing events may in turn materially and adversely affect our business, results of operations, financial condition and prospects.
Our legal right to some leased properties may be challenged.
As of the Latest Practicable Date, we leased 18 properties in the PRC, mainly as our offices. As of the Latest Practicable Date, we had not received copies of the real property title certificates from the lessors or the property owners for five properties we leased. As advised by our PRC Legal Advisors, if the relevant lessor has no right to lease the leased property and a third party other than the parties to the relevant lease contracts have legal title to such leased property, such third party may claim that the relevant lease contracts are null and void or have no effect thereto, or request us to cease our use and move out of such leased property.
As of the Latest Practicable Date, 14 out of our 18 leased properties used as our offices had not been registered and filed with relevant land and real estate administration bureaus in the PRC. Under the Measures for Administration of Lease of Commodity Properties (《商品房屋租赁管理办法》), which was promulgated by the Ministry of Housing and Urban-Rural Development of the PRC on December 1, 2010 and became effective on February 1, 2011, both lessors and lessees are required to file the lease agreements for registration and obtain property leasing filing certificates for their leases. We cannot assure you that we will be able to comply with the relevant laws and regulations by completing all required filings of our existing and future lease agreements in China. We may be required by relevant government authorities to file future lease agreements for registration within a time limit and may be subject to a fine ranging from RMB1,000 to RMB10,000 for each non-registered lease exceeding such time limit.
We may from time to time be subject to claims, disputes, lawsuits and other legal and administrative proceedings.
We may be subject to claims and various legal and administrative proceedings. Claims arising out of actual or alleged violations of law, breach of contract or torts could be asserted against us by customers, business partners, suppliers, competitors, employees or governmental entities in investigations and legal proceedings.
Regardless of the merit of any particular claim, legal and administrative proceedings may be expensive, time-consuming or disruptive to our operations and may divert management's attention from our business operations. If a judgment, a fine or a settlement involving a payment of a material sum of money were to occur, or injunctive relief were issued against us, it may result in significant monetary liabilities and may materially disrupt our business and operations, and our business, results of operations, financial condition, prospects and reputation could be materially and adversely affected. In recognition of these considerations, we may enter into agreements to settle litigation and resolve such disputes. There is no assurance that such agreements can be obtained on acceptable terms or that litigation will not occur. These agreements may also significantly increase our operating expenses. Legal or administrative proceedings and claims may arise in the future, which may cause us to incur legal costs, divert our managements' attention, suffer reputational damage and financial losses, and our business, results of operations, financial condition and prospects could be materially and adversely affected.
If we fail to obtain and maintain the requisite licenses and approvals required under the regulatory environment applicable to our business, or if we are required to take actions that are time consuming or costly in order to obtain and maintain such licenses and approvals, our business, results of operations, financial condition and prospects may be materially and adversely affected.
Under the current PRC regulatory scheme, a number of governmental authorities including but not limited to the MIIT, the MPS and the CAC, jointly regulate major aspects of our industries. We cannot – 52 –
assure you that we can successfully renew the licenses required for our business in a timely manner or that these licenses are sufficient to conduct all of our present or future business. The interpretation and implementation of existing and future laws and regulations governing our business activities may change from time to time in the future. If we fail to complete, obtain or maintain any of the required licenses or approvals or make the necessary filings, we may be subject to various penalties, such as confiscation of the revenue that was generated through the implicated operations, the imposition of fines and the discontinuation or restriction of our operations. Any such penalties may disrupt our business operations and materially and adversely affect our business, results of operations, financial condition and prospects.
We may be unable to expand and maintain existing strategic partnerships with academic institutions.
We have formed strategic partnerships with universities around the world to establish joint research laboratories. See "Business—Research and Development—Talent." Through these joint laboratories, we identify and engage highly skilled and qualified candidates to enrich our talent pool, gain deeper industry insight and establish stronger brand recognition. However, there can be no assurance that these institutions will continue to collaborate with us on commercially reasonable terms, or at all. We also cannot assure you that we will be able to establish new joint laboratories, or extend existing relationships with these institutions when our agreements with them expire. If we are unable to maintain our relationships with these institutions, or any of our collaborations with these institutions are terminated, we may not be able effectively attract qualified talents and enhance our brand awareness through these collaborations, which could adversely affect our business, results of operations, financial condition and prospects.
We may be subject to supplemental contributions of social insurance and housing provident funds required by relevant governmental authorities.
PRC laws and regulations require us to participate in various government sponsored employee benefit plans. These benefit plans include social insurance, housing provident funds and other welfare-oriented payment obligations. If we are deemed to be not in compliance, competent authorities may require us to make supplementary payments for social insurance shortfalls. See "Business—Employees." As of the Latest Practicable Date, we had not received any notice for payment of penalties of social insurance premium and housing provident funds from the competent authorities, nor had we been subject to any material administrative penalties during the Track Record Period and up to the Latest Practicable Date. However, if the relevant PRC authorities hold a different view with us and, for instance, determine that we shall make supplemental contributions, that we are not in compliance with labor laws and regulations or that we are subject to fines or other administrative penalties, our results of operation, financial conditions, business and prospects may be adversely affected.
Our technology infrastructure and information technology system may experience unexpected failure, interruption, inadequacy, security breaches or cyberattacks. Our reputation, business, results of operations, financial condition and prospects may be harmed by service disruptions or by our failure to timely and effectively scale up and adapt our existing technology and infrastructure.
Our technology infrastructure may encounter disruptions or other outages caused by problems or defects in our own technologies and systems, such as malfunctions in software or network overload, and by damages from fires, floods, earthquakes and other natural disasters, telecommunication failures, power loss, human error or other accidents. Our infrastructure and systems may be breached if any vulnerabilities therein are exploited by unauthorized third parties. We cannot assure you that any applicable recovery system, security protocol, network protection mechanisms or other defense procedures are, or will be, adequate to prevent such network or service interruptions, system failures or data losses. The occurrence of unanticipated problems that affect our technological infrastructure could result in interruptions in the availability of our services. It may be difficult for us to respond to such interruptions in a timely manner, or at all.
Furthermore, we rely on information technology networks and systems across various aspects of our operations including R&D and operations. These information technology systems may be susceptible to damage, disruptions or shutdowns due to failures during the process of upgrading or replacing software, databases or components, power outages, hardware failures, telecommunication failures, user errors or catastrophic events. If our information technology systems suffer damage, disruption or shutdown, we may incur substantial costs in repairing or replacing these systems. In addition, our information technology systems may be susceptible to computer viruses and attacks by computer hackers. Such instances may result in system failures, operational disruptions, or unauthorized usage of our computing power.
Any such disruption or inadequacy that causes interruptions to our operations, or failure to maintain the network and server or solve such problems in a timely manner, could affect the ability of customers to use our services and reduce our customer satisfaction. Any actual or perceived attack or security breach may damage our reputation and brand, expose us to risks of potential litigation and liabilities, and require us to expend significant capital and other resources to alleviate problems caused by such attacks or security breaches. As a result, our reputation, business, results of operations, financial condition and prospects could be adversely affected.
Our insurance coverage may not be sufficient to cover all losses or potential claims by our customers which would affect our business, results of operations, financial condition and prospects.
Our insurance may not be adequate to fully compensate for all kinds of losses we may suffer in the future. For example, we do not maintain key man life insurance, insurance policies covering damages to our network infrastructures or information technology systems or any insurance policies for our properties. In addition, our insurers will review our policies every year and we cannot guarantee that our policies can be renewed on similar or other acceptable terms or at all. Furthermore, if we suffer unexpected severe losses or losses that far exceed the policy limits, it could materially and adversely affect our business, results of operations, financial condition and prospects.
We are in the process of prudently expanding our international operations, which exposes us to significant regulatory, economic and political risks, the failure to handle which may adversely affect our business, results of operations, financial condition and prospects.
Although substantially all of our revenue during the Track Record Period was generated in China, we are in the process of expanding our international outreach. In the six months ended June 30, 2025, we derived RMB18.8 million from overseas markets, accounting for 9.8% of our total revenue during the period. Leveraging the advanced performance and safety of our models, we are continuously expanding our global reach and actively participating in building national and municipal foundation model platforms in countries and regions such as China and Southeast Asia. We may adapt to and develop strategies to address international markets but there is no guarantee that such efforts will have the desired effect. As a result, we may be required to devote significant management attention and financial resources. In connection with such expansion, we may face difficulties including increased competition, uncertain enforcement of our intellectual property rights, unfamiliar market conditions, credit and collectability risk on our trade receivables, and the complexity of compliance with Chinese and foreign laws and regulations such as those regarding cybersecurity, data security and personal information, potential adverse movement of currency exchange rates, tariffs and trade barriers, a variety of regulatory or contractual limitations on our ability to operate, political risks and a geographically and culturally diverse workforce and customer base. Furthermore, changes in foreign tax rules, regulations and other requirements, such as changes in tax rates and statutory and judicial interpretations of tax laws may adversely affect our global operation. In addition, operating in the AI industry, we are especially susceptible to the changes in geopolitical situations and laws and regulations relating to data security and personal information, especially those in jurisdictions where we do business. Failure to overcome any of these difficulties could harm our business. In some cases, compliance with the laws and regulations of one country could violate the laws and regulations of another country. We cannot assure you that we are able to fully comply with the legal requirements of each foreign jurisdiction and successfully adapt our business models to local market conditions.
We may experience any future occurrence of force majeure events, natural disasters or outbreaks of contagious diseases.
Any future occurrence of force majeure events, natural disasters or outbreaks of epidemics and contagious diseases, including COVID-19 pandemic, avian influenza, severe acute respiratory syndrome, H1N1 influenza or Ebola virus, may materially and adversely affect our business, results of operations, financial condition and prospects. An outbreak of an epidemic or contagious disease could result in a widespread health crisis and restrict the level of business activities in affected areas, which may, in turn, materially and adversely affect our business. Moreover, China has experienced natural disasters such as earthquakes, floods and droughts in the past few years. Any future occurrence of severe natural disasters in China may materially and adversely affect its economy and therefore our business.
We are also vulnerable to natural disasters and other calamities because our properties and information systems are susceptible to damage or disruption from fire, floods, typhoons, earthquakes, power loss, telecommunications failures, break-ins, war, riots, terrorist attacks, or similar events. Any of the foregoing events may give rise to interruptions, damage to our property, breakdowns, system failures, technology platform failures, or internet failures, which could result in disruptions to our business operations and adversely affect our business, results of operations, financial condition and prospects.
We may not be able to obtain or maintain adequate intellectual property rights protection for our business, or the scope of such intellectual property rights protection may not be sufficiently broad.
We consider our patents, software copyrights, trademarks and other intellectual property rights to be critical to our business operations. We rely on a combination of patent, copyright, trademark and trade secret laws to protect our intellectual property rights. However, the steps we take to protect our intellectual property might not be adequate to prevent or deter infringement or other misappropriation of our intellectual property by competitors, former employees or other third parties. We have been protecting the proprietary technologies that we consider commercially important by, among others, filing patent applications and registering software copyrights in the PRC and other jurisdictions. As of the Latest Practicable Date, we had 86 registered patents in China, among which 84 were invention patents, and 232 patent applications in China. In addition, as of the Latest Practicable Date, we had 160 copyrights in China. See "Business—Intellectual Property."
The patent application process may be expensive and time consuming, and we may not be able to file and prosecute all necessary or desirable patent applications at a reasonable cost or in a timely manner, or at all. In addition, we may however fail to identify patentable aspects of our R&D outputs before it is too late to obtain patent protection. As a result, we may not be able to prevent competitors from developing and commercializing competitive solutions in all such fields. Specifically, patents may be invalidated, and patent applications may not be granted for several reasons, including known or unknown prior deficiencies in the patent application or the lack of novelty of the underlying invention or technology. Our patent applications may not be granted in the end. As such, we do not know the degree of future protection that we will have on our proprietary technologies, if any, and we may not be able to obtain adequate intellectual property protection with respect to our solutions.
Further, the application of laws governing intellectual property rights in China is evolving, and could involve substantial risks to us. In the event that we have to resort to litigation and other legal proceedings to enforce our intellectual property rights, such action, litigation or other legal proceedings could result in substantial costs and diversion of our management's attention and resources and could disrupt our business. There is no assurance that we will be able to enforce our intellectual property rights effectively or otherwise prevent others from the unauthorized use of our intellectual property. Failure to adequately protect our intellectual property could materially and adversely affect our brand and reputation, and our business, financial condition and results of operations.
Any of the foregoing could materially and adversely affect our business, results of operations, financial condition, competitive position and prospects.
Unauthorized use of our intellectual property by third parties may harm our brand and reputation and may materially and adversely affect our business.
Competitors may infringe, misappropriate or violate our intellectual property rights. Unauthorized use of our intellectual properties by third parties may harm our brand and reputation and may materially and adversely affect our business. In addition, to counter infringement or unauthorized use, litigation may be necessary in the future to enforce or defend our intellectual property rights, to protect our trade secrets or to determine the validity and scope of our own intellectual property rights. This can be expensive and time-consuming. Any claims that we assert against perceived infringers could also provoke these parties to assert counterclaims against us alleging that we infringe their intellectual property rights. Many of our current and potential competitors have the ability to dedicate substantially greater resources to enforce and/or defend their intellectual property rights than we do. We may also encounter difficulties in protecting and defending our intellectual property rights in overseas jurisdictions. Accordingly, we may not be able to prevent third parties from infringing upon or misappropriating our intellectual property. Competitors may use our technologies in jurisdictions where we have not obtained intellectual property protection to develop their own technologies.
An adverse result in any litigation proceeding could put our intellectual properties, as well as any intellectual properties that may issue in the future from our pending intellectual property applications, at risk of being invalidated, held unenforceable or interpreted narrowly.
Furthermore, because of the substantial amount of discovery required in connection with intellectual property litigation, some of our confidential information could be compromised by disclosure during this type of litigation. Defendant counterclaims alleging invalidity or unenforceability are commonplace, and can be asserted on numerous grounds. Third parties may also raise similar claims before administrative bodies in China or abroad, even outside the context of litigation. Such proceedings could result in revocation or amendment to our intellectual properties in such a way that they no longer cover and protect our technologies. The outcome following legal assertions of invalidity and unenforceability is unpredictable. If a defendant were to prevail on a legal assertion of invalidity and/or unenforceability, we would lose at least part, and perhaps all, of the intellectual property protection on our business. Such a loss of intellectual property protection could materially and adversely affect our business.
Obtaining and maintaining our intellectual property protection depends on compliance with various procedural, documentary, fee payment and other requirements imposed by governmental agencies, and our intellectual property protection could be reduced or eliminated for noncompliance with these requirements.
Obtaining and maintaining intellectual properties can be challenging. Failure to respond to official actions within prescribed time limits, non-payment of periodic maintenance fees and failure to properly legalize and submit formal documents, can result in abandonment or lapse of the intellectual property or intellectual property application, leading to partial or complete loss of intellectual property rights in the relevant jurisdiction. In any such event, our competitors might be able to enter the market, which would materially and adversely affect our business.
We may be unable to protect the confidentiality of our trade secrets, and we may be subject to claims that our employees or third parties have wrongfully used or disclosed alleged trade secrets owned by others.
In addition to patents and copyrights, we rely on trade secrets, including unpatented know-how, technology and other proprietary information, to protect our business and thus maintain our competitive
position. We protect these trade secrets, in part, by entering into non-disclosure and confidentiality agreements, non-compete covenants or include such undertakings in the agreements with parties that have access to them. We also enter into employment agreements with our employees that include undertakings regarding assignment of inventions and discoveries. Nevertheless, there can be no guarantee that an employee or a third party will not make an unauthorized use or disclosure of our proprietary confidential information intentionally or inadvertently. It is possible that a competitor will gain access to such information and make use of such information, and that our competitive position will be compromised, in spite of any legal action we might take against persons making such unauthorized disclosures. In addition, to the extent that our employees or business partners use intellectual property owned by others in their work for us, disputes may arise as to the rights in related or resulting know-how and inventions.
Trade secrets are difficult to protect. Our employees or business partners might intentionally or inadvertently disclose our trade secret information to competitors, or our trade secrets may otherwise be misappropriated. Enforcing claims that third parties have misappropriated and are using our trade secrets may be costly and time-consuming, and the outcome of such proceedings is inherently uncertain.
If third parties claim that we infringe upon their intellectual property rights, we may incur liabilities and may have to redesign or discontinue selling the solutions involved.
Some of our competitors may claim that the commercial use of our solutions has infringed their intellectual properties. Therefore, our competitors may initiate legal proceedings alleging that we are infringing, misappropriating or violating their intellectual property rights in connection with the commercialization of the relevant solutions.
Our competitors may use intellectual property litigation to gain a competitive advantage. Infringements of the intellectual property involve analysis of complex legal and factual issues, the determination of which is often uncertain. We may hire employees who have previously worked for our competitors. We cannot guarantee that such employees will not use their previous employers' proprietary know-how or trade secrets in their work for us, which could result in litigation against us. Our competitors may also have filed for patent protection which is not as yet a matter of public knowledge or claim trademark rights that have not been revealed through our searches of relevant public records. Our efforts to identify and avoid infringing on third parties' intellectual property rights may not always be successful. Any claims of patent or other intellectual property infringement, regardless of their merit, could:
• forbid us from making or providing solutions that incorporate the challenged intellectual property;
• require us to enter into royalty or licensing agreements in order to obtain the right to use a third party's intellectual property, such agreements may not be available on terms acceptable to us or at all;
• result in customers terminating, deferring or limiting their use of the affected solutions until resolution of the litigation.
In addition, new intellectual properties obtained by our competitors could threaten the continued life of the solution in the market even after it has already been introduced.
We may not be able to sustain our historical growth rates, and our historical growth may not be indicative of our future growth or financial results.
We achieved growth during the Track Record Period. Our revenue increased by 116.9% from RMB57.4 million in 2022 to RMB124.5 million in 2023 and further increased by 150.9% to RMB312.4 million in 2024. Our revenue increased by 325.0% from RMB44.9 million in the six months ended June 30, 2024 to RMB190.9 million in the six months ended June 30, 2025. However, there is no assurance that we will be able to maintain our historical growth rates in future periods. Our growth was primarily driven by the commercialization of our general-purpose large models, which could be affected by the development of the AI and the related industry, accumulation of AI experts in China, awareness of enterprises to deploy AI applications, our investment in technology innovation and AI solutions and our ability to create value for users with our advanced large model solutions. We cannot assure you that we will be able to effectively manage our growth or implement our business strategies. If the market for our solutions does not develop as we expect or if we fail to address the needs of this dynamic market, our business, results of operations and financial condition will be materially and adversely affected.
We have incurred accumulated losses during the Track Record Period and incurred net losses in 2022, 2023 and 2024, the six months ended June 30, 2024 and 2025 and may continue to experience net losses in the foreseeable future.
We have incurred, and may continue to incur, significant research and development expenses, selling expenses, administrative expenses and other expenses related to our ongoing operations. We recorded accumulated losses during the Track Record Period mainly because, we made significant investments in research and development to support the development of advanced models and the ongoing improvement of our foundation models, and investments in sales and marketing activities to increase our customer base. For the years ended December 31, 2022 and 2023 and 2024 and the six months ended June 30, 2024 and 2025, we had net loss of RMB143.7 million, RMB788.0 million, RMB2,958.0 million, RMB1,235.6 million and RMB2,357.9 million, respectively. See "Financial Information—Results of Operations" for a discussion of our financial performance during the Track Record Period. Our ability to generate revenue will depend primarily on our ability to sell our solutions, as well as the overall development of the AI and the related industry and awareness of enterprises to deploy AI applications, which is subject to significant uncertainty. We expect to continue to incur significant expenses and losses for the foreseeable future. We anticipate that our expenses will increase significantly as we continue to advance performance.
The amount of our future net losses will depend, in part, on our future expenses resulted from costs and expenses incurred by our research and development and in relation to our operations and our ability to generate revenues if we cannot leverage our MaaS platform to deliver intelligence to clients in the most effective and efficient form, or fails to achieve market acceptance among enterprise clients and developers, our business may not become profitable. Even if we achieve profitability in the future, we may not be able to sustain profitability in subsequent periods thereafter. Our prior losses and expected future losses have had, and will continue to have, an adverse effect on business, financial position and results of operations.
We had net current liabilities and net liabilities and recorded net operating cash outflows historically which may continue into the foreseeable future and expose us to liquidity risk.
We recorded net current liabilities and net liabilities throughout the Track Record Period. As of December 31, 2022, 2023 and 2024 and June 30, 2025, we had net current liabilities of RMB232.3 million, RMB1,745.4 million, RMB4,822.2 million and RMB7,080.6 million, respectively, and net liabilities of RMB180.2 million, RMB982.6 million, RMB3,955.1 million and RMB6,150.8 million, respectively. We were in such positions primarily due to (i) an increase in financial instruments issued to investors, and (ii) a decrease in cash and cash equivalents. We expect that the financial instrument issued to investors will be
reclassified from liabilities to equity, because the redemption rights granted to the Pre-IPO investors will terminate upon Listing and the relevant financial instrument issued to investors will be reclassified to equity. A net current liabilities position can expose us to the risk of shortfalls in liquidity, in which case our ability to raise funds, obtain bank loans and declare and pay dividends will be materially and adversely affected. Also, we recorded net cash outflow from operating activities of RMB68.2 million, RMB648.0 million, RMB2,244.9 million, RMB994.7 million and RMB1,327.2 million for the years ended December 31, 2022, 2023 and 2024 and the six months ended June 30, 2024 and 2025, respectively. See "Financial Information—Liquidity and Capital Resources—Cash Flows." Negative operating cash flow may require us to obtain additional financing to meet our financing needs and obligations and support our expansion plans. In the event that we are unable to generate sufficient cash flow from our operations or otherwise obtain sufficient external funds to finance our business, our liquidity and financial condition may be materially and adversely affected and we may not be able to expand our business as expected. We cannot assure you that we will have sufficient cash from other sources to fund our operations. If we resort to other financing activities, we will incur additional financing costs, and we cannot guarantee that we will be able to obtain the financing on terms acceptable to us, or at all. As a result, our business, financial condition and results of operations may be materially and adversely affected. We cannot guarantee that prospective business activities of our Group and/or other matters beyond our control, such as market competition and changes to the macroeconomic environment, will not adversely affect our operating cash flow and lead to net operating cash outflows in the future. If we encounter long-term and continuous net operating cash outflow in the future, we may not have sufficient working capital to cover our operating costs, and our business, financial position and results of operations may be materially and adversely affected.
We are subject to credit risk related to delay in payment and defaults of customers or third parties, which would adversely affect our liquidity and financial condition.
Our cash flow and business sustainability are subject to the timely settlement of payments by our customers. Our customers primarily include enterprises, public sector entities and individual users. As of December 31, 2022, 2023 and 2024 and June 30, 2025, our trade and other receivables amounted to RMB27.9 million, RMB416.4 million, RMB666.8 million and RMB453.4 million, respectively, and we recorded provision for trade receivables of RMB29.0 thousand, RMB1.0 million, RMB9.0 million and RMB25.9 million. For more details on our trade and other receivables, see "Financial Information—Discussion of Selected Items of Consolidated Statements of Financial Position—Trade and Other Receivables." We cannot assure you that we will be able to collect all or any of our trade receivables or collect the amount for any unbilled work on time, or at all, after meeting the agreed contractual milestones. Our customers may face unexpected circumstances, including, but not limited to, long payment cycle, adverse operating condition or financial condition of our customers, and our customers' inability to pay caused by their end customers' delay in payment. Our customers may delay or even default in their payment obligation. As a result, we may not be able to receive such customer's payment of uncollected debts in full, or at all, and we may need to make provisions for trade and notes receivables. The occurrence of such event would materially and adversely affect our financial condition and results of operations.
The changes in the carrying amounts of financial instruments issued to investors may adversely affect our financial condition and results of operations.
During the Track Record Period, we recognized the financial instruments issued to Pre-IPO Investors as financial liabilities. The Pre-IPO Investors were granted redemption rights to require us to redeem all of the instruments upon certain redemption or liquidation events. See "History, Development and Corporate Structure—Pre-IPO Investments." The financial liabilities are initially recognized and subsequently measured at the present value of the redemption amount, which represents the settlement that would be triggered by the event with the highest settlement outcome. Changes in the carrying amounts of the financial liabilities are recognized as "changes in carrying amounts of financial instruments issued to investors" in profit or loss. We recognized losses in the changes in the carrying amounts of financial instruments issued to investors of RMB45.2 million, RMB161.5 million, RMB468.9 million and RMB429.3 million in 2022,
2023, 2024 and the six months ended June 30, 2025, respectively. Any significant fluctuations in the changes in the carrying amounts of financial instruments issued to investors may materially affect our financial condition and results of operations.
We may not be able to raise adequate capital to finance our business operations, expansion plans or R&D strategies, or we may be able to do so only on unfavorable terms.
The implementation of our business and R&D strategies requires a substantial outlay of capital. In 2022, 2023 and 2024 and the six months ended June 30, 2024 and 2025, our net cash used in operating activities amounted to RMB68.2 million (人民币6,820万元), RMB648.0 million (人民币6.48亿元), RMB2,244.9 million (人民币22.449亿元), RMB994.7 million (人民币9.947亿元) and RMB1,327.2 million (人民币13.272亿元), respectively. Our capital expenditures were RMB31.7 million (人民币3,170万元), RMB508.6 million (人民币5.086亿元), RMB132.8 million (人民币1.328亿元), RMB104.5 million (人民币1.045亿元) and RMB23.9 million (人民币2,390万元) in 2022, 2023 and 2024 and the six months ended June 30, 2024 and 2025, respectively. As we pursue our business and R&D strategies and seek to respond to developments in our business and opportunities and trends in our industry, our actual capital expenditure may differ from our expected capital expenditures. No assurance can be given that our available funds and cash flow from operations will be sufficient to meet our cash needs for the future, or that we will not require additional equity or debt financing. If we determine we need to obtain additional funds through external financing and are unable to do so, we may be prevented from fully implementing our business or R&D strategy.
We may record impairments of intangible assets and goodwill.
We may record impairments of intangible assets and goodwill, which may adversely affect our financial condition and results of operations. Impairment tests on goodwill are undertaken annually or when there is an indication of possible impairment at each reporting date. Other non-financial assets are subject to impairment tests whenever events or changes in circumstances indicate that their carrying amount may not be recoverable. When these events occur, we measure impairment by comparing the carrying value of the asset to the recoverable amount of such asset, which is the greater of the fair value less costs of disposal and the value in use. If the recoverable amount is less than the carrying amount of such asset, we recognize an impairment loss based on the recoverable amount of such asset.
The application of impairment test to our intangible assets and goodwill requires management's judgment, including an estimate of the recoverable amount which is the higher of its value in use and its fair value less costs of disposal. We use the value in use of the cash-generating unit to which the goodwill is allocated to determine the recoverable amount. The cash flow projections used to determine the value in use of a cash-generating unit is based on assumptions, such as revenue growth rates, long term growth rate, gross profit margin rates, and discount rate applied to the projected cash flows. These assumptions may be affected by unexpected changes in future market or economic conditions. If our estimates and judgments are inaccurate, the recoverable amount determined could be inaccurate and the impairment recognized may not be adequate, and we may need to record additional impairments in the future.
We have granted, and may continue to grant, certain awards under our share incentive plans, which may result in increased share-based payment expenses, affect our business, results of operations, financial condition and prospects, and potentially dilute the shareholding of our existing shareholders.
We adopted share incentive plans including share-based payments for the benefit of our Directors and employees to incentivize and reward the eligible persons who have contributed to our success. In 2022, 2023 and 2024 and the six months ended June 30, 2024 and 2025, we incurred share-based payment expenses of RMB1.0 million (人民币100万元), RMB5.5 million (人民币550万元), RMB23.6 million (人民币2,360万元), RMB4.2 million (人民币420万元) and RMB158.9 million (人民币1.589亿元) respectively. We believe the granting of share-based payments is of significant importance to our ability to attract and retain key personnel and employees. Nevertheless, share-based payment expenses would potentially dilute the shareholding of existing shareholders. We may continue to grant share-based payments to employees in the future. As a result, our share-based payment expenses may increase, which may affect
我们的财务状况和经营业绩。我们可能会不时重新评估股权激励计划下授予的归属时间表、锁定期或其他关键条款。如果我们选择这样做，在本次全球发售之后的报告期内，我们的股份支付费用可能会发生重大变化。
我们及我们的某些中国子公司依据现行中国税法享受各类优惠税收待遇。我们被认定为高新技术企业的中国子公司适用15%的企业所得税优惠税率。此外，我们某些被认定为小型微利企业的中国子公司适用20%的企业所得税优惠税率。上述优惠税收待遇可能发生变更或被终止。如果我们的优惠税收待遇被撤销、不再适用，或者中国税务机关成功对我们的税务责任计算提出质疑，我们所享受的各类优惠税收待遇中任何一项的终止均可能对我们的经营业绩产生重大不利影响。
我们还从地方政府获得补助，该等补助具有酌情性质，且逐年有所不同。2022年、2023年、2024年及截至2025年6月30日止六个月，我们确认为递延收益的政府补助金额分别为人民币1,030万元、人民币2,970万元、人民币3,480万元及人民币3,620万元。地方政府未来可能决定对上述补助进行调整。我们无法保证我们部分中国子公司目前所享受的政府补助将持续可得。政府补助的任何减少、取消或须予偿还均可能对我们的业务、财务状况及经营业绩产生不利影响。
Changes in China's economic, political and social conditions, as well as government policies, could have a material adverse effect on our business and prospects.
Our business, financial condition and results of operations may be influenced by the general political, economic and social conditions in China, where we operate and conduct our R&D activities. Governments worldwide have implemented, and may continue to introduce, among others, various policies and measures to encourage the economic growth and guide the allocation of resources. The AI industry in general is affected by macro-economic factors, including international, national, regional and local economic conditions, consumer demand and discretionary spending. Any changes in these factors may have material and adverse effect on our business, financial condition and prospects.
We may be subject to the approval or other requirements of the China Securities Regulatory Commission or other PRC governmental authorities in connection with future security activities.
As the PRC laws and regulations in relation to overseas issuance and listing of shares develop, we are required to make filings with or report to CSRC or other PRC regulatory authorities for our future capital raising activities. On February 17, 2023, the CSRC promulgated the Trial Administrative Measures of Overseas Securities Offering and Listing by Domestic Companies (《境內企業境外發行證券和上市管理試行辦法》) (the "Overseas Listing Trial Measures") and their implementation guidelines. The Trial Measures, which came into effect on March 31, 2023, mainly provide the scope of activities subject to the filing requirement, the entities subject to filing obligations, and the filing procedures. See "Regulatory Overview—Regulations Relating to Overseas Securities Offering, Listing and Full Circulation of H Shares—Regulations on Overseas Securities Offering and Listing." We are required to file with the CSRC in accordance with the Overseas Listing Trial Measures after our application for the Offering is submitted. If a domestic company fails to complete the filing procedure or conceals any material fact or falsifies any major content in its filing documents, such domestic company may be subject to administrative penalties, such as
order to rectify, warnings, fines, and its controlling shareholders, actual controllers, the person directly in charge and other directly liable persons may also be subject to administrative penalties, such as warnings and fines. In addition, such failure may restrict our ability to complete the proposed Offering and to finance the development of our business and may have a material and adverse effect on our business, financial condition and prospects.
On February 24, 2023, the CSRC, the MOF, the National Administration of State Secrets Protection of China, and the National Archives Administration of China published the Provisions on Strengthening Confidentiality and Archives Administration of Overseas Securities Offering and Listing by Domestic Companies (《关于加强境内企业境外发行证券和上市相关保密和档案管理工作的规定》) (the "Archives Rules"), which came into effect on March 31, 2023. The Archives Rules require that, in relation to the overseas securities offering and listing activities of domestic enterprises, either in direct or indirect form, such domestic enterprises, as well as securities companies and securities service institutions providing relevant securities services, are required to strictly comply with relevant requirements on confidentiality and archives management, establish a sound confidentiality and archives system, and take necessary measures to implement their confidentiality and archives management responsibilities. Any failure to comply with Archives Rules may materially and adversely affect our business, results of operations, financial condition and prospects.
We are closely monitoring how they will affect our operations and our future financing. In addition, if the CSRC or other PRC regulatory authorities in the future promulgate new rules or explanations imposing further requirements that we obtain their approvals or complete the required filing or other regulatory procedures for this Global Offering or future capital raising activities, there can be no assurance that we will be able to obtain a waiver of such requirements, if and when procedures are established to obtain such a waiver. Any unforeseen situations or negative publicity regarding such approval, filing or other requirements could materially and adversely affect our business, financial condition, prospects and the trading price of our Shares.
It may be difficult to effect service of process, enforce foreign judgments or bring original actions against us, our Directors, Supervisors and senior management residing in China.
We are a company incorporated under the PRC laws and substantially all of our assets and subsidiaries are located in China. Substantially all of our Directors and senior management reside within China. As a result, it may not be possible for you to effect service of legal process within China on us or our Directors or senior management. Judgments of courts of another jurisdiction may be reciprocally recognized or enforced if the jurisdiction has a treaty on that with China. Currently, China does not have treaties providing for the reciprocal enforcement of judgments in civil and commercial matters by courts with Japan, the United States, the United Kingdom or most other western countries. On July 14, 2006, Hong Kong and China entered into the Arrangement on Reciprocal Recognition and Enforcement of Judgments in Civil and Commercial Matters by the Courts of the Mainland and of the Hong Kong Special Administrative Region Pursuant to Choice of Court Agreements Between Parties Concerned (《关于内地与香港特别行政区法院相互认可和执行当事人协议管辖的民商事案件判决的安排》) (the "Arrangement"), pursuant to which reciprocal recognition and enforcement of the judgment may be possible between these two jurisdictions provided that the judgment is rendered by a designated court of these two jurisdictions and the parties has an expressly written choice of court. On January 18, 2019, the Supreme People's Court and the government of the Hong Kong Special Administrative Region entered into the Arrangement on Reciprocal Recognition and Enforcement of Judgments in Civil and Commercial Matters by the Courts of the Mainland and of the Hong Kong Special Administrative Region (《关于内地与香港特别行政区法院相互认可和执行民商事案件判决的安排》) (the "New Arrangement"), which became effective on January 29, 2024, which seeks to establish a mechanism with further clarification on and certainty for reciprocal recognition and enforcement of judgments in a wider range of civil and commercial matters between Mainland China and Hong Kong. Although the New Arrangement has been signed, the outcome and effectiveness of any action brought under the New Arrangement may still be uncertain. In addition, Hong Kong has no arrangement for
reciprocal enforcement of judgments with the United States and certain other jurisdictions. As a result, you may encounter difficulty in enforcing foreign judgments against us or our Directors or senior management.
Laws and regulations over foreign currency conversion and on the remittance of Renminbi into and out of China may affect our utilization of our revenue and our ability to remit dividends.
The PRC government imposes laws and regulations on the convertibility of the Renminbi into foreign currencies and, in certain cases, the remittance of Renminbi into and out of China. Under the existing PRC foreign exchange regulations, foreign exchange transactions under the current account conducted by us, including the payment of dividends, can be made in foreign currencies without prior approval of SAFE by complying with certain procedural requirements and conduct such transactions at designated foreign exchange banks within China that have the licenses to carry out foreign exchange business. Foreign exchange transactions under the capital account, however, normally need to be approved by or registered with the SAFE or its local branch unless otherwise permitted by law. Any insufficiency of foreign exchange may restrict our ability to obtain sufficient foreign exchange for dividend payments to shareholders or satisfy any other foreign exchange obligation. If we do not meet the procedural approvals in respect of the foreign exchange administration, our potential offshore capital expenditure plans and even our business may be materially and adversely affected.
Fluctuations in exchange rates of Renminbi against Hong Kong dollar, U.S. dollar or other foreign currencies could affect our business, results of operations, financial condition, and the value of your investment.
Fluctuations in the exchange rate of Renminbi against Hong Kong dollar, U.S. dollar and other foreign currencies are affected by, among other things, the changes in China's and international political and economic conditions. The proceeds from the Global Offering will be denominated in Hong Kong dollars. As a result, any appreciation of Renminbi against U.S. dollar, Hong Kong dollar or any other foreign currencies may result in a decrease in the value of our foreign currency-denominated assets and our proceeds from the Global Offering. Conversely, any depreciation of Renminbi may adversely affect the value of, and any dividends payable on our H Shares in foreign currencies. We have not utilized, and may not in the future utilize, any instrument to reduce our foreign currency risk exposure. All of these factors could affect our business, results of operations, financial condition and prospects, and could affect the value of, and dividends payable on, our H Shares in foreign currency terms.
We are a mainland China enterprise and we are subject to mainland China tax on our global income and any gains on the sales of H Shares and dividends on the H Shares may be subject to mainland China income taxes.
Under the PRC EIT Law and its implementation rules, subject to any applicable tax treaty or similar arrangement between the mainland China and a non-mainland China investor's jurisdiction of residence that provides for a different income tax arrangement, mainland China withholding tax at the rate of 10% is normally applicable to dividends from mainland China sources payable to investors that are non-mainland China resident enterprises, which do not have an establishment or place of business in mainland China, or which have an establishment or place of business in mainland China if the relevant income is not effectively connected with such establishment or place of business. Any gains realized on the transfer of shares by such investors are subject to a 10% mainland China income tax rate if such gains are regarded as income from sources within mainland China unless a treaty or similar arrangement provides otherwise.
Under the PRC Individual Income Tax Law (《中華人民共和國個人所得稅法》) and its implementation rules, dividends from sources within mainland China paid to foreign individual investors who are not mainland China residents are generally subject to a mainland China withholding tax at a rate of 20% and gains from mainland China sources realized by such investors on the transfer of shares are generally subject to a 20% mainland China income tax rate, in each case, subject to any reduction or exemption set forth in
applicable tax treaties and laws in mainland China. Pursuant to the Circular on Questions Concerning the Collection of Individual Income Tax Following the Repeal of Guo Shui Fa [1993] No. 045 (《关于国税发[1993]045号文件废止后有关个人所得税征管问题的通知》) (Guo Shui Han [2011] No. 348) (国税函[2011]348号) dated June 28, 2011, issued by the SAT, dividends paid to non-mainland China resident individual holders of H Shares are generally subject to individual income tax of mainland China at the withholding tax rate of 10%, in which the non-mainland China resident individual holder of H Shares resides as well as the tax arrangement between mainland China and Hong Kong. Non-mainland China resident individual holders who reside in jurisdictions that have not entered into tax treaties with mainland China are subject to a 20% withholding tax on dividends received from us. However, pursuant to the Circular Declaring that Individual Income Tax Continues to be Exempted over Income of Individuals from Transfer of Shares (《关于个人转让股票所得继续暂免征收个人所得税的通知》) issued by the MOF of mainland China and the SAT on March 30, 1998, gains of individuals derived from the transfer of listed shares of enterprises may be exempt from individual income tax. In addition, on December 31, 2009, the MOF, the SAT and the CSRC jointly issued the Circular on Relevant Issues Concerning the Collection of Individual Income Tax over the Income Received by Individuals from Transfer of Listed Shares Subject to Sales Limitation (《关于个人转让上市公司限售股所得征收个人所得税有关问题的通知》) (Cai Shui [2009] No. 167) which states that individuals' income from the transfer of listed shares on certain domestic exchanges shall continue to be exempted from individual income tax, except for the relevant shares which are subject to sales restrictions as defined in the Supplementary Circular on Relevant Issues Concerning the Collection of Individual Income Tax over the Income Received by Individuals from Transfer of the Listed Shares Subject to Sales Limitations (《关于个人转让上市公司限售股所得征收个人所得税有关问题的补充通知》) (Cai Shui [2010] No. 70). As of the Latest Practicable Date, the aforesaid provision has not expressly provided that individual income tax shall be collected from non-mainland China resident individuals on the sale of shares of mainland China resident enterprises listed on overseas stock exchanges.
If mainland China income tax is imposed on gains realized from the transfer of our H Shares or on dividends paid to our non-mainland China resident investors, the value of your investment in our H Shares may be affected. Furthermore, our Shareholders whose jurisdictions of residence have tax treaties or arrangements with mainland China may not qualify for benefits under such tax treaties or arrangements.
Payment of dividends is subject to restrictions under PRC law.
Under PRC law, dividends may be paid only out of distributable profits. Distributable profits are defined as our profits after taxes as determined under PRC GAAP less any recovery of accumulated losses and appropriations to statutory and other reserves that we are required to make. As a result, we may not have sufficient, if any, distributable profits to enable us to make dividend distributions to our Shareholders in the future, including periods for which our financial statements indicate that our operations have been profitable. Any distributable profits not distributed in a given year are retained and available for distribution in subsequent years.
Moreover, because the calculation of distributable profits under PRC GAAP is different from the calculation under IFRSs in certain respects, our subsidiaries may not have distributable profits as determined under PRC GAAP, even if they have profits for that year as determined under IFRSs, or vice versa. Accordingly, we may not receive sufficient distributions from our subsidiaries. Failure by our subsidiaries to pay dividends to us could have a negative impact on our cash flow and our ability to make dividend distributions to our Shareholders in the future, including those periods in which our financial statements indicate that our operations have been profitable.
Any failure to comply with relevant regulations regarding the registration requirements for employee share incentive plans may subject our share incentive plan participants or us to fines and other legal or administrative sanctions.
In February 2012, SAFE promulgated the Notices on Issues Concerning the Foreign Exchange Administration for Domestic Individuals Participating in Stock Incentive Plan of Overseas Publicly Listed
Company (《关于境内个人参与境外上市公司股权激励计划外汇管理有关问题的通知》) (Hui Fa [2012] No. 7), replacing earlier rules promulgated in 2007. Pursuant to these rules, PRC citizens and non-PRC citizens who reside in China for a continuous period of not less than one year and participate in any stock incentive plan of an overseas publicly listed company, subject to a few exceptions, are required to register with SAFE through a domestic qualified agent and complete certain other procedures. In addition, an overseas-entrusted institution must be retained to handle matters in connection with the exercise or sale of stock options and the purchase or sale of shares and interests. We and our executive officers and other employees who are PRC citizens or who reside in China for a continuous period of not less than one year and who have been granted options will be subject to these regulations when our company becomes an overseas-listed company upon the completion of the Global Offering. Failure to complete SAFE registrations may subject them to fines and legal sanctions. In light of the above, we cannot assure you that we will continuously adopt additional incentive plans for our directors, executive officers and employees under PRC law.
In addition, SAT has issued certain circulars concerning employee share options and restricted shares. Under these circulars, our employees working in China who exercise share options or are granted restricted shares will be subject to PRC individual income tax. We have obligations to file documents related to employee share options or restricted shares with relevant tax authorities and to withhold individual income taxes of those employees who exercise their share options. If our employees fail to pay or we fail to withhold their income taxes according to relevant laws and regulations, we may face sanctions imposed by the tax authorities or other PRC governmental authorities.
There has been no prior public market for our H Shares, an active trading market for our H Shares may not develop following the Global Offering and the liquidity and market price of our H Shares may be volatile.
Prior to the Global Offering, there was no public market for our H Shares. We cannot assure you that a public market for our H Shares with adequate liquidity and trading volume will develop and be sustained following the completion of the Global Offering. The Offer Price range for our H Shares was the result of negotiations between us, the Overall Coordinator and the Global Coordinator on behalf of the Underwriters, and the Offer Price may differ significantly from the market price for our H Shares following the Global Offering. If an active public market for our H Shares does not develop following the completion of the Global Offering, the market price and liquidity of our H Shares could be materially and adversely affected.
The price and trading volume of our H Shares may be highly volatile. Several factors, some of which are beyond our control, such as variations in our prospects, changes in our pricing policy, the emergence of new technologies, strategic alliances or acquisitions, the addition or departure of key personnel, changes in profit forecast or recommendations by financial analysts, changes in ratings by credit rating agencies, litigation or the removal of the restrictions on share transactions, could cause large and sudden changes to the volume and price at which our H Shares will trade. In addition, the Stock Exchange and other securities markets have, from time to time, experienced significant price and volume volatility that is not related to the operating performance of any particular company.
You will incur immediate and significant dilution if the Offer Price of our H Shares is substantially higher than the net tangible book value per H Share, and may experience further dilution if we issue additional Shares in the future.
The Offer Price of the Shares is higher than the net tangible book value per Share immediately prior to the Global Offering. Therefore, purchasers of the Shares in the Global Offering will experience an immediate dilution in pro forma net tangible book value, and our existing Shareholders will receive an increase in the pro forma adjusted consolidated net tangible asset value per Share of their Shares. In addition, holders of our Shares may experience further dilution of their interests if the Underwriters exercise the Over-allotment Option or if we issue additional shares in the future to raise additional capital.
We cannot assure you when, whether and in what form or size we will pay dividends in the future.
Our ability to pay dividends will depend on whether we are able to generate sufficient earnings. Distribution of dividends shall be decided by our Board of Directors at their discretion and will be subject to the approval of the general meeting. A decision to declare or to pay dividends and the amount thereof depends on various factors, including but not limited to our prospects, cash flows and financial position, operating and capital expenditure requirements, distributable profits, our Articles of Association and other constitutional documents, the PRC Company Law and any other applicable PRC laws and regulations, market conditions, our strategy and projection for our business, contractual restrictions and obligations, taxation, regulatory restrictions and any other factors from time to time deemed by our Board of Directors as relevant to the declaration or suspension of dividends. As a result, there can be no assurance whether, when and in what form we will pay dividends in the future. Subject to any of the above constraints, we may not be able to pay dividends in accordance with our dividend policy. See "Financial Information—Dividend Policy."
Certain facts, forecasts and other statistics obtained from government publications contained in this prospectus may not be reliable in terms of accuracy, competence or reliance.
Certain facts, forecasts and other statistics contained in this prospectus relating to China, the PRC economy and the industry in which we operate have been derived from various official government publications. We have no reason to believe that such information is false or misleading or that any fact has been omitted that would render such information false or misleading. However, we cannot assure you of the accuracy or completeness of information obtained from these sources. The information from official government sources has not been independently verified by us, the Sole Sponsor, the Overall Coordinator, the underwriters or any other party involved in the Global Offering and no representation is given as to its accuracy. For these reasons, the information from various government publications contained in this prospectus may not be accurate and should not be given undue reliance as a basis for making your investment in our H Shares.
You should read the entire prospectus carefully and we strongly caution you not to place any reliance on any information contained in press articles and other media regarding us and the Global Offering.
Prior to the publication of this prospectus, there has been and there may also be, subsequent to the date of this prospectus but prior to the completion of the Global Offering, press and media coverage regarding us, our business, our industry and the Global Offering, which contain, among other things, certain financial information, projections, valuations and other forward-looking information about us and the Global Offering. We have not authorized the disclosure of any such information in the press or media and do not accept responsibility for the accuracy or completeness of such press articles or other media coverage. We make no representation as to the appropriateness, accuracy, completeness or reliability of any of such projections, valuations or other forward-looking information about us. To the extent such statements are inconsistent with, or conflict with, the information contained in this prospectus, we disclaim responsibility for them. Accordingly, prospective investors are cautioned to make their investment decisions on the basis of the information contained in this prospectus only and should not rely on any other information.
根據《上市規則》第8.12條及第19A.15條，發行人須在香港維持足夠的管理層駐場，且在一般情況下，發行人至少須有兩名執行董事在香港通常居住。
本公司所有執行董事目前居住於中國內地，且日後亦不會在香港通常居住。由於本集團大部分核心業務均在香港以外地區營運、管理及開展，本集團目前並無，且在可預見的將來亦不會在香港維持足夠的管理層駐場，以符合《上市規則》第8.12條及第19A.15條的規定。
(a) 根據《上市規則》第3.05條，本公司已委任兩名授權代表（「授權代表」），分別為本公司執行董事兼董事會主席劉德兵博士（Dr. Liu Debing）及本公司公司秘書鄭程傑先生（Mr. Cheng Ching Kit），彼等將擔任本公司與聯交所之間的主要溝通渠道。各授權代表可應聯交所要求在合理時間內親赴香港與聯交所會面，並可隨時通過電話、傳真及╱或電子郵件（如適用）與聯交所聯絡。各授權代表均獲授權代表本公司與聯交所進行溝通；
(b) 各授權代表均具備在任何時候迅速聯絡本公司所有董事（包括獨立非執行董事）的方式，以便聯交所在任何事項上需要聯絡本公司董事時使用。非在香港通常居住的本公司董事均持有或可申請有效旅行證件前往香港，並能在被要求時於合理時間內與聯交所會面。各董事已向授權代表提供其各自的手提電話號碼、辦公室電話號碼、傳真號碼及╱或電子郵件地址（如適用）。如某位董事預計將外出公幹，彼╱她將盡力向授權代表提供其下榻處的電話號碼，或保持手提電話暢通。各董事及授權代表均已向聯交所提供其手提電話號碼、辦公室電話號碼、傳真號碼及╱或電子郵件地址（如適用）；
(c) 根據《上市規則》第3A.19條，本公司已委任麥士資本有限公司（Maxa Capital Limited）為本公司的合規顧問，該合規顧問可隨時聯絡本公司的授權代表、董事、高級管理層及其他人員，並將於上市後擔任聯交所與本公司之間額外的溝通渠道；及
(d) 聯交所與本公司董事之間的會面可通過授權代表或合規顧問安排，或直接與本公司董事在合理時間內安排。本公司將按照《上市規則》的規定，在切實可行的情況下盡快知會聯交所有關授權代表、董事及╱或合規顧問的任何變動。
WAIVERS FROM STRICT COMPLIANCE WITH THE REQUIREMENTS UNDER THE LISTING RULES WAIVER UNDER RULE 10.04 AND CONSENT UNDER PARAGRAPH 1C(2) OF APPENDIX F1 TO THE LISTING RULES IN RESPECT OF SUBSCRIPTIONS OF OFFER SHARES BY CLOSE ASSOCIATES OF EXISTING SHAREHOLDERS AS CORNERSTONE INVESTORS
Rule 10.04 of the Listing Rules provides that a person who is an existing shareholder of the issuer may only subscribe for or purchase any securities for which listing is sought which are being marketed by or on behalf of a new applicant either in his or its own name or through nominees if the conditions set out in Rules 10.03(1) and (2) of the Listing Rules are fulfilled.
Paragraph 1C(2) of Appendix F1 to the Listing Rules provides, inter alia, that no allocations will be permitted to an applicant's existing shareholders or their close associates, whether in their own names or through nominees, unless the conditions set out in Rules 10.03 and 10.04 are fulfilled, without the prior written consent of the Stock Exchange.
Chapter 2.5 of the Guide provides that (i) given the likely significant funding needs of Specialist Technology Companies (as defined under Chapter 18C of the Listing Rules) and the importance of existing shareholders in meeting the funding needs of these companies, existing shareholders or their close associates may participate in the initial public offering ("IPO") of a Specialist Technology Company provided that the applicant complies with MB Rules 8.08(1)/19A.13A, 18C.08 and 8.08A/19A.13C. An existing shareholder holding 10% or more of the shares in the Specialist Technology Company prior to the IPO must subscribe for shares in the IPO as a cornerstone investor; and an existing shareholder holding less than 10% of the shares in the Specialist Technology Company prior to the IPO may subscribe for shares in the IPO as either a cornerstone investor or a placee. In the case of subscription as a placee, the applicant and its sponsors must confirm that no preference in allocation was given to the existing shareholder; and in the case of subscription as a cornerstone investor, the applicant and its sponsors must confirm that no preference was given to the existing shareholder other than the preferential treatment of assured entitlement at the IPO price and the terms are substantially the same as other cornerstone investors.
As further described in "Cornerstone Investors," each of JSC International Investment Fund SPC (acting for and on behalf of Qizhi SP) ("Qizhi SP"), JinYi Capital Multi-Strategy Fund SPC Ltd. (acting for and on behalf of Structured Credit SP Fund) ("Structured Credit SP Fund") and Luster LightTech International Limited (凌雲光技術國際有限公司) ("Luster LightTech International") has entered into a cornerstone investment agreement as a cornerstone investor with the Company and the Sponsor-Overall Coordinator to subscribe for Offer Shares.
As further described in "Cornerstone Investors," Qizhi SP is indirectly wholly owned by Jingquan Qizhi (Beijing) Equity Investment Fund Partnership (Limited Partnership) (璟泉啟智(北京)股權投資基金合夥企業(有限合夥)) ("JSC Qizhi (Beijing)"), whose general partner is JSC Management Consulting (Beijing) Co., Ltd. (璟泉善誠管理諮詢(北京)有限公司) ("JSC Management Consulting (Beijing)"). JSC Management Consulting (Beijing) is controlled by Beijing Financial Holdings Group Co., Ltd. (北京金融控股集團有限公司), which is wholly owned by State-owned Assets Supervision and Administration Commission of the Beijing Municipal People's Government (北京市人民政府國有資產監督管理委員會) ("Beijing SASAC"). Beijing Jingneng Green Energy M&A Investment Fund (Limited Partnership) (北京京能綠色能源並購投資基金(有限合夥)) ("Beijing Jingneng") is a limited partner of JSC Qizhi (Beijing) holding 30.69% partnership interest therein. Beijing Jingneng Tongxin Investment Management Co., Ltd. (北京京能同鑫投資管理有限公司) ("Beijing Jingneng Tongxin") is the general partner of Beijing Jingneng, and the sole limited partner of Beijing Jingneng is Beijing Energy Group Co., Ltd. (北京能源集團有限責任公司) ("Beijing Energy Group"), holding 88.00% partnership interest therein. Both Beijing Jingneng Tongxin and Beijing Energy Group are wholly owned by Beijing SASAC.
As further described in "History, Development and Corporate Structure," each of (i) Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業(有限合夥)) ("Zhongguancun Science City"), (ii) AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限合夥)) ("AI Fund"); and (iii) Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發展基金合夥企業(有限合夥)) ("Daxing Industrial Fund") was a Shareholder of the Company as of the Latest Practicable Date. Zhongguancun Science City is owned as to 1% by Beijing Zhongguancun Science City Technology Investment Management Co., Ltd. (北京中關村科學城科技投資管理有限公司) as its general partner, which is ultimately controlled by State-owned Assets Supervision and Administration Commission of the People's Government of Haidian District, Beijing (北京市海淀區人民政府國有資產監督管理委員會) ("Haidian SASAC"). AI Fund is owned as to 0.5% by Beijing Jingguoguan Property Management Co., Ltd. (北京京國管置業管理有限公司) as one of its general partners, which is wholly owned by Beijing State-owned Capital Operation Management Co., Ltd. (北京國有資本運營管理有限公司) and ultimately controlled by Beijing SASAC. Daxing Industrial Fund is owned as to 0.10% by Beijing North Business Capital Management Co., Ltd. (北商資本管理(北京)有限公司) as its general partner, which is ultimately controlled by State-owned Assets Supervision and Administration Commission of the People's Government of Daxing District, Beijing (北京市大興區人民政府國有資產監督管理委員會) ("Daxing SASAC").
On the basis that (i) Zhongguancun Science City, AI Fund and Daxing Industrial Fund are ultimately controlled by, among others, Haidian SASAC, Beijing SASAC and Daxing SASAC, respectively, each of which is a PRC governmental body in Beijing and (ii) Qizhi SP is ultimately controlled by Beijing SASAC, Qizhi SP is a close associate of each of Zhongguancun Science City, AI Fund and Daxing Industrial Fund.
Luster LightTech International is a wholly owned subsidiary of Luster LightTech Co., Ltd. (凌雲光技術股份有限公司) ("Luster"). As further described in "History, Development and Corporate Structure," Luster was a Shareholder of our Company as of the Latest Practicable Date. As further described in "Cornerstone Investors," as Luster LightTech International is wholly owned by Luster, it is a close associate of Luster.
JinYi Capital Multi-Strategy Fund SPC Ltd. is a segregated portfolio company incorporated in the Cayman Islands. The funding of JinYi Capital Multi-Strategy Fund SPC Ltd.—Structured Credit SP Fund, which is participating in the Global Offering, is from Tsinghua University Education Foundation (清華大學教育基金會). The objects of the Tsinghua University Education Foundation include improving quality of education and academic standards in China. Its activities include financial support for improving educational and research facilities, research projects and publications, engaging prominent global scholars to lecture and hold teaching positions in China and establishing scholarships, bursaries and teaching awards. Tsinghua University has substantially benefitted from Tsinghua University Education Foundation and is accordingly beneficially interested in Tsinghua University Education Foundation and therefore Structured Credit SP Fund.
As further described in "History, Development and Corporate Structure," Tsinghua Control Technology Transfer Co., Ltd. (華控技術轉移有限公司) ("Tsinghua Technology") was a Shareholder of the Company as of the Latest Practicable Date. Tsinghua Technology is wholly owned by Tsinghua University Asset Management Co., Ltd. (清華大學資產管理有限公司), which is in turn wholly owned by Tsinghua University.
On the basis that Tsinghua University has a beneficial interest in both Structured Credit SP Fund and Tsinghua Technology, Structured Credit SP Fund is a close associate of Tsinghua Technology.
We have applied for a waiver under Rule 10.04 of the Listing Rules and consent under paragraph 1C(2) of Appendix F1 to the Listing Rules to permit Qizhi SP, Structured Credit SP Fund and Luster LightTech International to participate as cornerstone investors in the Global Offering to subscribe for the Offer Shares under the International Offering. The Stock Exchange has agreed to grant the requested consent subject to the conditions that:
(a) our Company will comply with the public float requirements of Rules 19A.13A and 18C.08 of the Listing Rules and the free float requirements of Rule 19A.13C of the Listing Rules;
(b) our Company and the Sole Sponsor confirm that no preferential treatment has been, nor will be, directly or indirectly, given to Qizhi SP, Structured Credit SP Fund and Luster LightTech International, as cornerstone investors by virtue of their relationship with the Company in any allocation in the Global Offering, other than the preferential treatment of assured entitlement under the cornerstone investments at the Offer Price and the terms are substantially the same as other cornerstone investors; and
(c) details of the subscription of the Offer Shares by Qizhi SP, Structured Credit SP Fund and Luster LightTech International as cornerstone investors under the Global Offering are disclosed in this prospectus, and details of the allocation will be disclosed in the allotment results announcement of our Company. For further information about the relevant cornerstone investments, see "Cornerstone Investors."
CONSENT IN RESPECT OF THE PROPOSED SUBSCRIPTION OF H SHARES BY CERTAIN CORNERSTONE INVESTOR WHO IS A CONNECTED CLIENT
《上市规则》附录F1第1C(1)段规定，未经联交所事先书面同意，不得向整体协调人、任何辛迪加成员（整体协调人除外）或任何分销商（辛迪加成员除外）（统称"分销商"，各称"分销商"）的"关联客户"进行分配。
《上市规则》附录F1第1B段规定，就交易所参与者而言，"关联客户"指与该交易所参与者属同一企业集团成员的任何客户。中国国际金融（香港）交易有限公司（"中金交易"）已与本公司及中国国际金融香港证券有限公司（"中金香港证券"）订立基石投资协议。中金交易与中国国际金融股份有限公司（"中金股份"）将与各自的最终客户（"中金交易最终客户（高毅）"）（均为独立第三方）订立一系列跨境delta-one场外掉期交易（"高毅场外掉期"），据此，中金交易将以非全权委托方式持有发售股份以对冲高毅场外掉期，而相关发售股份的经济风险及回报则转移至中金交易最终客户（高毅）。中金交易与中金香港证券（全球发售的独家保荐人、保荐人兼整体协调人及整体协调人兼承销商之一）同属同一企业集团成员。因此，中金交易为中金香港证券的关联客户。
(c) 在全球发售的任何发售股份分配中，中金交易并无亦不会因其与中金香港证券的关系而获得优待，惟相关基石投资协议项下的保证配额除外，且该协议遵循《指引》第4.15章所载原则，即中金交易的基石投资协议所载重要条款对其而言并不比其他基石投资协议中的条款更为优惠；
(f) 基石投资的详情、证券最终实益拥有人的身份及相关结构性产品的详情已于本招股章程中披露，而最终分配的详情将于配发结果公告中披露。
本招股章程由本公司董事（包括招股章程中列名的任何候任董事）共同及个别承担全部责任，其内容符合《公司（清盘及杂项条文）条例》、《证券及期货（股票市场上市）规则》（香港法例第571V章）及《上市规则》的规定，旨在就本集团向公众提供资料。本公司董事经作出一切合理查询后确认，就其所知所信，本招股章程所载资料在各重要方面均属准确完整，并无误导或欺骗成分，亦无其他遗漏事项会令本招股章程任何陈述产生误导。
我们于2025年7月3日就H股在联交所上市、未上市股份转换及全球发售向中国证券监督管理委员会（"中国证监会"）提交备案申请。中国证监会于2025年12月15日确认我们已完成备案。
DEALINGS IN H SHARES ARE EXPECTED TO COMMENCE AT 9:00 A.M. ON JANUARY 13, 2026.
SHARES WILL BE TRADED IN BOARD LOTS OF 200 H SHARES EACH.
Each person acquiring the Hong Kong Offer Shares under the Hong Kong Public Offering will be required to, or be deemed by his/her acquisition of the Hong Kong Offer Shares to, confirm that he/she is aware of the restrictions on offers and sales of the Shares described in this prospectus.
No action has been taken to permit a public offering of the Offer Shares or the general distribution of this prospectus in any jurisdiction other than in Hong Kong. Accordingly, and without limitation to the following, this prospectus may not be used for the purposes of, and does not constitute, an offer or invitation in any jurisdiction or in any circumstances in which such an offer or invitation is not authorized or to any person to whom it is unlawful to make such an offer or invitation for subscription. The distribution of this prospectus and the offering and sale of the Offer Shares in other jurisdictions are subject to restrictions and may not be made except as permitted under the applicable securities laws of such jurisdictions and pursuant to registration with or authorization by the relevant securities regulatory authorities or an exemption therefrom. In particular, the Offer Shares have not been offered and sold, and will not be offered and sold, directly or indirectly, in the PRC.
Persons applying for or purchasing the Offer Shares under the Global Offering are deemed, by their making an application or purchase, to have represented that they are not associates of any of our Directors or existing Shareholders or a nominee of any of the foregoing.
We have applied to the Stock Exchange for the granting of the listing of, and permission to deal in, our H Shares to be issued pursuant to the Global Offering (including any H Shares which may be issued pursuant to the exercise of the Over-allotment Option) and the H Shares to be converted from the Unlisted Shares.
No part of our Shares or loan capital is listed or dealt in on any other stock exchange and no such listing or permission to list is being or proposed to be sought on any other stock exchange as of the date of this prospectus. All the Offer Shares will be registered on the H Share register of members of the Company in Hong Kong in order to enable them to be traded on the Stock Exchange.
Under section 44B(1) of the Companies (Winding Up and Miscellaneous Provisions) Ordinance, any allotment made in respect of any application will be invalid if the listing of, and permission to deal in, the H Shares on the Stock Exchange is refused before the expiration of three weeks from the date of the closing of the application lists, or such longer period (not exceeding six weeks) as may, within the said three weeks, be notified to the Company by or on behalf of the Stock Exchange.
Assuming that the Hong Kong Public Offering becomes unconditional in Hong Kong at or before 8:00 a.m. in Hong Kong on Thursday, January 8, 2026, it is expected that dealings in the H Shares on the Stock Exchange are expected to commence at 9:00 a.m. on Thursday, January 8, 2026. The H Shares will be traded in board lots of 100 H Shares each. The stock code of the H Shares will be 2513.
Subject to the granting of the listing of, and permission to deal in, the H Shares on the Stock Exchange and compliance with the stock admission requirements of HKSCC, the H Shares will be accepted as eligible securities by HKSCC for deposit, clearance and settlement in CCASS with effect from the date of commencement of dealings in the H Shares on the Stock Exchange or on any other date as determined by HKSCC. Settlement of transactions between participants of the Stock Exchange is required to take place in CCASS on the second settlement day after any trading day. All activities under CCASS are subject to the General Rules of HKSCC and HKSCC Operational Procedures in effect from time to time. – 73 –
All necessary arrangements have been made enabling the H Shares to be admitted into CCASS. Investors should seek the advice of their stockbrokers or other professional advisers for details of the settlement arrangements as such arrangements may affect their rights and interests.
You should consult your professional advisors if you are in any doubt as to the taxation implications of subscribing for, purchasing, holding or disposing of, or dealing in, the H Shares or exercising any rights attaching to the H Shares. We emphasize that none of us, the Relevant Persons, any of our or their respective directors, officers or representatives or any other person involved in the Global Offering accepts responsibility for any tax effects or liabilities resulting from your subscription, purchase, holding or disposing of, or dealing in, the H Shares or your exercise of any rights attaching to the H Shares.
All of the H Shares issued pursuant to applications made in the Global Offering will be registered on our H Share register of members to be maintained in Hong Kong by our H Share Registrar, Tricor Investor Services Limited. Our principal register of members will be maintained by us at our head office in the PRC.
Dealings in the H Shares registered in our H Share register of members will be subject to Hong Kong stamp duty. See "Statutory and General Information—E. Other Information—6. Taxation of Holders of H Shares" in Appendix VI to this prospectus. Investors should seek professional tax advice for further details of Hong Kong stamp duty.
Unless determined otherwise by our Company, dividends payable in respect of our H Shares will be paid to the Shareholders listed on the H Share register of members of our Company in Hong Kong, by ordinary post, at the Shareholders' risk, to the registered address of each Shareholder of our Company.
Solely for your convenience, this prospectus contains translations among certain amounts denominated in Renminbi, Hong Kong dollars and U.S. dollars. No representation is made that the amounts denominated in one currency could actually be converted into the amounts denominated in another currency at the rates indicated or at all. Unless indicated otherwise, (i) the translations between Renminbi and U.S. dollars were made at the rate of RMB7.0572 to US$1.00, (ii) the translations between U.S. dollars and Hong Kong dollars were made at the rate of HK$7.7808 to US$1.00, and (iii) the translations between Renminbi and Hong Kong dollars were made at the rate of RMB0.9070 to HK$1.00. Any discrepancies in any table between totals and sums of amounts listed therein are due to rounding.
If there is any inconsistency between the English version of this prospectus and the Chinese translation of this prospectus, the English version of this prospectus shall prevail unless otherwise stated. The English translation of the names of the PRC entities, enterprises, nationals, facilities, regulations in Chinese included in this prospectus is for identification purposes only. To the extent there is any inconsistency between the Chinese names of the PRC entities, enterprises, nationals, facilities, regulations and their English translations, the Chinese names shall prevail. In addition, if there is any inconsistency between the names of any of the entities mentioned in the English version of this prospectus which are not in the English language and their English translations, the names in their respective original language shall prevail.
Certain amounts and percentage figures included in this prospectus have been subject to rounding adjustments. Accordingly, figures shown as totals in certain tables or charts may not be an arithmetic aggregation of the figures preceding them and figures rounded to the nearest thousand, million or billion may not be identical to figures that have been rounded differently to them. – 74 –
| Name | Address | Nationality | |------|---------|-------------| | **Executive Directors** | | | | Dr. Liu Debing (劉德兵) | Unit 1304, Unit 3, Building 3, District 2, Chaoxin Jiayuan Xili, Chaoyang District, Beijing, PRC | Chinese | | Dr. Zhang Peng (張鵬) | Unit 201, 6th Floor, Building 5, Courtyard 5, Liqing Road, Chaoyang District, Beijing, PRC | Chinese | | Ms. Zhang Xiaohan (張笑涵) | Room 603, No. 17, Lane 300, Qinzhou South Road, Xuhui District, Shanghai, PRC | Chinese | | **Non-executive Directors** | | | | Dr. Li Juanzi (李涓子) | Apartment 1701, Unit 4, Building 10, Xueqingyuan, Xueqing Road, Haidian District, Beijing, PRC | Chinese | | Mr. Li Jiaqing (李家慶) | Room 302, No. 8, Lane 800, Jinxiu Road, Pudong New Area, Shanghai, PRC | Chinese | | Mr. Wang Meng (王盟) | | |
Room 101, Unit 2, Building 11 One Sino Park No. 2 Qijia Village South Street Xingshikou Road Haidian District, Beijing PRC
Xiajiayuan Subdistrict Office Chaoyang District, Beijing PRC Note: In accordance with the applicable PRC laws and regulation, the Supervisor will resign as Supervisor of the Company with effect from the completion of Listing.
For further information regarding our Directors and Supervisor, see "Directors, Supervisor and Senior Management" in this prospectus.
China International Capital Corporation Hong Kong Securities Limited 29/F, One International Finance Centre, 1 Harbour View Street Central, Hong Kong
China International Capital Corporation Hong Kong Securities Limited 29/F, One International Finance Centre, 1 Harbour View Street Central, Hong Kong
Guotai Junan Securities (Hong Kong) Limited 27/F, Low Block, Grand Millennium Plaza 181 Queen's Road Central Hong Kong
China Merchants Securities (HK) Co., Limited 48/F, One Exchange Square 8 Connaught Place Central Hong Kong – 76 –
China International Capital Corporation Hong Kong Securities Limited 29/F, One International Finance Centre, 1 Harbour View Street Central, Hong Kong
Guotai Junan Securities (Hong Kong) Limited 27/F, Low Block, Grand Millennium Plaza 181 Queen's Road Central Hong Kong
China Merchants Securities (HK) Co., Limited 48/F, One Exchange Square 8 Connaught Place Central Hong Kong
Futu Securities International (Hong Kong) Limited 34/F, United Centre No. 95 Queensway Admiralty Hong Kong
Zheshang International Financial Holdings Co., Limited 1703-1706, 17/F, Infinitus Plaza 199 Des Voeux Road Central Sheung Wan Hong Kong
As to PRC laws: Tian Yuan Law Firm Unit 509, Tower A Corporate Square 35 Financial Street Xicheng District Beijing PRC
As to Singaporean data protection laws: Bayfront Law LLC 79 Robinson Road #14-01 CapitaSky Singapore 068897
As to International Sanctions laws: King & Wood Mallesons 10F, Building B4, Xinchen Lin-gang Center Lane 9, North Yunjuan Road, Shengang Street Pudong New District, Shanghai PRC – 78 –
As to Hong Kong and U.S. laws: Baker & McKenzie 14/F, One Taikoo Place 979 King's Road Quarry Bay Hong Kong
As to PRC laws: Commerce & Finance Law Offices 12-14th Floor, China World Office 2 No. 1 Jianguomenwai Avenue Beijing PRC
KPMG Certified Public Accountants Public Interest Entity Auditor registered in accordance with the Accounting and Financial Reporting Council Ordinance 8th Floor, Prince's Building 10 Chater Road, Central Hong Kong
Frost & Sullivan (Beijing) Inc., Shanghai Branch Co. Room 2504, Wheelock Square 1717 Nanjing West Road Jing'an District Shanghai PRC
Bank of Communications (Hong Kong) Limited Unit B B/F & G/F, Unit C G/F, 1-3/F, 16/F Room 01 & 18/F, Wheelock House 20 Pedder Street Central, Hong Kong
Mr. Cheng Ching Kit (鄭程傑) (an associate member of The Hong Kong Chartered Governance Institute and The Chartered Governance Institute in the United Kingdom) 40/F, Dah Sing Financial Centre 248 Queen's Road East, Wanchai Hong Kong
Dr. Liu Debing (劉德兵) 10th Floor, Building 9 Yard 1, Zhongguancun East Road Haidian District Beijing PRC
| Year | Revenue (RMB billions) | |------|----------------------| | 2022 | 93.7 | | 2023 | 121.0 | | 2024 | 160.7 | | 2025E | 218.9 | | 2026E | 304.4 | | 2027E | 430.4 | | 2028E | 586.4 | | 2029E | 775.5 | | 2030E | 993.0 |
Note: The size of the AI market in China represents the market players' revenue derived from the provision of AI solutions.
本節所載若干資料及統計數據摘錄自各官方政府刊物、市場數據提供商，以及由我們委託獨立第三方弗若斯特沙利文（Frost & Sullivan）編制的報告。來自官方政府來源的資料未經我們、獨家保薦人、保薦人-整體協調人、整體協調人、聯席全球協調人、聯席賬簿管理人、聯席牽頭經辦人、資本市場中介機構、包銷商或其各自董事、高級職員、僱員、顧問或代理人或參與全球發售的任何其他方獨立核實，且不就其準確性、公平性及完整性作出任何陳述。
第四次工業革命代表著一場深刻的全球性技術變革，其特點是物理世界、數字世界與生物世界的融合。與由蒸汽、電力及計算機和互聯網驅動的前三次工業革命不同，第四次工業革命正以更快的速度、更廣泛的規模，以及對社會和經濟更深遠的影響持續推進。人工智能（AI）是這場革命最關鍵的驅動力和定義性力量之一。它在賦能其他數字技術、實現智能自動化、創造新商業模式和新興產業，以及融合物理與數字領域方面發揮著不可替代的作用。
人工智能是計算機科學的重要分支，是指模擬和延伸人類智能的技術，使機器能夠感知、理解、推理、學習並自主決策。這些能力使機器能夠在無需人工干預的情況下執行複雜任務。人工智能的發展正深刻改變經濟增長、商業運營和人類生活。預計到2030年，人工智能將賦能全球至少20%的日常商業決策，並使全球至少80%的消費者主流智能設備具備人工智能功能，創造超過20萬億美元的人工智能影響經濟體量。
• 算力。算力是人工智能模型訓練和推理的基礎，對於擴展和商業化人工智能技術至關重要。隨著模型日益複雜，模型參數呈指數級增長，對計算資源的需求持續上升。
• 算法。算法是使人工智能能夠模擬乃至超越人類智能的基本構建模塊。算法的優化和創新直接影響學習效率、通用能力和適應性。從早期使用的較簡單算法（如決策樹和支持向量機）到如今的先進方法（如深度神經網絡和基於Transformer的算法架構），算法的演進推動了多模態交互和基於智能體的任務執行等領域的突破。卓越的算法能提升模型性能，並顯著降低對算力的依賴，從而提高人工智能應用的可行性和成本效益。
• 數據。數據是訓練和持續優化人工智能模型的基本要素。人工智能能力的提升有賴於大量高質量的數據輸入。通過深度挖掘數據和識別規律，人工智能模型在感知、理解和決策任務中的表現得到逐步提升。因此，數據的規模、質量和多樣性直接決定了人工智能系統的學習成果和智能水平。
中國人工智能市場規模從2022年的人民幣937億元增長至2024年的人民幣1,607億元，2022年至2024年的複合年增長率為31.0%，預計到2030年將進一步增長至人民幣9,930億元，2024年至2030年的複合年增長率為35.5%。
| 年份 | 收入（人民幣十億元） | |------|------------------| | 2022 | 93.7 | | 2023 | 121.0 | | 2024 | 160.7 | | 2025E | 218.9 | | 2026E | 304.4 | | 2027E | 430.4 | | 2028E | 586.4 | | 2029E | 775.5 | | 2030E | 993.0 |
註：中國人工智能市場規模代表市場參與者提供人工智能解決方案所取得的收入。
AI technology is undergoing a significant shift from traditional discriminative AI to large language models (LLMs). Traditional discriminative AI focuses primarily on recognition and judgment tasks, such as classification, regression and object detection. At its core, it learns the mapping relationship between inputs and labeled outputs to help machines make accurate decisions. LLMs are very large deep learning models that are pre-trained on vast amounts of data and are built upon neural networks with billions to hundreds of billions of parameters, making them capable of understanding and generating natural language and other types of contents to perform a wide range of tasks. As a result, they are capable of handling tasks that traditional discriminative AI methods struggle with. The main differences between discriminative AI and LLMs can be summarized as follows.
| | LLMs | Discriminative AI | |---|---|---| | Main Functions | • Capable of performing various generative tasks including text, image, audio and code generation, and content summarization, translation and style transformation. The core lies in generating semantically coherent content based on contextual understanding. • With autonomy, perception, decision-making and execution capabilities, LLMs can act independently in response to external changes and set goals. | • Focuses on classifying input data into predefined categories or making predictions based on labeled data. Emphasizes recognition and abstraction of existing knowledge. Mainly applied in tasks such as image recognition, voice interaction and intelligent recommendation. |
| | LLMs | Discriminative AI | |---|---|---| | Application Examples | • Generative tools for text, images, audio, video, code and AI agents. | • Facial recognition, speech recognition and media content recommendation. | | Computing Power | • Requires high-performance GPUs or specialized AI chips, and optimized hardware architecture to support large-scale training and inference. Especially during training, it demands massive distributed computing clusters. | • Typically has fewer parameters and can be trained or run using general-purpose CPUs or GPUs. Usually does not require large-scale distributed computing clusters. | | Algorithms | • Primarily based on Transformer architectures with billion or more parameters, along with reinforcement learning techniques. | • Typically uses techniques such as logistic regression, support vector machines, decision trees, random forests and smaller neural networks. | | Data | • Relies mainly on large-scale unstructured data (especially in pre-training), with relatively low dependency on labeled data (mostly needed in later fine-tuning stages). | • Heavily dependent on structured, high-quality labeled data. |
Compared with LLMs, discriminative AI has clear limitations in generalization, semantic understanding and multimodal processing as well as autonomous decision-making and interaction capabilities, making it difficult to meet increasingly complex and diverse application demands.
• Limited semantic understanding and multimodal capabilities. Discriminative AI struggles with deep semantic comprehension, making it difficult to handle multi-turn dialogs and process multimodal information effectively.
• Limited generalization. Discriminative AI is typically designed for specific tasks, making it less adaptable to new tasks and lacking in flexibility and general-purpose applicability.
• Lack of autonomous decision-making. Discriminative AI primarily operates in a passive and reactive manner, lacking the ability to make independent decisions.
AI is currently transitioning from Artificial Narrow Intelligence (ANI), which is limited to specific tasks, toward Artificial General Intelligence (AGI), which represents a sophisticated level of artificial intelligence that matches and even surpasses human capabilities across all cognitive tasks. LLMs are at the heart of this shift and are emerging as the key driving force behind this new era in AI development. With continuous breakthroughs in parameter scale, semantic understanding, multimodal fusion and self-evolution
capabilities, LLMs have overcome the limitations of fragmented traditional discriminative AI applications. They are beginning to show technical potential to approach general intelligence.
• Pre-training stage. Capable of understanding, writing and speaking natural language, with basic language interaction abilities such as text-based conversations, it represents the earliest form of linguistic intelligence.
• Alignment and reasoning stage. With multimodal understanding and output capabilities, these models are aligned with human intentions and are further developed to reason and plan. This enhances safety, reduces hallucination and extends alignment capabilities beyond text to include images, videos, audio, and actions.
• Self-learning stage. Capable of using tools and calling external resources (such as APIs, standard software or physical devices) to solve real-world problems, these models can plan and execute multi-step tasks through self-critique, self-reflection and rumination, marking a transition from closed models to open ecosystem collaboration.
• Self-perception stage. Operating independently of human oversight, these models form their own attitudes and simulated emotions by observing and interpreting behavior on their own.
在商业化进程方面，中国大型语言模型（LLM）消费市场仍处于早期阶段。尽管AI助手和AIGC工具等面向消费者的应用正在兴起，但用户为LLM付费的意愿仍然较低。相比之下，企业场景是中国LLM市场增长的主要驱动力。企业对LLM部署具有明确而强烈的实际需求，并表现出更高的支付能力和更强的业务运营效率。因此，目前中国LLM商业化的重点集中于机构客户的采用。
意识阶段（Consciousness stage）。这些模型对其内部流程和外部环境均具备一定形式的意识，能够探索科学规律并回答哲学问题。这些模型展现出系统性思维和组织能力，可能将自身融入复杂的社会结构，或发展出自组织系统。
根据弗若斯特沙利文（Frost & Sullivan）的研究，目前尚无官方标准，但人工智能行业（无论是中国还是全球）对迈向通用人工智能（AGI）的发展阶段已形成高度趋同的观点。该序列与认知科学中观察到的认知发展阶段密切对应——智能被视为从基本感知与语言能力，逐步发展至推理与问题解决能力的习得，再到自我意识与抽象思维的涌现。这一类比为将人工智能发展理解为反映复杂性与自主性不断提升的多个阶段提供了理论依据。
领先机构进一步验证了这一方法。例如，全球领先的人工智能公司之一OpenAI已阐明了一个类似框架，用以描述人工智能能力的演进——截至本招股说明书日期，该框架同样围绕五个发展层级构建。第一级（Level 1）为对话式人工智能（Conversational AI），指通过自然对话语言与用户交互的系统。第二级（Level 2）为推理式人工智能（Reasoning AI），指能够以相当于博士学历水平执行基本问题解决任务的系统。第三级（Level 3）为自主式人工智能（Autonomous AI），涉及能够独立采取行动、代表用户完成任务的智能体。第四级（Level 4）为创新式人工智能（Innovating AI），涵盖无需人工干预即可生成新想法与创新成果的系统。第五级（Level 5）为组织式人工智能（Organizational AI），代表有时被描述为"超级人工智能"的阶段，人工智能能够执行通常由整个组织承担的全部职能。学术界与科技公司之间日益增强的思想融合表明，五阶段分类体系提供了一个有用且被广泛认可的路线图。
LLM的商业化涉及为企业提供模型能力、工具链支持及训练或微调服务，以及为个人用户提供AIGC（人工智能生成内容）应用。通过云端部署和本地部署等方式，LLM供应商使企业能够在文本生成、语义理解、逻辑推理和多模态交互等方面构建智能化能力。
在商业化进程方面，中国大型语言模型（LLM）消费市场仍处于早期阶段。尽管AI助手和AIGC工具等面向消费者的应用正在兴起，但用户为LLM付费的意愿仍然较低。相比之下，企业场景是中国LLM市场增长的主要驱动力。企业对LLM部署具有明确而强烈的实际需求，并表现出更高的支付能力和更强的业务运营效率。因此，目前中国LLM商业化的重点集中于机构客户的采用。
LLM正日益融入企业业务工作流程，以提升多项职能的效率和生产力。在客户服务方面，由LLM驱动的聊天机器人和虚拟客服基于客户档案和交互历史提供个性化沟通。在营销和销售活动方面，LLM支持自动化营销活动内容生成和个性化消息推送，以提高参与度和转化率。在生产力和工作流程自动化方面，LLM被嵌入生产力软件和工具中，支持电子邮件起草、会议记录生成和文档处理等任务，从而减少在例行工作上花费的时间。通过利用LLM，企业可以自动化重复性、文字密集型流程，做出更明智的数据驱动决策，并生成高度个性化且与情境高度相关的内容。
在收入方面，2024年中国LLM市场规模为人民币53亿元（RMB5.3 billion），其中机构客户贡献人民币47亿元（RMB4.7 billion），个人客户贡献人民币6亿元（RMB0.6 billion）。随着LLM技术的持续进步以及机构和个人客户需求的不断增长，预计市场规模将于2030年增长至人民币1,011亿元（RMB101.1 billion），2024年至2030年的复合年增长率（CAGR）为63.5%。机构客户仍将是主要增长驱动力，预计中国企业LLM市场将于2030年达到人民币904亿元（RMB90.4 billion），2024年至2030年的复合年增长率为63.7%。
| | 2022-2024年复合年增长率 | 2024-2030E年复合年增长率 | |---|---|---| | 合计（Total） | 264.0% | 63.5% | | 企业（Enterprise） | 261.3% | 63.7% | | 消费者（Consumer） | 192.8% | 61.6% |
| 年份 | 企业（Enterprise） | 消费者（Consumer） | 合计 | |---|---|---|---| | 2022 | 0.4 | 0.1 | — | | 2023 | 2.0 | 0.2 | — | | 2024 | 4.7 | 0.6 | 5.3 | | 2025E | 8.5 | 1.1 | 9.6 | | 2026E | 14.7 | 1.9 | 16.6 | | 2027E | 24.5 | 3.1 | 27.6 | | 2028E | 39.4 | 4.9 | 44.3 | | 2029E | 60.9 | 7.4 | 68.3 | | 2030E | 90.4 | 10.7 | 101.1 |
随着LLM技术的持续演进和企业需求的快速增长，中国企业LLM市场的商业化路径日益清晰，并呈现出多元化格局。目前，企业LLM的主要商业模式包括两种方式：云端部署（cloud-based deployment）和本地部署（on-premise deployment），各自针对企业在成本、数据安全和实时业务需求方面的不同需要。云端部署解决方案提供LLM工具链、应用开发平台及模型服务（通常通过API交付）
INDUSTRY OVERVIEW by LLM providers to enterprises through the cloud. On-premise deployment solutions, by contrast, are customized software and services installed within an enterprise's own IT environments. These solutions include locally deployed LLMs, toolchains and tools for model training and fine-tuning. This setup allows organizations to build, train and use LLMs tailored to their needs, all within their own infrastructure.
| | Cloud-based Deployment | On-Premise Deployment | |---|---|---| | Pricing Model | • Subscription-based / Pay-per-use • Embedded in software licensing / hardware bundling | • Based on the specific products and/or services provided to clients | | Typical Clients | • Software application service providers, smart hardware manufacturers and small and medium-sized enterprises | • Large enterprises | | Core Features | • Standardized, easy to integrate, pay-as-you-go and highly flexible • Suitable for lightweight deployment and rapid testing | • Can be customized based on specific requirements, strong data control and high level of security • Suitable for industries with strict data privacy and system stability requirements | | Use Cases | • Enterprises integrate general LLM capabilities via API access | • Enterprises deploy LLMs locally to build proprietary model systems |
In terms of revenue, the size of the enterprise LLMs market in China was RMB4.7 billion in 2024, of which the cloud-based deployment market accounted for RMB0.9 billion and the on-premise deployment market accounted for RMB3.8 billion. By 2030, the enterprise LLMs market in China is estimated to increase to RMB90.4 billion, with cloud-based deployment market accounting for RMB21.3 billion and on-premise deployment market accounting for RMB69.1 billion. The cloud-based enterprise LLMs market in China is estimated to grow at a CAGR of 69.4% from 2024 to 2030, outpacing the average CAGR of the overall enterprise LLMs market in China of the same period.
INDUSTRY OVERVIEW Market Size of Enterprise LLMs Market in China, in terms of revenue (RMB billions)
| | CAGR 2022–2024 | CAGR 2024–2030E | |---|---|---| | Total | 261.3% | 63.7% | | Cloud-based | 163.1% | 69.4% | | On-premise | 306.5% | 62.2% |
| Year | Cloud-based | On-premise | Total | |---|---|---|---| | 2022 | 0.1 | 0.2 | 0.4 (approx.) | | 2023 | 0.5 | 1.5 | 2.0 | | 2024 | 0.9 | 3.8 | 4.7 | | 2025E | 1.7 | 6.8 | 8.5 | | 2026E | 3.1 | 11.6 | 14.7 | | 2027E | 5.4 | 19.1 | 24.5 | | 2028E | 9.0 | 30.4 | 39.4 | | 2029E | 14.2 | 46.7 | 60.9 | | 2030E | 21.3 | 69.1 | 90.4 |
Key participants in the LLM industry value chain include upstream participants, comprising infrastructure and data providers, midstream participants, represented by independent LLM providers, such as the Company, and non-independent LLM providers, and downstream participants, comprising of enterprises and individual consumers. The following chart sets forth key participants in the LLM industry value chain.
| Upstream | Midstream | Downstream | |---|---|---| | Infrastructure and Data Providers | LLM Providers | Enterprises and Consumers | | Infrastructure Software • Database • Others | Independent LLM Providers (*Where the Group is at) • Providers that are characterized by being natively built around LLM technology, products and business models from the early stages of their operations. | Enterprises • Internet • Financial services • Education • Others | | Infrastructure Hardware • Servers • Storages • Network | Non-independent LLM Providers • Providers that already have an established portfolio of products and services before entering the LLM market. They use their existing advantages in other business areas to expand and commercialize their LLM offerings. | Consumers • Individual users | | Data | | |
As enterprises in China advance their digital transformation, cloud-based deployment is gradually becoming an important path for integrating LLM capabilities. Compared to on-premises deployment, cloud-based deployment offers "ready-to-use" and "scalable" features, enabling software vendors, device manufacturers of AI devices (such as smartphones, embodied AI robots and AI glasses) and various enterprises to access LLM services more efficiently and conveniently, thus meeting diverse business needs.
For software and device providers, cloud-based deployment offers an efficient way to embed LLM functions. Through standardized APIs, they can integrate capabilities such as text generation, image
INDUSTRY OVERVIEW generation and voice generation into existing products such as customer service platforms, marketing content generators and service robots, or even create new products based on capabilities of LLMs.
For end-use enterprises, training and deploying LLM typically involves intensive investments in computing resources and complex engineering works, making it hard for most to build end-to-end solutions independently. Under the cloud-based model, LLM providers handle centralized training, deployment, fine-tuning and inference, while enterprises access these capabilities through standard APIs and pay by usage or subscription, significantly reducing the barriers of adoption and upfront investment. Cloud-based deployment improves accessibility and shortens the cycle from planning to commercial deployment.
Demand for highly customized AI solutions. Vertical industries such as manufacturing, law, finance and others have complex operational scenarios with highly specialized requirements that general LLMs often cannot fully meet. On-premise deployment enables enterprises to deeply customize and fine-tune models based on their own business processes, industry knowledge bases and internal data, resulting in more accurate and efficient outputs. This high level of customization is difficult to achieve with standardized cloud-based services. On-premise deployment thus enables enterprises to build competitive AI capabilities tailored to their unique needs.
Growing preference for integrated hardware and software solutions. As enterprise demands for on-premise deployment increase, the market is shifting toward integrated hardware and software solutions that combine model capabilities, computing infrastructure, and deployment tools into unified packages. This trend simplifies the technical complexity and operational burden for enterprises, allowing them to deploy high-performance AI capabilities without deep technical expertise in infrastructure management.
Expanding use cases in specialized and regulated industries. As on-premise LLM capabilities mature, adoption is expected to expand significantly in regulated industries such as finance, healthcare, government, and defense. These sectors are particularly sensitive to data sovereignty and operational control, making on-premise deployment the preferred choice. Over time, as more industry-specific models and tools become available, on-premise deployment is expected to penetrate deeper into enterprise workflows, moving beyond pilot programs to become core operational infrastructure.
The AI-driven enterprise services market encompasses a broad range of applications where artificial intelligence technologies are integrated into business processes to enhance productivity, decision-making, and operational efficiency. This market includes AI-powered tools for customer service, internal knowledge management, software development, data analysis, and workflow automation, among others.
The rapid commercialization of LLMs has catalyzed the growth of this market by making advanced AI capabilities accessible to enterprises of all sizes. Cloud-based deployment models have further lowered barriers to adoption, enabling even small and medium-sized enterprises to integrate AI into their operations without significant upfront investment.
AI-Powered Customer Engagement and Service: AI technologies are increasingly deployed to enhance customer-facing operations, including intelligent customer service chatbots, virtual assistants, and personalized recommendation systems. These tools help enterprises reduce service costs, improve response times, and enhance customer satisfaction.
AI-Powered Knowledge Management and Internal Productivity: Enterprises are leveraging AI to build intelligent knowledge bases, automate document processing, and enhance internal search and retrieval capabilities. These applications help employees access relevant information more efficiently, reducing time spent on repetitive tasks and improving overall productivity.
AI-Powered Software Development Assistance: AI coding assistants and automated testing tools are transforming software development workflows by accelerating code generation, identifying bugs, and improving code quality. These tools enable development teams to deliver software faster and with greater reliability.
AI-Powered Data Analysis and Decision Support: AI tools are being used to process large volumes of structured and unstructured data, extract actionable insights, and support management decision-making. Applications include financial analysis, market intelligence, risk assessment, and operational optimization.
AI-Powered Workflow Automation: AI-driven automation tools are enabling enterprises to streamline repetitive and rule-based processes across functions such as finance, human resources, procurement, and compliance. These tools reduce manual effort, minimize errors, and free up human resources for higher-value tasks.
需要行业特定的定制化。大语言模型应用通常高度专业化，涉及智能客户服务、法律文件审查、金融分析和医疗诊断等领域。本地部署允许企业利用专有数据和特定领域知识库对模型进行私有训练和微调，从而更好地契合业务工作流程，显著提升实际应用效果和价值。
• 开源与商业化的双引擎模式。开源模型与闭源模型将共存并协同发展。开源模型因其灵活性和可及性，很可能持续对企业保持吸引力。它们有助于降低创新成本，使企业更易于开发定制化应用；其开放性也推动技术快速扩散与协作创新，加速模型普及。与此同时，由专业技术团队开发和维护的闭源模型，提供更稳定高效的性能及更完善的支持服务。未来，开源模型将在推动技术创新和促进社区协作方面发挥关键作用，而闭源模型将在商业应用和企业服务领域占据主导地位。这种双引擎发展模式将为企业提供更广泛的选择，以满足不同行业和场景的多元化需求。
• 大语言模型价值链的完善与生态建设。中国企业AI市场的快速增长得益于日趋成熟的价值链支撑——从上游的数据收集、标注及计算资源基础设施，到中游的模型开发与算法优化，再到下游的行业特定应用与商业化落地。该生态系统的协作性不断增强，价值链各环节之间的合作日益深化，共同推动市场的可持续增长。
• 通用基础模型与垂直生态系统并存。市场将演进为通用基础模型与垂直生态系统相结合的结构。通用模型通常由科技巨头提供，将作为数字基础设施，为通用语义理解与生成提供标准化服务；与此同时，面向特定行业的垂直模型与解决方案将相继涌现。例如，在医疗健康领域，大语言模型将与医学影像和电子健康档案融合，构建垂直医疗生态系统；在金融领域，大语言模型将聚焦于风险分析、投资决策等场景，打造金融生态系统。这种并存结构既为企业提供通用模型的广泛能力，又兼顾垂直解决方案的深度功能，推动市场向更加多元化和专业化的方向发展。
中国大语言模型市场的参与者可分为独立服务商和非独立服务商。独立服务商的特点是从早期运营阶段起便围绕大语言模型技术、产品和商业模式原生构建，其业务通常不与行业生态系统中的客户业务产生竞争。相比之下，非独立服务商在进入大语言模型市场之前已拥有完善的产品和服务体系，借助其在其他业务领域的既有优势来拓展和商业化其大语言模型业务。
按2024年收入计算，本公司是中国最大的独立大语言模型服务商，以及中国第二大整体大语言模型服务商。
| 排名 | 公司 | 类型 | 收入（十亿人民币） | 市场份额 | |------|------|------|------------------|---------| | 1 | A公司 | 非独立 | 0.44 | 9.4% | | 2 | 本公司 | 独立 | 0.31 | 6.6% | | 3 | B公司 | 非独立 | 0.30 | 6.4% | | 4 | C公司 | 非独立 | 0.29 | 6.1% | | 5 | D公司 | 非独立 | 0.22 | 4.7% |
消费者服务、智慧城市业务及企业AI解决方案等在内的多种服务，采用项目制及产品销售等多种商业化模式。截至2024年12月31日，该公司拥有约5,000名员工。
2) B公司成立于1999年，总部位于杭州，为同时在香港联合交易所和纽约证券交易所上市的公众公司。该公司向机构客户和个人客户提供包括云计算及AI服务、物流服务、本地生活服务、娱乐服务和电子商务服务等在内的多种服务，采用项目制及佣金抽成等多种商业化模式。截至2024年12月31日，该公司员工人数不足20万人。
3) C公司成立于2014年，总部位于香港，为在香港联合交易所上市的公众公司。该公司是一家主要采用计算机视觉技术的人工智能企业，主要向机构客户提供计算机视觉AI解决方案、汽车解决方案及计算基础设施解决方案等，采用项目制商业化模式。截至2024年12月31日，该公司员工人数不足5,000人。
built an extensive developer community by offering open source foundational model weights, comprehensive model development frameworks and model development tools, attracting developers to build upper-level applications based on their models and tools. This further forms a virtuous cycle where the larger developer base enables rapid technological iteration, which in turn attracts more developers. At the same time, top enterprises have built a rich ecosystem of independent software vendors (ISVs) to enable the joint deployment of their models in diverse vertical industries and application scenarios, further consolidating their market position. Additionally, some leading enterprises form strategic alliances with other companies in the AI industry chain, including hardware manufacturers and cloud providers, to ensure optimal model training and inference performance, which in turn creates system-level barriers.
• Technical barriers. Leading vendors rely on independently developed LLM pre-training frameworks to build a comprehensive and multi-layered model portfolio with both breadth and depth, which forms a complete matrix that ranges from lightweight edge models to flagship models with hundreds of billions of parameters, supporting various application scenarios such as text generation, image understanding, code generation, multimodal interaction, retrieval-augmented generation and video synthesis. These architectures, which cover multi-dimensional task capabilities, not only meet the increasingly diverse needs of clients but also continuously enhance model general capability and training efficiency through iterative development. In addition, leading companies can quickly customize and optimize models based on real client feedback, driving the evolution of foundational technologies through accumulated scenario data and industry demands. This forms a positive feedback loop featuring low training costs, strong model performance and high adaptability to applications. The systematic synergy from the foundational framework and model design to application feedback significantly raises the technical difficulty for new entrants to replicate or surpass, forming the core technological barrier in the LLMs market in China.
• Flexible business models and delivery strategies. Enterprises differ greatly in terms of industry characteristics, data sensitivity, computing power infrastructure and budget levels. Therefore, the ability of vendors to provide flexible and customizable business models as well as diversified delivery strategies has become a key factor in determining their breadth of client coverage and market penetration capabilities. Leading LLM vendors typically offer a range of business models, such as on-demand usage, subscription-based payments and one-time deployment options, to meet the multi-tiered needs of both small and medium enterprises and large enterprises.
• Ecosystem building capabilities. Building a large-scale and well-integrated ecosystem has become a crucial barrier to sustainable competitive advantage for leading vendors. Top enterprises have
built an extensive developer community by offering open source foundational model weights, comprehensive model development frameworks and model development tools, attracting developers to build upper-level applications based on their models and tools. This further forms a virtuous cycle where the larger developer base enables rapid technological iteration, which in turn attracts more developers. At the same time, top enterprises have built a rich ecosystem of independent software vendors (ISVs) to enable the joint deployment of their models in diverse vertical industries and application scenarios, further consolidating their market position. Additionally, some leading enterprises form strategic alliances with other companies in the AI industry chain, including hardware manufacturers and cloud providers, to ensure optimal model training and inference performance, which in turn creates system-level barriers.
developed comprehensive ecosystem networks encompassing developer communities, hardware partners, industry clients and public sector users. This expands the application scope and deployment efficiency of their model capabilities, and significantly raises the barriers for new entrants. In the developer ecosystem, some vendors have taken the lead in open-sourcing their models, continuously iterating them while building active communities, thereby accumulating a vast developer base and extending their technological reach. Upstream, leading vendors work closely with major computing chip manufacturers to enable high-efficiency adaptation across multiple hardware platforms and optimize model inference and training performance. Downstream, they collaborate with independent software vendors and key clients in industries such as finance, healthcare, government and manufacturing to jointly develop intelligent solutions, accelerating real-world adoption. This "multi-directional expansion and closed-loop feedback" ecosystem model, which connects open-source communities, software-hardware synergy and industry applications, not only accelerates technical optimization and product iteration but also achieves a deep integration of resources, capabilities and clients, forming a unique ecosystem barrier that is difficult to replicate.
• Talent barriers. The LLMs market in China has a high demand for talent, especially for experts with deep technical backgrounds and extensive experience. Leading industry players have attracted top-tier talent and built strong technical teams, while new entrants face intense competition in acquiring skilled professionals.
In connection with the Global Offering, we have engaged Frost & Sullivan to conduct a detailed analysis and prepare an industry report on the markets in which we operate (the "Frost & Sullivan Report"). Services provided by Frost & Sullivan include market assessments, competitive benchmarking, and strategic and market planning for a variety of industries. We have agreed to a total of RMB600,000 in fees and expenses for the preparation and use of the Frost & Sullivan Report. The payment of such an amount was not contingent upon our successful Listing or on the results of the Frost & Sullivan Report. Apart from the Frost & Sullivan Report, we have not commissioned any other industry report in connection with the Global Offering.
We have extracted certain information from the Frost & Sullivan Report in this section, as well as in the sections headed "Summary," "Risk Factors," "Business," "Financial Information" and elsewhere in this prospectus to provide our potential investors with a more comprehensive presentation of the industries in which we operate. Unless otherwise noted, all of the data and forecasts contained in this section are derived from the Frost & Sullivan Report, various official government publications and other publications. Frost & Sullivan prepared its report based on its in-house database, independent third-party reports and publicly available data from reputable industry organizations. Where necessary, Frost & Sullivan contacts companies operating in the industry to gather and synthesize information in relation to the market, prices and other relevant information. Frost & Sullivan believes that the basic assumptions used in preparing the Frost & Sullivan Report, including those used to make future projections, are factual, correct and not misleading. Frost & Sullivan has independently analyzed the information, but the accuracy of the conclusions of its review largely relies on the accuracy of the information collected. Frost & Sullivan's research may be affected by the accuracy of these assumptions and the choice of these primary and secondary sources.
REGULATORY OVERVIEW REGULATIONS ON INFORMATION INDUSTRY Regulations on the Application of Artificial Intelligence Technologies
Based on our understanding of the regulatory regime of artificial intelligence technologies and as advised by our PRC Legal Advisor, we are subject to the following AI-related laws and regulations ("AI-related Laws and Regulations").
On November 25, 2022, the CAC, the MIIT and the Ministry of Public Security jointly promulgated the Administrative Provisions on Deep Synthesis in Internet-based Information Services (《互联网信息服务深度合成管理规定》), which took effect on January 10, 2023. These provisions impose certain compliance obligations on service providers that use deep synthesis technology to provide internet-based information services, including but not limited to establishing databases to identify illegal or malicious information, labeling information generated from using deep synthesis technologies and verifying users' real identities before allowing them to use deep synthesis information publishing services. Deep synthesis technology refers to the use of generative synthetic algorithms such as deep learning and virtual reality to create online content, including text, images, audios, videos, virtual scenes and other related content. In addition, deep synthesis service providers with public opinion attributes or social mobilization capabilities shall conform with filing requirements, modification procedures or cancelation of filing in accordance with the Administrative Provisions on Recommendation Algorithms in Internet-based Information Services (《互联网信息服务算法推荐管理规定》). Where deep synthesis service providers develop new products, new applications and new functions with public opinion attributes or social mobilization capabilities, they shall conduct a security assessment in accordance with the relevant provisions. As our cloud-based AI model and agent solutions would be considered as internet-based information services using deep synthesis technology, we are subject to the compliance requirements of these provisions.
On July 10, 2023, the CAC and six other ministries jointly published the Interim Administrative Measures on Generative AI Services (《生成式人工智能服务管理暂行办法》) ("Interim Measures on Generative AI"), which came into effect on August 15, 2023. The Interim Measures on Generative AI apply to the provision of generative content such as text, images, audios and videos to the public within the territory of China by utilizing generative AI technology. As our AI model and agent solutions are generative AI services, we are subject to the Interim Measures on Generative AI.
The Interim Measures on Generative AI impose various obligations on generative AI service providers, from content filtering and control to protecting the personal information of users. Specifically, (i) service providers shall be responsible as producers of the online information. Upon discovering any "illegal content," service providers shall promptly take appropriate measures, such as suspending generation, halting transmission, removing relevant content, implement rectification measures, including model optimization training, and report the foregoing to competent authorities. Service providers shall also add marks to generated content such as images and videos; (ii) service providers shall sign service agreements with users, protect users' personal information and fulfill the obligation to protect users' input information and usage records in accordance with the law. Service providers shall not collect unnecessary personal information, unlawfully retain input information or usage records that can identify users' identities, or unlawfully provide users' input information or usage records to others; (iii) service providers shall specify and publicly disclose the applicable users, scenarios and purposes of their services. Service providers shall guide users to scientifically and rationally understand and use generative AI technology in compliance with applicable laws. Service providers shall implement effective measures to prevent minors from developing excessive dependence on or addiction to generative AI services; and (iv) service providers shall, upon discovering that users are engaging in illegal activities through generative AI services, promptly implement appropriate measures, including issuing warnings, restricting service functions, suspending or terminating services in compliance with relevant laws or as agreed, preserve relevant records and report to the competent authorities.
The Interim Measures on Generative AI provide special requirements for data training activities such as pre-training and fine-tuning. In particular, (i) service providers shall use data and base models from lawful sources; (ii) service providers shall not infringe others' legally vested intellectual property rights; (iii) service providers shall obtain consents from individuals or ensure that other conditions provided by laws and regulations are met if personal information is used; (iv) service providers shall implement effective measures to enhance training data quality, ensuring authenticity, accuracy, objectivity and diversity of the training data; and (v) service providers shall comply with the requirements of relevant laws.
In addition, the Interim Measures on Generative AI require service providers offering generative AI services with public opinion attributes or social mobilization capabilities to conduct security assessments and comply with filing formalities, modification procedures and filing cancellation procedures with the CAC in accordance with the Administrative Provisions on Recommendation Algorithms in Internet-based Information Services (《互聯網信息服務算法推薦管理規定》). To comply with the Measures on Generative AI Services, we have completed eight generative AI filings during the period from August 2023 to June 2025. Furthermore, we have taken measures including but not limited to (i) protecting users' personal information and privacy by providing user agreements and privacy policy which clarify the rights and obligations between parties; (ii) establishing complaint and reporting mechanism to set up a convenient channel for users to file complaints and reports with us; and (iii) taking reasonable disposal measures for illegal information identified during the generative AI service. To identify illegal information during the generative AI service, we established a comprehensive sensitive vocabulary database combined with multilevel review protocols to conduct automated content review. We utilized our models to improve contextual comprehension, enabling detection of illegal information. We also implemented a two-stage review process for high-risk content, including automated initial screening followed by sampled manual re-examination.
On December 31, 2021, the CAC, the MIIT, the Ministry of Public Security and the State Administration for Market Regulation (the "SAMR") jointly promulgated the Administrative Provisions on Internet Information Service Algorithm Recommendation (《互聯網信息服務算法推薦管理規定》), which became effective on March 1, 2022. The Administrative Provisions on Internet Information Service Algorithm Recommendation implement classification and grading management for algorithm recommendation service providers based on various criteria and stipulates that algorithm recommendation service providers with public opinion attributes or social mobilization capabilities shall file with the CAC within ten business days from the date of providing such services. We have completed 14 filings of Internet information services algorithm between June 2023 and November 2025.
On March 7, 2025, the CAC, the MIIT, the Ministry of Public Security and the National Radio and Television Administration jointly promulgated the Measures for the Identification of AI-Generated and Synthesized Content (《人工智能生成合成內容標識辦法》), which became effective on September 1, 2025. The Measures for the Identification of AI-Generated and Synthesized Content requires that all AI-generated text, images, audio, video and virtual scenes be clearly identified with both explicit and implicit markings. Explicit markings, such as visible text prompts and visual indicators, must be presented prominently to users. Implicit markings, such as metadata and digital watermarks, must embed key production details within the content itself. Service providers, platforms and application distributors are responsible for ensuring compliance. Users must declare when content is generated by AI and tampering with any identification marks is expressly prohibited. The measures provides exemptions for non-public research and development activities. Breaches may result in penalties. As we generate and synthesize content utilizing generative AI technologies, we are subject to such regulations. We make explicit and implicit markings on all AI-generated and synthesized content which is subject to such requirements to make sure that such contents are identified pursuant to the foregoing Measures.
As (i) we have completed relevant filings required under AI-related Laws and Regulations, including algorithms filings and Generative AI filings, and (ii) we have not been subject to any administrative penalties, litigations or criminal liabilities or any material regulatory inquiries, formal notice, warnings, sanctions or penalties in related to AI-related Laws and Regulations during the Track Record Period and up to the Latest Practicable Date, our Directors are of the view, and as confirmed by our PRC Legal Advisors, that during the – 96 –
Track Record Period and up to the Latest Practicable Date, (a) there was no material discrepancy between requirements under AI-related Laws and Regulations and our practices from PRC legal prospective, (b) we were not in violations of any applicable AI-regulated Laws and Regulations in all material respects, (c) we were not involved in any material investigations by the CAC or received any material regulatory inquiries, formal notice, warnings, sanctions or penalties in relation to AI-related Laws and Regulations. We do not expect the AI-related Laws and Regulations to have any material adverse impact on our business operation and financial performance. Based on the foregoing and the due diligence work performed by the Sole Sponsor, nothing material has come to their attention that contradicts such view of the Directors.
In accordance with the National Catalog for Guidance on Industrial Restructuring (2024 Version) (《产业结构调整指导目录(2024年本)》), which was promulgated by the National Development and Reform Commission (the "NDRC") on December 27, 2023 and came into effect on February 1, 2024, big data, cloud computing, software and information technology services and blockchain information services are classified as encouraged categories to the extent permitted under relevant PRC laws and regulations.
In accordance with the Decisions on Protection of Internet Security (《关于维护互联网安全的决定》), which were enacted by the SCNPC on December 28, 2000 and amended on August 27, 2009, the following activities conducted through the internet, if constitute a crime according to PRC laws, are subject to criminal punishment: (i) intrusion into a strategically significant computer or system; (ii) intentionally inventing and disseminating destructive programs, such as computer viruses, to attack the computer system and the communications network, thereby destroying the computer system and the communications networks; (iii) violating national regulations, suspending the computer networks or the communication services without authorization; (iv) leaking state secrets; (v) spreading false commercial information; or (vi) infringing intellectual property rights through the internet.
In accordance with the Provisions on Technical Measures for the Internet Security Protection (《互联网安全保护技术措施规定》), which were promulgated by the Ministry of Public Security on December 13, 2005, internet service providers shall take proper measures including anti-virus, data back-up, record keeping of certain information such as the login-in and exit time of users and other related measures, keep records of certain information about their users for at least 60 days and detect illegal information.
In accordance with the Cybersecurity Law of the People's Republic of China (《中华人民共和国网络安全法》), which was promulgated by the SCNPC on November 7, 2016 and came into effect on June 1, 2017, a graded system for cybersecurity protection is adopted, under which network operators are required to perform the obligations of security protection to ensure that the network is free from interference, disruption or unauthorized access, and prevent network data from being disclosed, stolen or tampered. In addition, network operators shall formulate emergency plans and promptly respond and handle security risks, initiate emergency plans, take appropriate remedial measures and report to regulatory authorities in the event comprising cybersecurity threats and provide technical assistance and support to public security and national security authorities for protection of national security and facilitate criminal investigations in accordance with the law.
In accordance with the Administrative Measures for the Classified Protection of Information Security (《信息安全等级保护管理办法》), which was jointly promulgated by the Ministry of Public Security, the State Secrecy Administration, the State Cryptography Administration and the Information Office of the State Council on June 22, 2007 and came into effect on the same date, and the Guide for the Classification
of Information Security and Cybersecurity (《信息安全技術網絡安全等級保護定級指南》), which was promulgated by Standardization Administration of the PRC on April 28, 2020 and came into effect on November 1, 2020, the classified protection of the information security at the national level shall follow the principle of "independent grading and independent protection." Accordingly, the security protection grade of the information system shall be determined by entities operating and using an information system in accordance with the applicable rules.
In accordance with the State Security Law of the People's Republic of China (《中華人民共和國國家安全法》), which was promulgated by the SCNPC on February 22, 1993 and most recently amended on July 1, 2015, the government shall establish a network and information security assurance system, enhance cybersecurity protection capabilities, strengthen innovation in network and information technologies and achieve secure and controllable mastery of core network technologies, critical infrastructure, major information systems and data. The government shall intensify network governance, prevent, suppress and lawfully punish cybercrimes including cyberattacks, network intrusions, cyber espionage and dissemination of illegal or harmful information, thereby safeguarding national cyberspace sovereignty, security and developmental interests.
In accordance with the Criminal Law of the People's Republic of China (《中華人民共和國刑法》), which was promulgated by the National People's Congress on July 1, 1979 and most recently amended on December 29, 2023, a network service provider is subject to criminal liability if such network service provider fails to fulfill its information network security management obligations prescribed by applicable laws or administrative regulations, and refuse to rectify upon being ordered by regulatory authorities if they meet specified circumstances.
In accordance with the Data Security Law of the People's Republic of China (《中華人民共和國數據安全法》), which was promulgated by the SCNPC on June 10, 2021 and came into effect on September 1, 2021, the government safeguards data-related rights and interests of individuals and organizations, encourages the reasonable and effective utilization of data in accordance with the law, ensures the orderly and lawful free flow of data and promotes the development of the digital economy with data as a critical enabler. The government establishes a data classification and grading protection system and data security review system, and conducts national security reviews of data processing activities that affect or may affect national security. Decisions rendered in accordance with the law through such reviews shall be final. Data processing activities shall comply with laws and regulations by establishing and improving a whole-process data security management system, organizing data security education and training and implementing technical measures and other necessary safeguards to ensure data security. When conducting data processing activities via the internet or other information networks, such obligations shall be fulfilled on the basis of compliance with the cybersecurity graded protection system. Any organization or individual that engages data processing activities in violation of the Data Security Law of the People's Republic of China shall bear corresponding civil, administrative or criminal liability depending on the specific circumstances.
In accordance with the Regulations on Protection of Critical Information Infrastructure (《關鍵信息基礎設施安全保護條例》), which were promulgated by the State Council on July 30, 2021 and became effective on September 1, 2021, a critical information infrastructure (the "CII") refers to important network facilities or information systems in important industries or fields such as public communication and information service, energy, communications, water conservation, finance, public services, e-government affairs and national defense science, which may endanger national security, people's livelihood and public interest in case of damage, function loss or data leakage. In addition, competent authorities and administration departments of each important industry and field shall be responsible for formulating determination rules and determining the critical information infrastructure operator in the respective critical industry or field. The result of the determination of critical information infrastructure operator shall be communicated to the operator and reported to the public security department of the State Council.
On December 28, 2021, the CAC and twelve other PRC regulatory authorities jointly revised and promulgated the Cybersecurity Review Measures (《網絡安全審查辦法》), which came into effect on
February 15, 2022 and replaced the previous version. The Cybersecurity Review Measures provide that the purchase of cyber products and services by critical information infrastructure operators (the "CIIOs") and the network platform operators (the "Network Platform Operators") which engage in data processing activities that affect or may affect national security shall be subject to the cybersecurity review by the Cybersecurity Review Office, the department which is responsible for the implementation of cybersecurity review under the CAC. Specifically, CIIOs that purchase network products and services shall anticipate the potential national security risk of products and services after they enter operation. If they affect or may affect national security, a cybersecurity review shall be reported to the Cybersecurity Review Office. In addition, Network Platform Operators holding personal information of more than one million users that seek listing in a foreign country are obliged to apply for a cybersecurity review by the Cybersecurity Review Office.
On July 7, 2022, the CAC promulgated the Measures on Security Assessment of Cross-border Data Transfer (《數據出境安全評估辦法》) (the "Security Assessment Measures"), which came into effect on September 1, 2022. The Security Assessment Measures outline the requirements and procedures for security assessments on outbound transfer of important data or personal information collected within the territory of mainland China. More specifically, the security assessment shall be required provided that: (i) the data transferred out of mainland China is important data; (ii) a critical information infrastructure operator or data processor that processes personal information of more than one million individuals intends to provide personal information overseas; (iii) the data processor who has provided personal information of 100,000 individuals or sensitive personal information of 10,000 individuals to overseas recipients, in each case as calculated cumulatively, since January 1 of the previous year, intends to provide personal information overseas; or (iv) under other circumstances as stipulated by the CAC. The assessment result is valid for two years. In the case of changes to the transfer circumstances, the recipient country's data laws, or other major situational changes, the data processing entity may also need to re-submit an application.
On December 8, 2022, the MIIT issued the Administrative Measures for Data Security in the Industrial and Information Technology Field (Trial Implementation) (《工業和信息化領域數據安全管理辦法(試行)》), which came into effect on January 1, 2023. According to these administrative measures, industrial and telecommunication data processors shall implement data classification and grading on a regular basis. Furthermore, industrial and telecommunication data processors are required to establish and improve a comprehensive data classification management system, implement protective measures corresponding to the types and levels of the data risk level, and apply the highest level of required protection when processing different levels of data simultaneously if it is difficult to take separate protection measures. The measures also impose certain obligations on industrial and telecommunication data processors regarding the implementation of data security work system, key management administration, data collection, storage, usage, transmission, provision, publicity, destruction, safety audits, and emergency planning, among other aspects.
On February 17, 2023, the China Securities Regulatory Commission (the "CSRC") issued the Trial Administrative Measures of the Overseas Securities Offering and Listing by Domestic Companies (《境內企業境外發行證券和上市管理試行辦法》) (the "Overseas Listing Trial Measures") and five supporting guidelines, which came into effect on March 31, 2023. According to the Overseas Listing Trial Measures, (i) provision of personal information, important data to overseas parties in relation to overseas offering and listing of domestic companies shall be in compliance with applicable laws, administrative regulations and relevant state rules; and (ii) overseas offering and listing by domestic companies shall be made in strict compliance with relevant laws, administrative regulations and rules concerning national security in spheres of cybersecurity and data security, and duly fulfill their obligations to protect national security. If the proposed overseas offering and listing necessitates a national security review, relevant security review procedures shall be completed according to law before the application for such offering and listing is submitted to any overseas parties such as securities regulatory agencies and trading venues.
On March 22, 2024, the CAC promulgated the Provisions on Promoting and Regulating Cross-Border Data Flows (《促進和規範數據跨境流動規定》), effective on the date of promulgation. The provisions
provide several exemptions from undergoing data security assessment, obtaining personal information protection certification, or entering into standard contract for outbound transfer of personal information for businesses. These exemptions include, among others, scenarios where a data processor, other than CIIO, has cumulatively transferred overseas the personal information (excluding sensitive personal information) of fewer than 100,000 individuals since January 1 of the current year. A data processor, other than critical information infrastructure operator, shall enter into a standard contract with overseas recipients for the cross-border transfer of personal information, or obtain certification for personal information protection if since January 1 of the current year, the data processor has cumulatively transferred to overseas recipients personal information (excluding sensitive personal information) of more than 100,000 but less than 1,000,000 individuals, or sensitive personal information of less than 10,000 individuals. The provisions also explicitly state that data processors are not required to conduct data security assessment for cross-border transfers of data if the data has not been notified or published as important data by relevant departments or regions.
On September 24, 2024, the Cyber Data Security Regulations (《网络数据安全管理条例》) were promulgated by the State Council, which came into effect on January 1, 2025. The Cyber Data Security Regulations implemented and reiterated the general requirements on data security management from the Cybersecurity Law, the Data Security Law and the Personal Information Protection Law. The specific provisions include, but are not limited to, the following: (i) data processors must establish and improve data security protection, and adopt technical measures such as encryption, backup and access control to protect data from being tampered with, damaged, leaked, illegally obtained or illegally used; (ii) a network data processor shall inform an individual in accordance with the law by way of formulating rules for the processing of personal information prior to the processing of personal information; and (iii) where personal information and important data are provided to third parties, the data processor must agree on the purpose, method, scope, and security protection obligations by contract or other means. The Cyber Data Security Regulations also aligns with the existing relevant laws and regulations on cybersecurity review, the protection of important data and cross-border security management of network data. Network data processors that conduct network data processing activities affecting or may affect national security shall undergo national security review in accordance with relevant state regulations. For data recognized as important data, the relevant regions and departments shall promptly notify or publicly release to the network data processors. Notably, if the data has not been notified or publicly released as important data by relevant regions or departments, there is no need to apply for an outbound important data transfer security assessment before transferring the data overseas.
In accordance with the Decisions on Strengthening the Protection of Online Information (《关于加强网络信息保护的决定》), issued by the SCNPC in 2012 and the Protection Provisions for the Personal Information of Telecommunications and Internet Users (《电信和互联网用户个人信息保护规定》) promulgated by the MIIT in 2013, telecommunication business operators and internet service providers are required to set up their own rules for collecting and use of internet users' information and are prohibited from collecting or using such information without consent from users. Moreover, they shall strictly keep users' personal information confidential and shall not divulge, tamper with, damage, sell or illegally provide others with such information.
On May 28, 2020, the SCNPC issued the PRC Civil Code (《中华人民共和国民法典》), effective on January 1, 2021, which provides that the personal information of natural person shall be protected by law, and the processing of personal information shall be subject to the principles of legitimacy, lawfulness and necessity, with no excessive processing.
The PRC Minor Protection Law (《中华人民共和国未成年人保护法》), which came into effect on June 1, 2021 and last amended in April 2024, provides that online service providers for products and services such as social networking are required to establish special management systems of user duration, access authority and consumption for the minors.
On August 20, 2021, the SCNPC promulgated the PRC Personal Information Protection Law (《中华人民共和国个人信息保护法》), which became effective on November 1, 2021. Pursuant to the PRC Personal Information Protection Law, the handling of personal information shall follow the principles of lawfulness, legitimacy, necessity, shall be for a specific and reasonable purpose and shall be limited to the minimum scope for achieving the purpose. The PRC Personal Information Protection Law specified the rules for handling sensitive personal information. Personal information handlers shall bear responsibility for their personal information handling activities and adopt the necessary measures to safeguard the security of the personal information they handle. If personal information handlers do not handle personal information in accordance with the law, they will be ordered to correct, suspend or terminate the provision of services, and face confiscation of illegal income, fines or other penalties.
On June 27, 2022, the CAC promulgated the Provisions on the Administrative of Account Information of Internet Users (《互联网用户帐号信息管理规定》), which came into effect on August 1, 2022. These provisions apply to the registration, use, and management of internet users' account information by internet information service providers. The provisions stipulate that internet information service providers must, among others, equip themselves with professional and technical capabilities appropriate to the scale of services, establish, improve and strictly implement the authentication of real identity information, verification of account information, security of information content, ecological governance, emergency responses, protection of personal information and other management systems. The internet information service providers shall set up a convenient portal for complaints and whistleblowing at an eye-catching position and timely handle the complaints and whistleblowing of users and the public.
On February 12, 2025, the CAC published the Administrative Measures for the Compliance Audit of Personal Information Protection (《个人信息保护合规审计管理办法》), effective on May 1, 2025. Personal information processors that process personal information of more than ten million individuals shall carry out a compliance audit of personal information protection at least once every two years. Where a personal information processor conducts a compliance audit on personal information protection by itself, it shall have its internal department, or a commissioned professional institution regularly conduct compliance audits on its compliance with laws and administrative regulations in processing personal information. In certain circumstances, national cyberspace administration authorities and other departments performing personal information protection duties may require personal information processors to commission professional institutions to conduct compliance audits on personal information processing activities. When conducting compliance audits, personal information processors shall follow the Guidelines for Personal Information Protection Compliance Audits outlined in these Measures.
The Administrative Provisions on Mobile Internet Application Information Services (《移动互联网应用程序信息服务管理规定》), issued by the CAC on June 28, 2016 and amended on June 14, 2022, provides that mobile internet application providers are prohibited from utilizing applications to engage in activities that may threaten national security, public interests or infringe on the legitimate rights of individuals or organizations. Mobile internet application providers shall fulfill the responsibility for information content management, establish and improve information content security management, information content ecological governance, network data security, personal information protection, protection of minors and other management systems to ensure content security, foster a healthy online environment, and strengthen the protection of user rights. In addition, mobile internet application providers offering information services through mobile internet applications are required to verify the identity of the registered users. If an internet information service provider violates relevant laws and regulations and service agreement with the mobile application distribution platforms, mobile app stores may issue warnings, suspend services, or remove the applications from distribution, and report the violations to governmental authorities.
The Announcement of Conducting Special Supervision against the Illegal Collection and Use of Personal Information by App (《关于开展App违法违规收集使用个人信息专项治理的公告》) jointly issued
by the CAC, the MIIT, the Ministry of Public Security and the SAMR on January 23, 2019, pursuant to which, (i) application operators are prohibited from collecting any personal information irrelevant to their services; (ii) information collection and usage policies should be presented clearly and simply, with voluntary user consent; and (iii) user authorization should not be obtained through coercive methods, such as default settings, bundling clauses, or making consent a prerequisite for service access. Applications operators violating such rules may face corrective orders within a specified period of time, public reporting, suspension of operations for rectification, or revocation of business licenses or operational permits.
The Notice on Further Special Rectification of App Infringing upon Users' Personal Rights and Interests (《關於開展縱深推進App侵害用戶權益專項整治行動的通知》), issued by the MIIT on July 22, 2020, mandates inspections of app service providers for practices including, among others, (i) collecting or using personal information without user consent, collecting or using personal information beyond the scope necessary for providing services, or forcing users to receive advertisements; (ii) excessively or frequently requesting user permissions or launching third-party apps; and (iii) deceiving or misleading users into downloading apps or sharing personal information. The notice specifies a regulatory inspection period, requiring non-compliant entities to rectify violations within five business days, failing which the MIIT may publicly announce violations, remove the apps from the app stores, or impose other administrative penalties.
The Notice from the Ministry of Industry and Information Technology on Further Improving the Service Capabilities of Mobile Internet Applications (《工業和信息化部關于進一步提升移動互聯網應用服務能力的通知》), issued by the MIIT on February 6, 2023 and came into effect on the same day, requires app providers to inform users of personal information processing rules in a concise, clear and easy-to-understand manner. Any changes to these rules must be promptly communicated to users. The data processors shall clearly outline the purpose, method, and scope of sensitive personal information processing activities, maintain a list of collected personal information, and shall not use methods such as default ticking, reducing text size, or lengthy texts to induce users' consent.
The Notice on the Filing of Mobile Internet Applications (《關於開展移動互聯網應用程序備案工作的通知》), which was promulgated by the MIIT on July 21, 2023 and came into effect on the same day, mandates that app operators providing internet information services in China complete filing procedures as required by the PRC Anti-telecommunications Network Fraud Law (《中華人民共和國反電信網絡詐騙法》) and the Administrative Measures on Internet Information Service (《互聯網信息服務管理辦法》). App operators shall complete filing procedures with the provincial-level communications administration bureau where they are registered. The filing process is conducted online via the National Internet Basic Resources Management System (國家互聯網基礎資源管理系統), with applications submitted through network access service providers and app distribution platforms, including those for mini-programs and quick applications, for review and approval.
In March 2019, the Foreign Investment Law of PRC (《中華人民共和國外商投資法》) was promulgated by National People's Congress and came into effect on January 1, 2020, which replaced the Sino-Foreign Equity Joint Venture Enterprise Law of PRC, the Sino-Foreign Cooperative Joint Venture Enterprise Law of PRC and the Wholly Foreign-owned Enterprise Law of PRC, and became the legal foundation for foreign investment in the PRC.
According to the Foreign Investment Law of the PRC, the State Council shall promulgate or approve a list of special administrative measures for access of foreign investments. The Foreign Investment Law grants national treatment to foreign-invested entities, except for those investing in the industries specified in the Negative List.
The NDRC and the Ministry of Commerce of the People's Republic of China (the "MOFCOM") jointly issued the Special Administrative Measures (Negative List) for Foreign Investment Access (2024
version) (《外商投資准入特別管理措施 (負面清單) (2024年版)》) in November 2024 (the "2024 Negative List"). The 2024 Negative List sets out the industries in which foreign investments are prohibited or restricted. Pursuant to the Foreign Investment Law and the 2024 Negative List, foreign investors shall not make investments in prohibited industries as specified in the 2024 Negative List, while foreign investments must satisfy certain conditions stipulated in the 2024 Negative List for investment in restricted industries. Industries not listed in the 2024 Negative List are generally deemed "permitted" for foreign investments.
On December 30, 2019, the MOFCOM and SAMR promulgated the Measures for the Reporting of Foreign Investment Information (《外商投資信息報告辦法》) (the "Reporting Measures"), which came into effect on January 1, 2020. The Reporting Measures regulate information reporting relating to foreign investment in the PRC. Pursuant to the Reporting Measures, foreign investors and foreign-invested enterprises who directly or indirectly carry out investment activities in the PRC shall report investment information to the competent departments of commerce by submitting initial reports, change reports, cancelation reports and annual reports.
On December 19, 2020, the NDRC and the MOFCOM jointly promulgated the Measures on the Security Review of Foreign Investment (《外商投資安全審查辦法》), effective on January 18, 2021, setting forth provisions concerning the security review mechanism on foreign investment, including the types of investments subject to review, review scopes and procedures, among others. The following categories of foreign investment shall require foreign investors or domestic relevant parties to proactively declare the security review prior to implementation: (i) investments in national defense security-related sectors such as military industry, military supporting industries, or within geographical areas adjacent to military facilities and military industrial installations; and (ii) investments in major sectors affecting national security, including critical agricultural products, energy and resources, major equipment manufacturing, critical infrastructure, significant transportation services, essential cultural products and services, major information technology and internet products/services, critical financial services, key technologies and other vital domains, where the investor acquires actual control over the invested enterprise.
The Telecommunications Regulations of the PRC (《中華人民共和國電信條例》) (the "Telecommunications Regulations") as promulgated by the State Council in 2000 and most recently amended in 2016, requires telecommunications service providers to obtain operating licenses prior to the commencement of their operations. The Telecommunications Regulations distinguish "infrastructure telecommunications services" from "value-added telecommunications services." According to the Telecommunications Regulations, operators of value-added telecommunications services shall obtain value-added telecommunications business operation licenses from the MIIT or its provincial branches prior to the commencement of such services.
Moreover, the Administrative Measures on Telecommunications Business Operating Licenses (《電信業務經營許可管理辦法》), revised by the MIIT in July 2017, set forth more provisions to specify the types of licenses required to operate value-added telecommunications services, the qualifications and procedures for applying such licenses and the administration and supervision of such licenses. According to the Catalog of Telecommunications Business (《電信業務分類目錄》) most recently amended by the MIIT on June 6, 2019, internet information services fall within the value-added telecommunications services.
Pursuant to the Regulations for the Foreign-Invested Telecommunications Enterprises (《外商投資電信企業管理規定》), which was promulgated by the State Council in 2001 and most recently amended on March 29, 2022, the ultimate foreign equity ownership in a foreign-invested value-added
telecommunication enterprise is subject to a cap of 50%, except as otherwise stipulated by the state. The 2024 Negative List further states that the equity ratio of foreign investment in the value-added telecommunications enterprises operating telecommunication services opened up pursuant to the PRC WTO Accession Commitments shall not exceed 50% except for the investment in e-commerce operation business, domestic multi-party communication business, information storage and re-transmission business or call center business. As we provide Internet information services to our users through our mobile apps and websites, we are classified as a value-added telecommunication enterprise under the category of Internet content provider, which is restricted from foreign investments under the 2024 Negative List. Therefore, the ultimate foreign equity ownership in our Company is subject to and in compliance with a cap of 50%.
2006年，工业和信息化部（现称"工信部"）发布了《信息产业部关于加强外商投资经营增值电信业务管理的通知》（the Circular on Strengthening the Administration of Foreign Investment in and Operation of Value-added Telecommunications Business），根据该通知，在中国电信服务行业进行投资的外国投资者必须设立外商投资企业并申请电信业务经营许可证，同时，持有增值电信业务经营许可证的中国境内企业被禁止以任何方式向外国投资者出租、转让或出售该许可证，亦不得向在中国非法开展增值电信业务的外国投资者提供任何协助，包括提供资源、场地或设施。
《互联网信息服务管理办法》（the Administrative Measures on Internet Information Services）由国务院于2000年9月颁布，并于2024年12月6日最新修订，就互联网信息服务的提供制定了相关指导原则。该办法将互联网信息服务分为经营性互联网信息服务和非经营性互联网信息服务，并规定经营性互联网内容提供服务的经营者必须向相应电信主管部门申请取得经营范围含互联网信息服务的增值电信业务经营许可证。
根据《中华人民共和国专利法》（the Patent Law of the PRC）及其实施细则，专利分为发明专利、外观设计专利和实用新型专利。发明专利权、外观设计专利权和实用新型专利权的期限分别为20年、15年和10年，均自申请日起计算。发明或实用新型须满足新颖性、创造性和实用性三项标准方可获得专利授权。国家知识产权局负责审查和批准专利申请。未经专利权人授权实施专利的行为构成侵犯专利权，须向专利权人承担赔偿责任，并可能被处以罚款，甚至承担刑事责任。
根据《中华人民共和国商标法》（the Trademark Law of the PRC）及其实施细则，注册商标的有效期为10年，商标所有人可申请续展，每次续展期限为10年。商标许可协议须向商标局备案。侵犯注册商标专用权的行为包括但不限于未经注册商标权利人许可，在相同或类似商品上使用与注册商标相同或近似的商标，侵权方将被责令立即停止侵权行为，并可能被处以罚款。
The infringing party may also be held liable for the right holder's damages, which will be equal to gains obtained by the infringing party or the losses suffered by the right holder as a result of the infringement, including reasonable expenses incurred by the right holder for stopping the infringement.
According to the Copyright Law of the PRC (《中華人民共和國著作權法》) and implementation rules, Chinese citizens, legal persons, or other organizations shall, whether published or not, own copyright in their works, which include, among others, works of literature, art, natural science, social science, engineering technology and computer software. Copyright owners of protected works enjoy personal rights and property rights with respect to publication, authorship, alteration, integrity, reproduction, distribution, lease, exhibition, performance, projection, broadcasting, dissemination via information network, production, adapting, translation, compilation and other rights.
Pursuant to the Regulation on Computer Software Protection (《計算機軟件保護條例》) and the Measures for the Registration of Computer Software Copyright (《計算機軟件著作權登記辦法》), the National Copyright Administration is mainly responsible for the registration and management of software copyright in China and recognizes the China Copyright Protection Center as the software registration organization. The China Copyright Protection Center shall grant certificates of registration to computer software copyright applicants in compliance with the regulations of the Measures for the Registration of Computer Software Copyright and the Regulation on Computers Software Protection.
The Measures on Administration of Internet Domain Names (《互聯網域名管理辦法》) provide that the MIIT shall supervise the domain names services nationwide and publicize the PRC domain name system. After completion of the registration procedures, the applicant will become the holder of the relevant domain name. In November 2017, the MIIT promulgated the Notice on Regulating the Use of Domain Names in Providing Internet-based Information Services (《工業和信息化部關于規範互聯網信息服務使用域名的通知》), which became effective on January 1, 2018. Pursuant to the notice, the domain name used by an internet-based information service provider in providing internet-based information services must be registered and owned by such provider in accordance with the law. If the internet-based information service provider is an entity, the domain name registrant must be the entity (or any of the entity's shareholders), or the entity's principal or senior manager.
The major PRC laws and regulations that govern employment relationship are the Labor Law of the PRC (《中華人民共和國勞動法》), the Labor Contract Law of the PRC (《中華人民共和國勞動合同法》) and the Implementation Rules of the Labor Contract Law of the PRC (《中華人民共和國勞動合同法實施條例》). Pursuant to the aforementioned laws and regulations, labor relationships between employers and employees must be executed in written form. The laws and regulations above impose stringent requirements on the employers in relation to entering into fixed-term employment contracts, hiring of temporary employees and dismissal of employees. As prescribed under the laws and regulations, employers shall ensure their employees have the right to rest and the right to receive wages no lower than the local minimum wages. Employers must establish a system for labor safety and sanitation that strictly abides by state standards and provide relevant education to its employees. Violations of the Labor Contract Law of the PRC and the Labor Law of the PRC may result in the imposition of fines and other administrative liabilities and/or incur criminal liabilities in the case of serious violations.
Pursuant to the Social Insurance Law of the PRC (《中華人民共和國社會保險法》), which was promulgated by the SCNPC on October 28, 2010 and amended on December 29, 2018, enterprises and institutions in the PRC shall provide their employees with welfare schemes covering pension insurance, unemployment insurance, maternity insurance, occupational injury insurance, medical insurance and other welfare plans. The employer shall apply to the local social insurance agency for social insurance registration within 30 days from the date of its establishment. And it shall, within 30 days from the date of employment, apply to the social insurance agency for social insurance registration for the employee. Employer shall make contribution to the social insurance in full and on time, and failure to do so will result in late fees and fines. Meanwhile, the Interim Regulation on the Collection and Payment of Social Insurance Premiums (《社會保險費徵繳暫行條例》) promulgated by the State Council on January 22, 1999 and revised on March 24, 2019 prescribes the details concerning the collection and payment of social insurance.
According to the Regulation on the Administration of Housing Provident Fund (《住房公積金管理條例》), which was implemented on April 3, 1999 and most recently revised on March 24, 2019, any newly established entity shall make deposit registration at the housing accumulation fund management center within 30 days as at its establishment. After that, the entity shall open a housing accumulation fund account for its employees in an entrusted bank. Within 30 days as at the date an employee is recruited, the entity shall make deposit registration at the housing accumulation fund management center and go through the formalities of opening housing provident fund accounts on behalf of its employees. Any entity that fails to make deposit registration of the housing accumulation fund or fails to open a housing accumulation fund account for its employees shall be ordered to complete the relevant procedures within a prescribed time limit. Any entity failing to complete the relevant procedure within the time limit will be fined RMB10,000 to RMB50,000. Any entity that fails to make payment of housing provident fund within the time limit or has a shortfall in payment of housing provident fund will be ordered to make the payment or make up the shortfall within the prescribed time limit, otherwise, the housing provident management center is entitled to apply for compulsory enforcement with the People's Court.
The Overseas Listing Trial Measures adopt a filing and regulatory regime to regulate the direct and indirect overseas listing of securities of PRC domestic enterprises. If a domestic enterprise fails to comply with the filing procedures as required, or if it is listed outside of PRC despite being prohibited from doing so, the CSRC may order the domestic enterprise to rectify the situation, issue a warning and impose a fine of not less than RMB1,000,000 and not more than RMB10,000,000. A warning may be given to the directly responsible officer and other directly responsible persons and a fine of not less than RMB500,000 and not more than RMB5,000,000 shall be imposed. If the controlling shareholder or the actual controller of the domestic enterprise organizes or instructs to engage in the above illegal acts, he may be liable to a fine of not less than RMB1,000,000 and not more than RMB10,000,000. The Overseas Listing Trial Measures also stipulate that in the event of any material events such as change of control, investigation or punishment by the overseas securities supervisory authority or relevant authorities, change of listing status or transfer of listing segment, or termination of listing on its own initiative or compulsory termination of listing after the issuer's overseas listing, the issuer shall report the specific circumstances to the CSRC within three working days from the date of the announcement of the relevant event.
On February 24, 2023, the CSRC, the Ministry of Finance, the National Administration of State Secrets Protection and the National Archives Administration of China jointly promulgated the Provisions on Strengthening the Confidentiality and Archives Administration Concerning the Overseas Securities Offering and Listing by Domestic Enterprises (《關于加强境內企業境外發行證券和上市相關保密和檔案管理工作的規定》) (the "Confidentiality and Archives Administration Provisions"), which became effective on March 31,
2023. According to the Confidentiality and Archives Administration Provisions, if a domestic joint stock company with a direct overseas listing or a domestic operating entity with an indirect overseas listing provides or publicly discloses through its overseas listed entity, documents or information involving state secrets or secrets of the work of state organs, or other documents or information the disclosure of which would adversely affect national security or public interests, the corresponding procedures shall be strictly complied with in accordance with the relevant state regulations.
"Full circulation" represents listing and circulating on the Stock Exchange of the domestic unlisted shares of an H-share listed company, including unlisted domestic shares held by domestic shareholders prior to overseas listing, unlisted domestic shares additionally issued after overseas listing, and unlisted shares held by foreign shareholders. On November 14, 2019, CSRC announced the Guidelines on Application for "Full Circulation" of Domestic Unlisted Shares of H-share Companies (《H股公司境內未上市股份申請全流通業務指引》) ("Full Circulation Guidelines"), which were amended on August 10, 2023. As regulated in the Full Circulation Guidelines, shareholders of domestic unlisted shares have the flexibility to jointly decide the amount and proportion of shares that will be included in the circulation application. This decision should be reached through mutual consultation, ensuring compliance with relevant laws, regulations and policies governing state-owned asset administration, foreign investment and industry regulation. Meanwhile, the H-share listed company corresponding to these shares may be authorized to file for "full circulation" with the CSRC. An unlisted domestic joint stock company may file with the CSRC for "full circulation" at the time of its offering and listing overseas. After domestic unlisted shares are listed and circulated on the Stock Exchange, they may not be converted back domestically. Pursuant to the Overseas Listing Trial Measures, for a domestic company directly offering shares and listing overseas, shareholders of its domestic unlisted shares applying to convert such shares into shares listed and offering on an overseas trading venue shall conform to relevant regulations promulgated by the CSRC. Additionally, they are required to authorize the domestic company to submit the conversion application to the CSRC on their behalf. On December 31, 2019, China Securities Depository and Clearing Corporation Limited and Shenzhen Stock Exchange jointly announced the Measures for Implementation of H-share "Full Circulation" Business (《H股全流通業務實施細則》) (the "Measures for Implementation"). The businesses of cross-border share transfer registration, maintenance of deposit and holding details, transaction entrustment and instruction transmission, settlement, management of settlement participants, services of nominal holders, and other details in relation to the H-share "full circulation" business are subject to the Measures for Implementation.
According to the PRC Enterprise Income Tax Law (《中華人民共和國企業所得稅法》) promulgated by the National People's Congress on March 16, 2007, and most recently amended on December 29, 2018 and the Enterprise Income Tax Implementation Regulations (《中華人民共和國企業所得稅法實施條例》) promulgated by the State Council on December 6, 2007, and most recently amended on December 6, 2024 and effective from January 20, 2025, enterprises are categorized as resident enterprises and non-resident enterprises. Resident enterprises refer to enterprises that are established in accordance with the law within the territory of China, or enterprises that are established in accordance with foreign (regional) laws but have their actual place of management within the territory of China. Non-resident enterprises refer to an enterprise that is established in accordance with foreign (regional) laws and whose actual place of management is not within the territory of China, but which has established establishments or premises within the territory of China, or enterprises that have not established establishments or premises within the territory of China but derive income from sources within China. Resident enterprises are subject to a uniform 25% enterprise income tax rate on their worldwide income. The enterprise income tax rate is reduced by 20% for qualifying small low-profit enterprises. The high-tech enterprises that need full support from the PRC government will enjoy a 15% tax rate reduction for enterprise income tax.
Pursuant to the Arrangement Between Mainland China and the Hong Kong Special Administrative Region for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with respect to Taxes on Income (《内地和香港特别行政区关于对所得避免双重征税和防止偷漏税的安排》) and relevant protocols, which were promulgated by the State Taxation Administration on August 21, 2006, came into effect on December 8, 2006, the withholding tax rate of 5% applies to dividends paid by a PRC company to a Hong Kong company if such Hong Kong company directly holds at least 25% of the equity interests in a PRC company; otherwise the 10% withholding tax rate applies.
Pursuant to the Administrative Measures on Entitlement of Non-resident Taxpayers to Preferential Treatment under Tax Treaties (《非居民纳税人享受协定待遇管理办法》), which was promulgated by the State Taxation Administration on October 14, 2019, and came into effect on January 1, 2020, non-resident taxpayers who enjoy treaty benefits shall follow the procedure of "self-assessment, declaration and enjoyment, and retention of documentation for inspection." If a non-resident taxpayer self-assesses that it meets the conditions for enjoying treaty benefits, it may declare and enjoy such benefits during its own tax filing, or through the withholding agent during withholding declarations. Such taxpayer must simultaneously collect and retain relevant documentation for inspection in accordance with such measures, and is subject to subsequent administration by the tax authorities.
Pursuant to the Provisional Regulations on Value-added Tax of the PRC (《中华人民共和国增值税暂行条例》), which was promulgated by the State Council on December 13, 1993 and most recently amended on November 19, 2017, and the Detailed Rules for the Implementation of the Interim Regulations of the PRC on Value-added Tax (《中华人民共和国增值税暂行条例实施细则》), which was promulgated by the Ministry of Finance on December 25, 1993 and most recently amended on October 28, 2011, all entities or individuals in the PRC engaged in the sale of goods, processing services, repair and replacement services, and the provision of services, sales of intangible assets, real estate and importation of goods are required to pay value-added tax ("VAT"). Unless otherwise provided, taxpayers engaged in the provision of services and sales of intangible assets are subject to a tax rate of 6%.
According to the Circular on Policies for Simplifying and Consolidating Value-added Tax Rates (《关于简并增值税税率有关政策的通知》), announced by the Ministry of Finance and the State Taxation Administration on April 28, 2017, the structure of value-added tax rates will be simplified from July 1, 2017 and the 13% VAT rate will be canceled. The scope of goods with an 11% tax rate and the provisions for deducting input tax are specified. According to the Circular on Adjusting Value-added Tax Rates of Ministry of Finance and the State Administration of Taxation (《财政部、国家税务总局关于调整增值税税率的通知》) announced by the Ministry of Finance and the State Taxation Administration on April 4, 2018, effective from May 1, 2018, with respect to VAT taxable sales or imported goods of a taxpayer, the previously applicable 17% and 11% tax rates are adjusted to 16% and 10% respectively. According to the Announcement on Relevant Policies for Deepening Value-Added Tax Reform (《关于深化增值税改革有关政策的公告》) announced by the Ministry of Finance, the State Taxation Administration and the General Administration of Customs on March 20, 2019, effective from April 1, 2019, with respect to VAT taxable sales or imported goods of a VAT general taxpayer, the originally applicable VAT rate of 16% shall be adjusted to 13%, and the originally applicable VAT rate of 10% shall be adjusted to 9%.
The principal regulation governing foreign currency exchange in China is the Foreign Exchange Administration Regulations of the PRC (《中华人民共和国外汇管理条例》), which was promulgated by the State Council on January 29, 1996 and most recently amended on August 5, 2008. Pursuant to this regulation and other PRC rules and regulations on currency conversion, Renminbi is freely convertible for
payments of current account items, such as trade and service-related foreign exchange transactions and dividend payments, but not freely convertible for capital account items, such as direct investment, loan or investment in securities outside China unless prior approval or registration of the State Administration of Foreign Exchange (the "SAFE") or its local counterpart is obtained.
According to the Notice on Relevant Issue Concerning the Administration of Foreign Exchange for Overseas Listing (《关于境外上市外汇管理有关问题的通知》) issued by the SAFE on December 26, 2014, the domestic companies shall register the overseas listing with the foreign exchange control bureau located at its registered address in 15 working days after completion of the overseas listing and issuance. The funds raised by the domestic companies through overseas listing may be repatriated to China or deposited overseas, provided that the intended use of the fund shall be consistent with the contents of the public disclosure documents.
According to the Guidelines for the Foreign Exchange Business under the Capital Account (2024) (《资本项目外汇业务指引(2024年版)》) issued by SAFE on April 3, 2024, in principle, the funds raised by overseas listings of domestic companies should be repatriated to China in a timely manner, and can be repatriated in RMB or foreign currency. The use of funds shall be consistent with the relevant contents listed in the publicly disclosed documents. Domestic companies using the funds raised from overseas listings to carry out overseas direct investment, overseas securities investment, overseas lending and other businesses shall comply with the relevant foreign exchange management regulations.
The Notice on Simplifying Direct Investment-related Foreign Exchange Administration Policies (《关于进一步简化和改进直接投资外汇管理政策的通知》), which was issued by SAFE on February 13, 2015 and was amended on December 30, 2019, allowing entities and individuals to apply for foreign exchange registrations through qualified banks. Under SAFE's supervision, these banks can directly review applications. On March 30, 2015, SAFE released the Circular on Reforming Settlement Management of Foreign Capital in Foreign-invested Enterprises (《关于改革外商投资企业外汇资本金结汇管理方式的通知》). This circular mandates discretionary foreign exchange settlement for foreign-invested enterprises, enabling them to settle foreign exchange capital based on operational needs, subject to document verification. The circular emphasizes authentic and self-use principles within the enterprise's business scope, barring use for payments beyond business scope, securities investment (unless specified), Renminbi entrust loans, inter-enterprise borrowings, or real estate expenses (except for self-use by foreign-invested real estate enterprises).
The Circular on Reforming and Standardizing the Foreign Exchange Settlement Management Policy of Capital Account (《关于改革和规范资本项目结汇管理政策的通知》) (the "SAFE Circular 16"), was promulgated by SAFE on June 9, 2016 and was amended on December 4, 2023. Pursuant to the SAFE Circular 16, foreign currency earnings in capital account that relevant policies of willingness exchange settlement have been clearly implemented on (including the recalling of raised capital by overseas listing) may undertake foreign exchange settlement in the banks according to actual business needs of the domestic institutions. The tentative percentage of foreign exchange settlement for foreign currency earnings in capital account of domestic institutions is 100%, subject to adjustment by SAFE in due time in accordance with international revenue and expenditure situations.
On January 26, 2017, SAFE promulgated the Circular on Further Improving Reform of Foreign Exchange Administration and Optimizing Genuineness and Compliance Verification (《关于进一步推进外汇管理改革完善真实合规性审核的通知》), which stipulates several capital control measures with respect to the outbound remittance of profit from domestic entities to offshore entities, including: (i) banks should check board resolutions regarding profit distribution, the original version of tax filing records, and audited financial statements pursuant to the principle of genuine transactions; and (ii) domestic entities should hold income to account for previous years' losses before remitting the profits. Moreover, pursuant to this circular, domestic entities should make detailed explanations of the sources of capital and utilization arrangements, and provide board resolutions, contracts, and other proof when completing the registration procedures in connection with an outbound investment.
The Notice for Further Advancing the Facilitation of Cross-border Trade and Investment (《關於進一步促進跨境貿易投資便利化的通知》) was promulgated by the SAFE on October 23, 2019, and was amended on December 4, 2023. Among others, it allows all FIEs to use Renminbi converted from foreign currency denominated capital for equity investments in China, as long as the equity investment is genuine, does not violate applicable laws, and complies with the negative list on foreign investment.
According to the Circular of the State Administration for Foreign Exchange on Optimizing Foreign Exchange Administration to Support the Development of Foreign-related Business (《國家外匯管理局關于優化外匯管理支持涉外業務發展的通知》) promulgated with effect from April 10, 2020 by the SAFE, the reform of facilitating the payments of incomes under the capital accounts shall be promoted nationwide. Under the prerequisite of ensuring true and compliant use of funds and compliance and complying with the prevailing administrative provisions on use of income from capital accounts, enterprises which satisfy the criteria are allowed to use income under the capital account, such as capital funds, foreign debt and overseas listing, for domestic payment, without the need to provide proof materials for veracity to the bank beforehand for each transaction.
In accordance with the Company Law of the People's Republic of China (《中華人民共和國公司法》), which most recently amended on December 29, 2023, and the Foreign Investment Law of the People's Republic of China (《中華人民共和國外商投資法》), foreign-invested enterprises in the PRC may pay dividends only out of their accumulated profit, if any, determined in accordance with PRC accounting standards and regulations. A PRC company, including foreign-invested enterprise, is required to set aside as general reserves at least 10% of its after-tax profit, until the cumulative amount of such reserves reaches 50% of its registered capital, and shall not distribute any profits until any losses from prior fiscal years have been offset.
According to the PRC Anti-Unfair Competition Law (《中華人民共和國反不正當競爭法》), which was adopted by the SCNPC on September 2, 1993 and most recently amended on June 27, 2025 with effect from October 15, 2025, unfair competition refers to that the operator disrupts the market competition order and damages the legitimate rights and interests of other operators or consumers in violation of the provisions of the PRC Anti-Unfair Competition Law in the production and operating activities. Pursuant to the PRC Anti-Unfair Competition Law, operators must abide by the principle of voluntariness, equality, impartiality, integrity and adhere to laws and business ethics during market transactions. Operators in violation of the PRC Anti-Unfair Competition Law should bear corresponding civil, administrative or criminal liabilities depending on the specific circumstances.
On May 6, 2024, the SAMR issued Interim Provisions Against Unfair Competition in Cyberspace (《網絡反不正當競爭暫行規定》), which took effect on September 1, 2024. Such provisions provide a regulatory basis for preventing and curbing unfair competition acts in cyberspace, maintaining the market order of fair competition, encouraging innovation, protecting the legitimate rights and interests of business operators and consumers, and promoting the standardized, sustainable and sound development of the digital economy.
The U.S. government imposes export controls for national security, foreign policy and other various policy reasons. One of the primary U.S. export control regimes is governed by the EAR, which are administered and enforced by the BIS. BIS is responsible for regulating the export, reexport or transfer (in-country) of a diverse range of goods, software and technology (collectively, "items") including most commercial items, "dual-use" items (i.e., those items having both commercial and military or proliferation applications) and less-sensitive military items.
BIS regulates the export, reexport and in-country transfer of items that are "subject to the EAR." The following items are subject to the EAR: (i) all U.S.-origin items wherever they are located in the world; (ii) any item physically in or moving in transit through the United States or U.S. Foreign Trade Zone (including items of foreign origin); (iii) any foreign-made item containing more than a de minimis amount of certain controlled U.S.-origin content; and (iv) certain foreign made items that are the "direct product" of certain highly-controlled U.S.-origin software or technology (or are the direct product of a plant or major plant component that is itself the direct product of such U.S.-origin software or technology). Generally, foreign-made items that incorporate controlled U.S.-origin content accounting for less than 25% of the value of such items are not subject to the EAR when exported, reexported or transferred (in-country) to any country except for Cuba, Iran, North Korea or Syria (for which the de minimis threshold is 10%), unless the controlled content is of a certain type for which there is no de minimis threshold. For purposes of the de minimis analysis, any item that by itself requires a destination-based license to be exported to, reexported to or transferred (in-country) within the country at issue is considered to be a controlled item.
Items that are subject to the EAR may require a license from the BIS prior to the export, reexport or transfer (in-country) of the item. Violations of U.S. export controls may have serious consequences including, but not limited to, civil monetary penalties of up to $374,474 in 2025 (as periodically adjusted for inflation) or twice the value of the transaction, whichever is greater; or criminal penalties of up to $1 million per violation and/or up to 20 years in prison; loss of access to items subject to the EAR; inclusion on one or more BIS lists of parties of concern; and/or reputational harm. To determine whether a license is required for the export, reexport or transfer (in-country) of an item subject to the EAR, it is necessary to review the following:
• Classification of the Item. BIS maintains the Commerce Control List ("CCL," Supplement 1 to 15 C.F.R. Part 744), which provides descriptions of items under Export Control Classification Numbers ("ECCNs"). Items that are not described under an ECCN on the CCL, but that are nevertheless subject to the EAR, are designated EAR99;
• Country-based License Requirements. Each ECCN identifies reasons for control, which indicate licensing requirements to certain destinations based on a review of the Commerce Country Chart (Supplement 1 to 15 C.F.R. Part 738). EAR99 items do not have specific reasons for control and do not require a license for most destinations, except for countries subject to comprehensive U.S. sanctions. Items on the CCL may also be subject to restrictions imposed by comprehensive U.S. sanctions;
• Restricted Parties. BIS maintains lists of companies, organizations and individuals that may be subject to additional license requirements, regardless of the classification of the item; the Entity List is one of such lists; and
• Prohibited End Uses. BIS may also impose license requirements, regardless of an item's classification, if the item will be used in certain end uses, which are often related to certain proliferation activities.
Our Company and our nine subsidiaries (together, the "Listed Entities") were added to the Entity List on January 16, 2025 (the "Entity List Addition"). The addition of the Listed Entities to the Entity List restricts our ability to purchase or otherwise access certain items that are subject to the EAR. However, the Entity List restrictions do not apply to other entities within the Group that are legally distinct from the Listed Entities. Public guidance in the form of a Frequently Asked Question ("FAQ") No. 134 issued by BIS has clarified that "[s]ubsidiaries, parent companies, and sister companies are legally distinct from listed entities [and,] . . . [t]herefore, the licensing and other obligations imposed on a listed entity by virtue of its being listed do not per se apply to its subsidiaries, parent companies, sister companies, or other legally distinct affiliates that are not listed on the Entity List." Similarly, BIS has advised that "[t]he Entity List license requirements do not extend to parent companies unless the applicable listing for the company so
| Subsidiary | Place of incorporation | Date of incorporation | Registered capital | Principal activities | |---|---|---|---|---| | Beijing Knowledge Xingyao | PRC | September 24, 2024 | RMB300,000,000 | Provision of large model related services | | Tianjin Knowledge Atlas | PRC | October 25, 2024 | RMB950,000,000 | Provision of large model related services |
states" (see BIS FAQ 136). In order to address EAR-related risks after the Entity List Addition, we have put in place a series of export control compliance measures for the entire Group, in abundance of caution.
Pursuant to Section 744.16(a) of the EAR, a person "may not, without a license from the BIS, export, reexport or transfer (in-country) any items included in the License Requirement column of an entity's entry on the Entity List . . . when that entity is a party to a transaction as described in §748.5(c) through (f) of the EAR," i.e., the purchaser, intermediate consignee, ultimate consignee or end-user. As specified on the Entity List, the license requirement for exports, reexports or transfers (in-country) to entities listed on the Entity List applies to "all items subject to the EAR." A party (e.g., a supplier to the Company) that exports, reexports or transfers (in-country) an item that is subject to the EAR is strictly liable for violations related to such activity. Any other party to a transaction, including the buyer (e.g., the Company), must also comply with the EAR. Specifically, the EAR provides a basis for liability for activities, including, but not limited to, the following: (i) causing, aiding or abetting a violation; (ii) soliciting or attempting a violation; (iii) conspiring to bring about or engage in a violation; (iv) misrepresenting or concealing facts to the U.S. government in connection with activities subject to the EAR; (v) acting with the intent to evade the EAR; (vi) failing to comply with recordkeeping requirements of the EAR; and (vii) acting with "knowledge" that a violation has occurred or is about to occur. The EAR defines "knowledge" as including "positive knowledge that the circumstance exists or is substantially certain to occur," as well as "an awareness of a high probability of its existence or future occurrence," which is "inferred from evidence of the conscious disregard of facts known to a person and is also inferred from a person's willful avoidance of facts." 15 C.F.R. § 772.1.
We are a leading AI company in China, dedicated to pursuing innovation toward artificial general intelligence (AGI).
Our history can be traced back to June 2019, when our Company was founded by a team with deep academic roots at Tsinghua University, including Dr. Liu, Dr. Li, Dr. Tang, Dr. Xu, Dr. Zhang and Mr. Wang. Our Company draws on the academic rigor and innovative spirit of its founding team to advance cutting-edge developments in AI. For the background and the relevant industry experience of our founders, see "Directors, Supervisor and Senior Management" in this prospectus.
The following table sets forth the key milestones of our corporate and business development.
| Year | Milestone events | |---|---| | 2019 | • Our Company was established. | | 2020 | • We commenced development of General Language Model (GLM), our pre-training framework. | | 2021 | • We released GLM-10B, our first ten-billion-parameter pre-trained large model. | | 2022 | • We launched our Model-as-a-Service (MaaS) product development and commercialization platform. • We released GLM-130B, an open-source large model. • We released CodeGeeX, our high-performance coding model, in September. | | 2023 | • We released billion-parameter foundation chat model, ChatGLM, in March and an open-source ChatGLM-6B, in March. • We released Zhipu QingYan, which was one of the first domestic large model products to complete regulatory filing in China. | | 2024 | • We released GLM-4, our foundation model with agent orchestration capabilities. • We released GLM-4V, our visual comprehension foundation model. • We released CogVideoX, our next-generation video generation model, and our Zhipu Qingying mobile app. • We released GLM-4-Plus, the latest and most advanced foundation model in the GLM-4 series, and introduced the AI video call function in Zhipu QingYan. • We released GLM-4-Voice, our end-to-end emotional voice generation model, and AutoGLM, our foundational agent model for autonomous device operation. | | 2025 | • We released GLM-Z1, our reflection model for advanced reasoning tasks. • We released GLM-Realtime, an end-to-end model supporting real-time audiovisual interaction, humming, two-minute memory retention, and "Function Call" features. • We released AutoGLM Rumination, our first AI agent combining deep research and operational execution, enabling autonomous multi-step reasoning and action. |
| Year | Milestone events | |---|---| | | • We released GLM-4.5, GLM-4.5V, an updated version of AutoGLM (also known as "AutoGLM 2.0") and GLM-4.6. |
The following table sets forth the information on our principal subsidiaries that have made material contribution to our operating results and financial position during the Track Record Period.
| Subsidiary | Place of incorporation | Date of incorporation | Registered capital | Principal activities | |---|---|---|---|---| | Beijing Knowledge Xingyao | PRC | September 24, 2024 | RMB300,000,000 | Provision of large model related services | | Tianjin Knowledge Atlas | PRC | October 25, 2024 | RMB950,000,000 | Provision of large model related services |
截至最後實際可行日期，我們的其他附屬公司主要從事大模型相關服務等業務。我們其他附屬公司的主要業務摘要載於本招股書附錄一所載會計師報告附註1。
2019年6月11日，本公司由北京聯排、劉博士、唐博士、李博士、許博士、張博士、北京中科創星硬科技創業投資合夥企業（有限合夥）（"中科創星"）及華控技術轉移有限公司（"華控技術"）依據中國法律共同出資設立，登記資本為人民幣1,000萬元，公司形式為有限責任公司。北京聯排由劉博士以普通合夥人身份持有約92.70%的權益，王少蘭先生、張鈸（我們的首席科學家）及孟兆軍分別以有限合夥人身份持有約4.21%、2.65%及0.44%的權益。孟兆軍為獨立第三方，在本公司早期進行了財務投資，並通過北京聯排繼續持有少數股東股份。中科創星及華控技術均為獨立第三方。
成立時，本公司由北京聯排、劉博士、唐博士、李博士、許博士、張博士、中科創星及華控技術分別持有約49.38%、0.80%、21.97%、4.00%、0.80%、0.40%、10.67%及11.99%的股權。
2021年8月，北京聯排將其在本公司的登記資本人民幣80萬元及人民幣40萬元分別以約人民幣80萬元及人民幣40萬元的對價轉讓予我們的員工持股平台匯匯及智登。
上述轉讓完成後，本公司由北京聯排、劉博士、唐博士、李博士、許博士、張博士、匯匯、智登、中科創星及華控技術分別持有約37.38%、0.80%、21.97%、4.00%、0.80%、0.40%、8.00%、4.00%、10.67%及11.99%的股權。
有關匯匯及智登的詳情，請參閱下文"——員工持股平台"。
為資助我們的戰略增長並擴大股東基礎，我們已進行多輪上市前投資。詳情請參閱"——上市前投資"。
2025年3月26日，本公司由有限責任公司整體變更為股份有限公司。根據本集團截至2025年1月31日的經審計淨資產，本公司按71.9047597:1的比例將有限責任公司的全部出資額折合為股份有限公司的股份。整體變更完成後，本公司登記資本為人民幣36,224,375元，分為36,224,375股，每股面值人民幣1.00元，由全體股東按其於整體變更前在本公司的各自權益比例認購。
根據股東於2025年6月28日通過的決議，股份將於上市前即時按一拆十的比例進行拆分，股份面值將由每股人民幣1.00元變更為每股人民幣0.10元。股份拆分完成後，本公司登記股本將為人民幣40,281,069元，共402,810,690股，每股面值人民幣0.10元。
根據我們的中國法律顧問的意見，截至最後實際可行日期，本公司已就上述股權轉讓及增加以及股份的發行、轉讓及拆分在所有重大方面向相關地方市場監督管理局辦理了所有必要的登記或備案手續，且上述股權轉讓及增加以及股份的發行、轉讓及拆分在所有重大方面均符合適用的中國法律法規。
我們於2022年1月完成天使輪融資（"天使輪融資"），北京創新智源科技有限公司（"創新智源"）以評估價值約為人民幣2,037萬元（估值基準日為2021年3月31日）的知識產權作為對價，認購本公司人民幣526,316元的登記資本。
| 認購方 | 認購登記資本（人民幣元） | 對價（人民幣元） | |---|---|---| | 棗莊通智股權投資合夥企業（有限合夥）（"通智投資"） | 706,767 | 47,000,000 | | 深圳市達晨創鴻私募股權投資企業（有限合夥）（"達晨創鴻"） | 556,391 | 37,000,000 | | 北京華控產業投資基金（有限合夥）（"北京華控"） | 300,752 | 20,000,000 | | 北京榮品投資管理有限公司（"榮品投資"） | 300,752 | 20,000,000 | | 北京將門創業投資中心（有限合夥）（"將門創投"） | 150,376 | 10,000,000 | | 凌雲光技術股份有限公司（"凌雲光"） | 150,376 | 10,000,000 | | 南京圖靈一期創業投資合夥企業（有限合夥）（"南京圖靈"）(1) | 75,188 | 5,000,000 | | 深圳市財智創贏私募股權投資企業（有限合夥）（"財智創贏"） | 45,113 | 3,000,000 | | 合計 | 2,285,715 | 152,000,000 |
Note: (1) In August 2023, Nanjing Turing transferred all registered capital in our Company held by it to Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業（有限合夥)) ("Xinglian Zhaoji"). Such capital transfer was conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company.
We completed the series B1 financing (the "Series B1 Financing") in February 2023. Details of our Series B1 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業（有限合夥)) ("Junlian Xiangdao") | 878,017 | 125,000,000 | | Suzhou Qiming Rongqian Equity Investment Partnership (Limited Partnership) (蘇州啟明融乾股權投資合夥企業（有限合夥)) ("Qiming Rongqian")(1) | 343,628 | 48,921,079 | | Kunshan Qiming Rongkai Equity Investment Partnership (Limited Partnership) (昆山市啟明融凱股權投資合夥企業（有限合夥)) ("Qiming Rongkai")(1) | 239,375 | 34,078,921 | | Total | 1,461,020 | 208,000,000 |
Note: (1) In February 2023, Innovation Zhiyuan transferred registered capital of our Company in the amount of RMB78,662 and RMB54,797 to Qiming Rongqian and Qiming Rongkai, respectively. Such capital transfers were conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company.
We completed the series B2 financing (the "Series B2 Financing") in August 2023. Details of our Series B2 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Stage 1 (completed in July 2023)** | | | | Tianjin Sankuai Technology Co., Ltd. (天津三快科技有限公司) ("Tianjin Sankuai") | 1,721,731 | 300,000,000 | | **Stage 2 (completed in August 2023)** | | | | Junlian Xiangdao | 167,333 | 31,383,890 | | Xinglian Zhaoji | 146,551 | 27,486,153 | | Qingdao Huakong Growth Equity Investment Partnership (Limited Partnership) (青島華控成長股權投資合夥企業（有限合夥)) ("Qingdao Huakong") | 67,706 | 12,698,553 | | Qiming Rongqian | 65,847 | 12,349,835 |
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | Dachen Chuanghong | 80,462 | 15,091,018 | | Qiming Rongkai | 45,870 | 8,603,079 | | Caizhi Chuangying | 13,330 | 2,500,000 | | Total | 2,308,830 | 410,112,528 |
We completed the series B3 financing (the "Series B3 Financing") in January 2024. Details of our Series B3 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Stage 1 (completed in November 2023)** | | | | Trend Mega Limited ("Trend Mega") | 1,134,991 | 255,304,308 | | Shanghai Yunya Enterprise Management Consulting Co., Ltd. (上海雲玡企業管理諮詢有限公司) ("Shanghai Yunya")(1)(2) | 666,846 | 150,000,000 | | Junlian Xiangdao | 140,787 | 31,668,557 | | Hangzhou Guanghe II Venture Capital Partnership (Limited Partnership) (杭州光合貳期創業投資合夥企業（有限合夥)) ("Hangzhou Guanghe") | 86,629 | 20,000,000 | | **Stage 2 (completed in January 2024)** | | | | Tianjin Heyuan Youze Yihao Venture Capital Partnership (Limited Partnership) (天津合遠優擇壹號創業投資合夥企業（有限合夥)) ("Tianjin Heyuan") | 346,515 | 80,000,000 | | Xinglian Zhaoji | 160,264 | 37,000,158 | | Qingdao Huakong | 151,600 | 34,999,900 | | Total | 2,687,632 | 608,973,038 |
Notes: (1) In October 2023, CAS Star transferred registered capital of our Company in the amount of RMB342,314 to Shanghai Yunya. Such capital transfer was conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company. (2) Shanghai Yunya (i) transferred registered capital of our Company in the amount of RMB171,157 and RMB333,423 to Hangzhou Duoxiang Network Technology Co., Ltd. (杭州多項網絡科技有限公司) ("Duoxiang Network") in October 2023 and December 2023,
respectively; and (ii) transferred registered capital of our Company in the amount of RMB138,971 to Shanghai Feiya Technology Co., Ltd. (上海飛玡科技有限公司) ("Shanghai Feiya") in August 2024. Such capital transfers were conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company.
We completed the series B4 financing (the "Series B4 Financing") in August 2024. Details of our Series B4 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | Guangxi Tencent Venture Capital Co., Ltd. (廣西騰訊創業投資有限公司) ("Tencent Investment") | 694,854 | 200,000,000 | | Beijing Shunying Equity Investment Partnership (Limited Partnership) (北京順贏股權投資合夥企業（有限合夥)) ("Beijing Shunying") | 521,141 | 150,000,000 | | TAL Education (Beijing) Co., Ltd. (欣欣相融教育科技（北京）有限公司) ("TAL") | 347,427 | 100,000,000 | | Beijing Xiaofeng Technology Co., Ltd. (北京小鋒科技有限公司) ("Xiaofeng Technology") | 347,427(1) | 100,000,000 | | Xiamen Yaheng Venture Capital Investment Fund Partnership (Limited Partnership) (廈門雅恆創業投資基金合夥企業（有限合夥)) ("Xiamen Yaheng") | 347,427 | 100,000,000 | | Ningbo Meishan Free Trade Port Zone Mingheng Enterprise Management Consulting Partnership (Limited Partnership) (寧波梅山保稅港區明恒企業管理諮詢合夥企業（有限合夥)) ("Ningbo Mingheng") | 347,427 | 100,000,000 | | Xinglian Zhaoji | 334,372 | 96,242,408 | | Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金（北京）合夥企業（有限合夥)) ("Social Security Zhongguancun Innovation Fund") | 330,056 | 95,000,000 | | Dachen Chuanghong | 186,488 | 53,676,779 | | Qingdao Huakong | 175,282 | 50,451,444 | | Shanghai Yunya | 173,714 | 50,000,000 | | Duoxiang Network | 173,714 | 50,000,000 | | Tianjin Heyuan | 116,175 | 33,438,672 | | Qiming Rongqian | 102,387 | 29,470,000 | | Qiming Rongkai | 71,327 | 20,530,000 | | Total | 4,269,218 | 1,228,809,303 |
Note: (1) The registered share capital was subscribed by Beijing Huapin Borui Network Technology Co., Ltd. (北京華品博睿網絡技術有限公司) ("Huapin Borui"), the parent company of Xiaofeng Technology.
We completed the series B5-1 financing (the "Series B5-1 Financing"), series B5-2 financing (the "Series B5-2 Financing") and series B5-3 financing (the "Series B5-3 Financing", together with Series B5-1 Financing and Series B5-2 Financing, the "Series B5 Financing") in November 2024. Details of our Series B5 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Series B5-1 Financing** | | | | Shanghai Feiya | 465,090 | 200,000,000 | | Wuxi Yunhui Digital Economy Investment Management Partnership (Limited Partnership) (無錫雲暉數字經濟投資管理合夥企業 (有限合夥)) ("Wuxi Yunhui") | 232,545 | 100,000,000 | | Junlian Xiangdao | 155,038 | 66,670,000 | | Shenzhen Zhaoshang Shuke Innovation Private Equity Fund Partnership (Limited Partnership) (深圳市招商數科創新私募股權投資基金合夥企業 (有限合夥)) ("Zhaoshang Shuke") | 139,527 | 60,000,000 | | Social Security Zhongguancun Innovation Fund | 77,507 | 33,330,000 | | Xiaofeng Technology | 69,763(1) | 30,000,000 | | Xiamen Yaheng | 34,882 | 15,000,000 | | **Series B5-2 Financing** | | | | P7 China Holdings PCC Limited ("Prosperity7") | 495,390 | 213,030,000 | | AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金 (有限合夥)) ("AI Fund") | 465,090 | 200,000,000 | | Shanghai Feiya | 465,090 | 200,000,000 | | Social Security Zhongguancun Innovation Fund(2) | 118,474 | 50,946,934 | | Beijing Lianrong Zhiyuan Equity Investment Partnership (Limited Partnership) (北京聯融致遠股權投資合夥企業 (有限合夥)) ("Lianrong Zhiyuan") | 116,272 | 50,000,000 | | Suzhou Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯錦帆創業投資合夥企業 (有限合夥)) ("Junlian Jinfan")(2) | 116,173 | 49,957,107 |
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Series B5-3 Financing** | | | | Hubei Yangtze CITIC Technology Mobile Communication Industry Investment Fund Partnership (Limited Partnership) (湖北長江中信科移動通信技術產業投資基金合夥企業 (有限合夥)) ("5G Fund") | 209,290 | 90,000,000 | | **Total** | **3,160,131** | **1,358,934,041** |
Notes: (1) The registered share capital was subscribed by Huapin Borui. In November 2024, Huapin Borui transferred all registered capital of our Company held by it to Xiaofeng Technology as its intra-group transaction. Such capital transfer did not involve any increase of registered capital by our Company. (2) In November 2024, (i) Dr. Liu transferred registered capital of our Company in the amount of RMB14,397 and RMB14,683 to Junlian Jinfan and Social Security Zhongguancun Innovation Fund, respectively; and (ii) Beijing Lianpai (which was controlled by Dr. Liu as its general partner) transferred registered capital of our Company in the amount of RMB44,671 and RMB45,555 to Junlian Jinfan and Social Security Zhongguancun Innovation Fund, respectively. Such capital transfers were conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company.
We completed the series B6-1 (the "Series B6-1 Financing") and series B6-2 financing (the "Series B6-2 Financing") in December 2024. Details of our Series B6-1 Financing and Series B6-2 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Series B6-1 Financing** | | | | Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業 (有限合夥)) ("Zhongguancun Science City") | 826,211 | 500,000,000 | | Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發展基金合夥企業 (有限合夥)) ("Daxing Industrial Fund") | 495,726 | 300,000,000 | | Xinglian Zhaoji | 160,285 | 97,000,000 | | Junlian Xiangdao | 89,700 | 54,284,213 | | Social Security Zhongguancun Innovation Fund | 34,231 | 20,715,787 | | **Series B6-2 Financing** | | | | Tianjin Haihe Fuxin Youda Venture Capital Fund Partnership (Limited Partnership) (天津海河富新優達創業投資基金合夥企業 (有限合夥)) ("Haihe Fuxin Youda Fund") | 1,569,800 | 950,000,000 | | **Total** | **3,175,953** | **1,922,000,000** |
We completed the series B6-3 (the "Series B6-3 Financing"), series B6-3+ (the "Series B6-3+ Financing") and series B6-4 financing (the "Series B6-4 Financing", together with Series B6-1 Financing, Series B6-2 Financing, Series B6-3 Financing, Series B6-3+ Financing, the "Series B6 Financing") in May 2025. Details of our Series B6-3 Financing, Series B6-3+ Financing and Series B6-4 Financing are set out below:
| Subscriber | Registered capital subscribed for (RMB) | Consideration (RMB) | |---|---|---| | **Series B6-3 Financing** | | | | | 578,347 | 350,000,000 | | | 495,726 | 300,000,000 | | | 240,784 | 145,715,755 | | | 165,242 | 100,000,000 | | | 48,390 | 29,284,213 | | | 330,484 | 200,000,000 | | | 214,815 | 130,000,000 | | | 826,211 | 500,000,000 | | | 165,242 | 100,000,000 | | | 165,242 | 100,000,000 |
Series B6-4 Financing Zhihui Linghang Venture Capital Partnership (Limited Partnership) (上海浦東智慧領航創業投資合夥企業 (有限合夥)) ("Zhihui Linghang") | 826,211 | 500,000,000
Notes: (1) On September 3, 2024, the Company, Tianchuang Capital, Dr. Liu and Dr. Tang entered into a convertible note agreement, pursuant to which the Company agreed to issue a convertible note in the principal amount of RMB130 million to Tianchuang Capital. (2) The Company, Dr. Liu and Dr. Tang entered into convertible note agreements respectively with Lenovo Venture Capital on February 7, 2025, Xiarui Investments on January 25, 2025 and Zhuhai Huafa on January 26, 2025, pursuant to which, the Company agreed to issue convertible notes to each of Lenovo Venture Capital, Xiarui Investments and Zhuhai Huafa in the principal amount of RMB100 million, respectively. On March 17, 2025, the Company, Dr. Liu and Dr. Tang entered into an additional convertible note agreement with Zhuhai Huafa for a principal amount of RMB400 million, on substantially the same terms as the above. (3) Pursuant to the convertible note agreements entered into between the Company and each of Tianchuang Capital, Lenovo Venture Capital, Xiarui Investments and Zhuhai Huafa, the notes are convertible into the Shares of the Company at the subscription price under the Series B6 Financing or Series B6-3 Financing (as the case may be) upon the completion of the Company's conversion into a joint stock company on March 26, 2025. Pursuant to the capital increase agreements entered into between the Company and each of Tianchuang Capital, Lenovo Venture Capital, Xiarui Investments and Zhuhai Huafa on May 13, 2025, the conversion of the convertible notes issued by each of Tianchuang Capital, Lenovo Venture Capital, Xiarui Investments and Zhuhai Huafa shall be effected by way of issuance of newly issued Shares at the agreed conversion prices, with each of Tianchuang Capital, Lenovo Venture Capital, Xiarui Investments and Zhuhai Huafa receiving equity interests in the Company corresponding to the principal amounts of the notes converted and enjoying all rights of the other Pre-IPO Investors in the relevant round. The conversion was completed on May 29, 2025.
In August 2025, each of Beijing Lianpai, Zhaoshang Shuke, Lianrong Zhiyuan, Tianjin Heyuan, Hangzhou Guanghe, Innovation Zhiyuan, Rongpin Investments, Xiaofeng Technology and Xinglian Zhaoji transferred the Shares in our Company held by them respectively to certain Pre-IPO Investors as set forth as follows. Such capital transfers were conducted based on arm's length negotiation between the Shareholders and did not involve any increase of registered capital by our Company.
| Transferor | Transferee | Number of Shares transferred | Consideration (RMB) | |---|---|---|---| | Beijing Lianpai | Shanghai Shanchuang Zhizhi Venture Capital Partnership (Limited Partnership) (上海杉創智至創業投資合夥企業（有限合夥）) ("Shanchuang Zhizhi") | 42,920 | 20,000,000 | | Beijing Lianpai | Guangdong Hengqin Shanzhi Investment Partnership (Limited Partnership) (廣東橫琴杉智投資合夥企業（有限合夥）) ("Hengqin Shanzhi") | 42,920 | 20,000,000 | | Zhaoshang Shuke | Zhongxiao Ruizheng (Shanghai) Venture Capital Partnership (Limited Partnership) (中小銳正（上海）創業投資合夥企業（有限合夥）) ("Zhongxiao Ruizheng") | 40,303 | 20,000,000 |
| Transferor | Transferee | Number of Shares transferred | Consideration (RMB) | |---|---|---|---| | Lianrong Zhiyuan | Shandong Fuhong New Energy Industry Investment Fund Partnership (Limited Partnership) (山東孚弘新能源產業投資基金合夥企業（有限合夥）) ("Shandong Fuhong") | 30,210 | 14,999,602 | | Tianjin Heyuan | Nanjing Rongjia Xingpu Equity Investment Partnership (Limited Partnership) (南京融嘉星譜股權投資合夥企業（有限合夥）) ("Rongjia Xingpu") | 72,967 | 35,325,956 | | Tianjin Heyuan | Qingdao Lingtou Future Venture Capital Fund Partnership (Limited Partnership) (青島瓴投未來創業投資基金合夥企業（有限合夥）) ("Lingtou Future") | 68,548 | 33,186,520 | | Hangzhou Guanghe | Rongjia Xingpu | 13,662 | 6,614,044 | | Hangzhou Guanghe | Lingtou Future | 12,834 | 6,213,480 | | Innovation Zhiyuan | Anhui Jia'an Qixin Venture Capital Partnership (Limited Partnership) (安徽嘉岸啟信創業投資合夥企業（有限合夥）) ("Jia'an Qixin") | 165,242 | 80,000,000 | | Rongpin Investments | Hainan Hezun Investment Co., Ltd. (海南何尊投資有限公司) ("Hainan Hezun") | 300,752 | 20,000,000 | | Xiaofeng Technology | Beijing Beijiao United Lingyue No.4 Equity Investment Centre (Limited Partnership) (北京北交聯合羚躍肆號股權投資中心（有限合夥）) ("Lingyue No.4") | 125,997 | 61,000,000 | | Xiaofeng Technology | Beijing Beijiao United Lingyue No.5 Equity Investment Centre (Limited Partnership) (北京北交聯合羚躍伍號股權投資中心（有限合夥）) ("Lingyue No.5") | 39,245 | 19,000,000 | | Xiaofeng Technology | Huahai Jinpu Venture Capital (Jinan) Partnership (Limited Partnership) (華海金浦創業投資（濟南）合夥企業（有限合夥）) ("Huahai Jinpu") | 103,276 | 50,000,000 |
| | Series Angel Financing | Series A Financing | Series B1 Financing | Series B2 Financing | Series B3 Financing | Series B4 Financing | Series B5 Financing | Series B6 Financing | |---|---|---|---|---|---|---|---|---| | Date of agreements | April 8, 2021 | April 8, 2021 April 28, 2021 | January 12, 2022 February 15, 2022 | March 8, 2023 April 19, 2023 | May 12, 2023 May 15, 2023 July 6, 2023 | February 24, 2022 June 28, 2023 January 2, 2024 January 24, 2024 June 21, 2024 | November 28, 2023 August 8, 2024 February 2, 2024 December 8, 2024 April 18, 2024 December 31, 2024 January 7, 2025 February 7, 2025 March 19, 2025 May 13, 2025 May 23, 2025 | May 29, 2025 | | Date of last settlement of consideration | July 8, 2021 | April 8, 2021 | February 24, 2022 | June 28, 2023 | January 2, 2024 January 24, 2024 June 21, 2024 | August 8, 2024 December 8, 2024 December 31, 2024 February 7, 2025 March 19, 2025 May 13, 2025 May 23, 2025 | May 29, 2025 | | | Amount of consideration paid to the Company | RMB20.37 million(6) | RMB152.00 million | RMB208.00 million | RMB410.11 million | RMB1,228.81 million | RMB1,358.93 million | RMB4,377.00 million | | | Approximate pre-money valuation of our Company(2) | RMB387.02 million | RMB700.00 million | RMB1,900.00 million | RMB2,580.00 million | RMB3,850.00 million | RMB6,000.00 million | RMB11,999.99 million | RMB19,999.99 million | | Approximate post-money valuation of our Company(3) | RMB407.39 million | RMB852.00 million | RMB2,108.00 million | RMB608.97 million RMB3,210.11 million | RMB4,572.00 million | RMB7,228.82 million | RMB13,358.93 million | RMB24,376.99 million | | Cost per Share(4) | RMB3.87 | RMB6.65 | RMB14.24 | RMB17.42 Stage 2: RMB18.76 | RMB22.49 Stage 2: RMB23.09 | RMB28.78 | RMB43.00 | RMB60.52 | | Discount to the Offer Price(5) | 96.33% | 93.69% | 86.49% | Stage 1: 83.47% Stage 2: 82.20% | Stage 1: 78.66% Stage 2: 78.09% | 72.69% | 59.20% | 42.58% |
For details of the shareholding of our Company upon completion of the Pre-IPO Investments, See "—Capitalization of Our Company" below.
The consideration for the Pre-IPO Investments which involved increase of registered capital and/or issuance of new Shares in which the Company was a party was determined based on arm's length negotiations between the Company and the Pre-IPO Investors, after taking into consideration various factors, including but not limited to, (i) the timing of the investments; (ii) the status of milestones and prospects of commercialization of our products and our technology advancement; (iii) our expansion capacity and R&D management system; (iv) strategic layout, execution efficiency and other factors of our Group; and (v) the expected market value and prospects of our business.
To the best knowledge of our Company, for the Pre-IPO Investments which involved the transfer of existing registered capital or Shares to the Pre-IPO Investors, the considerations were determined among the relevant then Shareholders of our Company and the relevant Pre-IPO Investors upon their respective arm's length negotiations.
The proceeds received by us from the Pre-IPO Investments which involved increase of registered capital and/or issuance of new Shares by our Company amounted to approximately RMB8,364.20 million, of which approximately 58% of the funds raised from such Pre-IPO Investments had been utilized as of the Latest Practicable Date. All of such proceeds were utilized for the R&D, capital expenditures and general working capital needs of our Group in accordance with the annual consolidated budget of the Company approved by the Pre-IPO Investors.
No proceeds were received by our Company from the Pre-IPO Investments that involved transfers by the then Shareholders of existing registered capital or Shares by our Company to the Pre-IPO Investors.
Our Pathfinder SIIs will be subject to the disposal restrictions pursuant to Chapter 18C.14 of the Listing Rules. See "—Lock-up Periods" below for further details.
In addition, all existing Shareholders (including the Pre-IPO Investors) are subject to a lock-up period of 12 months following the Listing Date according to the applicable PRC law.
We are of the view that (i) our Group would benefit from the additional capital provided by the Pre-IPO Investors; (ii) our Group could benefit from the Pre-IPO Investors' knowledge and experience and take advantage of their industry resources and networks, while at the same time broaden our shareholder base; and (iii) the Pre-IPO Investors' investment demonstrated their confidence in our Group and served as an endorsement of our performance, strengths and prospects.
(5) The discount to the Offer Price is calculated based on the foreign exchange rate as set out in this prospectus.
(6) This represents the appraised value of the intellectual property rights as of the valuation benchmark date of March 31, 2021 transferred in kind as consideration for the Series Angel Investment.
历史、发展及公司架构 我们公司估值变动与上市前投资相关的原因 我们公司在数轮上市前投资中估值大幅提升的主要原因如下： (1) 从天使轮融资到A轮融资的估值提升，主要由于本公司早期知识图谱相关产品开始产生稳定收入，且我们的研发团队主动开始预训练大语言模型。 (2) 从A轮融资到B1轮融资的估值提升，主要由于2021年9月我们推出了GLM-10B，即我们首个百亿参数预训练大模型，显著提升了我们的市场知名度。 (3) 从B1轮融资到B2轮融资的估值提升，主要由于2022年8月我们推出了开源大模型GLM-130B，以及2022年9月推出了高性能编程模型CodeGeeX，扩展了我们的产品组合和客户覆盖范围。 (4) 从B2轮融资到B3轮融资的估值提升，主要由于2023年3月我们同步发布了亿级基础模型ChatGLM及开源版ChatGLM-6B，受到客户的热烈欢迎。 (5) 从B3轮融资到B4轮融资的估值提升，主要由于ChatGLM及ChatGLM-6B的成功发布所带来的市场影响力持续增长。 (6) 从B4轮融资到B5轮融资的估值提升，主要由于2024年8月智谱清言的发布，该产品是国内首批完成监管备案的大模型产品之一，展示了巨大的商业化潜力，并提升了我们在市场中的地位。 (7) 从B5轮融资到B6轮融资的估值提升，主要由于GLM-4及GLM-4V的发布，展示了我们领先的研发能力及显著的增长潜力。
(1) 上市前投资包括以下两类：(i) 上市前投资者认购本公司额外注册资本或新股，本公司为该等上市前投资的一方，并从该等上市前投资者处收取认购款项，相关详情载于本表；及(ii) 向上市前投资者转让现有注册资本或股份，本公司并非该等上市前投资的一方，亦未从该等上市前投资者处收取任何款项，有关该等转让的进一步详情，请参阅上文"——A轮融资"、"——B1轮融资"、"——B3轮融资"及"——B5轮融资"。
(2) 投前估值系指就相应轮次上市前投资向本公司支付的每股成本，乘以紧接相应轮次上市前投资完成前本公司的总注册股本。
(3) 投后估值系指就相应系列上市前投资向本公司支付的每股成本，乘以相应轮次上市前投资完成后紧接本公司的总注册资本。我们公司在每轮上市前投资中估值的提升，源于本集团的业务发展及经营状况，尤其是本集团实现的关键里程碑、我们专业技术产品的发布及商业化、我们研发工作的推进，以及投资完成时的市场情绪。详情请参阅"——关键里程碑"及"——我们公司估值变动与上市前投资相关的原因"。
(4) 每系列上市前投资的每股成本，系以相关上市前投资者认购的相应系列上市前投资中增加的注册资本总额除以对价总额计算，并参考本公司于2025年3月由有限责任公司转制为股份有限公司时的股份转换比例进行调整（假设股份细分已完成）。详情请参阅"——本公司的设立及主要股权变动——2025年转制为股份有限公司"及"——本公司的设立及主要股权变动——上市前股份细分"。
上市前投資者的特殊權利 Pre-IPO投資者已獲授慣常特殊權利，包括但不限於清算優先權、優先認購權、優先購買權、隨售權、知情權、贖回權、反稀釋權及董事提名權。
根據2025年6月27日簽訂的股東特殊權利終止協議，為配合全球發售，(a)所有贖回權及撤資權已於本公司首次向聯交所提交上市申請（「首次申請」）之前即時終止並停止行使，惟若出現以下最早發生的情況，該等權利將予恢復：(i)本公司撤回向聯交所提交的上市申請；(ii)相關監管機構或聯交所最終不接受、拒絕、退回、否決、終止審核、不予登記或不予備案上市申請；或(iii)本公司未能於首次申請後12個月內完成上市；及(b)所有其他特殊權利將於上市日期上市前即時自動終止，以符合新上市申請人指引第4.2章的規定。
獨家保薦人確認 基於以下原因：(i) Pre-IPO投資的代價已於上市日期前不少於120個清楚日悉數結清；及(ii)授予Pre-IPO投資者的贖回權及撤資權已於向聯交所提交上市申請前終止，且所有其他特殊權利將於上市時終止，獨家保薦人確認Pre-IPO投資符合聯交所頒布的新上市申請人指引第4.2章的規定。
有關我們Pre-IPO投資者的資料 我們已從三名成熟獨立股東獲得投資。其中兩名成熟獨立股東，即聯想資本SII及美團SII，為我們的領航SII，各自在首次申請前至少12個月已投資於本公司。根據新上市申請人指引第2.5章，聯想資本SII及美團SII各自持有本公司於首次申請日期及首次申請前12個月期間已發行股本逾3%，合計逾10%。有關各成熟獨立投資者於本公司股本中的持股百分比詳情，請參閱「—本公司的資本化」。除(i)作為本公司的成熟獨立股東及(ii)聯想資本（其並非且從未為李家慶先生的緊密聯繫人）根據聯想資本的董事提名權提名李家慶先生為董事外，據本公司所知、所悉及所信，截至最後實際可行日期，我們各成熟獨立投資者均與本公司任何核心關連人士相互獨立且並無關連，且與本公司任何核心關連人士均無一致行動安排。
截至最後實際可行日期，我們的成熟獨立投資者合計持有本公司已發行總股本約13.49%，並假設本公司於上市時的預期市值將超過300億港元，合計將持有本公司已發行總股本不少於10%。
據本公司所知、所悉及所信，除本文件所披露者外，各Pre-IPO投資者彼此相互獨立，且各Pre-IPO投資者均為獨立第三方。
以下載列我們成熟獨立投資者（包括領航SII）及其他Pre-IPO投資者的描述。
君聯向道為根據中國法律成立的有限合夥企業，從事股權投資、投資管理、資產管理及其他活動。君聯向道的權益構成如下：(i)約1.18%由其普通合夥人拉薩君祺企業管理有限公司（「拉薩君祺」）持有，拉薩君祺由君聯資本管理股份有限公司（「LCM」）全資擁有；及(ii)約98.82%由其他29名有限合夥人持有，其中任何一名有限合夥人的合夥權益均不超過30%。LCM由北京君誠合眾投資管理合夥企業（有限合夥）（「君誠合眾」）持有80.00%權益。君誠合眾由其普通合夥人北京君祺嘉睿企業管理有限公司控制，該公司由陳浩先生持有40%權益，其他合夥人均不持有30%或以上的權益。
君聯金帆為根據中國法律成立的有限合夥企業，主要從事創業投資。君聯金帆的權益構成如下：(i)約1.40%由其普通合夥人拉薩君祺持有；(ii)約70.15%由廈門建發新興產業股權投資拾壹號合夥企業（有限合夥）作為最大有限合夥人持有；
which is ultimately controlled by Xiamen Municipal People's Government State-owned Assets Supervision and Administration Commission (廈門市人民政府國有資產監督管理委員會), and (iii) approximately 28.45% by Junlian Xiangdao as its limited partner.
Social Security Zhongguancun Innovation Fund is a limited partnership established under the laws of the PRC, which is principally engaged in investment activities and venture capital. Social Security Zhongguancun Innovation Fund is owned as to (i) approximately 1.96% by Beijing Jun Chuang Li Xin Venture Capital Partnership (Limited Partnership) (北京君創勵新創業投資合夥企業(有限合夥)) as its general partner, in which Lhasa Junqi holds 50% partnership interest as the general partner, and (ii) approximately 98.04% by National Social Security Fund Council (全國社會保障基金理事會) as its sole limited partner.
As the fund manager of each of Legend Capital SIIs is LCM and the general partner of each of the Legend Capital SIIs is ultimately controlled by LCM, the different shareholding entities are purely different funds managed by the same fund manager and should be aggregated as one SII pursuant to Chapter 2.5 of the Guide for New Listing Applicants. LCM is a leading professional investment institution in the PRC with a focus on early-stage venture capital investments and growth-stage private equity investments. LCM has made investments in companies across various industries, such as the technology, healthcare, consumer, enterprise service and intelligent manufacturing industries. The assets under management ("AUM") of LCM was more than HK$15 billion as of December 31, 2021 (being a date not more than six months prior to the date on which Legend Capital SIIs signed the first definitive agreement for its investment in our Company) and more than HK$15 billion as of December 31, 2024, respectively. As of the Latest Practicable Date, Legend Capital SIIs held approximately 6.73% of the total issued Shares. In compliance with Rule 18C.05 of the Listing Rules, Legend Capital SIIs held approximately 6.73% and 6.63% of the total issued Shares as of June 30, 2025 (being the date of the First Filing) and June 30, 2024 (being the commencement date of the 12-month period prior to the First Filing), respectively.
Tianjin Sankuai is a limited liability company established under the laws of the PRC, which is principally engaged in retail business. Tianjin Sankuai is wholly owned by Meituan (美团), the class B shares of which are listed on the Main Board of the Stock Exchange (stock codes: 3690 (HKD counter) and 83690 (RMB counter)).
Meituan has a diverse investment portfolio and its portfolio companies include Li Auto Inc., a company listed on the Stock Exchange (stock code: 2015) and NASDAQ (ticker symbol: LI); Hesai Group, a company listed on NASDAQ (ticker symbol: HSAI); H World Group Limited, a company listed on the Stock Exchange (stock code: 1179) and NASDAQ (ticker symbol: HTHT) and Maoyan Entertainment, a company listed on the Stock Exchange (stock code: 1896). The value of the diverse investment portfolio held by Meituan was approximately RMB33.98 billion as of December 31, 2022 (being a date not more than six months prior to the date on which Meituan SII signed the relevant definitive agreement for its investment in our Company) and RMB41.31 billion as of December 31, 2024.
As of the Latest Practicable Date, Meituan SII held approximately 4.27% of the total issued Shares. In compliance with Rule 18C.05 of the Listing Rules, Meituan SII held approximately 4.27% and 8.26% of the total issued Shares as of June 30, 2025 (being the date of the First Filing) and June 30, 2024 (being the commencement date of the 12-month period prior to the First Filing), respectively.
Qiming Rongqian is a limited partnership established under the laws of the PRC. Qiming Rongqian is owned as to (i) approximately 1.01% by Suzhou Qikun Venture Capital Partnership (Limited Partnership)
(蘇州啟坤創業投資合夥企業(有限合夥)) ("Suzhou Qikun") as its general partner, which is ultimately controlled by Yu Jia (于佳) and Xu Jing (徐靜) as to 50% and 50%, respectively, and (ii) approximately 98.99% by other 35 limited partners, none of which holds more than 30% partnership interest therein.
Qiming Rongkai is a limited partnership established under the laws of the PRC. Qiming Rongkai is owned as to (i) approximately 2.89% by Suzhou Qikun as its general partner, (ii) 30% by Kunshan Industrial Development Investment Fund Partnership (Limited Partnership) (昆山市產業發展投資母基金合夥企業(有限合夥)) as its limited partner, which is ultimately controlled by Kunshan Municipal Government State-owned Assets Supervision and Administration Office (昆山市政府國有資產監督管理辦公室), and (iii) approximately 67.11% by other 13 limited partners, none of which holds more than 30% partnership interest therein.
As each of Qiming Rongqian and Qiming Rongkai is a venture capital fund operated under Qiming Venture Partners, the different shareholding entities are purely different funds managed by the same fund manager and should be aggregated as one SII pursuant to Chapter 2.5 of the Guide for New Listing Applicants. Qiming Venture Partners is a leading venture capital firm in China, which is focused on investing in outstanding companies at their early and growth stages in the technology and healthcare industries. The AUM of Qiming Venture Partners was approximately US$6.0 billion as of December 31, 2021 (being a date not more than six months prior to the date on which Qiming Venture SIIs signed the first definitive agreement for its investment in our Company), and approximately US$9.0 billion as of December 31, 2024, respectively.
As of the Latest Practicable Date, Qiming Venture SIIs held approximately 2.49% of the total issued Shares. Qiming Venture SIIs held approximately 2.49% and 3.97% of the total issued Shares as of June 30, 2025 (being the date of submission of the First Filing) and June 30, 2024 (being the commencement date of the 12-month period prior to the First Filing), respectively.
We have also received investments from the other major Pre-IPO Investors, each of which held 1.00% or more interest in the Company as of the Latest Practicable Date.
Xinglian Zhaoji is a limited partnership established under the laws of the PRC, which is principally engaged in software development services, computer system services, data processing, market research, economic and trade consulting, enterprise management consulting and public relations services. Xinglian Zhaoji is owned as to (i) approximately 0.002% by Beijing Zhiqiao Management Consulting Co., Ltd. (北京智橋管理諮詢有限公司) as its general partner, which is ultimately controlled by Ye Xiaobin (葉曉斌), and (ii) approximately 99.998% by other ten limited partners, none of which holds more than 30% partnership interest therein.
Tongzhi Investments is a limited partnership established under the laws of the PRC, which is principally engaged in equity investments, investment management, asset management, and other activities through private placement funds. Tongzhi Investments is owned as to (i) approximately 0.02% by Hainan Zhiqiao Private Equity Fund Management Partnership (Limited Partnership) (海南智橋私募基金管理合夥企業(有限合夥)) as its general partner, which is ultimately controlled by Ye Xiaobin (葉曉斌), and (ii) approximately 99.98% by other 14 limited partners, none of which holds more than 30% partnership interest therein.
Shanghai Yunya is a limited liability company established under the laws of the PRC, an investment vehicle of Ant Group Co., Ltd. (螞蟻科技集團股份有限公司) ("Ant Group"). None of the shareholders of Ant Group holds more than one third of the equity interest therein.
Shanghai Feiya is a limited liability company established under the laws of the PRC. Shanghai Feiya is wholly owned by Accelerator VIII Ltd., an exempted company incorporated in the Cayman Islands with limited liability and wholly owned by Ant Unicorn Fund, L.P. Ant Unicorn Fund, L.P. is an exempted limited partnership registered in the Cayman Islands, and managed by its general partner, Ant Unicorn Ltd., an indirect wholly owned subsidiary of Ant Group.
海河富鑫優達基金 (Haihe Fuxin Youda Fund) is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment, investment management, and asset management through private equity funds. Haihe Fuxin Youda Fund is owned as to (i) approximately 0.20% by Beijing Shang Finance Co., Ltd. (北京尚融資本管理有限公司) as its general partner, which is ultimately controlled by Wei Lidong (尉立東), (ii) approximately 33.27% by Tianjin Wuqing District Innovation and Entrepreneurship Investment Co., Ltd. (天津市武清區創新創業投資有限公司) as its limited partner, which is ultimately controlled by the State-owned Assets Supervision and Administration Commission of the People's Government of Wuqing District, Tianjin (天津市武清區人民政府國有資產監督管理委員會) ("Wuqing SASAC"), (iii) approximately 33.16% by Tianjin Jingjin New Town Technology Development Co., Ltd. (天津市京津新城科技發展股份有限公司) as its limited partner, which is ultimately controlled by Wuqing SASAC, and (iv) approximately 33.37% by other three limited partners, none of which holds more than 30% partnership interest therein.
Trend Mega was incorporated in the British Virgin Islands. Trend Mega is indirectly owned as to approximately 99.75% by Capital Today Evergreen Fund, L.P., which is controlled by its general partner, Capital Today Evergreen GenPar Ltd., which is in turn controlled by Xu Xin.
達晨創鴻 (Dachen Chuanghong) is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment. Dachen Chuanghong is owned as to (i) approximately 4.25% by Shenzhen Fortune Venture Capital Co., Ltd. (深圳市達晨財智創業投資管理有限公司) ("Fortune Capital") as its general partner, which is ultimately controlled by Hunan Provincial State-owned Cultural Assets Supervision and Administration Committee (湖南省國有文化資產監督管理委員會), and (ii) approximately 95.75% by other forty-nine limited partners, none of which holds more than 30% partnership interest therein.
財智創贏 (Caizhi Chuangying) is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment. Caizhi Chuangying is owned as to (i) approximately 0.18% by Fortune Capital, and (ii) approximately 99.82% by other 30 limited partners, none of which holds more than 30% partnership interest therein.
北京華控 (Beijing Huakong) is a limited partnership established under the laws of the PRC, which is principally engaged in investment, investment management and consulting for non-securities businesses. Beijing Huakong is owned as to (i) 1% by Beijing Huakong Investment Consulting Co., Ltd. (北京華控投資顧問有限公司) as its general partner, which is wholly owned by Beijing Huakong Investment Management Group Co., Ltd. (北京華控投資管理集團有限公司) and ultimately controlled by Zhang Yang (張揚), and (ii) 99% by other eight limited partners, none of which holds more than 30% partnership interest therein.
青島華控 (Qingdao Huakong) is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment, investment management, and asset management through private equity funds.
Qingdao Huakong is owned as to (i) approximately 0.03% by Horgos Huakong Venture Capital Co., Ltd. (霍爾果斯華控創業投資有限公司) as its general partner, which is ultimately controlled by Zhang Yang (張揚), and (ii) approximately 99.97% by other 21 limited partners, none of which holds more than 30% partnership interest therein.
Zhongguancun Science City is a limited partnership established under the laws of the PRC, which is principally engaged in equity investments, investment management, asset management, and other activities through private equity funds. Zhongguancun Science City is owned as to (i) 1% by Beijing Zhongguancun Science City Technology Investment Management Co., Ltd. (北京中關村科學城科技投資管理有限公司) as its general partner, which is ultimately controlled by State-owned Assets Supervision and Administration Commission of the People's Government of Haidian District, Beijing (北京市海淀區人民政府國有資產監督管理委員會) ("Haidian SASAC"), and (ii) 99% by Beijing Haidian District State-owned Assets Investment Group Co., Ltd. (北京市海淀區國有資產投資集團有限公司) as its sole limited partner, which is wholly owned by State-owned Capital Operation Company of Haidian District, Beijing (北京市海淀區國有資本運營有限公司) and ultimately controlled by Haidian SASAC.
Zhuhai Huafa is a limited partnership established under the laws of the PRC, which is principally engaged in venture capital, equity investment, investment management, and asset management activities. Zhuhai Huafa is owned as to (i) approximately 0.13% by Zhuhai Kechuang Haisheng Venture Capital Fund Management Co., Ltd. (珠海科創海盛創業投資基金管理有限公司) as its general partner, which is wholly owned by Zhuhai Technology and Entrepreneurship Investment Co., Ltd. (珠海科技創業投資有限公司) and ultimately controlled by Zhuhai Municipal Government State-owned Assets Supervision and Administration Commission (珠海市人民政府國有資產監督管理委員會) ("Zhuhai SASAC"), (ii) 62.5% by Zhuhai Huafa Group Co., Ltd. (珠海華發集團有限公司) ("Huafa Group") as its largest limited partner, which is in turn ultimately controlled by Zhuhai SASAC, and (iii) approximately 37.37% by Zhuhai Huafa Technology Industry Group Co., Ltd. (珠海華發科技產業集團有限公司) as its limited partner, whose largest shareholder is Huafa Group.
Zhihui Linghang is a limited partnership established under the laws of the PRC, which is principally engaged in venture capital. Zhihui Linghang is owned as to (i) approximately 0.20% by Bokang Co-Winning Equity Investment Fund Management Co., Ltd. (博康共贏股權投資基金管理有限公司) as its general partner, which has Bokang Holding Group Co., Ltd. (博康控股集團有限公司) as its largest shareholder and ultimately controlled by Zhang Tao (張滔), (ii) approximately 79.84% by Shanghai Pudong Leadership Area Investment Center (Limited Partnership) (上海浦東引領區投資中心(有限合夥)) as its largest limited partner, which is ultimately controlled by State-owned Assets Supervision and Administration Commission of Pudong New District, Shanghai (上海市浦東新區國有資產監督管理委員會) ("Pudong SASAC"), and (iii) approximately 19.96% by Shanghai Zhangjiang Technology Venture Capital Co., Ltd. (上海張江科技創業投資有限公司) as its limited partner, which is wholly owned by Shanghai Zhangjiang (Group) Co., Ltd. (上海張江(集團)有限公司) and ultimately controlled by Pudong SASAC.
AI Fund is a limited partnership established under the laws of the PRC, which is principally engaged in equity investments, investment management, asset management, and other activities through private equity funds. AI Fund is owned as to (i) 0.5% by Beijing Jingguoguan Property Management Co., Ltd. (北京京國管置業管理有限公司) as its general partner, which is wholly owned by Beijing State-owned Capital Operation Management Co., Ltd. (北京國有資本運營管理有限公司) and ultimately controlled by Beijing
Municipal People's Government State-owned Assets Supervision and Administration Commission (北京市人民政府國有資產監督管理委員會) ("Beijing SASAC"), (ii) 0.5% by Beijing Qiou Management Consulting Partnership (Limited Partnership) (北京啟歐管理諮詢合夥企業(有限合夥)) as its general partner, which is ultimately controlled by Qiming Venture, and (iii) 99% by Beijing Municipal Government Investment Guidance Fund (Limited Partnership) (北京市政府投資引導基金(有限合夥)) as its sole limited partner, which is ultimately controlled by Beijing SASAC.
Tencent Investment is a limited liability company established in the PRC on January 6, 2020. Tencent Investment is wholly owned by Shenzhen Tencent Ruijian Investment Co., Ltd. (深圳市騰訊睿見投資有限公司), which is in turn wholly owned by Shenzhen Tencent Ruitou Enterprise Management Co., Ltd. (深圳市騰訊睿投企業管理有限公司), a company principally engaged in enterprise management consulting. Shenzhen Ruitou Enterprise Management Co., Ltd is controlled each as to 50% by Shenzhen Tengyuan Enterprise Management Partnership (Limited Partnership) (深圳市藤遠企業管理合夥企業(有限合夥)) ("Shenzhen Tengyuan") and Shenzhen Tenglv Enterprise Management Partnership (Limited Partnership) (深圳市藤綠企業管理合夥企業(有限合夥)) ("Shenzhen Tenglu"). The general partner of both Shenzhen Tengyuan and Shenzhen Tenglu is Shenzhen Tengqing Enterprise Management Co., Ltd. (深圳市藤青企業管理有限公司), which is controlled as to 80% by Ma Huateng (馬化騰), the founder of Tencent, and held as to 20% by Xu Chenye (許晨曄). Each of Shenzhen Tengyuan and Shenzhen Tenglu has two limited partners, namely, Xu Chenye, holding 20% partnership interests therein, and Lu Shan (盧山), holding 5% partnership interests therein.
Duoxiang Network is a limited company incorporated under the laws of the PRC, which is wholly owned by Hangzhou Zhenxi Investment Management Co., Ltd. (杭州臻希投資管理有限公司), which is in turn owned by Hangzhou Zhensheng Investment Management Partnership (Limited Partnership) (杭州臻晟投資管理合夥企業(有限合夥)) ("Hangzhou Zhensheng") and Hangzhou Zhenqiang Investment Management Partnership (Limited Partnership) (杭州臻強投資管理合夥企業(有限合夥)) ("Hangzhou Zhenqiang") as to 50% and 50%, respectively. The general partner of Hangzhou Zhensheng and Hangzhou Zhenqiang is Hangzhou Zhenyue Enterprise Management Co., Ltd. (杭州臻悅企業管理有限公司), which is owned by each of Zheng Junfang (鄭俊芳), Wu Zeming (吳澤明), Jiang Fang (蔣芳) and Shao Xiaofeng (邵曉鋒) as to 25%, respectively. Each of Duoxiang Network and its ultimate beneficial owners is an Independent Third Party.
Hangzhou Chengtou Industrial Fund is a limited partnership established under the laws of the PRC, which is principally engaged in business management consulting and equity investment. Hangzhou Chengtou Industrial Fund is owned as to (i) 0.02% by Hangzhou Chengchuang Investment Management Co., Ltd. (杭州城創投資管理有限公司) as its general partner, which is wholly owned by Hangzhou Chengtou Capital Group Co., Ltd. (杭州城投資本集團有限公司) and ultimately controlled by Hangzhou Municipal People's Government State-owned Assets Supervision and Administration Commission (杭州市人民政府國有資產監督管理委員會), and (ii) 99.98% by Hangzhou Urban Construction Investment Group Co., Ltd. (杭州市城市建設投資集團有限公司) as its sole limited partner, which is ultimately controlled by Hangzhou Municipal People's Government (杭州市人民政府).
Beijing Shunying is a limited partnership established under the laws of the PRC. As of the Latest Practicable Date, Beijing Shunying is owned as to (i) approximately 0.05% by Beijing Shunzhong Shunying Enterprise Management Partnership (Limited Partnership) (北京順眾順贏企業管理合夥企業(有限合夥)) as its general partner, whose general partner is Beijing Shunzhong Enterprise Management Co., Ltd. (北京順眾企業管理有限責任公司) ("Beijing Shunzhong EMC"), which is ultimately owned by Lei Jun (雷軍), Ma Wenjing (馬文靜), Cao Liping (曹莉平) and Cheng Tian (程天), (ii) approximately 33.32% by Beijing Shunmi Shunying Enterprise Management Partnership (Limited Partnership) (北京順米順贏企業管理合夥企業(有限合夥)) as its limited partner, (iii) approximately 33.32% by Beijing Shunjin Shunying Enterprise Management Partnership (Limited Partnership) (北京順金順贏企業管理合夥企業(有限合夥)) as its limited
partner, and (iv) approximately 33.31% by other three limited partners, none of which holds more than 30% partnership interest of Beijing Shunying.
Prosperity7 is a protected cell company established in Guernsey, and is wholly-owned by Aramco Ventures Company, a company established under the laws of the Kingdom of Saudi Arabia.
Daxing Industrial Fund is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment, investment management, and asset management through private equity funds. Daxing Industrial Fund is owned as to (i) 0.10% by Beijing North Business Capital Management Co., Ltd. (北商資本管理(北京)有限公司) as its general partner, which is ultimately controlled by State-owned Assets Supervision and Administration Commission of the People's Government of Daxing District, Beijing (北京市大興區人民政府國有資產監督管理委員會) ("Daxing SASAC") and (ii) 99.90% by Beijing Daxing Development Guidance Fund (Limited Partnership) (北京市大興發展引導基金(有限合夥)) as its sole limited partner, which is ultimately controlled by Daxing SASAC.
High-tech Orinno is a limited partnership established under the laws of the PRC, which is principally engaged in equity investment, investment management, and asset management through private equity funds. High-tech Orinno is owned as to (i) 1.00% by Chengdu High-tech New Economy Venture Capital Co., Ltd. (成都高新新經濟創業投資有限公司) as its general partner, (ii) 50.00% by Chengdu High-tech Jicui Technology Co., Ltd. (成都高新集萃科技有限公司) as its limited partner, (iii) 44.00% by Chengdu High-Tech Investment Group Co., Ltd. (成都高新投資集團有限公司) as its limited partner, and (iv) 5% by Chengdu High-tech Zone Finance and State-Owned Assets Bureau (成都高新技術產業開發區國資金融局). Each of Chengdu High-tech New Economy Venture Capital Co., Ltd., Chengdu High-tech Jicui Technology Co., Ltd. and Chengdu High-Tech Investment Group Co., Ltd. is ultimately controlled by Chengdu High-tech Zone Finance and State-Owned Assets Bureau.
During the Track Record Period, our Company did not conduct any acquisition of business for which any of the applicable percentage ratios as defined under the Listing Rules exceeds 25% which would require disclosure pursuant to Rule 4.05A of the Listing Rules. We also did not conduct any other acquisitions, disposals or mergers material to our Group during the Track Record Period and up to the Latest Practicable Date.
During the Track Record Period, our Company acquired additional equity interest in Beijing Lingxin Intelligent, which is mainly engaged in providing artificial intelligence services. Our Company held approximately 5.02% interest in Beijing Lingxin Intelligent prior to such acquisitions, and held 100% interest in Beijing Lingxin Intelligent upon completion of such acquisitions. The acquisitions were made as part of the Group's strategy to expand its market share of artificial intelligence in the PRC.
In September 2023, the Company entered into transfer agreements with then shareholders of Beijing Lingxin Intelligent, namely Nantong Xinfang Technology Development Center (Limited Partnership) (南通欣房科技發展中心(有限合夥)), Sanya Lianxing Shangzhi No.1 Equity Investment Fund Partnership (Limited Partnership) (海南三亞連星尚智壹號股權投資基金合夥企業(有限合夥)), Nanjing Turing, Infinite Qihang Venture Investment (Taiyuan) Partnership (Limited Partnership) (無限啟航創業投資(太原)合夥企業(有限合夥)), Huang Minlie (黃民烈), Tsinghua Technology, Beijing Beiqingxin Intelligent Technology Center (Limited Partnership) (北京倍傾心智能科技中心(有限合夥)), Shanghai Xindiao Enterprise Management Partnership (Limited Partnership) (上海心雕企業管理事務所(有限合夥)) and Shanghai Yiniansha Enterprise Management Consulting Center (Limited Partnership) (上海依念薩企業管理諮詢中心(有限合夥)), to acquire
合共约94.98%的北京灵心智能股权，总现金对价为人民币73,004,600元。就本公司所知、所悉及所信，上述交易中各转让方均为独立第三方。
上述收购的对价经本公司与相关转让方经公平磋商厘定，参考了截至2023年5月31日北京灵心智能净资产值的独立估值，并已于2024年10月24日悉数结清。就上述收购而言，上市规则所界定的任何适用百分比率均未超过25%（该比率须根据上市规则第4.05A条作出披露）。根据中国法律顾问的意见，上述各项收购均已妥善合法地完成及结清，并符合适用中国法律法规的所有重要方面，且已取得所有必要的监管批准。
为表彰核心员工的贡献并激励其进一步推动本公司发展，本公司于2021年至2025年采纳了员工激励计划，以向员工激励计划项下的合资格参与者授予员工持股平台的合伙权益。于最后实际可行日期，汇汇及智登已作为本公司的员工持股平台而设立。员工激励计划的条款不受上市规则第17章条文的约束，原因是本公司于上市后并不涉及任何购股权或奖励的授予或任何新股份的发行。
根据员工激励计划及相关授予协议，合资格参与者获授员工持股平台的合伙权益。员工持股平台的所有管理权及投票权均由各自唯一普通合伙人根据各自的合伙协议行使。员工持股平台的合资格参与者仅享有其中的经济利益。
汇汇于2021年6月23日根据中国法律设立，由刘博士（本公司执行董事）担任其唯一普通合伙人，持有约30.33%的合伙权益。于本招股说明书日期，汇汇共有426名有限合伙人，均为本集团的现任及前任员工，其中张博士及张晓涵女士（本公司执行董事）分别作为有限合伙人持有约20.98%及0.46%的合伙权益。汇汇的任何有限合伙人均未持有30%或以上的合伙权益。
智登于2021年6月23日根据中国法律设立，由刘博士（本公司执行董事）担任其唯一普通合伙人，持有约39.01%的合伙权益。于本招股说明书日期，智登共有25名有限合伙人，均为本集团的现任员工及顾问，其中张博士（本公司执行董事）作为有限合伙人持有约4.63%的合伙权益。有限合伙人中的全部16名顾问均为全职实习生，系本集团聘用的算法专家，各人均为独立第三方。该等顾问的主要职责包括（其中包括）代码开发、研究及视觉语言模型的预训练以及算法开发。智登的任何有限合伙人均未持有30%或以上的合伙权益。
汇汇及智登的任何有限合伙人均无权根据相关合伙协议享有任何特殊权利，而该等权利将影响刘博士作为唯一普通合伙人对汇汇及智登的控制权。
有关本公司员工激励计划的进一步详情，请参阅本招股说明书「附录六——D.员工激励计划」。
As of the Latest Practicable Date and immediately after completion of the Share Subdivision, the Conversion of Unlisted Shares and the Global Offering (assuming the Over-allotment Option is not exercised), the summary of the capitalization of our Company is set out as follows:
As of the Latest Practicable Date | | Upon completion of the Share Subdivision, the Conversion of Unlisted Shares and the Global Offering | | | | |---|---|---|---|---|---| Shareholder | Number of Shares | Percentage of shareholding | Number of H Shares | Number of Unlisted Shares | Total number of Shares | Percentage of shareholding | Whether the H Shares count towards public float or not |
Our Controlling Shareholders | Beijing Lianpai | 3,403,839 | 8.45% | — | 34,038,390 | 34,038,390 | 7.73% | — | | Dr. Liu | 92,515 | 0.23% | — | 925,150 | 925,150 | 0.21% | — | | Dr. Tang | 2,683,533 | 6.66% | — | 26,835,330 | 26,835,330 | 6.10% | — | | Dr. Li | 336,776 | 0.84% | — | 3,367,760 | 3,367,760 | 0.76% | — | | Dr. Xu | 79,904 | 0.20% | — | 799,040 | 799,040 | 0.18% | — | | Dr. Zhang | 39,952 | 0.10% | — | 399,520 | 399,520 | 0.09% | — | | Huihui | 3,948,271 | 9.80% | 5,922,407 | 33,560,303 | 39,482,710 | 8.97% | No | | Zhideng | 2,719,633 | 6.75% | 4,079,450 | 23,116,880 | 27,196,330 | 6.18% | No |
Our Sophisticated Independent Investors Legend Capital SIIs | Junlian Xiangdao | 1,866,775 | 4.63% | 18,667,750 | — | 18,667,750 | 4.24% | Yes | | Junlian Jinfan | 175,241 | 0.44% | 1,752,410 | — | 1,752,410 | 0.40% | Yes |
Social Security | Zhongguancun Innovation Fund | 668,896 | 1.66% | 6,688,960 | — | 6,688,960 | 1.52% | Yes |
Qiming Venture SIIs | Qiming Rongqian | 590,524 | 1.47% | 2,952,620 | 2,952,620 | 5,905,240 | 1.34% | Yes | | Qiming Rongkai | 411,369 | 1.02% | 2,056,845 | 2,056,845 | 4,113,690 | 0.93% | Yes |
Other Shareholders Xinglian Zhaoji and Tongzhi Investments | Xinglian Zhaoji | 951,934 | 2.36% | 6,696,856 | 2,822,484 | 9,519,340 | 2.16% | Yes | | Tongzhi Investments | 706,767 | 1.75% | 2,120,300 | 4,947,370 | 7,067,670 | 1.61% | Yes |
Shanghai Yunya and Shanghai Feiya | Shanghai Yunya | 539,323 | 1.34% | 5,393,230 | — | 5,393,230 | 1.23% | Yes | | Shanghai Feiya | 1,069,151 | 2.65% | 10,691,510 | — | 10,691,510 | 2.43% | Yes |
As of the Latest Practicable Date | | Upon completion of the Share Subdivision, the Conversion of Unlisted Shares and the Global Offering | | | | |---|---|---|---|---|---| Shareholder | Number of Shares | Percentage of shareholding | Number of H Shares | Number of Unlisted Shares | Total number of Shares | Percentage of shareholding | Whether the H Shares count towards public float or not |
| Shareholder | Percentage of shareholding | Number of Shares | Upon completion of the Share Subdivision, the Conversion of Unlisted Shares and the Global Offering | | | | | |---|---|---|---|---|---|---|---| | | | | Number of H Shares | Number of Unlisted Shares | Total number of Shares | Percentage of shareholding | Whether the H Shares count towards public float or not |
Haihe Fuxin Youda Fund . . . . . . . . . . . . . | 1,569,800 | 3.90% | 4,709,400 | 10,988,600 | 15,698,000 | 3.57% | Yes |
Fortune Capital Dachen Chuanghong . . | 823,341 | 2.04% | 2,470,023 | 5,763,387 | 8,233,410 | 1.87% | Yes |
Huakong Capital Beijing Huakong . . . . . | 300,752 | 0.75% | 1,503,760 | 1,503,760 | 3,007,520 | 0.68% | Yes | Qingdao Huakong . . . . . | 528,047 | 1.31% | 2,640,235 | 2,640,235 | 5,280,470 | 1.20% | Yes |
Zhongguancun Science City . . . . . . . . . . . . . | 826,211 | 2.05% | 2,478,633 | 5,783,477 | 8,262,110 | 1.88% | Yes | Zhuhai Huafa . . . . . . . . | 826,211 | 2.05% | 2,478,633 | 5,783,477 | 8,262,110 | 1.88% | Yes | Zhihui Linghang . . . . . | 826,211 | 2.05% | 4,131,055 | 4,131,055 | 8,262,110 | 1.88% | Yes | AI Fund . . . . . . . . . . . . | 795,574 | 1.98% | 3,977,870 | 3,977,870 | 7,955,740 | 1.81% | Yes | Tencent Investment . . . | 694,854 | 1.73% | 6,948,540 | — | 6,948,540 | 1.58% | Yes | Duoxiang Network . . . | 678,294 | 1.68% | 6,782,940 | — | 6,782,940 | 1.54% | Yes | CAS Star . . . . . . . . . . . | 591,227 | 1.47% | 1,773,681 | 4,138,589 | 5,912,270 | 1.34% | Yes |
Hangzhou Chengtou Industrial Fund . . . . | 578,347 | 1.44% | 1,735,041 | 4,048,429 | 5,783,470 | 1.31% | Yes | Beijing Shunying . . . . . | 521,141 | 1.29% | 5,211,410 | — | 5,211,410 | 1.18% | Yes | Prosperity7 . . . . . . . . . | 495,390 | 1.23% | 4,953,900 | — | 4,953,900 | 1.13% | Yes |
Daxing Industrial Fund . . . . . . . . . . . . . | 495,726 | 1.23% | 1,735,041 | 3,222,219 | 4,957,260 | 1.13% | Yes | High-tech Orinno . . . . | 495,726 | 1.23% | 1,982,904 | 2,974,356 | 4,957,260 | 1.13% | Yes | Xiamen Yaheng . . . . . . | 382,309 | 0.95% | 1,911,545 | 1,911,545 | 3,823,090 | 0.87% | Yes |
Guanghe Tianjin Heyuan . . | 321,175 | 0.80% | 3,211,750 | — | 3,211,750 | 0.73% | Yes | Hangzhou Guanghe . . . | 60,133 | 0.15% | 601,330 | — | 601,330 | 0.14% | Yes |
TAL . . . . . . . . . . . . . . . | 347,427 | 0.86% | 3,474,270 | — | 3,474,270 | 0.79% | Yes | Ningbo Mingheng . . . . | 347,427 | 0.86% | 1,042,281 | 2,431,989 | 3,474,270 | 0.79% | Yes | Hainan Hezun . . . . . . . | 300,752 | 0.75% | 902,256 | 2,105,264 | 3,007,520 | 0.68% | Yes |
Lingyue Lingyue No.4 . . . . | 208,618 | 0.52% | 625,854 | 1,460,326 | 2,086,180 | 0.47% | Yes | Lingyue No.5 . . . . | 39,245 | 0.10% | 117,735 | 274,715 | 392,450 | 0.09% | Yes |
Wuxi Yunhui . . . . . . . . | 232,545 | 0.58% | 2,325,450 | — | 2,325,450 | 0.53% | Yes | Tianchuang Capital . . . | 214,815 | 0.53% | 1,074,075 | 1,074,075 | 2,148,150 | 0.49% | Yes | 5G Fund . . . . . . . . . . . . | 209,290 | 0.52% | 1,046,450 | 1,046,450 | 2,092,900 | 0.48% | Yes |
Shangcheng Linghang . . . . . . . . . | 165,242 | 0.41% | 1,321,936 | 330,484 | 1,652,420 | 0.38% | Yes | Xiarui Investments . . . | 165,242 | 0.41% | 826,210 | 826,210 | 1,652,420 | 0.38% | Yes |
| Shareholder | As of the Latest Practicable Date | | Upon completion of the Share Subdivision, the Conversion of Unlisted Shares and the Global Offering | | | | | |---|---|---|---|---|---|---|---| | | Number of Shares | Percentage of shareholding | Number of H Shares | Number of Unlisted Shares | Total number of Shares | Percentage of shareholding | Whether the H Shares count towards public float or not | | Lenovo Venture Capital . . . . . . . . . . . | 165,242 | 0.41% | 660,968 | 991,452 | 1,652,420 | 0.38% | Yes | | Jia'an Qixin . . . . . . . . . | 165,242 | 0.41% | 660,968 | 991,452 | 1,652,420 | 0.38% | Yes | | Jiangmen Venture Capital . . . . . . . . . | 150,376 | 0.37% | 451,128 | 1,052,632 | 1,503,760 | 0.34% | Yes | | Luster . . . . . . . . . . . . . . | 150,376 | 0.37% | 451,128 | 1,052,632 | 1,503,760 | 0.34% | Yes | | Xiaofeng Technology . . . . . . . | 148,672 | 0.37% | 1,486,720 | — | 1,486,720 | 0.34% | Yes | | Huahai Jinpu . . . . . . . . | 103,276 | 0.26% | 1,032,760 | — | 1,032,760 | 0.23% | Yes | | Zhaoshang Shuke . . . . | 99,224 | 0.25% | 992,240 | — | 992,240 | 0.23% | Yes | | Innovation Zhiyuan . . . | 94,156 | 0.23% | 282,468 | 659,092 | 941,560 | 0.21% | Yes | | Rongjia Xingpu . . . . . . | 86,629 | 0.22% | 433,145 | 433,145 | 866,290 | 0.20% | Yes | | Lianrong Zhiyuan . . . . | 86,062 | 0.21% | 860,620 | — | 860,620 | 0.20% | Yes | | Lingtou Future . . . . . . . | 81,382 | 0.20% | — | 813,820 | 813,820 | 0.18% | — | | Shanchuang Zhizhi . . . | 42,920 | 0.11% | 429,200 | — | 429,200 | 0.10% | Yes | | Hengqin Shanzhi . . . . . | 42,920 | 0.11% | 429,200 | — | 429,200 | 0.10% | Yes | | Zhongxiao Ruizheng . . . . . . . . . | 40,303 | 0.10% | 201,515 | 201,515 | 403,030 | 0.09% | Yes | | Shandong Fuhong . . . . | 30,210 | 0.07% | 151,050 | 151,050 | 302,100 | 0.07% | Yes | | Public Shareholders taking part in the Global Offering . . . . | — | — | 37,419,500 | — | 37,419,500 | 8.50% | Yes | | Total . . . . . . . . . . . . . . | 40,281,069 | 100% | 215,701,705 | 224,528,485 | 440,230,190 | 100.00% | |
由汇惠及智登（均为本公司控股股东集团成员）持有的合计10,001,857股由非上市股份转换而来的H股（计及股份细分后），占截至最后实际可行日期本公司已发行股份总数约2.48%，或于上市时（假设超额配股权未获行使）占本公司已发行股份总数约2.27%，根据《上市规则》第19A.13A(1)条，该等H股将不计入本公司的公众持股量。
此外，合计224,528,485股（计及股份细分后）不会转换为H股的非上市股份，占截至最后实际可行日期本公司已发行股份总数约55.74%，或于上市时（假设超额配股权未获行使）占本公司已发行股份总数约51.00%，将不计入本公司的公众持股量。
据本公司董事所知及所获悉，计及股份细分及非上市股份于上市时转换为H股后，205,699,848股H股将根据《上市规则》第19A.13A(1)条计入本公司的公众持股量，于上市时（假设超额配股权未获行使）占本公司已发行股份总数约46.73%。
假设超额配股权未获行使，按每股发售股份116.20港元的发售价计算，本公司于上市时的预期市值约为512亿港元，届时公众人士持有的H股预期市值达45亿港元所对应的百分比约为8.80%。据此，根据《上市规则》第19A.13A(1)条适用于本公司H股的最低规定公众持股量百分比为10%。因此，本公司将能够满足《上市规则》第19A.13A条规定的最低公众持股量要求，因为于全球发售完成后，逾10%的本公司已发行股份总数将由公众人士持有。
根据适用中国法律，所有现有股东（包括本公司上市前投资者）均被禁止在上市日期后12个月内出售其持有的任何股份。
| 人士 | 身份 | 于股份细分及全球发售完成后须遵守出售限制的股份数目 | 于全球发售完成后（假设超额配股权未获行使）须遵守出售限制的持股比例 | 商业公司禁售期 | |---|---|---|---|---| | 刘博士 | 联合创始人、执行董事兼董事会主席 | 925,150 | 0.21% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 北京联湃(1) | 由刘博士控制的持股平台 | 34,038,390 | 7.73% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 汇惠(2) | 由刘博士控制的员工激励平台 | 39,482,710 | 8.97% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 智登(2) | 由刘博士控制的员工激励平台 | 27,196,330 | 6.18% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 唐博士 | 联合创始人 | 26,835,330 | 6.10% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 李博士 | 联合创始人兼非执行董事 | 3,367,760 | 0.76% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 许博士 | 联合创始人 | 799,040 | 0.18% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 张博士 | 联合创始人兼执行董事 | 399,520 | 0.09% | 自本招股章程日期起至上市日期起计12个月届满为止 | | 探路者SII 君联资本SII | 探路者SII | 27,109,120(3) | 6.16% | 自本招股章程日期起至上市日期起计6个月届满为止 | | 美团SII | 探路者SII | 17,217,310(4) | 3.91% | 自本招股章程日期起至上市日期起计6个月届满为止 |
(1) 截至最后实际可行日期，王少兰先生（本公司联合创始人）及张钹（本公司首席科学家）为北京联湃的有限合伙人，须根据《上市规则》第18C.14条遵守出售限制。
(2) 截至最后实际可行日期，汇惠及智登均作为本公司的员工持股平台而设立，符合条件的参与者（包括本公司执行董事刘博士、张博士及张晓晗女士以及本集团其他核心研发员工）获授予员工持股平台的合伙权益。就《上市规则》第18C.14条而言，本公司的核心研发员工包括：
• 杜正晓博士，算法专家。
详情请参阅「—员工持股平台」。本公司上述所有执行董事、高级管理层成员及关键人士须根据《上市规则》第18C.14条遵守出售限制。
(3) 指于全球发售完成后由君联向导、君联金帆及社保中关村创新基金持有的股份。
(4) 指于全球发售完成后由天津三快持有的股份。
Pursuant to Rule 19A.13C(1) of the Listing Rules, our Company must ensure that a portion of the total number of its issued shares listed on the Stock Exchange with a market capitalization of at least HK$600,000,000 are not subject to any disposal restrictions (whether under contract, the Listing Rules, applicable laws or otherwise) at the time of Listing.
On the basis that (i) no Offer Shares will be allocated under the Global Offering to any core connected person of our Company or person which is not regarded as a member of the public under Rule 8.24 of the Listing Rules and (ii) all Offer Shares to be issued to the cornerstone investors are excluded for the purpose of satisfying the free float requirement, upon completion of the Global Offering (assuming the completion of Share Subdivision and Conversion of Unlisted Shares and the Over-allotment Option is not exercised), it is expected that 11,737,900 H Shares will not be subject to any disposal restrictions (whether under contract, the Listing Rules, applicable laws or otherwise) at the time of the Listing, representing a market value of approximately HK$1,363.9 million based on the Offer Price of HK$116.20 per Offer Share at the time of Listing, which is over HK$600 million and will satisfy the free float requirement under Rule 19A.13C(1)(b) of the Listing Rules.
The following chart sets forth our corporate and shareholding structure immediately before the completion of the Global Offering:
Dr. Liu (1) | Dr. Tang(1) | Huihui(1) | Beijing Lianpai(1) | Dr. Li(1) | Dr. Xu(1) | Dr. Zhang(1) | Zhideng(1) | Other Shareholders | Legend Capital SIIs(2) | Meituan SII(2) | Qiming Venture SIIs(2) 8.45% | 0.23% | | | 9.80% | 6.75% | 6.66% | 0.84% | 0.20% | 6.73% | 4.27% | 2.49% | 0.10% | 53.27%
| 100% | 100% | 70% | 100% | 100% | 100% | 100% | 100% | | Tianjin Knowledge Atlas | Beijing Knowledge Xingyao | Beijing Knowledge Huixing(3) | Nanjing Knowledge Atlas | Shenzhen Knowledge Atlas | Beijing Knowledge Linghang | Beijing Lingxin Intelligent | Shenzhen Lingxin Intelligent |
| 100% | 100% | 100% | 100% | 100% | 100% | 100% | 100% | | Hangzhou Knowledge Atlas | Beijing Knowledge Qingyan | Shanghai Knowledge Huanyu | Beijing Knowledge Qingying | Beijing Knowledge Future | Beijing Knowledge Haiying | Zhejiang Knowledge Xinpian | Chengdu Knowledge Atlas |
| 100% | 100% | 100% | 100% | 100% | 100% | 100% | | Zhuhai Knowledge Future | Zhuhai Knowledge Linghang | Jingsheng Hengxing | Huangshi Knowledge Atlas | Jincheng Yaoda | Xiangtai Ruifeng | ZYNIX LIMITED |
Notes: (1) Beijing Lianpai, Dr. Liu, Dr. Tang, Dr. Li, Dr. Xu, Dr. Zhang, Huihui and Zhideng are parties acting in concert and are our group of Controlling Shareholders. See "—Concert Party Arrangement and Our Controlling Shareholders" for details. Beijing Lianpai, Huihui and Zhideng are controlled by Dr. Liu by virtue of Dr. Liu serving as their respective general partner. (2) For the details of the background information of Legend Capital SIIs, Meituan SII, Qiming Venture SIIs and other Pre-IPO Investors, see "—Pre-IPO Investments" for details. (3) The remaining 30% equity interest of Beijing Knowledge Huixing is held by Hainan Hezun, an Independent Third Party.
The following chart sets forth our corporate and shareholding structure immediately after completion of the Global Offering (assuming the Over-allotment Option is not exercised):
Dr. Liu(1) | Dr. Tang(1) | Huihui(1) | Beijing Lianpai(1) | Dr. Li(1) | Dr. Xu(1) | Dr. Zhang(1) | Zhideng(1) | Other Shareholders | Legend Capital SIIs(2) | Meituan SII(2) | Qiming Venture SIIs(2) | Other Public Shareholders 7.73% | 0.21% | | | 8.97% | 6.18% | 6.10% | 0.76% | 0.18% | 6.16% | 3.91% | 2.28% | 0.09% | 48.74% | 8.50%
| 100% | 100% | 70% | 100% | 100% | 100% | 100% | 100% | | Tianjin Knowledge Atlas | Beijing Knowledge Xingyao | Beijing Knowledge Huixing(3) | Nanjing Knowledge Atlas | Shenzhen Knowledge Atlas | Beijing Knowledge Linghang | Beijing Lingxin Intelligent | Shenzhen Lingxin Intelligent |
| 100% | 100% | 100% | 100% | 100% | 100% | 100% | 100% | | Hangzhou Knowledge Atlas | Beijing Knowledge Qingyan | Shanghai Knowledge Huanyu | Beijing Knowledge Qingying | Beijing Knowledge Future | Beijing Knowledge Haiying | Zhejiang Knowledge Xinpian | Chengdu Knowledge Atlas |
| 100% | 100% | 100% | 100% | 100% | 100% | 100% | | Zhuhai Knowledge Future | Zhuhai Knowledge Linghang | Jingsheng Hengxing | Huangshi Knowledge Atlas | Jincheng Yaoda | Xiangtai Ruifeng | ZYNIX LIMITED |
Notes: (1)-(3) Please refer to the notes in "—Shareholding and Corporate Structure—Corporate Structure Immediately Before the Global Offering" above.
BUSINESS OVERVIEW Who We Are We are a leading AI company in China, dedicated to developing general-purpose large models. We were founded in 2019 on the bold idea of pursuing innovation toward artificial general intelligence (AGI) in China. We have solidly delivered advanced technology across the full spectrum of AI research and steadily scaled up its commercial application to achieve fast growth in revenue. In 2021, we launched GLM framework, China's first proprietary pre-trained large model framework, and debuted our Model-as-a-Service (MaaS) product development and commercialization platform, through which we provide our large model services. In 2022, we open-sourced our first 100 billion–scale model (GLM-130B). We operate in the large language model (LLM) market, a sub-segment of the broader AI market. We offer general-purpose large model services to institutional customers, including private enterprises and public sector entities, as well as individual users, including individual end-users and individual developers. Our models had empowered over eight thousand institutional customers as of June 30, 2025 and approximately 80 million devices as of the Latest Practicable Date. According to Frost & Sullivan, we ranked first among China's independent developers and second among all developers of general-purpose large models in terms of revenue in 2024.
How We Define AGI What does AGI look like and how do we get there? Although there is no official industry standard, we think it follows this roadmap:
• Pre-training stage. We teach machines to understand, write and speak human languages.
• Alignment and reasoning stage. We align machines with human intentions and teach them to reason and plan. This improves safety, reduces hallucination and enables alignment with images, videos, audios and actions.
• Self-learning stage. We teach machines to learn from what they thought and did through self-critique, self-reflection and rumination.
• Self-perception stage. Without human supervision, machines develop their own attitudes and emotions by observing their behavior and interpreting it themselves.
• Consciousness stage. Machines become aware of their internal and external existence, similar to human consciousness (subject to the complexities of that concept even for humans).
We have developed large models and agents across the first three stages, such as GLM-130B, China's first 100 billion-scale large model (at the pre-training stage), ChatGLM, China's first open-source large chat model (at the alignment and reasoning stage) and AutoGLM – Rumination, China's first AI agent with rumination capabilities (at the self-learning stage). We are advancing towards the self-perception stage and consciousness stage, standing on the frontier of AGI innovation.
Our MaaS Platform We aim to teach machines to think like humans and benefit humanity with reliable AI. Since we released our first model in 2020, as we continue to iterate, we have started to explore how our technology, so long as it is reliable, can be useful for enterprises and individuals. We commenced commercialization in 2021, two years earlier than scaled commercialization of general-purpose large models began in China
• 全面的模型产品组合。我们建立了一个全面的先进AI模型组合，在语言、多模态、智能体及代码能力方面展示了行业领先的性能。凭借我们广泛而强大的产品储备，客户和开发者始终能够找到最适合其特定需求的解决方案。
• 赋能企业与个人。我们MaaS商业模式的一项显著优势在于，我们有潜力借助机构客户触达其自身的客户群体，从而以间接但高效的方式将我们的影响力延伸至这些终端用户。根据弗若斯特沙利文的资料，截至2025年6月30日，在独立大模型服务提供商中，我们在中国赋能的设备数量最多；在所有大模型服务提供商中，我们赋能的设备数量位居第二。
• 便捷的定制化部署。我们的模型可托管于云端、通过应用程序接口（API）访问、本地化部署以计算私有数据集，或预装于设备端。我们还提供即插即用的模板和插件，以及用于模型微调、部署和智能体开发的标准化集成工具。上述方式支持快速的场景化、无代码或低代码模型开发，推动跨行业垂直领域的高度可扩展性。
• 安全性与可靠性。我们已构建安全、可扩展的架构，用以开发安全可靠的模型，并以领先的安全性能作为支撑。例如，根据斯坦福大学发布的《2025年人工智能指数报告》，经Hughes幻觉评估模型（HHEM-2.1-Open）评测，我们的GLM-4-9B模型在顶级模型中实现了最低的幻觉率之一（1.3%）。
• "MaaS闭环"。借助我们的MaaS平台，我们已建立起一个连接算力资源提供商、智能设备制造商、机构客户、开发者及个人客户的繁荣生态网络。我们从模型的实际部署中积累了大量洞察，使我们能够更好地理解用户在不同使用场景下实际使用和受益于我们AI的方式，并以更具针对性的方式优化训练策略，从而形成良性的洞察飞轮效应。
我们相信，随着模型能力的持续提升，AI赋能应用、AI智能设备及AI转型组织将得到更广泛的普及。
在业绩记录期间，我们的收入实现显著增长。2022年、2023年及2024年，我们的收入分别为人民币5,740万元、人民币1.245亿元及人民币3.124亿元，复合年增长率超过130%。截至2024年6月30日及2025年6月30日止六个月，我们的收入分别为人民币4,490万元及人民币1.909亿元。
China's Largest Independent general-purpose large model developer by revenue China's Second Largest Overall general-purpose large model developer by revenue
RMB57.4 million, RMB124.5 million, RMB312.4 million and RMB190.9 million Revenue in 2022, 2023, 2024 and the six months ended June 30, 2025
Notes: (1) According to Frost & Sullivan. (2) As of June 30, 2025. (3) Included smart phones, personal computers and smart vehicles. (4) As of the Latest Practicable Date. (5) Ranked in November 2025 by CodeArena, the latest industry-recognized global evaluation platform designed to assess models' coding capabilities.
Our Ecosystem We have fostered an ecosystem based on three pillars: open-source collaboration, robust computing infrastructure and diverse industry partnership. From foundational research to real-world application, participants in our ecosystem develop dynamic and interdependent relationships, creating an environment where innovation is continuously stimulated, reinforced and scaled across sectors.
• Developers – Advancing innovation through open source. We see open-sourcing as a key catalyst for the vitality of our ecosystem. We have broadly open-sourced our models, inviting global participation in a transparent governance structure. Developers (i.e. programmers) worldwide collaborate at multiple levels around our open-source offerings, ranging from foundational models to algorithm plugins and security enhancements to performance optimization and multilingual adaptation. As of June 30, 2025, our open-source models had been downloaded for over 45 million times in the global developer community, and more than 1,000 open-source projects had been created based on our models, through which developers customized the open-source models through model fine-tuning and incremental model training so that our models can serve their specific needs. For example, based on our open-source models, developers built an open-source project named "LangChain-Chatchat," which is a Q&A chatbot customized based on the user's local knowledge base.
基础设施提供商——促进广泛兼容性。我们致力于实现广泛的计算兼容性，使模型能够轻松部署于各类计算基础设施（"infra"），例如基于云的大规模集群、异构高性能服务器以及边缘嵌入式加速器。截至2025年6月30日，我们的模型已兼容全球超过40个主要芯片平台。在这一统一系统中，开发者可发起请求、优化性能并实现高效的资源利用。这种兼容性使先进的大型模型得以在异构基础设施环境中部署，同时也使我们的基础设施合作伙伴能够通过集成我们的模型拓展更广阔的商业前景。
业务合作伙伴——推动行业转型。我们的业务合作伙伴网络定义了生态系统的上限边界。凭借我们强大的基础模型、顶级的计算兼容性以及全面的开发工具，业务合作伙伴能够将人工智能能力迅速延伸至各行各业。每一次部署不仅为模型创新提供新鲜数据、为工具优化提供真实反馈，还向更广泛的社区贡献新的行业实践。这形成了持续改进的良性循环，将新的洞见不断反哺至我们的基础模型与共享工具之中。
我们憧憬一个人工智能成为向善变革力量的世界——在那个世界里，人工智能（最终演进为通用人工智能，即AGI）将赋能人类改善地球每一个角落的生活，并应对我们时代最严峻的挑战。
我们认为，此次人工智能革命具有两个方面的特质，使其将比人类历史上早期的工业革命更具变革性。一方面，人工智能将成为无处不在的基础资源，如同石油一般，随时随地驱动各项工作的运转，从而将人类从中解放出来。更为可贵的是，这种新资源的供给可以是无限的，因为它并非从地壳中开采而来，而是从人类最优秀的思维深处工程化提炼而成。另一方面，人工智能将不仅仅是一种工具性资源。凭借潜在的无限规模与超越人类的能力，人工智能或许能够实现人类独力无法完成的科学突破。若得到妥善驾驭，我们相信人工智能将带来难以想象的进步，推动人类文明达到前所未知的高度。
与此同时，创造模拟人类智能的机器系统本身蕴含着深刻的风险。若处置不当，人工智能可能造成严重危害，甚至引发灾难性后果。尤其是，人工智能将渗透入人类社会最细微的肌理与每个人生活最隐秘的角落，每个社区的文化与每个个体的福祉都将受其影响。因此，我们必须以极度审慎的态度对待人工智能。
作为基础人工智能技术领域的引领者，我们深知自身肩负的责任。我们所做的工作本质上涵盖两个方面：提升技术以使其更加实用，以及拓展应用以使其服务于更有意义的目的。为此，为确保我们的人工智能始终成为全人类的向善力量，我们致力于恪守核心人工智能安全与伦理原则，并以这些原则引领我们的组织及一切行动。安全是我们首要且最重要的优先事项。我们是唯一一家与国际同行共同签署2024年首尔人工智能峰会《前沿人工智能安全承诺》的中国人工智能公司。我们同样高度重视民主化、多样性、公平性、环境可持续性与透明度等伦理层面的考量。尤其是，我们将竭力确保人工智能不会侵蚀使我们成其为人的丰富文化多样性。详见"——我们的核心人工智能安全与伦理原则"。
负责任地行动，要求我们在每一步都具备前瞻性、保持勤勉，并怀有充分的谦逊。我们坦然承认自身知识的局限。人工智能的发展将踏入未知领域，提出我们尚无法回答的问题。在砥砺前行之际，我们承诺保持谦逊，持续审视自身的价值观、假设前提与方法论。当我们犯错时，我们将正视错误、加以纠正并从中汲取教训。通过将自身锚定于责任、透明与协作，我们将努力确保人工智能成为推动人类持续繁荣的载体。
根据弗若斯特沙利文的资料，我们是中国首家自研超千亿参数大模型的AI公司，并于2024年在通用大模型收入方面在中国独立开发者中排名第一，在所有开发者中排名第二。
创新基础架构。我们的技术领先地位建立在GLM框架之上，该框架是中国首个自主研发的预训练大模型框架。与主流的单向架构不同，我们的创新设计独特地融合了单向和双向注意力机制。这使我们的模型在复杂语言理解和细致生成方面均具备扎实能力，为我们所有模型奠定了稳固而多元的基础。
开创性模型创新。我们持续推出突破性模型，不断拓展大模型能力的边界。2022年，我们推出了GLM-130B，这是中国首个千亿参数规模的开源预训练模型。2024年10月，我们发布了设备控制智能体模型AutoGLM，该模型最初专为手机使用而设计，随后我们进一步推出了计算机操控AI智能体GLM-PC，与其他全球领先AI公司实质上同步发布。2025年3月，我们推出了AutoGLM-Rumination，这是前沿AI智能体的重大突破，能够自主执行需要持续数十分钟"边思考边工作"的智能体操作。2025年7月，我们推出了GLM-4.5，这是我们的旗舰基础模型，具备强大的智能体、推理和代码生成能力。根据弗若斯特沙利文的资料，基于2025年7月对十二项行业标准基准测试的综合评估，GLM-4.5全球排名第三，中国排名第一，全球开源模型中排名第一。
早期商业化战略与先发优势。我们的技术实力伴随着商业化方面的先发优势。我们于2021年开始商业化，根据弗若斯特沙利文的资料，比中国通用大模型规模化商业化的起始时间早两年。我们已通过MaaS平台及行业合作伙伴关系成功构建起规模庞大且可持续扩展的业务。这一市场领先地位因大量用户的采用和参与而得到进一步巩固，从而形成真实世界洞察、模型优化与持续增长的良性循环。
根据弗若斯特沙利文的资料，基于我们自主研发的大模型预训练框架GLM，我们构建了全面的模型产品组合。这使我们能够为客户提供多样化的基础模型，并具备快速定制的能力。我们的模型矩阵涵盖轻量级端侧模型、高性价比模型及千亿参数规模旗舰大模型，每类模型均专为满足特定客户需求而构建。该产品组合支持全面的功能，包括对话、通用智能体、代码生成、图像理解及文生图/视频生成（含语音能力），实现了对所有主要大模型应用场景的全面覆盖。
语言模型——高级智能的可靠基础。我们的旗舰模型GLM-4.5的能力可与全球最先进的模型相媲美。根据弗若斯特沙利文（Frost & Sullivan）于2025年7月对十二项行业标准基准测试的综合评估，GLM-4.5全球排名第三，中国排名第一，全球开源模型中排名第一。GLM-4.5发布后仅48小时内，便在全球最大开源模型平台Hugging Face的热门榜上荣登全球第一，充分彰显了其在全球用户中的广泛影响力。
智能体模型——复杂任务与设备控制的自动化。我们的智能体模型专为驱动通用智能体而设计，能够代表人类理解、思考并执行复杂任务。该能力建立在两个核心模型之上：GLM-Z1深思（GLM-Z1-Rumination）作为智能体的"大脑"，负责深度推理与自主规划；AutoGLM则充当智能体的"双手"，能够感知并操作任意用户界面。AutoGLM在AgentBench基准测试中达到了SOTA性能——AgentBench是斯坦福大学发布的2024年AI指数所认可的智能体AI基准。
多模态模型——重新定义内容创作。我们的领先地位覆盖多模态AI的全谱系。在文本生成图像领域，我们的CogView4在DPG-Bench基准测试中排名第一，同时也是全球首个能够准确理解并生成中文文本的开源模型。在视频生成领域，CogVideoX表现卓越，在综合性SuperCLUE-I2V基准测试中位居领先行列。GLM-4V作为中国首个开源双语多模态对话模型，以及GLM-4-Voice作为中国首个端到端超真实感语音模型，进一步巩固了我们全面的多模态领先地位，彰显了我们在所有主要模态上的开创性能力。
编程模型——变革编程体验。CodeGeeX于2023年发布并持续迭代，始终展现出卓越的性能表现，并在开发者社区中获得了热烈的广泛采用。截至2025年6月30日，CodeGeeX平均每日生成代码超过1亿行。
Supported by our comprehensive model portfolio, our customers can flexibly access and integrate a wide range of model capabilities to meet their diverse needs. Specifically, our MaaS platform provides an agent workspace, which encompasses a variety of agent templates and scenario-based solutions. Through this agent workspace, our customers can swiftly customize agents through streamlined model fine-tuning, incremental model training and prompt engineering. See also "—All-in-one MaaS platform maximizing model commercialization" below for details. They also provide valuable feedback on industry applications and scenario demands. The collaboration helps us refine our large model pre-training strategies, enhance development efficiency and lower training costs. Ultimately, it creates a positive cycle that strengthens the general capabilities of our foundation models and expands both the breadth and depth of our model portfolio. This allows us to better serve the diverse needs of a broader customer base and drive continuous expansion of large model applications.
我们处于人工智能研究的前沿，在全球学术界获得广泛认可。2024年5月，国际顶级学术期刊《自然》（NATURE）刊载了我们的ChatGLM模型，将其列为中国基础大模型的杰出代表。我们还在全球顶级学术会议上发表研究成果，包括国际学习表征会议（ICLR）和国际万维网大会（WWW）。截至2025年6月30日，我们的精英研究团队及学术顾问团队已发表约500篇具有重要影响力的论文，被引用总次数逾58,000次。
We are at the forefront of AI research with significant recognition in the global academic community. In May 2024, NATURE, a leading international academic journal, featured our ChatGLM models, citing them as prominent representatives of Chinese foundational large models. We have also presented our research at top global conferences, including the International Conference on Learning Representations (ICLR) and the International World Wide Web Conference (WWW). As of June 30, 2025, our elite research and academic advisory teams had published approximately 500 highly influential papers, which had been cited over 58,000 times collectively.
We are, at the core, a company of data scientists and engineers, with R&D ingrained in every aspect of what we do. We are singularly focused on elevating the intelligence of our foundation models, driving incremental progress toward more advanced AI. This is achieved first and foremost through our people. As of June 30, 2025, over 74% of our employees are dedicated to research and development, with core members bringing experience from leading research institutions and universities such as Tsinghua University. Our team specializes in areas such as natural language processing, advanced decision-making in complex systems and multimodal semantic analysis.
Our core technology team maintains a close and enduring partnership with the Knowledge Engineering Group (KEG) at Tsinghua University, forming a distinctive and robust system for technological exchange. This association, rooted in a shared academic lineage, helps instill continuity and coherence in our technological roadmap. It enables us to systematically collaborate with KEG, one of China's leading AI laboratories, to achieve cutting-edge research outcomes. We actively engage in extensive research collaborations with other top-tier universities and AI research institutions as well, which also allows us to identify and nurture promising AI scientists at an early stage and secure a steady inflow of high-caliber talent.
Our all-in-one MaaS platform is built on our advanced model capabilities. The platform provides access to a matrix of models and a suite of agentic tools—including agent development and model fine-tuning platforms—with flexible deployment options. Any institutional customer, regardless of its in-house AI expertise, can efficiently and cost-effectively use and fine-tune models, develop applications, and harness the agentic capabilities across diverse industry scenarios through our MaaS platform.
Our comprehensive model portfolio empowers customers with foundational capabilities to solve general-purpose tasks. Our extensive suite of agentic tools, such as scenario templates and plugin libraries, further enable quick model customization and application development. Customers can easily build dedicated models and solutions tailored to specific sectors or scenarios. This approach allows us to achieve greater operational efficiency and economies of scale, surpassing traditional project-based AI companies. Within our intelligent platform, customers can seamlessly access and deploy a range of AI agent capabilities to address their unique requirements and workflows, satisfying diverse deployment needs and driving broad AI transformation for enterprises and developers.
Our MaaS platform has earned a solid customer base. This vibrant and dedicated community provides a solid foundation for ongoing advancements in our model and agent AI capabilities and accelerates our commercialization trajectory.
We have developed a diverse ecosystem involving developers, computing infrastructure providers and industry business partners, which creates a positive dynamic that accelerates our progress.
We were the first Chinese AI company to open-source proprietary large models. In 2023, we were named the fifth most popular open-source large model provider globally by Stanford University, the only Chinese company to earn this recognition. As of June 30, 2025, we had open-sourced over 50 models, and our open-source large models had been downloaded over 45 million times. Our open-source strategy has formed a robust and dynamic developer community. Through this community, we continuously enhance our industry influence and global brand presence and secure mindshare among our customers.
We maintain close partnerships with upstream leading computing power infrastructure providers. Our model capabilities allow us to define the computing power required. In turn, advances in computing power infrastructure drive the development of next-generation models. Through these partnerships, we have significantly enhanced the compatibility of our models with both domestic and international mainstream hardware chips. As of June 30, 2025, our models were compatible with more than 40 major hardware chips. By developing proprietary operator libraries and related technologies, we are able to ensure efficient model training across diverse hardware environments. We can also provide model-chip-integrated solutions tailored to specific needs of our institutional customers, which expands our network of hardware partners.
We also collaborate with lighthouse customers and specialized industry partners to create benchmark use cases across industries. Leveraging our MaaS platform, we empower our partners with AI agent capabilities to enhance their products and deliver greater value to end users. Through our industry network development program Project Z, we empower our industry business partners to explore opportunities in long-tail market scenarios. Participants in Project Z are typically early-stage start-ups seeking to implement large models within specific industry sectors. They often have limited technical capabilities or lack the resources to develop large models independently. We provide these companies with access to our large models and infrastructure and offer tailored technical support. This approach enables our industry business partners to scale their operations efficiently, while concurrently facilitating the expansion of our large models into new industry scenarios. In this way, our business growth is supported not only by our own efforts, but also by the collective success of our business partners. In addition, we place great importance on regional collaboration in AI development, actively participating in building national and municipal foundation model platforms in countries and regions such as China and Southeast Asia.
Our management and scientific advisory teams bring together a wealth of expertise in AI, large models and internet-sector applications. With proven track records in technological advancement and successful commercialization across AI, big data and related fields, they continue to steer our pursuit of technological leadership and sustainable business growth. Their global perspective and forward-thinking research agenda have shaped our long-term strategic direction.
We have a management team with elite academic and industry backgrounds. Dr. Liu Debing, our co-founder, executive Director and chairman of the Board, brings extensive expertise in large language models, machine learning and data mining. He has led or participated in more than thirty major scientific research projects, collaborating with institutions such as the Ministry of Science and Technology and Chinese Academy of Engineering. Dr. Liu plays a leading role in driving our technical innovation. Dr. Zhang Peng, our co-founder, executive Director and CEO, specializes in knowledge graphs and large-scale pre-trained models and is a core contributor to the development of our GLM model series and AMiner. With over ten publications in leading conferences such as International Conference on Machine Learning (ICML) and International Semantic Web Conference (ISWC), Dr. Zhang is a prominent figure in translating AI research into impactful real-world applications.
We benefit profoundly from the expertise of and guidance from our esteemed academic advisors. Professor Zhang Bo, our chief scientist, is an academician of the Chinese Academy of Sciences. Professor Zhang has played a pivotal role in both foundational AI theory and practical applications such as pattern recognition, knowledge engineering and robotics. In addition, we have collaborated closely with the Department of Computer Science and Technology at Tsinghua University. KEG, a research group of the department, focuses on original innovation in AI driven by both data and knowledge, conducting research in areas such as AI, LLMs, knowledge graphs, data mining and social networks. KEG is led by Professor Li Juanzi, a leading expert in knowledge graphs, semantic content management and social network mining. Her insights help us bridge foundational AI research with real-world system deployment. Since 2018, with the internationally acclaimed leading scientists in artificial intelligence such as Dr. Tang Jie, KEG has pioneered research in cognitive intelligence and spearheaded the development of the GLM large model framework, representing a milestone in China-originated large model innovation.
With the leadership of our management team, we have assembled a world-class team of scientists and engineers, combining cutting-edge technical expertise with extensive experience in translating research into tangible industry impact.
We plan to implement the following strategies to further develop our business.
We are continuously developing foundation models which can explore the upper bounds of intelligence. We are committed to consolidating our technological leadership by investing in the iterative development of our foundation models, key algorithms and large-scale training and inference infrastructures. Specifically:
• Enhance large model capabilities and optimize training infrastructure. We will continue to invest in building versatile, powerful large models, focusing on designing new model architecture, and optimizing training infrastructure to reduce training costs and boost performance. This would include new attention and memory mechanisms, infinity context, test-time and online learning, deep reasoning algorithms, self-refinement and self-evolution paradigms. By these means, our
models will be able to autonomously explore, reflect and operate in both online and real-world environments, enabling them to tackle complex tasks and continuously improve their performance through self-learning and reflection.
Strengthen our GLM framework and optimize data processing platform. We will continue to improve our GLM framework. Specifically, we plan to continue investing in the research and development of our technology infrastructure to improve model performance, increase the efficiency of underlying computing resources and ensure that computing resources provided by our computing resource partners are optimally suited to our models and sufficient for our scalable training and model deployment across diverse platforms and hardware. We plan to upgrade our data processing platforms to support high-quality storage and analysis of datasets, providing a solid foundation for further expanding the frontiers of large model capabilities.
Continue to develop AI agents. We are building models that can drive innovation and collaboratively work as multiple agents in an organization. We plan to further upgrade our agent workspace that enables customers to easily and seamlessly integrate diverse model applications and tools. This will facilitate the efficient deployment of AI agent solutions deeply tailored to specific industries, regions and scenarios, and drive advancements in intelligent automation. We are building foundation models for embodied intelligence which can interact with the physical world. In the future, we will push the boundaries of large foundation models to understand emotion and consciousness such that we can build a personalized, superintelligent assistant for everyone.
We are committed to optimizing our MaaS platform and deepening our industry engagement. Through a combination of technical innovation, industry-focused expansion and targeted marketing initiatives, we aim to make our large models accessible to more industry sectors and users. Specifically:
Further diversify our model offerings through the MaaS platform. We are strategically expanding our model portfolio across a spectrum of parameter scales, ensuring optimal performance on various computing resources and devices. By engineering models that adapt to differing computational capabilities, we reduce hardware barriers and democratize access to advanced AI technologies. At the application level, we remain focused on delivering convenient, user-friendly experiences that make large model technology more accessible, enabling companies and organizations of all sizes, as well as individual customers, to unlock large models' potential.
Empower a wider range of institutional customers, end-users and business partners. We will continue to enhance the capabilities of our foundation models, especially their adaptability with diverse industry sectors and application scenarios. In addition, through our scalable MaaS platform, we aim to leverage our institutional customers to reach their end users on a significant magnitude. When our technology helps our institutional customers better serve their consumer clients, we ourselves reach and extend our influence to these end users in an indirect but efficient way. Specifically, we have collaborated with several leading global consumer electronics providers, deploying our models on their flagship products to support on-device intelligent interactions, facilitate multilingual conversations and provide personalized recommendations. As consumer electronics companies increasingly recognize the transformative potential of AI, we anticipate their willingness to adopt such technologies will continue to grow. Looking ahead, we intend to further expand the role of AI in consumer electronics, aiming to enable cross-device collaborations and adaptive edge intelligence (i.e. on-device AI systems that can autonomously adjust its behavior in response to dynamic changes in the environment). We have also formed partnerships with several leading Chinese internet companies (e.g., a prominent social and lifestyle platform), enhancing functions such as translation, data analytics, content creation and
search with our AI capabilities. As the strategic value of AI becomes more widely appreciated among internet companies, we aim to expand the application of our models to an even broader array of use cases, such as advertising recommendations, intelligent customer service and knowledge analysis.
In addition, we will continue to invest in Project Z, our industry network development program, which we believe will facilitate our expansion into new industry sectors. Participants in Project Z are typically early-stage start-ups seeking to implement large models within specific industry sectors. They often have limited technical capabilities or lack the resources to develop large models independently. We provide these companies with access to our large models and infrastructure and offer tailored technical support. This approach enables our industry business partners to scale their operations efficiently, while concurrently facilitating the expansion of our large models into new industry scenarios. In this way, our business growth is supported not only by our own efforts, but also by the collective success of our business partners.
As a company built by scientists and engineers, talent serves as the bedrock of our sustained success and long-term growth. We have recruited many of the most talented scientists and professionals across the world who are driven by a shared ambition to advance general-purpose AI.
We are committed to attracting and retaining top-tier R&D professionals as well as drawing specialists from a wide range of disciplines through comprehensive talent development programs and competitive incentive structures. We actively foster an environment where scientific inquiry is valued and innovation is encouraged at every level. By cultivating a strong sense of belonging and loyalty among our team, we empower each team member to contribute to our collective mission. This unwavering focus on talent enables us to build and sustain a highly skilled workforce and maintain a robust foundation that drives our ongoing development and secures our leadership in the industry.
As we commercialize our technology to seize the tremendous market opportunity presented by advanced AI, we organize our offerings around our all-in-one MaaS platform. Through this product development and commercialization platform, we deliver intelligence to institutional customers, developers and individual customers in the most suitable, sensible and scalable way despite great heterogeneity in computing infrastructure, devices and applications.
• 全面的模型组合。我们构建了一套全面的先进AI模型组合，在语言、多模态、智能体及编程能力方面展现出业界领先的性能。凭借我们广泛而强大的模型库，客户和开发者始终能够找到最适合其特定需求的解决方案。
• 可扩展的应用程序。我们的模型和智能体专为在多样化硬件、应用场景及业务流程中实现无缝运行而设计。它们能够处理复杂任务、支持AI原生、多模态及全方位对话，并执行深度推理。例如，我们的模型和智能体可协助机构客户优化业务流程、大规模处理和分析运营数据，并为决策提供支持。此外，我们的MaaS平台提供智能体工作空间，其中涵盖多种智能体模板及基于场景的解决方案。通过该智能体工作空间，客户可借助简化的模型微调、增量模型训练及提示词工程，快速定制智能体。
• 简易的基础设施适配性。我们与基础设施业务合作伙伴协同设计先进的计算基础设施，使我们的MaaS平台能够提供集计算、网络、训练通信及推理加速于一体的综合能力。此次合作还使我们的模型具备广泛的适配性，支持参数规模从15亿到2,300亿的模型，并可在云端和各类芯片组上进行大规模实时部署。特别是，这种适配性使我们的模型能够在手机、个人电脑和智能汽车等大众消费设备上进行扩展，惠及大量终端用户。
• 用户友好。我们的客户无需投入预训练资源。我们的MaaS平台提供丰富的智能体模板和工具，支持三步式专属模型微调及无代码应用开发。我们通过一体化MaaS平台将AI模型和智能体工具作为整合产品进行捆绑提供。该平台为客户提供统一访问AI模型矩阵的入口，以及一套智能体工具，包括智能体开发环境和模型微调平台。客户—— — 157 —
regardless of their in-house AI expertise, can efficiently and cost-effectively use and finetune AI models, develop customized applications and leverage agentic capabilities for a range of industry scenarios. Our offering accommodates two modes of bundling: customers may select from either pre-configured packages or customizable combinations, according to their operational requirements. A pre-configured package is a ready-made bundle to meet common needs, whereas a customizable combination allows customers to adjust the type and scale of models included to suit their specific needs. Customers can deploy and integrate these AI models and agentic tools into their own workflows using scenario templates, plugin libraries and user-friendly interfaces offered by the MaaS platform. This enables them to build and tailor solutions to their specific sector or operational needs. The platform supports seamless access, development and deployment, allowing enterprises and developers to address unique requirements and drive broad AI adoption.
成本高效（Cost-efficient）。我们研发了高性能模型，将较小的模型规模与较低的计算需求相结合，从而实现更广泛、更经济实惠的AI能力访问。例如，根据Frost & Sullivan的数据，我们的推理模型GLM-4-Air的模型规模仅为320亿参数，却能实现与竞争对手6,710亿参数模型相近的性能表现，将计算成本降低了20倍。
Cost-efficient. We have engineered high-performance models that combine smaller model sizes with lower computing requirements—enabling broader, more affordable access to AI capabilities. For example, our reasoning model GLM-4-Air, with a model size of only 32 billion parameters, delivers a performance similar to competitors' models with a size of 671 billion parameters, reducing computing costs by 20 times, according to Frost & Sullivan.
快速（Fast）。我们认识到处理速度对客户体验至关重要。根据Frost & Sullivan的数据，我们的GLM-Z1-32B-0414已实现每秒200个token的推理速度，跻身全球最快之列。
Fast. We recognize that processing speed is crucial for customer experience. According to Frost & Sullivan, our GLM-Z1-32B-0414 has achieved a reasoning speed of 200 tokens per second, which is among the fastest in the world.
安全（Secure）。我们的产品组合还提供无与伦比的安全性与可靠性。根据斯坦福大学发布的《2025年AI指数报告》，我们的GLM-4-9B模型在顶级模型中实现了最低幻觉率之一（1.3%），经Hughes幻觉评估模型（HHEM-2.1-Open）评测认证。
Secure. Our portfolio also offers unparalleled security and reliability. Our GLM-4-9B model achieved one of the lowest hallucination rates (1.3%) among top models, as evaluated by the Hughes Hallucination Evaluation Model (HHEM-2.1-Open), according to the 2025 AI Index Report published by Stanford University.
我们MaaS商业模式的一个显著优势在于，我们有能力借助机构客户触达其终端用户，且规模可观。当我们的技术帮助机构客户更好地服务其消费端客户时，我们自身也以间接但高效的方式触达并延伸至这些终端用户，而无需自行获取、筛选和维护这些客户关系。具体而言：
A notable benefit of our MaaS business model is the potential for us to leverage our institutional customers to reach their end users on a significant magnitude. When our technology helps our institutional customers better serve their consumer clients, we ourselves reach and extend our influence to these end users in an indirect but efficient way, without having to acquire, select and maintain these relationships. Specifically:
增长放大效应（Growth amplification）。当我们的机构客户发布由我们模型驱动的爆款应用程序时（例如，2025年初某领先生活方式社交媒体应用在意外迎来大量英语新用户时所采用的自动翻译工具），我们间接分享其商业成功。借助MaaS平台，我们不仅依赖自身销售努力实现增长，而是通过向机构客户提供优质模型助其取得更大成功，进而从由此产生的集体发展势能中获益。
Growth amplification. When our institutional customers release blockbuster applications powered by our models (for example, an auto translation tool adopted by a leading lifestyle-focused social media app in early 2025 when it received an unexpected influx of English-speaking new users), we share indirectly in their commercial success. Leveraging the MaaS platform, we do not rely solely on our own sales efforts to achieve growth; instead, we deliver good models to our institutional customers to make them more successful and benefit from the collective momentum that results.
通用能力服务于无数应用场景（General capability serving myriad application scenarios）。AI乃至最终AGI技术之所以具有潜在的深刻变革性，正是因为其可应用于每个行业垂直领域和每个应用场景。这正是我们致力于服务各行各业灯塔客户的原因，从而使我们的模型以最具影响力的方式造福人类。尽管应用场景无穷无尽，所需的核心模型能力本质上是相同的，例如文本生成、图像理解和代码编写。随着我们不断增强模型的通用能力，其对社会的影响力及相应的商业价值将呈指数级增长。
General capability serving myriad application scenarios. The reason why the AI—and ultimately AGI—technology is potentially so transformative is that it could be applied in every industry vertical and every use case. This is precisely the reason we strive to serve lighthouse customers in all industries across the board, so that our models benefit humanity in the most impactful way. Even though application scenarios are myriad, the core model capabilities required are essentially the same, for example text generation, image understanding and coding. As we enhance the general capabilities of our models, their impact on society and the attendant commercial value grow exponentially.
"循环中的MaaS"（"MaaS in the loop"）。借助我们MaaS平台的开放架构和先发优势，我们已建立起一个蓬勃发展的社区，将算力资源提供商、智能设备制造商、机构客户、开发者和终端用户连接在一起。我们积累了大量……
"MaaS in the loop." Leveraging the open architecture and first-mover advantages of our MaaS platform, we have established a thriving community that connects computing resource providers, smart device manufacturers, institutional customers, developers and end users. We have gathered substantial
insights from the real-world deployment of our models, which enable us to better understand how people actually use and benefit from our AI across different use cases and refine our training strategies in a more targeted manner. The result is a virtuous insight flywheel: as our customers' applications thrive, they produce more quality feedback that fuels further model improvements, which then allows us to optimize our models, and to attract even more institutional customers and application opportunities. This self-reinforcing loop strengthens our platform's core competitiveness.
Our large models can be seamlessly applied across a broad range of industries to address unique challenges and optimize workflows. By leveraging the advanced capabilities of our MaaS platform, our institutional customers can efficiently customize and deploy models tailored to their specific business needs. For example:
• Technology and internet sector. We support institutional customers in the technology and internet sector from various perspectives. For consumer electronics such as smartphones, our large models power on-device AI agents to support functions such as advanced audiovisual call interpretation, long-context memory for sustained dialogue and automated social media content creation. This allows smartphones to provide natural, real-time and context-aware AI experiences, making intelligent technology an accessible part of everyday smartphone interactions. Beyond consumer electronics, we enhance office software with intelligent content generation, enabling such software to autonomously generate and refine documents, presentations and speech drafts from short prompts. Our large models also help our institutional customers review documents and analyze operating data on a massive scale. For creative platforms, we enable their users to turn simple prompts into complex multimedia works (e.g., short videos and images) through multimodal generation, thereby broadening creative horizons and strengthening community engagement.
• Public service sector. We apply large-model technologies across various areas in the public service sector, such as public transportation management, municipal road and bridge maintenance and education. For example, we developed large models for municipal public transportation management that enable accurate monitoring of bus traffic flow and real-time estimation of arrival times. In road and bridge maintenance, our multimodal large models power an intelligent inspection system that collects real-time imagery and geographic data, allowing precise identification of structural hazards. For the academic and research community, we offer an AI-driven scientific intelligence platform that delivers advanced academic support for institutions and scholars worldwide. The platform allows users to search for academic articles using plain-language prompts and provides AI-assisted literature review with automated summarization and interactive Q&A features.
• Traditional corporate sector. Our large models facilitate the transformation of traditional corporate sectors, such as financial services and manufacturing. For example, trained on extensive financial datasets, our large models help financial institutions to analyze market trends, assess the impact of newly implemented regulations and evaluate the financial performance of listed companies. We also help clients in the manufacturing industry to establish intelligent safety monitoring systems. Our large models can accurately interpret safety monitoring footage, provide comprehensive assessments and identify safety risks or non-compliant behaviors.
See "—Our Commercial Use Cases" for the use cases of our large models applied in the sectors above.
Our MaaS platform offers flexible custom model deployment options to meet the diverse needs of businesses while maintaining efficiency, scalability and data security. We primarily offer two deployment approaches—on-premise and cloud-based deployment:
• For on-premise deployment, our models are hosted within the customer's own infrastructure. This approach allows organizations to utilize their proprietary or sensitive data to tailor AI models to their specific domains. On-premise deployment offers greater control over performance optimization and infrastructure configuration, making it suitable for complex or highly specialized application scenarios.
• For cloud-based deployment, our models are hosted on a scalable and reliable cloud infrastructure. This approach is sensible for businesses seeking agility and ease of implementation. By utilizing the cloud, customers eliminate the need for costly local infrastructure, allowing them to deploy AI solutions quickly and cost effectively.
The following table sets forth a breakdown of our revenue during the Track Record Period by means of deployment:
| | Year ended December 31, | | | | | | Six Months Ended June 30, | | | | |---|2022| |2023| |2024| |2024 (Unaudited)| |2025| | | | (RMB in thousands) | | | | | | | | | | | On-premise deployment | 54,815 | 95.5% | 112,614 | 90.4% | 263,930 | 84.5% | 26,806 | 59.7% | 161,777 | 84.8% | | Cloud-based deployment | 2,594 | 4.5% | 11,924 | 9.6% | 48,484 | 15.5% | 18,103 | 40.3% | 29,100 | 15.2% | | Total | 57,409 | 100.0% | 124,538 | 100.0% | 312,414 | 100.0% | 44,909 | 100.0% | 190,877 | 100.0% |
For on-premise deployment, we recognize revenue at the point in time when the large model and related services are delivered to the customer's designated location and accepted by the customer. For cloud-based deployment, we recognize revenue over the contract term. Specifically, for subscription-based contracts, we generally recognize revenue ratably over the contract term; for usage-based contracts, we recognize revenue based on the customer's utilization of the resources when the services are rendered to the customers. For details of our revenue recognition policies, see "Financial Information—Material Accounting Policies and Estimates—Material Accounting Policy Information—Revenue Recognition."
For on-premise deployment, our solutions are offered as a package, with pricing determined by the types and scale of models, the amount of computing resources included and implementation costs. The packaged price is charged on a one-off or annual basis. For cloud-based deployment, we offer both usage-based contracts, where customers are charged according to the volume of tokens consumed, and subscription-based contracts, where pricing is determined by the length of subscription, the types and scale of models and the amount of computing resources included. Individual users typically access our models and agents through our web portals and mobile applications (such as the Z.ai website and AutoGLM mobile app) free of charge. Certain advanced features within these web portals and mobile applications are available on a subscription basis, and the relevant revenue is categorized under the cloud-based deployment segment.
We regularly review and adjust our pricing strategies with the aim of further enhancing both our attractiveness to customers and our overall profitability. Specifically, we plan to optimize our pricing tiers by taking into account usage volume, performance and advanced functionalities. We intend to introduce
lower-priced plans that provide limited access to the most resource-intensive models, thereby attracting a broader customer base while encouraging clients who require premium performance to select higher-margin plans. In addition, we will continue to negotiate directly with high-value customers, developing bespoke agreements that reflect both the substantial value they receive and the costs incurred in providing those services. We expect that optimization of our pricing tiers will improve our overall profitability, as lower-priced plans enable us to expand our customer base and thus spread fixed costs more effectively, and higher-priced, bespoke plans increase our profit margins. Furthermore, we plan to offer usage analytics to clients, encouraging them to optimize their interactions with our services, which in turn should help us to reduce our operational burden.
| Infra Providers | Service Delivery | | |---|---|---| | | Customer Needs Analysis | | | | Customer Type | | | | Enterprises | Public service | | | Individual users | Individual developers | | Provide elastic cloud computing resources | Service Plan Design | | | | Design our service plan based on customers' needs. Evaluate scenario suitability and deployment method | | | | Core Customer Needs | | | | Data-related needs: | | | | Data volume / quality / compliance | | | | Application scenario-related needs: | | | | Single scenario / Multiple, integrated scenarios | | | | Determine Deployment Method | | | Provide high-performance infrastructure for on-premise training | On-premise deployment | Cloud-based deployment | | | Model customization | | | | Model fine-tuning | | | | Model incremental training | | | | Prompt engineering | | | Industry business partners | | | | Support industry-specific customization and service delivery | Continuous optimization | | | | Continuously optimize our services based on customer feedback | |
• **The upstream: collaboration with infrastructure providers.** Our computing infrastructure partners provide highly elastic cloud computing resources, allowing our customers to efficiently scale up or down the usage based on their actual demand. In addition, our computing infrastructure partners can also supply localized computing resources, such as high-performance computing servers, to support on-premise incremental model training and operations. We also collaborate with computing infrastructure partners to co-design infrastructure that offers broad adaptability to support large-scale, real-time deployment. See "—Easy infra adaptability" above.
• **Our service delivery.** To begin each engagement, we conduct in-depth discussions to understand each customer's specific requirements, focusing on two primary areas: data and application scenarios. For data-related needs, we assess the volume and type of data required for effective model fine-tuning and incremental training, with the goal of tailoring our models for each customer's specific industry and use cases. We also discuss data security and compliance requirements to ensure that our solutions meet regulatory and organizational standards. For application scenario–related needs, we work closely with each customer to determine whether they plan to implement large models in a single scenario or across multiple, integrated scenarios. This step guides us to determine the scale and sophistication of the large models we propose.
Building on our understanding of customer needs, we design and present a tailored service plan. At this stage, we further assess model suitability for the intended scenarios, refining our approach through ongoing discussions with the customer. We determine the deployment method—cloud-based or on-premise—based on the customer's preferences and requirements. See above for details of respective advantages of cloud-based and on-premise deployment. We then customize our large models for the customer's specific application scenarios using model fine-tuning, incremental training and prompt engineering. After deployment, we remain engaged, continuously improving model performance and optimizing our services based on customer feedback.
• The downstream: empowering industry business partners. Through Project Z, we empower our industry business partners to explore opportunities in long-tail market scenarios. Participants in Project Z are typically early-stage start-ups seeking to implement large models within specific industry sectors. They often have limited technical capabilities or lack the resources to develop large models independently. We provide these companies with access to our large models and infrastructure and offer tailored technical support. This approach enables our industry business partners to scale their operations efficiently, while concurrently facilitating the expansion of our large models into new industry scenarios. In this way, our business growth is supported not only by our own efforts, but also by the collective success of our business partners.
To teach machines to think like humans, we must empower AI with three core human faculties: deep thinking, cognition and tool use. We have developed our AI models accordingly, which can be grouped into three corresponding categories: reflection and rumination models, multimodal models and agent models. We have also developed coding models, which generate code autonomously and enhance programming efficiency. All four categories are developed upon our GLM (General Language Model) series of foundation models. Foundation models and reflection and rumination models belong to the broader category of language models.
| Deep Thinking | Cognition | Tool Use | Foundation Model | |---|---|---|---| | • GLM-Z1 | • CogView | • AutoGLM | • GLM-4.5 | | • GLM-Z1-Rumination | • GLM-4.5V | • Coco | | | | • CogVideoX | • CodeGeeX | | | | • GLM-Realtime | | | | | • GLM-4-Voice | | |
CogView (image generation) GLM-4.5V (visual comprehension and reasoning) CogVideoX (video generation) GLM-Realtime (realtime video call)
AutoGLM ("from chat to act" – agent for autonomous mission completion) AutoGLM – Rumination ("thinking while working" – agent for autonomous mission completion with deep thinking capabilities) CoCo (enterprise agent)
基础模型是经过预训练的大型语言模型，作为开发各类专业化模型的基础。我们的旗舰基础模型为GLM-4.5。
反思与沉思模型在生成答案前会额外花费时间进行"深度思考"，从而使其更适用于复杂推理任务。在我们基础模型的基础上，我们构建了反思模型（GLM-Z1）和沉思模型（GLM-Z1-Rumination）。基础模型与反思及沉思模型同属语言模型这一更广泛的类别。
多模态模型能够处理并整合来自不同模态的信息，如文本、图像、音频和视频。我们已开发出多种多模态模型，服务于不同功能场景，例如文生图、文生视频的生成，以及图像和视频的理解。
AI智能体融合了推理、规划与工具调用能力，能够在无需人类持续介入的情况下自主执行多步骤任务。我们的基础智能体模型为AutoGLM，并在此基础上进一步开发了AutoGLM-Rumination。此外，我们还开发了企业级智能体CoCo。
代码模型能够自主生成代码并提升编程效率。我们的代表性代码模型为CodeGeeX。
GLM-4.5 is our flagship foundation model, which we open-sourced upon launch. Through multi-stage training and comprehensive post-training with fine-tuning and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning and coding tasks. GLM-4.5 also supports multi-modal extensions and large context processing, allowing it to interpret high-level prompts and autonomously generate practical solutions. GLM-4.5 has a model scale of 355 billion parameters and we have also developed GLM-4.5-Air, a lightweight version with 106 billion parameters.
• Benchmark tests. Based on an evaluation across twelve industry-standard benchmark tests¹ in July 2025, GLM-4.5 ranked third globally, first in China and first among global open-source models. GLM-4.5 achieved a comprehensive score of 63.2 under these twelve benchmarks, compared with scores ranging from 46.3 to 65.0 for industry peer models.
• Global leaderboards. GLM-4.5 ranked fifth globally on Chatbot Arena and WebDev Arena in September 2025, which are industry-recognized global leaderboards that rank the overall capabilities and coding capabilities of large models, respectively.
• Token consumption volume. Since the launch of GLM-4.5 and up to early December 2025, our token consumption volume on OpenRouter, a leading global platform that provides API access to a wide range of large models, had consistently ranked among the top ten globally and the top three among Chinese companies. This sustained performance underscores the strong competitiveness and market recognition of GLM-4.5, demonstrating its advanced efficiency, scalability and real-world applicability.
• Popularity rankings. Within only 48 hours of its initial launch, GLM-4.5 ranked first globally on the trending board of Hugging Face, the world's largest platform for open-source models.
¹ Benchmark tests are structured, standardized evaluations that measure LLMs' capabilities across a range of tasks. The twelve industry-standard benchmark tests we used to evaluate GLM-4.5 include three categories: (i) agentic benchmarks, including TAU-Bench, BFCL V3 and BrowseComp. GLM-4.5 achieved a comprehensive score of 58.1 under these agentic benchmarks, compared with scores ranging from 45.0 to 61.1 for industry peer models; (ii) reasoning benchmarks, including MMLU-Pro, AIME 24, MATH-500, SciCode, GPQA, HLE and LCB (2407-2501). GLM-4.5 achieved a comprehensive score of 68.8 under these reasoning benchmarks, compared with scores ranging from 63.5 to 74.2 for industry peer models; and (iii) coding benchmarks, including SWE-Bench Verified and Terminal-Bench. GLM-4.5 achieved a comprehensive score of 50.9 under these coding benchmarks, compared with scores ranging from 36.7 to 55.5 for industry peer models.
幻觉率。根据检索增强生成（RAG）大语言模型幻觉排行榜（LLM Hallucination Leaderboard for Retrieval-Augmented Generation），2025年9月，GLM-4.5的幻觉率位居全球第二低、中国最低。该基准测试根据大型模型在回答故意具有误导性的问题时产生不存在答案（即幻觉）的频率对其进行评估。
2025年9月，我们发布了GLM-4.6，这是我们基础模型的进一步更新版本，主要特点是增强了编码能力。2025年11月，GLM-4.6在CodeArena上排名全球第一，CodeArena是业界公认的最新全球评估平台，专为评估模型的编码能力而设计。
预训练与中期训练。预训练是模型在通用数据集上进行训练以学习基本语言规律的初始阶段。我们的预训练语料库由多语言（主要为英文和中文）文档组成。完成预训练后，我们增加三个阶段以进一步提升模型在重要应用领域的性能。与在大规模通用数据集上进行传统预训练不同，这些训练阶段采用中等规模的特定领域数据集。我们使用从公共互联网资源自行收集的数据集以及从专业第三方供应商采购的数据集。根据弗若斯特沙利文（Frost & Sullivan）的资料，此类做法符合行业惯例。
后训练。后训练对模型进行进一步优化，使其与人类偏好保持一致，例如理解人类意图、遵循指令以及促进多轮对话。对齐通过监督微调（SFT）和推理强化学习（RL）实现。监督微调通过真实的人类提示和交互，在很大程度上使基础模型与人类偏好保持一致，而推理强化学习则进一步增强了模型在逻辑推理和结构化问题解决方面的能力。
卓越的智能体能力。GLM-4.5擅长处理复杂的智能体任务，展现出解读高层级提示并自主生成实用解决方案的高级能力。例如，在仅提供单句指令的情况下，GLM-4.5能够构建视频流媒体网站、设计并编写经典井字棋游戏，或创建精美、视觉效果出众的演示文稿。这些先进的智能体功能使GLM-4.5成为简化日常工作流程、提升自动化水平和生产力的有力工具。
混合专家（MoE）架构支持的高效性。GLM-4.5是我们首个采用混合专家（MoE）架构的模型。MoE是一种创新架构，利用多个专业子网络（即"专家"）处理复杂数据输入的不同部分。这类似于一组具有不同专业背景的医生共同协作，治疗复杂的医疗问题。门控网络（即路由器）决定针对特定输入激活哪些专家。MoE架构大幅提升了模型效率，并有助于产生更准确的输出结果。根据我们在相关基准测试下的评估，GLM-4.5在模型性能与模型规模之间实现了出色的平衡，并在同等规模的同类模型中表现领先。
快速响应速度与双模式设计。GLM-4.5的输出速度可达每秒100个token。这使GLM-4.5非常适合低延迟、高并发的应用场景，即多个计算需求同时涌现并需要实时快速响应的场景。GLM-4.5还具备双模式设计：思考模式适用于复杂的多步骤推理任务，非思考模式适用于需要快速响应的任务。用户可根据具体需求选择适当的模式。
Building upon our foundation model, GLM-4-Air-0414, we developed our reflection model (GLM-Z1) and rumination model (GLM-Z1-Rumination). Reflection model and rumination model both spend additional time "deep thinking" before generating an answer, which makes them better for complex reasoning tasks.
• Reflection models are designed to tackle problems with certainty, aiming for more precise and accurate solutions. For example, reflection models can solve a math equation using multiple approaches and plug the solution back in to confirm its correctness. Compared with traditional LLMs, reflection models significantly boost reliability and precision in solving structured, determinate tasks.
• Rumination models are designed to address problems with uncertainty, especially those open-ended, exploratory questions that require gathering and processing external information iteratively. For example, rumination models can address complex, opinion-driven tasks, such as "writing a comparative analysis of AI development in two cities and their future development plans." Rumination models require longer thinking time compared with reflection models.
GLM-Z1 is a reflection model featuring deep thinking capabilities. It was developed based on the foundation model through extended reinforcement learning and further training on tasks including mathematics, coding and logic. Compared to the foundation model, GLM-Z1 significantly improves mathematical abilities and the capability to solve complex tasks. Key features and advantages of GLM-Z1 include:
Robust reasoning capabilities. We evaluated GLM-Z1 on various commonly used benchmarks relating to math, reasoning, coding, instruction following, tool calling and science. GLM-Z1 demonstrated capabilities comparable to peer leading reflection models.
Fast inference speed and high cost-efficiency. GLM-Z1 is able to achieve an inference speed of up to 200 tokens per second, substantially surpassing comparable reflection models. Meanwhile, the cost of GLM-Z1 is substantially lower than comparable reflection models. Unlike other resource-intensive models, GLM-Z1 can run on a single consumer-grade GPU, reducing hardware-related barriers for users of our models.
Extended reinforcement learning. GLM-Z1 is equipped with extended reinforcement learning, which requires longer training durations, more sophisticated reward models and novel algorithms designed specifically to cultivate multi-step reasoning processes. The goal is to teach the model not just to provide plausible answers, but to follow logical steps, identify implicit assumptions and handle complex instructions.
GLM-Z1-Rumination is a deep reasoning model with rumination capabilities. Compared with reflection models, GLM-Z1-Rumination is capable of deeper and longer thinking and using tools to solve more open-ended and complex problems. GLM-Z1-Rumination is trained through scaled end-to-end reinforcement learning with responses graded by "ground-truth" answers (the most accurate answer against which the performance of a model is evaluated) or "rubrics" (a set of qualitative criteria).
以更深入、更持久的思考解决开放性问题。GLM-Z1-Rumination能够针对缺乏明确答案的任务进行迭代式、多步骤推理，例如细致入微的研究或比较分析。此类任务需要综合来自多种潜在来源的信息、构建复杂论点、作出细致判断并生成连贯的长篇文本。
深度思考过程中集成搜索工具。对于复杂的研究型任务，模型的内部知识可能不足以应对。通过在生成过程中集成搜索功能，GLM-Z1-Rumination可以主动检索最新信息、对其进行分析，并将其纳入其"沉思"（rumination）过程。这使该模型从静态知识库转变为动态研究助手。
CogView是我们的文本到图像生成模型系列。CogView-4是首个支持生成中文字符的开源文本到图像模型。它还在DPG-Bench基准测试的综合评分中排名第一，在开源文本到图像生成模型中达到了SOTA（最先进）性能。CogView-4的主要特性与优势如下所述：
SOTA性能。CogView-4在复杂语义对齐和指令遵循方面表现出强大能力。它支持中英文双语输入，对输入长度无限制。该模型可在指定范围内生成任意分辨率的图像。2025年3月，CogView-4在DPG-Bench基准测试中获得综合最高分，DPG-Bench是业界公认的文本到图像生成模型评测基准。
双语能力。CogView-4是首个支持生成中文字符的开源文本到图像模型。该模型使用双语文本-图像数据进行训练，能够有效处理两种语言的提示词。CogView-4能够将中英文字符无缝融入图像，使包含中文文本的海报和插图创作更加便捷。此外，CogView-4在理解和遵循中文提示词方面表现出色，例如能够捕捉并可视化中国古典诗词中所描绘的意境与情感。
seamlessly convert it into visual content by producing a four-panel comic strip. Each panel could depict key portions of the narrative, capturing the essence of the story while integrating text or imagery as needed, effectively demonstrating its strength in both storytelling and image generation tasks.
GLM-4.5V是我们专为通用视觉理解与推理而设计的基础视觉语言模型（VLM）。VLM将语言理解与视觉处理相结合。与仅依赖文本输入的传统语言模型不同，VLM将视觉数据（如图像、视频或图形）与自然语言相融合，以执行需要同时具备文本与视觉理解能力的任务。基于涵盖42项业界公认基准测试的综合评估，GLM-4.5V在同等规模开源模型中的几乎所有任务上均达到了最先进（SOTA）水平。
视觉定位。GLM-4.5V能够准确识别、分析并定位图像中的目标对象。该功能可广泛应用于安全检查、质量控制及遥感分析等场景。与传统视觉模型相比，GLM-4.5V具备更丰富的知识储备和更强大的推理能力，使其能够理解并执行更为复杂的定位指令。
网页复现。GLM-4.5V能够通过分析网页截图或屏幕录制来复现网页内容。它能够推断网页元素之间的逻辑关系、布局规则及用户意图，进而生成高度准确且完整的网页代码。与传统的基于图像的识别方式不同，GLM-4.5V可分析网页交互的屏幕录制内容，并生成能够同时还原静态与动态特征的可执行代码。用户还可通过简单指令对网页进行进一步优化。
基于图像的推理。GLM-4.5V堪称一位出色的"视觉侦探"，将敏锐的感知能力与推理技能相结合，用于图像识别与推断。例如，给定一张风景或街景照片，该模型能够检查植被、天气特征及建筑风格等细粒度线索，从而推断出可能的地理位置及大致经纬度，且无需借助任何外部搜索工具。
对图表密集型复杂文档的理解。GLM-4.5V能够阅读多页且含有大量图表的文档，从文字与图表中收集并汇总信息，同时还能根据所理解的信息提供评述。对于图表内容，与依赖独立光学字符识别（OCR）及文本分析的传统方法不同，GLM-4.5V能够像人类一样同步执行上述两个过程。这提升了其对图像数据的理解准确性，使其能够高效处理技术报告和演示文稿等图表密集型文档。
CogVideoX是我们的视频生成模型系列。凭借前沿的效率表现、先进的技术能力及多样化的功能，CogVideoX能够将创意概念转化为高质量视频。
CogVideoX在视频生成能力方面取得了显著进步，大幅提升了视频生成任务中的画面质量、美学表现、动态连贯性，以及对复杂提示词的语义理解能力。它支持4K超高清分辨率，可生成每秒60帧的10秒4K视频。CogVideoX能够精准解读并执行复杂的提示词，支持多种艺术风格，且呈现效果与用户预期高度吻合。该模型还支持多路并行生成，允许用户基于同一文本提示词或图像同时生成最多四个视频。
GLM-Realtime 是我们的多模态模型，支持视频与音频交互，能够实现无缝、实时的 AI 音视频通话体验。在研发过程中，我们已将 GLM-Realtime 集成至智能眼镜和陪伴机器人中，使用户能够体验智能化、近乎即时的交互。凭借其实时音视频能力，GLM-Realtime 为 AI 与硬件的融合树立了新标准，为下一代智能设备提供了坚实的基础。
GLM-Realtime 的一项核心功能是在视频通话过程中能够记忆长达两分钟的对话上下文，从而增强对话连贯性与用户交互体验。在基于语音的交互中，GLM-Realtime 支持完全端到端的体验，并引入了创新性的演唱功能，允许模型演唱旋律。
此外，GLM-Realtime 不仅限于实时交互，还融入了函数调用（function call）功能，使其能够动态访问外部知识和工具。这种灵活性将其能力扩展至内置知识范围之外，解锁了广泛的潜在使用场景和现实世界应用。
GLM-4-Voice 是一款端到端语音模型。与传统方法需要两步流程——先将语音转换为文本，再将文本转换回语音——不同，GLM-4-Voice 使用音频 token 并直接对语音生成过程进行建模。因此，语音的理解与生成在单一架构内同步进行，消除了与中间文本转换步骤相关的信息损失风险。
GLM-4-Voice 能够直接理解并生成中英文语音，进行实时语音对话，并可根据用户指令改变语音的情感、语调、语速、方言及其他属性。
为弥合文本与语音模态之间的差距，我们开展了大规模语音-文本预训练。这使 GLM-4-Voice 在语音语言建模和口语问答等多种任务中表现出色。为进一步增强 GLM-4-Voice 的对话能力，我们采用"流式思维"（streaming thoughts）模板，在高质量对话数据集上对模型进行微调。该模板交替输出文本与语音 token，提升了模型生成无缝、低延迟响应的能力，同时保持高质量性能。
BUSINESS OUR AI AGENTS AutoGLM – "From Chat to Act": autonomous mission completion AutoGLM represents a major step forward in the evolution of our AI universe—from "chat" to "act," bridging the gap between conversation-based AI and real-world task execution. Designed as a foundation agent tailored for autonomous control of digital devices through graphical user interfaces (GUIs), AutoGLM transforms human-like reasoning into concrete actions: AutoGLM can simulate human actions on a mobile phone or a computer, allowing it to autonomously complete tasks such as ordering food delivery, booking hotels and commenting on social media posts.
AutoGLM achieved SOTA performance under AgentBench, an agentic AI benchmark recognized by the 2024 AI Index published by Stanford University. Specifically, in the benchmark for phone use, AutoGLM outperformed prior leaders with a task success improvement of over 20%; in the benchmark for browser use, AutoGLM also outperformed advanced models such as GPT-4o and Claude-3.5-Sonnet.
In August 2025, we released an updated version of AutoGLM (also known as "AutoGLM 2.0"), which is powered by our then latest foundation model GLM-4.5 and visual comprehension and reasoning model GLM-4.5V. This updated version enables AutoGLM to simulate human actions across a broader range of mobile applications and websites. It can autonomously complete requested tasks on the cloud without occupying the user's mobile phone or computer, allowing users to continue using their own devices without interruption.
Developing foundation agents is very different from developing ordinary LLMs. Training foundation agents faces a critical challenge: the scarcity of decision-making data in existing pre-training sets. While the internet contains vast amounts of human knowledge, it primarily consists of static information that inadequately captures human decision-making and environmental interaction. Building capable foundation agents requires enriching them with dynamic knowledge, either through direct interaction with real-world environments or through learning from synthesized trajectories. Such foundation agents can then self-evolve in the digital world, iteratively improving to achieve genuine general intelligence. To address this challenge, we employed a comprehensive suite of training techniques, such as:
Intermediate interface design. It is essential to design an intermediate interface that disentangles planning and grounding behaviors in foundation GUI agents. They present distinct requirements—planning demands flexibility and error recovery, while "grounding," which is the process of ensuring that the model's outputs are solidly based on trustworthy sources rather than on speculation or fabricated content, emphasizes action accuracy. Their separation via the intermediate interface enables more agile development and enhanced performance.
Self-evolving online curriculum reinforcement learning. We recognize that error recovery is crucial for robust and deployable agent applications, yet it remains difficult to acquire through training alone. Additionally, the shortage of instructions and trajectories impedes training progress. We address this challenge through self-evolving reinforcement learning, implemented according to a progressive weak-to-strong curriculum.
High success rate of task execution. We evaluated AutoGLM by executing frequent tasks on seven common Chinese Android Apps, such as "order a cold coconut latte from the nearest coffee shop, with half sugar." According to Frost & Sullivan, AutoGLM performed decently on most evaluated apps, with success rates ranging between 70% and 100%. We also evaluated AutoGLM's abilities on AndroidLab, an interactive Android benchmark and development environment that supports reproducible evaluation.
跨多款移动应用执行超长任务。AutoGLM能够理解并执行超长指令和任务，且不会中断。此外，AutoGLM支持跨多款应用执行任务，无需用户在不同应用之间切换。作为用户与应用之间的调度层，这种跨应用功能是自动化领域的重要进步。例如，下图展示了AutoGLM如何协助安排与朋友的晚餐。
电脑使用版本（GLM-PC）。GLM-PC是AutoGLM的电脑使用版本，基于CogAgent模型开发，是中国首款电脑使用类AI智能体。GLM-PC是一款多功能工具，适用于多种实际应用场景。它可作为会议助手，负责安排、参与和总结会议，并支持文档管理任务，如内容的下载、分享和摘要。它能够从公共资源中搜索和汇总信息。GLM-PC还支持远程及定时操作，允许用户通过移动端指令或在预设时间执行任务。此外，其隐形屏幕功能使其能够在后台自主运行，在高效完成任务的同时解放用户的工作界面。
AutoGLM Rumination是AutoGLM的高级版本，是一款自主AI智能体，专为探索开放性问题并根据发现采取行动而设计。AutoGLM Rumination具备"边思考边工作"的特性——它借助GLM-Z1-Rumination模型所赋予的卓越推理能力，同时融合了AutoGLM的交互操作能力。
AutoGLM Rumination能够处理涉及深度推理、迭代研究并产出可执行成果的复杂任务。它可以深入探索开放性、深层次问题，动态驾驭多步骤研究流程，进行基于网络的信息搜索，甚至根据研究发现采取进一步行动。
AutoGLM Rumination弥合了从推理到执行之间的鸿沟，提供推理、研究与可执行能力的无与伦比的融合。例如，它可以对出行方案进行深度研究，并直接为所选行程订票；还能在广泛比价后推荐性价比最高的产品并完成购买。
critical resources, such as enterprise software, knowledge bases and databases. Leveraging advanced personalized memory, CoCo enables one-click automation of complex workflows. Its fundamental principle is to develop a deep understanding of both individual user needs and the overall business context to ensure the delivery of meaningful output. CoCo has the following key features and advantages:
Results-focused delivery. CoCo is designed to accompany and assist business processes end-to-end, ensuring that every task concludes with a tangible, measurable result. This outcome-based framework underscores CoCo's commitment to delivering value through actionable output.
Intelligent, personalized support. CoCo introduces a unique memory system, dynamically adjusted to individual employee's responsibilities and needs. It acts as a trusted personal assistant, proactively supporting users with tailored services and industry insights that align with their current priorities, providing timely and precise information at every step of the workflow.
Seamless integration with corporate infrastructure. CoCo embeds effortlessly within existing enterprise systems, bringing together the enterprise's data and operational tools under a unified platform. This cohesion allows CoCo to function as an AI assistant that not only understands the nuances of enterprise operations but also drives efficiency from within. With these enterprise-grade capabilities, CoCo is positioned to elevate business performance and deliver clear, measurable results.
CodeGeeX is a powerful coding model designed to enhance programming efficiency and streamline workflows. It enables developers to automatically generate code based on natural language descriptions or complete unfinished lines or blocks of code, significantly improving productivity. CodeGeeX also facilitates semantic-level code translation across multiple programming languages and can automatically add detailed line-level comments. Additionally, developers can ask CodeGeeX technical questions directly within their integrated development environment (IDE), eliminating the need to consult external search engines and ensuring uninterrupted focus. Supporting popular IDEs as well as prevalent programming languages, CodeGeeX is a versatile assistant that adapts to developers' needs. As of June 30, 2025, CodeGeeX generated on average more than 100 million lines of code on a daily basis.
The following table sets forth a breakdown of the number of our institutional customers by business segment during each period of the Track Record Period:
| | Year Ended December 31, | | | Six Months Ended June 30, 2025 | |---|---|---|---|---| | | 2022 | 2023 | 2024 | | | Number of institutional customers | | | | | | Cloud-based deployment | — | 2,812 | 5,457 | 3,061 | | On-premise deployment | 48 | 61 | 123 | 95 | | Total | 48 | 2,873 | 5,580 | 3,156 |
For use cases that demonstrate the applications of our models and agents for institutional customers across diverse industry sectors, see "—Our Commercial Use Cases" below.
Token consumption volume. Our average daily token consumption volume was 0.5 billion, 2.1 billion, 0.2 trillion and 4.6 trillion in December 2022, 2023 and 2024 and June 2025, respectively.
Hallucination rate. The hallucination rate of our GLM-4.5-Air (released in July 2025) and GLM-4.6 (released in September 2025) was 9.3% and 9.5%, respectively, as evaluated by the Hughes Hallucination Evaluation Model (HHEM-2.3) according to the LLM hallucination rate leaderboard published on GitHub.
Our models and agents can be seamlessly applied across a broad range of industries to address unique challenges and optimize workflows. By leveraging the advanced capabilities of our MaaS platform, our clients can efficiently customize and deploy models and agents tailored to their specific business needs. Specifically, our MaaS platform provides an agent workspace, which encompasses a variety of agent templates and scenario-based solutions. Through this agent workspace, our clients can swiftly customize agents through streamlined model fine-tuning, incremental model training and prompt engineering. In contrast to conventional AI companies—which often require substantial resources for from-scratch development—our MaaS platform implements a lightweight approach that allows us to provide scalable services with high efficiency and affordable cost. Our models and agents enable our clients to unlock new levels of innovation, efficiency and personalization. Below are some use cases demonstrating the applications of our models and agents across diverse sectors.
Our client is a world-leading consumer electronics producer. We partnered with the client to integrate our agentic models into its latest smartphone series. This marks a significant step in integrating advanced AI agent capabilities directly to consumer mobile devices. Given that mobile phones now serve as a primary touchpoint for people worldwide, this integration immediately put our AI into the pocket of a vast number of consumers, allowing them to experience and benefit from our AI at any time. Our comprehensive model capabilities allow us to provide users with a diverse suite of agentic functionalities, spanning from audiovisual calls to social media content generation:
On-device AI audiovisual call. Powered by our GLM-Realtime model, the smartphone features an innovative AI audiovisual call function. This capability allows the mobile AI agent to interpret images and videos in real time during calls, instantly recognizing both on-screen text and nuanced visual details. Our model's advanced streaming inference ensures that analyses and responses are delivered with minimal latency, supporting smooth, interactive and natural communication.
Long-context memory. As an intelligent voice assistant, the mobile AI agent is equipped with multi-turn long-context memory. This enables it to recall information from previous parts of the conversation or earlier sessions, supporting not only one-off responses but also maintaining coherent, ongoing dialog throughout a user's interactions.
System integration and function calling. Utilizing our AutoGLM foundation agent, the mobile AI agent supports direct system-level function calling. Users can operate various device features—such as adjusting settings, retrieving information and initiating tasks—simply by using natural language commands. This demonstrates the model's strength in enabling robust, context-aware system integration.
社交媒体内容生成。借助模型的多模态生成能力，移动端AI智能体使用户能够基于从相册中选取的图片，高效生成简洁且引人入胜的社交媒体内容。这一无缝工作流程自动化了内容创作过程，从而提升生产力与创造力。
我们的客户金山办公（Kingsoft Office）是中国领先的办公软件服务提供商。金山办公意图整合先进的AI能力，以提升生产力并变革用户与办公工具的交互方式。为此，我们协助金山办公提升其AI赋能办公软件WPS AI的性能。我们的解决方案使WPS AI能够生成更准确、更高质量的内容，并自主协助用户完成各类任务，显著提升了生产力与用户满意度。通过将我们的模型整合至金山办公的各类产品中，我们提升了可扩展性并降低了维护成本，从而能够高效、一致地交付更多智能功能。
• 智能化程度有限。由于技术限制，原有WPS AI无法提供真正智能且直觉化的用户体验。原有WPS所生成的内容往往无法满足用户的预期或意图。
• 可扩展性挑战。金山办公依赖于针对每个专项应用场景分别开发和训练小型模型。这种方式导致研发成本高昂、产品开发周期漫长、维护成本居高不下，从而限制了可实现的智能功能数量。
• 智能化增强。我们帮助WPS AI提升了所生成内容的质量。例如，用户只需提供一个主题，WPS AI便能够逐步生成大纲、幻灯片内容及演讲草稿，并自主对内容进行格式化。这种自主能力通过迭代使用我们的模型，对所生成的内容进行逐步生成与修改来实现。
• 跨产品整合。为解决可扩展性与维护问题，我们协助金山办公将我们的大模型整合至其多样化的办公软件产品中，大幅提升了用户体验并降低了运营成本。
我们的客户智联招聘（Zhaopin.com）是中国领先的人力资本生态平台，为雇主和求职者提供一站式专业人力资源服务。
我们客户所面临的挑战。凭借庞大的职位发布和候选人简历数据库，企业招聘人员面临着筛选效率低下的困境。提升候选人与岗位匹配的精准度，对于改善雇主和求职者双方的招聘流程至关重要。
我们的解决方案。我们的AI招聘助手使智联招聘能够为雇主和求职者双方提供多种功能。对于雇主而言，我们的AI助手通过对话式交互理解雇主的招聘需求，筛选简历并提供针对性推荐。在面试过程中，AI助手分析候选人的技能并提供全面的面试后评估报告。对于求职者而言，一旦其提供了自身优势的相关信息，我们的AI助手便能生成完整、专业的简历，并可根据求职者的偏好对简历作进一步润色。
我们的客户Nieta是中国一家快速成长的人工智能生成内容（AIGC）平台。该平台赋能普通用户将创意想法转化为插画、漫画、动画、小说、音乐乃至虚拟商品，同时通过互动社区功能，使用户能够分享、发现并协作创作。
客户面临的挑战。Nieta致力于满足用户不断演变的现实创作需求。其目标是降低普通创作者的技术门槛，使其能够快速、无缝地将简单的提示词转化为复杂的高质量多媒体内容。然而，这需要先进的基础设施，包括具备强大语言能力、多模态能力、智能体能力及代码能力的前沿人工智能模型，以优化用户的创作流程和社区互动体验。
我们的解决方案。依托我们的大模型能力，我们与Nieta合作，于2025年4月推出了一款智能短视频生成智能体。该智能体允许用户仅需以自然语言输入创意想法，系统即可自动将其转化为一段完整的短视频，涵盖画面、旁白及背景音乐。我们的技术为该智能体赋予了先进的大语言模型及视频生成能力，实现了从文本输入到完整多模态内容的无缝转化。这一高度易用的平台充分展示了大模型在文化创意产业中的创新潜力，并显著拓展了人工智能支持和赋能普通用户的方式。
我们的客户是一家专注于高端智能纯电动汽车的汽车制造商。我们的解决方案帮助该客户将智能座舱的用户体验从僵化的指令式操控转变为直觉化、具备情感智能且高度个性化的交互模式，能够准确理解用户的自然语言输入并适应个人偏好。这不仅以动态内容和趣味对话丰富了车内环境，更为高端电动车车主带来了更令人满意的、拟人化的用户体验。
客户面临的挑战。我们的客户希望对其智能座舱系统进行升级——该系统原依赖于严格且精确的用户指令来执行命令，导致用户体验较为被动、缺乏动态感。该系统在理解口语化语音指令方面存在困难，导致对用户意图的理解成功率偏低。
• 通过多轮引导对话提升用户意图理解能力。我们的解决方案引入了基于多轮引导对话的用户意图确认机制。这使系统能够更准确地解读用户需求，显著提升了对指令的理解能力，即便指令以自然或非正式的方式表达亦然。
• 增强情感智能。我们的解决方案支持在多种对话风格与人设之间动态切换，提供更具沉浸感和亲切感的对话体验，并可根据个人用户偏好量身定制。此外，该系统展现出高级情感智能能力，能够进行富有同理心、具备上下文感知能力的对话，实现更接近真人的交互体验。
Dynamic and engaging content generation. Our solution allowed the creation of entertaining and impromptu content such as jokes, stories and personalized responses, enriching the interactive quality of intelligent cabin communications based on both user inputs and specific scenarios. The enhanced intelligent cabin system also supported voice-controlled games such as riddles and trivia quizzes, fostering a more interactive and enjoyable user experience.
Use Case 6: Retail Sector – Mengniu Dairy's AI Nutritionist Mengmeng Our client, Mengniu Dairy, is a world-renowned dairy company. We assisted Mengniu Dairy in creating an AI nutritionist "Mengmeng" to deliver expert-level, personalized nutrition and health services to consumers worldwide. Our solution helped Mengniu bring greater value to its customers while strengthening brand loyalty and differentiation in the health and wellness market. Our client's challenges. As more families seek expert advice to improve their health, the desire for on-demand access to reliable nutrition information is surging. Building an AI-powered nutritionist platform required significant investment in domain-specific model training and technical resources, which is challenging for dairy companies. Our solution. We collaborated with Mengniu Dairy to create Mengmeng, allowing the client's diverse customer base to interact naturally with the AI nutritionist anytime to receive expert insights and personalized nutrition health services. In addition, Mengmeng includes an AI Health Planner feature. The AI Health Planner creates personalized health and nutrition plans based on individual assessments, offering real-time interaction and progress tracking. It employs adaptive planning, dynamically adjusting goals when deviations occur, and provides proactive reminders, coaching and empathetic motivation to help users achieve their objectives effectively.
Use Case 7: Traditional Corporate Sector – Integrated Large Model Solutions for a Corporate Conglomerate Our client is a conglomerate in China with a primary focus on financial services and a range of other diversified business segments. Our client is dedicated to advancing intelligent system upgrades, integrating LLMs across various business lines and enhancing internal management efficiency. We have delivered a diverse range of integrated, one-stop solutions for our client. For example: •
Financial capital management LLM. We developed an advanced LLM tailored for financial capital management for our client. It is capable of responding autonomously in various formats—including natural language, images and tables—to inquiries from various departments of our client's commercial banking business group regarding the latest financial regulations. The model facilitates data inquiries and analysis and provides robust support for data governance. Notably, this LLM achieves an AI automation rate exceeding 50% in Q&A on new financial regulations. Its introduction greatly accelerated the practical implementation of the latest financial regulations throughout our client's commercial banking business group, empowering employees to conduct efficient information search, data analysis and data management, and propelling the bank toward a new era of intelligent capital management.
Transportation and logistics LLM. We developed an LLM for the client's transportation and logistics business group. It is designed for various application scenarios, including order management, transportation planning, intelligent customer service and equipment operation and maintenance. The model effectively enhances operational efficiency within the logistics industry. With strong perceptual and conversational abilities, this LLM demonstrates excellence in logistics image recognition, text information extraction, information query, logistics text recommendation, customer credit assessment and tool/plugin integration. – 176 –
办公协作系统的智能化升级。为满足客户对更智能化内部办公协作系统的需求，在"文档 + AI"理念的指导下，我们开发了员工个人知识库。这为员工配备了支持文档理解、创作和应用的生产力工具，从而转变日常运营模式。借助 CoCo 企业智能体，我们高效构建了"知识库 + 工作流 + 智能体"一体化系统，并成功完成了基础设施及向量数据库的部署。
AMiner 是一个由人工智能驱动的科学情报与大数据挖掘平台。AMiner 旨在为全球科研机构和学者提供智能学术服务，利用先进的数据挖掘与分析技术提升科研效率。依托海量学术论文数据库，并以我们的 GLM 系列模型为驱动，AMiner 的核心功能包括：自然语言学术检索、详尽的学者画像，以及支持自动摘要生成与交互式问答的 AI 辅助文献分析。AMiner 还提供写作辅助工具，涵盖文献综述撰写、翻译及语言润色，并支持最新学术动态的个性化订阅。
通用问答。智谱清言能够就广泛的主题回答用户问题，支持网络搜索、图像理解及文档阅读等能力。值得一提的是，智谱清言集成了 GLM-Z1-Rumination 模型，通过深度推演过程提供更全面、更精准的答案。详情请参阅上文"——我们的模型"部分。
多模态理解与生成。智谱清言能够生成视频和图像，助力高效且富有创意的艺术创作。在旅行途中，它可即时提供景点背景介绍，并提供环境识别服务，对视障用户尤为有益。在艺术与设计等创意领域，它能激发用户灵感，支持原创艺术作品的生成。
可定制智能体。用户可通过无代码平台创建专属 AI 智能体，针对特定需求进行个性化定制。截至最后可行日期，智谱清言平台上活跃的自定义 AI 智能体数量已逾 80 万个。
我们本质上是一家由数据科学家和工程师组成的公司，研发工作贯穿我们业务的每一个方面。在日常工作中，我们高度专注于提升基础模型的智能水平；为更多行业垂直领域及其他应用场景改进并开发实用、高性价比的 AI 智能体；同时与行业业务伙伴协作，共同设计和优化计算基础设施，使我们的 MaaS 平台能够提供全面的能力支撑。我们实现上述目标的首要依托是我们的人才，以及我们完善的技术基础设施和严格的研发流程。
我们拥有深厚的人工智能人才储备。截至2025年6月30日，我们拥有一支657人的研发团队，成员在自然语言处理、复杂系统高级决策及多模态语义分析等相关领域具备丰富的背景与经验。我们的研发团队由首席科学家张波博士及其他技术领军人才带领。我们的技术领军人才在相关领域的平均从业经验超过12年。
• 张波博士：张博士是中国科学院院士，在人工智能基础理论及模式识别、知识工程、机器人技术等实际应用领域均发挥了举足轻重的作用。他尤以其在问题求解商空间理论方面的开创性工作而著称，提出了多粒度推理与计算复杂度降低的创新方法，这些方法已成为粒计算领域的核心内容。张博士发表论文逾400篇，并著有多部获奖专著，其研究成果荣获多项重要荣誉及国家级科学技术奖励。他还联合创建了智能技术与系统国家重点实验室，进一步引领了人工智能研究的方向及其在中国的实际应用影响。
• 刘德兵博士：刘博士在大语言模型、机器学习及数据挖掘领域具备深厚的专业知识。他主持或参与了30余项重大科研项目，与科学技术部、中国工程院等机构开展合作。刘博士师从中国工程院院士、人工智能研究领域著名专家高文，在国际主要学术会议上发表论文逾20篇，并在全球范围内持有50余项发明专利。刘博士在推动我们的技术创新与战略研究方向方面发挥着主导作用。
• 张鹏博士：张博士担任我们的首席执行官，是中国实践性人工智能创新的前沿人物。张博士专注于知识图谱及大规模预训练模型研究，在我们GLM系列模型的研发中发挥了核心作用。他在国际机器学习大会（ICML）、国际语义网络会议（ISWC）等顶级会议上发表论文逾十篇，致力于推动由知识与数据双轮驱动的人工智能框架。其实践专长涵盖大规模预训练模型的部署、语义大数据分析、智能问答及决策支持系统，使其成为将人工智能研究转化为具有影响力的现实应用的杰出人物。
• 颜兴宇先生：颜先生专注于数据挖掘及大模型的预训练与研发，在我们MaaS平台的研究与开发中发挥了核心作用。颜先生的创新成果显著提升了我们MaaS平台的性能与质量。颜先生持有卡内基梅隆大学硕士学位，曾任职于谷歌，担任高级研究与开发工程师。
• 顾晓涛博士：顾博士是数据挖掘与大语言模型领域的专家，曾在中国一家大型科技公司从事自然语言处理开发及大语言模型预训练的实践工作。他获得伊利诺伊大学厄巴纳-香槟分校博士学位及清华大学学士学位。顾博士在计算语言学协会年会（ACL）、计算机视觉与模式识别大会（CVPR）、国际学习表征大会（ICLR）、知识发现与数据挖掘大会（KDD）等国际主要学术会议上发表论文逾40篇。
• 杜正晓博士：杜博士专注于将机器学习算法应用于现实系统，研究领域涵盖信息检索、知识图谱及预训练语言模型，同时致力于机器学习与强化学习算法的研究。
Dr. Du completed both his PhD and undergraduate studies in computer science at Tsinghua University. Dr. Du has published over 30 papers at major international conferences, accumulating over 8,000 citations.
We retain key management and technical staff with competitive remuneration packages and welfare benefits. In the event of termination of employment requested by a key staff, we closely communicate with the staff for the reason of departure and feedback for us. We also recruit candidates with relevant knowledge and skills and invest in training initiatives to avoid the negative impact that could be caused by the departure of any key staff. The key terms of agreements with management and technical staff are set out below.
• Inventions arrangement. We own all rights, titles and interests (including patent rights, copyrights, trade secret rights and all other intellectual property rights of any sort throughout the world) relating to any and all inventions (whether or not patentable), designs, know-how, ideas and information made, conceived or reduced to practice, in whole or in part, by the employee during the term of the employment contract to the fullest extent allowed by applicable laws, and the employee shall promptly disclose all inventions to us.
• Proprietary information arrangement. All inventions and all other business, technical and financial information the employee develops, learns or obtains during the term of the employment contract that relate to us or our business or demonstrably anticipated business, or that are developed in whole or in part during the employment or using our equipment, supplies, facilities or confidential information, or that are received by or for us in confidence, constitute proprietary information. The employee shall hold in confidence and not disclose or, except within the scope of the employment, use any proprietary information.
• Confidentiality. During the employment, except as necessary to perform work duties, and for all time thereafter, the employee shall not, without our prior written consent, disclose, divulge, announce, publish, impart, transfer or otherwise make known to any third party, or in any way use any information, such as technical and trade secrets, belonging to us or belonging to any other party for which we have a duty of confidentiality.
• Non-competition. For certain staff, we have the right to unilaterally initiate a non-competition period of up to two years following the termination of employment. During the term of employment and the non-competition period initiated by us, the employee shall not engage in any competitive behavior.
In addition to our full-time R&D employees, we have an internship program through which university students are offered the opportunity to work with our R&D team. We enter into written internship agreements with confidentiality, non-compete covenants and intellectual ownership clauses with our interns to make sure that all intellectual property developed by them in connection with their work at our Company, are our property, even after their departure from our Company.
While we conduct a vast majority of our R&D activities in-house, we also partner with leading universities around the world to establish joint research laboratories, driving innovation in advanced AI technologies and constantly replenishing our talent pipeline. We do not outsource any R&D activities to third parties. Our in-house R&D activities primarily serve our the development and enhancement of our serve offerings and our independently initiated projects, with an emphasis on technological innovation and practical application. Our in-house R&D focus on fundamental principles of large models, model training, parameter tuning and algorithm optimization. In contrast, our collaborative R&D is centered around joint research projects with academic institutions, mainly exploring the application of large models in specific fields. These projects have a stronger academic focus and tend to be more fundamental in nature, serving the purpose of basic research rather than direct technological enhancement. As of June 30, 2025, we had
close collaboration with various universities and research institutions in China such as Tsinghua University, Zhejiang University, Tianjin University, Yanshan University and Zhongguancun Academy. For example, we have collaborated closely with the Department of Computer Science at Tsinghua University and co-established the Knowledge Engineering Group (KEG). KEG focuses on original innovation in AI driven by both data and knowledge, conducting research in areas such as AI, LLMs, knowledge graphs, data mining and social networks. KEG has participated in a number of leading national and international projects. We also partner with numerous industry leaders through KEG, including the following:
• Dr. Li Juanzi: Dr. Li is a leading expert in semantic content management and text and social network mining. She has led pioneering research with impactful applications in journalism, social networks and web services. Dr. Li is the project leader and one of the drafters for the national standard "Chinese News Markup Language," now adopted by a major news agency in China. Dr. Li has also published extensively in leading academic venues and her research has been recognized with major awards. Her work bridges foundational AI research with real-world system deployment, driving advances in semantic, multilingual and multi-modal knowledge technologies.
• Dr. Huang Minlie: Dr. Huang is a leading expert in AI and NLP. His research focuses on large language models, AI safety, social intelligence, and language generation. He has made significant contributions to dialog systems, automatic question answering and sentiment analysis. Dr. Huang has published over 400 papers in top venues, accumulating over 25,000 citations. Dr. Huang has served key roles for major conferences. He has received prestigious awards such as the Wu Wenjun Artificial Intelligence Science and Technology Award (吴文俊人工智能科学技术奖) and Qian Weichang Chinese Information Processing Science and Technology Award (钱伟长中文信息处理科学技术奖).
• Dr. Xu Bin: Specializing in knowledge graphs, data mining and AI, Dr. Xu has led projects in developing large-scale educational knowledge graphs and scalable technology for scientific intelligence mining and knowledge services. He has published over 100 papers in top conferences and journals and serves on the editorial board of Service Oriented Computing and Applications (SOCA). Dr. Xu chairs key committees in the Chinese Computer Federation and leads standardization efforts as Chair of the IEEE Knowledge Graph Working Group. His work has advanced the application of knowledge graph technologies, supporting intelligent knowledge discovery and integration at scale.
Together with Tsinghua University and Peking University, we also jointly established the Beijing Key Laboratory of Foundation Models. This laboratory focuses on fundamental research into large models, super-intelligence technologies, security and ethical alignment, and is committed to translating cutting-edge research into practical applications in areas such as finance, healthcare and education. In addition, as a founding partner of the Tsinghua University Foundation Model Research Center, we facilitate interdisciplinary collaboration, academic exchange and talent development in the field of AI. Through these collaborations, we are able to identify promising future AI scientists early on, ensuring a continuous supply of young talent with great potential.
These R&D collaborations are mutually beneficial, with third-party institutions and researchers involved benefiting from academic or industry recognition for their contribution, while we are primarily entitled to the intellectual property generated. For our R&D collaborations, we typically enter into written agreements with the collaborating parties, which set out the following key terms.
• Roles and responsibilities. We are typically responsible for providing the materials and data required for the R&D projects as well as monetary support. For collaborations with co-owned patents, we are typically responsible for preparation and filing of the patent applications and maintenance of the relevant patent rights. The collaborating academic institute is typically responsible for designating lead professor and researchers to the project and meeting the set targets of the project.
负责人员。各方须为联合研发分配足够的人员。对于某些重大合作，我们的协议还规定了研究机构须为项目指派的关键研发人员。
资金贡献。我们通常负责合作项目的资金支持及其他实物资助，以及与知识产权申请和知识产权维护相关的费用。
知识产权。一般而言，我们单独拥有通过研发合作所创造的知识产权及商业化知识产权的经济利益。在少数情况下，若我们与学术机构合作主要是为了支持其研究目标，我们将与合作学术机构签订专项协议，共同拥有通过合作所创造的知识产权。
保密。各方须对合作过程中获取的信息保密，且不得将所获信息用于其他目的。
我们的技术基础设施 我们的技术基础设施旨在满足运营需求、支持业务增长并确保运营的可靠性。我们持续致力于技术基础设施的开发，以低成本为训练和迭代强大人工智能模型提供计算资源。 训练和推理均属于计算密集型任务。我们先进的计算基础设施为模型开发提供了充足的计算资源。此外，我们与计算资源合作伙伴开展合作并共同设计计算基础设施，以确保此类合作伙伴提供的计算资源最适合我们的模型，并足以支持我们在不同平台和硬件上进行可扩展的训练和模型部署。我们现有技术架构和服务的兼容性使我们能够在不同计算基础设施之间实现快速切换和有效整合。因此，我们预计在切换计算基础设施提供商方面不会遇到任何重大技术问题或其他障碍。我们还通过与中国受信赖的云计算服务提供商合作，采用了高度可扩展的云端技术架构。我们的云端技术架构使我们能够实时处理大量数据，并确保在大规模情况下实现高速稳定的性能，以适应和支持我们业务运营日益增长的复杂性和多样性。我们一直在通过增加对第三方云计算服务的投入来提升技术架构，以确保我们的云架构能够有效满足不断增长的业务需求。 我们由分布于中国各地的服务器提供支持，这确保了我们技术基础设施的高可用性。此外，我们制定了一套全面的应急预案，以应对任何紧急情况或服务中断的潜在风险。例如，我们定期备份运营数据，负责数据库管理的员工每日对备份记录进行检查，以确保所有运营数据均得到妥善存档。我们还测试系统的数据恢复能力，以确保备份数据可被完整检索。在记录期间及直至最后实际可行日期，我们未曾经历任何重大服务中断。
of applications, while meeting critical standards for safety, performance and scalability. Generally, our foundation language models are developed through the following process.
• 架构设计。我们首先根据目标任务的复杂性、输入数据的特征以及部署环境的要求，选择并调整底层模型结构。虽然我们将Transformer架构作为大多数模型的基础，但我们也在研究各种改进Transformer架构的方法。这些方法包括：稀疏专家模型（如MoE）、结合多种架构风格以增强灵活性的混合编码器结构，以及多模态交叉注意力设计——使模型能够处理和交互来自不同类型数据（如文本和图像）的信息。我们的设计策略强调模块化，使模型能够处理多种输入类型，包括文本、计算机代码和结构化表格数据。为支持训练稳定性和可靠学习，我们整合了以下技术：前层归一化（在网络某些节点对数据进行标准化以保持一致行为）、残差缩放（调整快捷连接对信息流的影响方式），以及随机深度（在训练过程中随机跳过某些层以提高泛化能力）。我们还充分考虑模型将部署于其上的硬件，使设计选择与合适的并行处理策略相匹配。这些策略包括：张量并行（将模型内的操作拆分至多个计算单元）、流水线并行（将模型的不同部分分布于硬件上以序列方式处理数据），以及全分片数据并行（FSDP）——将数据和模型参数分散至多台机器，以实现高效的大规模训练。所有决策均充分考量部署基础设施的具体要求与限制。
• 预训练。预训练阶段专注于培养模型具备强大的通用语言理解与生成能力。为此，我们在来自公开渠道和第三方供应商的大规模多样化数据集上训练模型。我们的数据处理流程包含三个主要步骤。第一，通过去重处理删除近似重复文档，有助于提高训练数据的多样性，并降低模型单纯记忆重复信息的风险。第二，过滤低质量内容，包括含有冒犯性语言、无意义占位符、有缺陷的网页标记或损坏代码的内容，以确保数据集的完整性。第三，应用分词技术，将文本拆解为更小的单元（如子词），以便对不同语言和内容类型进行一致处理。我们通过对数据集进行分层和重新加权，进一步提升数据质量，使高质量来源在训练中获得更好的代表性。
我们的训练目标通常遵循自回归语言建模范式或基于跨度的空白填充策略。与仅专注于生成或词元预测的传统预训练方法不同，我们的方法使模型能够同时执行语言生成和理解任务，从而提升其在广泛下游应用中的效能。为在训练过程中最大化计算效率，我们采用以下方法：混合精度计算（在降低内存使用的同时加快处理速度）、梯度检查点（通过选择性存储中间计算结果来节省内存），以及余弦学习率调度（以技术最优方式逐步降低学习速度）。
language skills to support concrete applications, such as text summarization, answering questions or generating computer code. We achieve fine-tuning through gradient-based optimization, where the model's parameters are updated based on specific objectives related to the target task. We also adjust training settings, known as hyperparameters, such as the learning rate which determines how quickly the model learns, batch size, which is the amount of data processed at once, and details of the optimization algorithm. For projects seeking efficient use of computational resources, we employ lightweight tuning strategies, such as Low-Rank Adaptation (LoRA) or P-Tuning, which update only a small subset of the model's parameters while retaining strong task performance. Throughout fine-tuning, we rely on established evaluation metrics to measure text quality and assess accuracy, to ensure that the model's outputs meet both technical standards and business needs.
Post-training alignment. While pre-training and fine-tuning equip the model with technical skills, a post-training alignment phase is essential to ensure the model responds in ways that reflect human values, comply with safety standards and follow task instructions accurately. We achieve this alignment using a two-step process. First, through supervised fine-tuning (SFT), the model is trained on real examples of human-written instructions and their corresponding responses. Unlike artificial, template-based prompts, these genuine interactions improve the model's ability to follow instructions naturally and produce human-like conversations. Second, we apply reinforcement learning from human feedback (RLHF). In this stage, human reviewers assess and rate the model's responses on criteria such as safety, factual accuracy, usefulness and relevance. We then use such annotated data to train a reward model, which helps the main model refine its behavior. We use the reinforcement learning algorithm Proximal Policy Optimization (PPO) for its stability and efficiency. PPO restricts the size of behavioral adjustments made during training by using a "clipped" objective function, helping to prevent abrupt changes and ensuring reliable progress. For example, in our GLM-4-Plus model series, integrating PPO into the post-training phase leads to more consistent and reliable responses and better alignment with human preferences. We carefully check the quality of alignment data review through multi-layer audits.
Deployment optimization ("quantization"). To make our models suitable for deployment in production environments where computing resources are limited, we use various model compression and quantization techniques. Quantization involves reducing the precision of model calculations, such as converting model weights from 32-bit floating point to 16-bit floating point or 8-bit integer formats, which helps decrease the model's size and the amount of memory it uses, with minimal effect on accuracy. We apply both post-training quantization, which is performed after the initial model training, and quantization-aware training, where the model is trained to anticipate and accommodate lower precision from the outset. Additionally, we deploy our models using optimized software runtimes which support advanced methods like tensor parallel inference to further improve efficiency during real-time use. These approaches collectively reduce the model's memory requirements and speed up response times, enabling applications that require fast, low-latency performance, such as those running on edge devices or private cloud infrastructure, while still supporting high throughput and robust task execution.
Continuous evaluation and feedback. After deployment of models, we maintain a rigorous evaluation and monitoring system to ensure ongoing model reliability and safety. This framework continuously tracks key performance indicators in real time, such as latency, frequency of errors and activation of any safety measures. It helps us promptly identify issues such as distributional drift (changes in input patterns) or declines in model performance. When introducing new versions of a model, we use controlled canary rollouts, where we release updates to a small subset of users first, to monitor for unintended effects before broader deployment. We also gather anonymized user feedback, which provides valuable insights for improving future iterations.
To further ensure model quality and compliance, we use carefully designed synthetic test suites and adversarial prompt sets, which assess the model's robustness, fairness and adherence to relevant regulations or guidelines. Updates to the model or its systems are implemented gradually, with comprehensive checks for backward compatibility and regression testing to prevent the reintroduction of past errors, thus safeguarding both technical and operational integrity.
Our advancements in AI technologies depend on consistent innovation and groundbreaking research. The following table highlights our major research breakthroughs, reflecting our commitment to advancing advanced AI and addressing global challenges through responsible innovation.
| Date | Breakthrough | |---|---| | June 2021 | Completed training of our first ten-billion-parameter model, GLM-10B, based on our GLM pre-training framework. According to Frost & Sullivan, our GLM-10B was one of the first few ten-billion-parameter models trained in China, representing a breakthrough in both model size and technological sophistication. Our GLM-10B's optimization for Chinese language environments also sets a strong foundation for localization of our AI models. | | June 2021 | Completed training of a trillion-parameter sparse model using an MoE architecture. Such sparse model also reached "convergence," meaning our training algorithm has succeeded in finding an optimized set of parameters that enable the model to perform the given task. | | August 2022 | Developed and open-sourced the massive bilingual 130-billion-parameter pre-trained language model GLM-130B. One of our major goals for GLM-130B was to lower the hardware requirements for accessing trillion-parameter-scale LLMs without efficiency and effectiveness disadvantages. To allow GLM-130B to run on smaller systems without losing performance, we attempted to compress GLM-130B as much as possible via quantization techniques. Our GLM-130B also outperformed many then globally recognized models in English tasks and demonstrated solid abilities in Chinese-language tasks, making it a standout model for bilingual applications. Its reproducibility and support for multiple hardware platforms further enhance its practicality for researchers and developers worldwide. | | March 2023 | Released and open-sourced our ChatGLM dialog model, ChatGLM-6B, based on our GLM framework. Quantization significantly reduced the hardware requirements of ChatGLM-6B, so that the model can run efficiently on consumer-grade hardware. By open-sourcing ChatGLM-6B for both academic and commercial use, developers of all levels can access, study and build upon this model, not only making AI models more accessible, but also accelerating the development and adoption of AI across industries. | | January 2024 | Released our next-generation foundation model GLM-4 with significant performance improvements over the previous generation. Our GLM-4 can handle longer and more complex conversations, work effectively across multiple types of input, such as text and images, and perform tasks faster and more efficiently, supporting more users at the same time while reducing costs. Additionally, it has enhanced agent abilities, making it a powerful tool for a wide range of applications. |
Open-sourced GLM-4-9B, the smaller, more accessible version of GLM-4, and GLM-4V-9B, a visual model based on GLM-4. Despite their smaller size, GLM-4-9B and GLM-4V-9B delivered multimodal capabilities comparable to then globally recognized multimodal models, according to Frost & Sullivan.
Introduced CogVideoX, the video generation model built on the earlier CogVideo model with multiple key upgrades that improve performance, efficiency and usability. Using a cutting-edge 3D Variational Autoencoder (3D VAE), it compresses video data to just 2% of its original size, making training faster and less resource-intensive while generating videos that are more natural and fluid by effectively capturing relationships between frames. Its Transformer architecture and innovative 3D position encoding enhances video consistency and ensure smooth transitions across frames. This allows CogVideoX to generate higher-resolution videos (up to 1440×960) in just 30 seconds for a six-second clip—six times faster than its predecessor.
Released the upgraded GLM-4 model, GLM-4-Plus, achieving performance that rivals or even surpasses top international models in key areas such as understanding language, following instructions and processing long and complex texts. Our GLM-4-Plus incorporates PPO to enhance precision and adaptability, making it suitable for complex real-world applications.
Released the end-to-end bilingual voice model GLM-4-Voice. It represents a major breakthrough in voice interaction technology by offering more natural, emotional and efficient human-like communication.
Released our AI Agent AutoGLM, the foundation agents for autonomous control of digital devices through GUIs. Our AutoGLM represents a step toward building more capable AI tools that can support real-world problem-solving with minimal human input.
Released the upgraded AutoGLM Agent with GLM-PC, initiating research on computer use, where the AI can independently operate a computer like a human user does. GLM-PC demonstrates the potential of AI to transform how we interact with everyday technology, bringing us closer to automated and intelligent personal computing.
Unveiled GLM-Zero-Preview, our first reasoning model developed using advanced reinforcement learning technique. Our GLM-Zero-Preview is designed for handling complex tasks requiring mathematical logic, coding and deep reasoning. Using advanced reinforcement learning, GLM-Zero-Preview can break down problems, reflect on its methodology and adapt its approach—mimicking how humans rethink and optimize their solutions.
Released the end-to-end model GLM-Realtime that enables real-time video comprehension and voice interaction with low delay. It can understand and respond quickly to video and audio inputs, remember up to two minutes of conversation context and even perform singing. Our GLM-Realtime is also capable of "Function Call," allowing it to use external tools or data for more complex tasks.
Released AutoGLM Rumination, our first AI agent capable of deep research and tasks. Unlike traditional AI, which either thinks or executes tasks separately, our AutoGLM Rumination combines deep reasoning with practical action, and can handle complex responsibilities and drastically improve efficiency across industries such as research, education, finance and everyday services.
We envision a world where AI serves as a transformative force for good, a world where AI—ultimately AGI—empowers humanity to improve life in all corners of our planet and address the greatest challenges of our time.
At the same time, there are profound risks inherent in creating machine systems that emulate human intelligence. Handled in wrong ways, AI could lead to grave harm, even catastrophe. In particular, AI will seep into the finest fabric of human society and smallest crease of everyone's lives. The culture of every community and the wellbeing of every individual will be at stake. Therefore, we must approach AI with extreme care.
As a leader in foundational AI technologies, we are committed to adhering to our core AI safety and ethics principles and steering our organization and everything we do around these principles.
Safety as our fundamental priority. We pledge to conduct rigorous research and deploy robust safety mechanisms to ensure predictable, controlled AI behaviors under all conditions. Our safety measures include the following three levels:
• Model level. We will continuously refine our models to reduce hallucinations and biased outputs. We will also prevent offensive behaviors or unethical recommendations from the models.
• Personal level. We implement strong safeguards to prevent unauthorized access to sensitive information and protect individuals from manipulative or unethical actions that could harm mental health or exploit personal vulnerabilities.
• Sovereign level. We also see a crucial need for collaboration among stakeholders we call "sovereigns"—governments, industries and international organizations that represent the wills of their constituencies—to ensure safety as AI poses unprecedented questions on the social fabric of our societies at every level: local, tribal, national, global. We are dedicated to working closely with the broader community to develop AI best practices and global governance frameworks. We, together with the world's leading AI companies, signed the Frontier AI Safety Commitments in the 2024 AI Seoul Summit. These commitments include: (i) effectively identify, assess and manage risks when developing and deploying frontier AI models and systems; (ii) be accountable for safely developing and deploying frontier AI models and systems; and (iii) use approaches to frontier AI safety that are appropriately transparent to external actors.
Democratization, diversity and equity. AI should be advancement for all humanity, not a privilege reserved for a few. We are dedicated to ensuring that its benefits extend beyond the tech elite and the wealthy, and reach underserved communities and marginalized groups. Addressing power imbalances and ensuring fairness will remain central to our approach. We will design AI systems to respect human values, autonomy and dignity, embedding alignment processes to ensure these systems are inclusive and equitable.
Environmental sustainability. We are deeply aware of the environmental costs associated with training and deploying large-scale AI systems. To mitigate these impacts, we are committed to designing energy-efficient systems and using clean, renewable energy sources for our operations. Sustainability will be a core principle guiding us, ensuring that technological progress does not come at the expense of our planet.
透明度与问责制。我们认为，人工智能领域的信任只能通过透明度与问责制来赢得。我们不仅将传达我们的成功，也将坦诚面对挑战，直面风险并公开说明局限性。问责制意味着对我们的行为承担责任。如果发生错误或出现意外后果，我们将迅速采取行动加以应对，公开说明问题所在，并实施解决方案以防止再次发生。通过与更广泛的人工智能社区开展对话，我们旨在创造一个透明、协作的环境，使各方反馈能够推动形成更优质、更安全的人工智能实践。
• 我们与全球领先的人工智能企业共同在2024年人工智能首尔峰会上签署了《前沿人工智能安全承诺》。
• 我们是唯一一家在未来生命研究所（Future of Life Institute）发布的《2024年人工智能安全指数》中获得认可的中国人工智能企业。未来生命研究所是一家由顶尖人工智能专家创立的知名非营利组织。
• 根据斯坦福大学发布的《2025年人工智能指数报告》，依据Hughes幻觉评估模型（HHEM-2.1-Open）的评测，我们的GLM-4-9B模型在顶级大型语言模型中实现了最低幻觉率之一（1.3%）：
| 模型 | 幻觉率 | |---|---| | ai21labs/AI21-Jamba1.5-Mini | 3.5% | | openai/GPT-4 | 2.9% | | openai/GPT-4o-mini | 2.8% | | openai/GPT-4-Turbo | 2.6% | | openai/GPT-4o | 2.5% | | openai/o1-mini | 2.5% | | gemini-2.0-flash-exp | 2.4% | | openai/GPT-3.5-Turbo | 2.4% | | deepseek/deepseek-chat | 1.9% | | openai/o1 | 1.8% | | microsoft/Phi-3.5-MoE-instruct | 1.7% | | microsoft/Orca-2-13b | 1.7% | | Intel/neural-chat-7b-v3-3 | 1.5% | | Qwen/Qwen2.5-7B-Instruct | 1.4% | | THUDM/glm-4-9b-chat | 1.3% |
• 根据检索增强生成（RAG）大型语言模型幻觉排行榜，我们的GLM-4.5于2025年9月实现了全球第二低、中国最低的幻觉率。
我们高度重视数据安全与保护工作。我们建立了完善的个人信息管理体系，并制定了一系列技术标准与规范，以确保数据及个人信息在其全生命周期内的安全。我们的数据安全团队负责统筹推进数据安全与个人信息保护工作。
我们通过自营应用程序，在用户直接授权的前提下收集用户数据，用于提供相关服务。我们还从以下渠道获取数据：（i）第三方供应商，主要包括文本、图像、音频及视频数据，用于训练和优化我们的模型；（ii）开放网站、公共数据集或其他可公开访问的来源，主要包括公开可获取的文本、图像及类似内容，用于训练和优化我们的模型。我们主要依赖开源数据来训练模型的通用能力，例如广泛的语言理解与生成能力，以及基础逻辑推理能力。我们使用第三方供应商数据集主要用于特定行业的训练，因为高质量的专业数据在公开来源中相对稀缺，而从第三方供应商处采购的数据集通常具有更高的质量、准确性和可靠性。总体而言，我们不从任何渠道获取
our institutional customers or their end-users unless we are explicitly entrusted by our institutional customers to do so for the specific purpose of providing services. Details regarding such entrusted data processing are set out below.
We collect data from users who directly authorize us through our self-operated applications, such as mobile apps, WeChat mini programs and other online platforms. Such data are collected solely for the purposes of providing, maintaining and improving our services. We collect data from both domestic and overseas users. Our user privacy policies clearly describe our data collection, usage, sharing and processing practices and how users can exercise their rights in activities relating to the processing of personal information. In particular, we provide users with prior notice and obtain their consent as to what data are being collected, how such data will be used and undertake to manage and use the data collected in accordance with applicable laws before they use our services. The user data we collect, store and use generally include (i) users' basic information, such as mobile phone number; (ii) data from users' interactions with our AI product, including user inputs such as prompts, queries, images, audio files, documents and other information, as well as the content generated by our product; and (iii) attribute information of devices used to install and operate our mobile applications, such as device identifiers. The scope of usage is consistent with that being disclosed in our privacy policies and does not exceed the scope authorized by users. The data are collected and used mainly for the purposes of user registration, content generation, and user safety.
We collaborate closely with our clients for provision of our services and are deeply involved in the deployment with routine communications with the clients. Generally, we do not collect or otherwise process any personal data of our clients' end users unless specifically commissioned or directed by the client to do so. Only in very limited circumstances, where we are expressly entrusted by our clients, we may process the personal data of our clients' end-users, solely for the purpose of performing the agreed-upon services. We only process clients' end-users' personal data in limited scenarios, for which we act as a data processor of the clients. In these scenarios, we provide operational support for two mobile applications on behalf of the clients and, in doing so, obtain consent directly from the end-users through the privacy policies displayed within such applications, or as otherwise required by applicable laws and regulations. Save for such limited and specific entrusted services, we do not collect or otherwise process any personal data of our clients' end-users in the ordinary course of business.
The data we obtained from third party vendors have been provided after anonymization or desensitization pursuant to the relevant agreements with such vendors. We request the third-party vendors to explicitly confirm in their agreements or undertakings with us that: (i) they have acquired data from legitimate sources and that they have obtained the rights to use such data for the purpose specified in the agreements or undertakings; (ii) they comply with all applicable laws and regulations and does not infringe on any third-party rights; (iii) all services, data and content provided to us are free of illegal or harmful content and do not infringe upon third-party intellectual property rights, trade secrets, personality rights, personal information rights or any other legitimate rights or interests; and (iv) they are liable for any breach of these obligations and will compensate us for resulting losses.
For public websites, public datasets or other publicly accessible sources, we limit the scope of data so that such data we collect do not contain sensitive personal information or private personal information.
We process data on an as-needed basis for our services and model training during the ordinary course of our business operations.
For individual users' data, we strictly process data in a manner with the least impact on the rights of data subjects. We process data for specific, reasonable purposes within a minimal scope and inform the subjects of such personal information in advance. Data shall not be used for any purpose irrelevant to such purpose.
对于客户数据，我们通常在服务器端部署访问控制机制，不收集、存储、处理、使用或拥有客户终端用户的个人数据。我们仅按照客户的委托，依据客户的指示处理此类数据，以履行与客户合作协议的目的，且不超出协议规定的范围或客户明确指示的范围处理此类数据。为确保这一原则的落实，我们在不同场景中采取了具体措施。例如，对于选择本地部署的MaaS客户，我们的模型托管于客户自身的基础设施内，从而确保数据处理始终处于客户的控制之下。在此场景中，在产品维护过程（如模型微调）中，我们仅在获得客户同意的前提下，出于维护目的访问客户本地存储的数据，不进行任何其他数据处理活动，例如传输或下载数据。而对于偏好基于云端部署而非本地基础设施的MaaS客户，我们仅处理客户终端用户与我们AI产品之间的交互数据，且此类数据处理严格受与客户签订的服务协议条款约束。
对于AI模型训练，我们在处理前对输入数据进行脱敏处理。我们删除或使用类似技术处理个人及敏感信息，例如个人可识别信息，并在现有技术措施的限度内作出合理努力，确保我们的数据在用于训练目的时不附带个人信息。
除我们隐私政策中规定的有限目的及特殊情形外，我们不与第三方共享用户数据。根据我们的政策，我们仅出于合法、必要及特定目的授权第三方业务合作伙伴访问我们的用户数据，并向用户告知数据共享的目的、用途及范围。我们在相关协议下从经授权的业务合作伙伴处取得具有法律约束力的承诺，要求其在处理此类数据时遵守授权的目的、范围及安全措施。我们对个人可识别信息进行去标识化处理，且在未经用户同意的情况下，不与经授权的合作伙伴共享任何个人可识别数据。
我们已建立数据及个人信息的存储与删除机制及管理程序。我们存储用户个人信息的时间以实现收集目的所必需的最短时间为限。对于法律规定有强制保存期限的个人信息，我们遵守相关要求。例如，根据适用法律的规定，我们自交易完成之日起至少保留三年的在线交易信息，包括服务详情及支付记录。对于不受特定法定保存期限约束的个人信息，我们遵循《中华人民共和国个人信息保护法》第十九条规定的"最短必要期限"原则，确保我们仅在实现收集或处理该信息的特定目的所需的最短时间内保留个人信息。我们根据个人信息的类型确定保存期限，并综合考量处理的目的与必要性、所涉个人信息的敏感程度、处理相关的潜在风险以及我们具体的业务需求。在处理数据的目的已实现后、用户申请注销账户或删除数据时，或在适用法律法规另有规定的情况下，我们及时删除此类个人信息或对其进行匿名化处理。
我们将在国内业务运营过程中收集和产生的所有用户数据存储于中国境内。为确保数据的保密性与完整性，我们维护着一套全面而严格的数据安全体系。具体而言，我们对用户个人信息实施集中治理，通过API加密传输的方式，将各业务单元收集的机密个人信息统一存储于我们的数据中台。我们的数据中台采用加密技术对此类个人信息进行加密处理。
algorithms. Each business unit and data processing personnel can only access such information on a need-to-use basis with formal approval procedure.
As of the Latest Practicable Date, personal information collected or generated by our domestic operating entities in the PRC was stored within the PRC and has not been transferred overseas and personal information collected by our overseas operating entities through our Z.ai website was stored in Singapore. In accordance with our privacy policy, our overseas operating entity has conducted certain cross-border transfers of user data. Specifically, during the Track Record Period, certain data collected or generated by our overseas operating entity, primarily consisting of anonymous technical data, such as device type, operating system, browser version, session duration and on-site event interactions, were transmitted to our domestic entities solely for processing. As our domestic entities did not provide any personal information of PRC individuals or any important data to overseas recipients during such transfer and processing, such transfer and processing activities are not subject to PRC laws and regulations governing cross-border data transfers, according to our PRC Legal Advisors. In addition, in accordance with our privacy policy, our overseas entities use third-party vendors and service providers, such as analytics service providers, as necessary to provide or optimize our services. Accordingly, the collection of pseudonymous or anonymous technical and interaction data may result in cross-border transfer of data.
While our overseas entity may collect certain information submitted by users located in the PRC, based on the user information identifiable to us and our reasonable assessment, we have not engaged in any cross-border transfer of important data or PRC personal information that would trigger any thresholds or specific criteria under PRC laws and regulations on cross-border data transfer. First, such information has not been identified, notified or publicly designated by any competent governmental authorities as "important data" and therefore does not require a security assessment as important data, according to our PRC Legal Advisors. Second, our "reasonable assessment" primarily focused on whether the data collection activities of our overseas operating entity constitute a cross-border processing activity subject to PRC data export regulatory requirements, including whether our overseas online product is targeted at PRC individuals or involved the analysis or evaluation of their behavior. In particular, (i) the online product is provided in English and accessible globally without PRC-specific versions or localization; (ii) registration and login do not require PRC-specific identifiers such as PRC mobile numbers or identity card numbers; and (iii) the product contains no precise geo-targeting functions or mechanisms that select users located in the PRC. While PRC-based users may voluntarily access the overseas online product and submit information due to the global and open nature of online services, such possibility does not constitute purposeful collection of PRC individuals' personal information. Accordingly, as advised by our PRC Legal Advisors, we are not required under PRC laws and regulations to apply for a security assessment for cross-border data transfers, to complete the filing of standard contracts for such transfers or to obtain personal information protection certification.
Although we did not generate any revenue from our Z.ai website during the Track Record Period, we have implemented appropriate compliance measures, including disclosure and user consent through our privacy policy, completion of a self-assessment on data cross-border transfer, adoption of security measures for data transmission and storage. In line with industry practice, we engaged experienced data security legal advisors to draft the privacy policy. We have been monitoring our overseas entity's data processing practices, with readiness to fulfill any additional regulatory obligations should triggering conditions arise. In light of the foregoing, our PRC Legal Advisors advise that we comply with the current PRC regulatory requirements on cross-border data transfers. Based on our Singapore Data Counsel's diligence on our data protection policies and practices, our Singapore Data Counsel advise that our Singapore operating entity's data collection, storage, usage, disclosure and transfer (including cross-border data transfer) complied with the relevant data protection laws and regulations in Singapore, including the Personal Data Protection Act 2012 of Singapore, during the Track Record Period and up to the Latest Practicable Date. Based on the foregoing, our Directors are of the view that we are in compliance with applicable PRC and overseas data protection (including data storage and data transfer) laws and regulations in all material respects.
我们已就数据隐私及个人信息保护事宜制定并实施了相关政策及管理制度，包括以下原则。
• 第三方供应商及客户管理。我们已制定第三方供应商管理制度及客户管理制度，以开展必要的数据安全评估及数据安全能力审查。与相关第三方数据提供商签订的协议包含相关第三方就遵守相关法律法规所作出的陈述与保证，以及在该等数据提供商违约时我们可援引的补救及赔偿条款和争议解决机制。对于我们的运营应用程序（包括移动应用程序），我们在产品上线及重大更新前进行技术测试及隐私合规审查。我们亦定期聘请外部法律顾问对我们与隐私相关的实践活动的合规性进行审查与评估。对于委托我们在特定业务合作中处理其数据的客户，我们要求其确认已从合法来源获取相关数据，并已获得使用该等数据的权利，且其终端用户已就我们协议中规定的目的给予同意。我们仅将数据用于客户明确授权的目的，不将数据用于未经事先批准及同意的目的。我们持续监控与客户的数据处理合作情况，并定期审查该等合作的内容、合作协议的范围及协议的执行情况，以确保符合相关法律法规。
• 全面的数据及个人信息安全管理政策。我们已实施全面的员工保密政策、数据使用审批程序及数据追踪机制，以确保数据库的安全性。我们对可能接触用户个人数据的员工采用最小授权原则。我们的操作系统及数据库系统设有密码复杂度要求，采用堡垒机（一种专门指定的服务器，作为数据库管理访问的安全网关）进行远程管理，并严格限制对默认账户的访问。我们保存覆盖所有系统用户的完整系统审计记录。我们已根据相关规章制度制定相应的工作场所操作规程。作为数据处理者，我们已实施多项数据保护及网络安全措施，以确保妥善处理敏感数据，包括用于所有数据训练活动的数据脱敏技术。
• 网络安全应急预案及应急演练。在业绩记录期内，我们未发现重大数据安全违规问题。我们已制定网络安全应急预案，并开展培训及应急响应演练，以应对任何网络安全紧急事件。如我们的安全措施遭到破坏，我们将依照相关法律法规向主管部门报告，并及时通知受影响的用户。
我们已采用安全套接层（SSL）以确保数据的安全传输，并防止任何未经授权的用户或人员以非预期目的访问或使用我们的数据。数据根据其保密级别进行分类，并在由我们的人员处理之前进行去标识化或匿名化处理。数据的使用及检索须根据数据分类进行评估及审批，并保存操作日志。我们的应用系统设置了身份认证、用户身份唯一性验证、基于角色的访问控制及其他安全控制机制，并使用HTTPS协议进行安全通信。
We use firewalls, anti-malware, network security protection applications and various encryption technologies at both software and hardware levels to protect data privacy and securely store such data. To minimize the risk of data loss or leakage, we conduct regular data backup and data recovery tests. We audit and monitor all the user accounts for server operation. If we find any server operating system with any security loopholes, we will upgrade the security protection to ensure the security of all server systems and applications.
We have our own independent database and secure server system. Our server systems are protected with heightened levels of security. We regularly conduct user account auditing and monitoring of our server operations. Once we discover security issues with certain server systems, we will promptly upgrade such systems to ensure the security of our server systems and applications. We have a comprehensive personal information security and management system, covering security management of our data, source code, personal information, third-party personnel, cybersecurity incidents and infrastructure.
Through continuous investment in technology advancement, we have improved our overall security capabilities. Meanwhile, we have obtained multiple certifications, including the information security management system certification (ISO/IEC 27001:2013), the one-star personal information protection impact assessment certificate, the data management capability maturity (DCMM) Level II certification and the privacy information management system certification (ISO/IEC 27701:2019), and our core information systems have multi-level protection scheme (MLPS) Level III certification. Moreover, our service design puts great emphasis on data security, and our solutions must pass data privacy evaluations and security tests before launching or delivery to clients.
Considering that (i) during the Track Record Period and up to the Latest Practicable Date, we had not been subject to any material administrative penalties, mandatory rectifications or other sanctions by any competent PRC regulatory authorities in relation to cybersecurity, data security or personal data protection, nor had we been involved in or subject to any investigation, warning, litigation or other administrative or governmental proceedings pending or, to the best of our knowledge, threatened against us in this regard; (ii) during the Track Record Period and up to the Latest Practicable Date, we had not experienced any material information security incidents including data or personal data theft, leakage, damage or loss, nor had we received any material claims from any third party based on infringement of such party's rights under applicable PRC data protection laws and regulations; (iii) we have formulated and implemented a comprehensive set of internal policies, procedures and measures on cybersecurity, data security, personal data protection, including, among others, data access control, data encryption and anonymization, data backup and recovery, contingency plans for cybersecurity incidents, incident response procedures, as well as vendor and third-party data management measures, which provide detailed operational guidance and controls for data processing activities throughout the entire data lifecycle; (iv) our core systems have satisfied the security protection requirements under the Multi-Level Protection Scheme ("MLPS") and obtained MLPS Level III Certification; (v) we have established a committee responsible for cybersecurity and data compliance oversight, including formulating and enforcing internal rules and conducting compliance reviews; (vi) we have been continuously monitoring updates to laws, regulations and industry guidelines on cybersecurity, data security and personal data protection, while maintaining ongoing communication with the relevant regulatory authorities and proactively seeking their compliance guidance; and (vii) we promptly update our internal policies to reflect regulatory changes, so as to ensure timely enhancements to our operational practices in response to evolving requirements, our Directors are of the view and our PRC Legal Advisors advise that, during the Track Record Period and up to the Latest Practicable Date, we had not been and were not involved in any non-compliance incidents related to cybersecurity, data security and personal data protection which, individually or in the aggregate, had or were reasonably likely to have a material and adverse financial or operational impact on the Group, and as of the Latest Practicable Date, we were in compliance with applicable PRC laws and regulations on cybersecurity, data security and personal data protection in all material respects in the PRC. In addition, during the Track Record Period and up to the Latest Practicable Date, we had not experienced any material user complaint or subject to any pending or threatened litigation, arbitration or administrative proceedings with respect to our overseas data protection.
《网络安全审查办法》规定了运营者须申请网络安全审查的两种具体情形。具体而言，《网络安全审查办法》要求关键信息基础设施运营者采购网络产品和服务时，应预判该产品和服务投入使用后可能带来的国家安全风险。如影响或可能影响国家安全，运营者须向网络安全审查办公室申报网络安全审查。此外，掌握超过一百万用户个人信息的网络平台运营者赴境外上市的，须向网络安全审查办公室申报网络安全审查。主动申报网络安全审查的义务对我们不适用，理由如下：（i）根据《关键信息基础设施安全保护条例》，重要行业和领域的主管部门、监督管理部门负责按照认定规则，组织认定本行业、本领域的关键信息基础设施，并将认定结果通知运营者。因此，关键信息基础设施运营者的认定须由主管部门作出。截至最后实际可行日期，我们未收到相关监管部门关于将我们认定为关键信息基础设施运营者的任何通知；及（ii）根据我们的中国法律顾问的意见，"赴境外上市"不应包括"在香港上市"。因此，截至最后实际可行日期，我们无需根据《网络安全审查办法》申报网络安全审查。
通过我们的一体化MaaS平台，我们借助多种部署方式，以模型和智能体解决方案向各行业的广泛客户提供智能服务，满足其多元化需求。我们的模型和智能体解决方案可托管于云端、通过API访问、以私有数据集本地部署，或预装于终端设备。为扩大市场覆盖范围和渗透率，截至2025年6月30日，我们已组建一支由145名成员组成的专业销售与营销团队。我们的销售与营销团队按不同地区和行业划分，以把握新商机、拓展新客户和项目，并更好地了解客户需求及市场趋势变化。我们通过内部销售团队推广服务。销售团队按地区划分，分布于中国主要城市及香港、新加坡等海外地区。通过区域化和行业化的销售策略，我们深入目标市场，拓展并维护与目标市场的紧密联系，为客户提供服务。
在销售团队之外，我们设有客户支持团队，专门协助客户对我们的模型进行微调和部署，以增强其在特定业务环境中的能力。我们的销售人员与客户支持团队紧密协作，确保能够针对各行业的痛点提出最优解决方案，并最大化我们服务的附加价值。与此同时，我们的销售运营与管理中心负责统筹项目管理、财务流程及其他销售行政事务。
我们致力于为客户创造价值，与客户共享成功。我们不仅寻求与行业领军企业紧密合作，还积极与专注于特定行业应用的企业建立合作伙伴关系。通过锁定各行业中具有影响力和创新力的领军企业，我们可以共同开发和打造标杆案例，以展示我们服务的切实价值和能力。这些标杆案例亦可作为具有参考价值的范本，建立信任与信誉，促进同行业或相关行业其他主体的进一步采用，并加速我们在不同行业的进入与认可。另一方面，通过与成熟行业应用提供商的合作，我们将模型的核心能力整合至这些行业合作伙伴提供的解决方案中，助力提升其为终端用户提供的产品和服务。这一方式不仅使我们的合作伙伴能够更有效地满足独特的行业需求，同时也确保我们模型的优势在实践中得到充分发挥。通过合作伙伴关系，我们进一步扩大平台的覆盖范围——
and deliver practical model capabilities and value to users across different industries and consumers. In this way, we bring together industry-leading collaborations and technology enablement to drive both the adoption and the ongoing service extension of AI.
We typically establish our presence with entry projects for specific use cases, from which we create value for the early adopters of AI model and agent solutions in each sector and cultivate our word-of-mouth reputation. We believe that our brand equity and market recognition depend, in part, on our ability to add value for iconic corporate clients in each industry sector and increase our industry influence.
We have also increased our brand awareness through a variety of branding and marketing activities to reach potential customers, including in-person and online events, content marketing, partner marketing, developer outreach, search engine optimization, social media and public relations. We hosted and participated in various offline events, such as industry conferences, product launch events and developer forums to showcase client success stories and developer breakthroughs and to deepen industry connections. Such events allow us to demonstrate how AI can empower entities across different industries. Through establishing exhibition booths at these regional and global events, our potential customers around the world may experience how we digitalize cities and businesses with AI. In addition, we enhance awareness of our brand and promote our new and existing solutions through online channels. We have also been a significant contributor to numerous publications on the latest industry development and market trends and publications that educate the general public about AI.
Our intellectual property is critical to our innovation which underpins our success. We seek to protect our intellectual property through a combination of patents, copyrights, trademarks, domain names, trade secrets, confidentiality agreements and other measures. As of the Latest Practicable Date, we had 86 registered patents in China, among which 84 were invention patents, and 232 patent applications in China. Among the 86 registered patents, four were co-owned under our R&D collaboration with universities and research institutions. The table below sets forth our key patents in relation to our MaaS platform as of the Latest Practicable Date.
| Patent name | Jurisdiction | Patent number | Status | Application date | |---|---|---|---|---| | A method, device and medium for optimizing the instruction-following capability of LLMs | PRC | 2024115864607 | Granted | November 8, 2024 | | An intelligent extraction method and system for input text containing mathematical formulas | PRC | 2024103497310 | Granted | March 26, 2024 | | A training method for web navigation AI agents based on LLMs | PRC | 2024103093694 | Granted | March 19, 2024 | | A method, device, equipment and medium for dynamically adjusting the depth of LLMs | PRC | 2024102713774 | Granted | March 11, 2024 | | A training method and device for MoE models based on decision trees | PRC | 2024102713030 | Granted | March 11, 2024 | | An application-oriented LLM interface system, method, device and medium | PRC | 2024100405834 | Granted | December 25, 2023 | | An automatic parallelised language model text generation method | PRC | 2023118355360 | Granted | December 28, 2023 |
| Patent name | Jurisdiction | Patent number | Status | Application date | |---|---|---|---|---| | A method, device and storage medium for generating instruction fine-tuning data | PRC | 2023117958411 | Granted | December 25, 2023 | | A general text quality evaluation method based on LLMs | PRC | 2023116186705 | Granted | November 30, 2023 | | An alignment evaluation method for Chinese LLMs | PRC | 2023116210193 | Granted | November 30, 2023 | | A fair and efficient multi-dialog system evaluation system and method | PRC | 2023115438272 | Granted | November 20, 2023 | | A generative information extraction method and device based on pre-trained models | PRC | 2021110162958 | Granted | August 31, 2021 | | An evaluation method, device and electronic equipment for LLMs | PRC | 2023109676521 | Granted | August 2, 2023 | | A personality test method, device and electronic equipment based on human-machine dialog | PRC | 2023108615338 | Granted | July 13, 2023 | | A model optimisation training system, method and relevant device | PRC | 2023108092439 | Granted | July 4, 2023 | | A method, device, medium and computing equipment for initial dialog content generation | PRC | 2023106006354 | Granted | May 25, 2023 | | A safety evaluation method and relevant device based on LLMs | PRC | 2024119331870 | Granted | December 26, 2024 | | A method for extracting key information frames from surveillance videos | PRC | 201510062263X | Granted | March 13, 2015 |
As of the Latest Practicable Date, we had 160 copyrights, including 154 software copyrights. The table below sets forth our key software copyrights in relation to our MaaS platform as of the Latest Practicable Date.
| Copyright name | Jurisdiction | Copyright number | Registration date | |---|---|---|---| | AutoGLM software V1.1.01 | PRC | 14591247 | December 25, 2024 | | GLM embedded large model system V1.0 | PRC | 12267756 | December 19, 2023 | | Zhipu QingYan software V1.0 | PRC | 11783805 | December 8, 2023 | | ChatGLM-6B software V1.1 | PRC | 11479077 | August 4, 2023 | | Large model GLM3.5_130B platform V0.8 | PRC | 11279154 | June 20, 2023 | | Large-scale pre-trained model application platform V1.0 | PRC | 9568431 | May 20, 2022 | | Large-scale pre-trained model system V1.0 | PRC | 9568432 | May 20, 2022 |
As of the Latest Practicable Date, we also had 314 trademarks and 59 domain names in China. During the Track Record Period and up to the Latest Practicable Date, we did not have (i) any material co-owned IP rights and (ii) any licensing-in or licensing-out arrangements with third parties. We have designed and adopted comprehensive measures to protect our intellectual property. We enter into employment agreements with confidentiality, non-compete covenants and intellectual property ownership clauses with our employees, certain consultants and advisors. They acknowledge that the intellectual property developed by
them in connection with their employment with us, including our in-house developed content, is our property. During the Track Record Period and up to the Latest Practicable Date, we had not been subject to any material disputes or claims for infringement of third parties' trademarks, licenses and other intellectual property rights.
在业绩记录期间，我们的收入实现强劲增长。我们的收入从2022年的人民币5,740万元增长至2023年的人民币1.245亿元，并进一步增长至2024年的人民币3.124亿元，复合年增长率超过130%。我们的收入亦从截至2024年6月30日止六个月的人民币4,490万元大幅增长至截至2025年6月30日止六个月的人民币1.909亿元。在业绩记录期间，我们的毛利及毛利率持续增长。我们的毛利从2022年的人民币3,140万元增加至2023年的人民币8,050万元，并进一步增加至2024年的人民币1.759亿元，复合年增长率超过100%。我们的毛利亦从截至2024年6月30日止六个月的人民币2,200万元大幅增长至截至2025年6月30日止六个月的人民币9,540万元。
在实现持续业务增长的同时，我们于2022年、2023年、2024年及截至2024年6月30日及2025年6月30日止六个月分别录得年内亏损人民币1.437亿元、人民币7.880亿元、人民币29.58亿元、人民币12.356亿元及人民币23.579亿元。在剔除以权益结算的股份酬金开支、向投资者发行的金融工具账面金额变动及上市开支（此等项目并不反映我们的经营业绩）的影响后，我们于2022年、2023年、2024年及截至2024年6月30日及2025年6月30日止六个月的经调整净亏损（非国际财务报告准则计量）分别为人民币9,740万元、人民币6.210亿元、人民币24.656亿元、人民币10.302亿元及人民币17.520亿元。
我们在业绩记录期间的亏损主要源于我们在研究与开发方面的重大投入。我们的研发开支从2022年的人民币8,440万元增加至2023年的人民币5.289亿元，并进一步增加至2024年的人民币21.954亿元。我们的研发开支于截至2025年6月30日止六个月较截至2024年6月30日止六个月的人民币8.592亿元增加85.6%至人民币15.947亿元。
尽管我们的净亏损绝对金额在业绩记录期间有所增加，我们预期将通过提升收入及提高运营效率来扭转净亏损局面。
收入增长是实现盈利的关键。我们已建立强大的人工智能模型及智能体产品组合，以赋能广泛的行业，应对各行业的独特挑战并优化工作流程。借助大语言模型市场的巨大市场潜力以及我们的技术领先地位，我们具备良好条件，能够针对新兴市场机遇升级我们的人工智能模型及智能体，并持续实现收入增长。收入的增长将逐步覆盖相关成本及开支，从而总体上减少我们的净亏损。
根据弗若斯特沙利文的数据，到2030年，人工智能将赋能全球至少20%的日常商业决策，并使全球至少80%的消费者智能设备具备智能化能力，从而创造逾20万亿美元的人工智能驱动经济体量。受计算能力、算法架构以及数据量、质量和多样性提升的推动，人工智能市场预计将经历快速且显著的增长。
According to Frost & Sullivan, the AI market in China grew from RMB93.7 billion in 2022 to RMB160.7 billion in 2024, representing a CAGR of 31.0% from 2022 to 2024, and is estimated to further increase to RMB993.0 billion in 2030 with a CAGR of 35.5% from 2024 to 2030. While the LLM market in China is still at its early stage, with the continued advancement of LLM technologies and growing demand from both enterprises and consumers, the LLM market in China is estimated to grow to RMB101.1 billion by 2030, representing a CAGR of 63.5% from 2024 to 2030. See "Industry Overview." According to Frost & Sullivan, we were the first AI company in China to have self-developed large models at a scale of over 100 billion parameters. Benefiting from the growth potential of the AI market and the LLM sub-market, our comprehensive model portfolio will enable us to grasp the market potential and achieve continuous revenue growth.
Our core growth strategy centers on the strengthen of R&D capabilities in general-purpose large models for the continuous iteration and upgrade of its foundation models and agent frameworks to enhance model capabilities, scalability and adaptability. We have developed a comprehensive suite of AI models, ranging from flagship, large-scale general-purpose models to lightweight agents designed for desktop and mobile environments. This diverse model portfolio allows us to address diverse computing requirements and client scenarios. Moving forward, we intends to further improve our models' apprehension, generation and reasoning abilities, with a particular focus on performance in complex industry settings. By incorporating network optimization technologies, such as MoE architecture and sparse activation techniques, and leveraging high-quality data resources, including proprietary datasets, industry-specific annotated data and synthetic data, we implement a unified technology roadmap under which these measures are developed and deployed in an integrated manner to improve model performance, stability, industry-specific applicability and delivery efficiency. For example, by optimizing inference paths and distributed scheduling, we seek to continuously enhance resource utilization efficiency and response consistency across different use cases, enabling enterprise clients to support large-scale API calls with lower and more stable per-request computing resource consumption and more predictable latency. In parallel, by continuously expanding high-quality general-purpose corpora and licensed structured industry datasets and building industry knowledge bases and fine-tuning frameworks, we aim to materially improve knowledge coverage, language understanding and reasoning consistency, while reducing hallucination rates and enhancing instruction following, tool calling and agent-like capabilities. As a result, our models are expected to be reusable and verifiable in a greater number of mission-critical workflows, such as document processing, professional question-and-answer services, research and decision support and compliance review tasks, driving client progression from single-scenario to multi-scenario use, increasing client renewals and support larger-scale deployments. These integrated measures are designed to improve training quality and generalization capabilities, enhance the performance of our existing AI models in terms of accuracy, efficiency and stability, and support the development of iterated and upgraded AI models with enhanced capabilities and a broader range of application scenarios. We will also continue building our multi-modal capabilities, enabling enhanced comprehension and output across text, speech, images and video, maintaining a technological advantage in multi-modal interaction and content understanding. For AI agent solutions, we are developing our agent frameworks with autonomous planning, tool use, long-term memory and reflective learning abilities, empowering the agents to deconstruct complex tasks, autonomously call external tools and self-optimize their own execution processes, and thus gradually achieving continuous self-improvement.
We plan to drive client demand and stickiness through coordinated and significant performance upgrade of our models, which are intended to enhance user experience and increase client dependency. Improvements in our AI models and agents' performance in certain critical tasks, such as code generation, question answering and content creation, are expected to boost client usage and application scenarios. As our network optimization technologies improve inference efficiency and service stability, we expect unit computing resource consumption and, accordingly, unit inference cost to decrease, which is expected to contribute to improved gross margins for our cloud-based solutions and create headroom for more flexible
pricing strategies. At the same time, as our models' agentic, reasoning and multi-step task execution capabilities strengthen and a single foundation model can be deployed in a broader range of high-value scenarios and critical business processes, we expect clients to increase both the breadth and depth of their usage and to integrate our models more deeply into their daily operations, thereby increasing switching costs and reinforcing client "stickiness". In addition to significant upgrade, our regular model and agent upgrades will ensure ongoing product competitiveness and sustained revenue growth. As model capabilities and deployment efficiency advance, we expect to steadily narrow our operating loss, supporting our long-term growth and strengthening our profitability.
To continue our revenue growth, we intend to expand our customer base. During the Track Record Period, our client base continued to grow as we improved our brand awareness. The number of our institutional customers increased from 48 in 2022 to 3,156 in the six months ended June 30, 2025. Our growing client base reflects the scalability and strength of our MaaS platform. We have strategically positioned ourselves in industry sectors with complex application scenarios, such as technology and financial services. By continuously creating value for clients, accumulating industry insights and enhancing our service offerings, we are able to deepen our expertise and broaden our industry impact and thus attract new clients in the same or related sectors. In addition, as our models can be easily adapted to other companies in the same or similar industry sector, we expect to expand our client base at low costs. Through our partnerships with lighthouse clients in various fields, we jointly drive innovation in industry demand and simultaneously improve model performance and cross-sector applicability. We plan to deepen our industry penetration through our comprehensive and scalable MaaS platform and reach more end users through more IoT devices empowered by our AI models, further empowering a broad range of industries with our large model capabilities. For example, in the consumer electronics sector, we have established collaborations with several leading global consumer electronics providers, deploying our models on their flagship products to support on-device intelligent interactions, facilitate multilingual conversations and deliver personalized recommendations. We intend to further expand the role of AI in consumer electronics, aiming to enable cross-device collaborations and adaptive edge intelligence. In the internet sector, we have formed partnerships with several leading Chinese internet companies, enhancing functions such as translation, data analytics, content creation and search with our AI capabilities. As awareness of AI's strategic value rises among internet companies, we aim to expand the application of our models to an even broader array of use cases, such as advertising recommendations, intelligent customer service and knowledge analysis.
In addition, technological innovation is considered a crucial factor for our development. By consistently iterating our foundation models, key algorithms and large-scale training and inference infrastructures, and continuously developing versatile, powerful models by designing new model architecture and optimizing training infrastructure, our AI model and agent solutions will be able to tackle complex tasks and continuously improve their performance through self-learning and reflection. Therefore, we expect to attract new customers through our technological innovation and leading performance of our AI model and agent solutions. Furthermore, as we expand our model portfolio across a spectrum of parameter scales and engineer models that adapt to different computational capabilities, we reduce hardware barriers and make our large models more accessible, extending our customer reach to companies and organizations of all sizes, as well as individual customers.
With our MaaS platform, we are able to enter into new industry sectors in a time-efficient and cost-efficient way. We have been continuously expanding to new industry sectors. For example, we partnered with a world-leading appliance and consumer electronics corporation to integrate our agentic models into their latest smartphone series in 2024. In the future, we intend to continue working with clients that are
market leaders with deep industry knowledge and abundant scenario data, such as Kingsoft Office, Zhaopin.com and Mengniu Dairy. We plan to co-create application scenarios and undertake tailored model fine-tuning, enabling us to build high-precision, industry-specific models that directly address real-world operational needs, expanding our footprint to empower more industry sectors. By leveraging these market leaders' industry know-how such as application scenarios and integration methodologies, we plan to rapidly obtain high-quality real-world data on user preference and habits and operating environment that enhances the performance and expertise of our AI models, facilitating their adoption in sectors with high entry barriers. In particular, through our collaboration with the consumer electronics corporation, we expect to accumulate practical know-how regarding the performance of our models under different hardware and resource constraints, such as latency, power consumption, memory usage and stability, cross-device collaboration mechanisms and common failure modes and edge cases, without accessing or relying on end-users' personal information. Once validated in these specialized settings, we plan to standardize the solutions for broader use, deploying them via API, cloud and on-premise deployment. Over time, these integration methodologies and insights form a replicable framework that shortens deployment cycles, lowers implementation and maintenance costs and reduces technical risks for new clients, making it easier to roll out our models across additional device categories and industry scenarios. As a result, enterprises are more likely to integrate our models in their business processes, leading to sustained engagement and increased usage over time. We expect this collaborative approach to accelerate the commercial adoption of our AI models, drive up client usage and generates recurring revenue streams from both existing and new clients. As the collaborating client base and our engagement levels grow, we expect the demand for our AI model and agent solutions to increase, further accelerating our revenue growth and deepening our relationships within each vertical sector.
We are dedicated to creating super intelligent, widely accessible AGI-native applications that can empower and serve a wide range of customers and industries. We plan to further upgrade our agent workspace that enables customers to easily and seamlessly integrate diverse model applications and tools. This will improve the efficiency in deployment of our AI agent solutions deeply tailored to new industries, regions and scenarios, and drive advancements in intelligent automation. Furthermore, by tapping into the expansive intelligence network across industry verticals, we are able to fully leverage the adaptability of our foundation models to deliver wide-ranging benefits to end users of IoT devices. This enables us to accelerate the commercialization of our AI model and agent solutions. Looking ahead, we expect to develop an extended network that evolves from business-to-business to business-to-consumer applications, fostering a more diversified and resilient revenue structure.
In the next six months, we intend to prioritize the consumer electronics and IoT sectors, accelerating our model deployment and scenario validation. In the longer term, we plan to explore sectors with higher entry barriers, such as education and healthcare. For entry into each sector, we plan to incur R&D expenditure, including the computing service fees, of 10% to 15% of the revenue generated from such sector.
We adopt a structured strategy to enter new industry sectors, including education and healthcare, which typically feature higher regulatory, technical and trust entry barriers. We plan to actively collaborate with industry participants that possess deep domain-relevant expertise and long-standing operational experience but do not have the capability or resources to develop their own foundation models. In the education sector, for example, we established strategic partnerships and joint ventures with leading companies to co-develop products and services tailored to educational use cases. Under these partnerships, we contribute our AI model and agentic technologies, while our partners contribute their curriculum expertise, teaching methodologies, compliance know-how and access to real-world application scenarios. This division of roles enables us to quickly understand and encode domain-specific requirements, such as pedagogical logic, assessment standards and student interaction patterns, into our solutions, and to deliver deployable offerings that address concrete business needs in a high-barrier sector without needing to build industry capabilities from scratch. In addition, as many companies in these sectors prefer to integrate existing third-party foundation models as the underlying capability for their business-facing or consumer-facing applications,
we plan to provide our general-purpose and industry-adapted models to institutional customers on a business-to-business-to-business or business-to-business-to-consumer basis, where enterprises can develop industry-specific AI model and agent applications based on our models. This approach allows us to participate in the value chain of high-barrier sectors indirectly but at scale, broadening our reach across multiple institutions and use cases while our clients handle end-user operations, content, licensing and sector-specific compliance. Furthermore, we are continually upgrading our models to satisfy the stringent requirements of education and healthcare use cases, which place particular emphasis on reliability, safety and controllability. In practice, we focus on optimizing key technical indicators that are critical for these sectors, including model stability, hallucination control, instruction following and robustness in complex, multi-step task execution. For example, we are enhancing our models' ability to (i) execute multi-stage reasoning processes with consistent outputs, (ii) strictly comply with predefined prompts and safety policies, (iii) call external tools in a controllable manner and (iv) maintain high availability and predictable behavior in complex, high-stakes scenarios. These improvements are designed to ensure that, when our models are deployed in education and healthcare workflows, they can provide highly reliable support for tasks such as structured content generation, knowledge-based question answering, workflow assistance and preliminary information triage, within guardrails set by sector specialists. Together, we expect these measures to enable us to not only enter but also to establish and consolidate our position in education, healthcare and other high-barrier industries, supporting sustainable business expansion and revenue growth in these sectors.
提升运营效率也是实现盈利的重要因素。下表列示了所示年度我们的研发费用、销售及营销费用和行政费用占收入的百分比。
| | 截至12月31日止年度 (Year Ended December 31,) | | | 截至6月30日止六个月 (Six Months Ended June 30,) | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (%) | | | (未经审计 / unaudited) | | | 占收入百分比 (As a percentage of revenue:) | | | | | | | 研发费用 (Research and development expenses) | 147.0 | 424.7 | 702.7 | 1,913.2 | 835.4 | | 一般及行政费用 (General and administration expenses) | 56.3 | 53.2 | 42.8 | 114.6 | 97.0 | | 销售及营销费用 (Selling and marketing expenses) | 26.4 | 81.3 | 124.0 | 321.1 | 109.3 | | 合计 (Total) | 229.7 | 559.2 | 869.5 | 2,348.9 | 1,041.7 |
我们的运营费用占总收入的百分比从2022年的229.7%增加至2023年的559.2%，并进一步增加至2024年的869.5%，主要归因于我们在研发及营销费用方面的大幅增加投入。尽管我们预计在不久的将来将继续产生大量研发费用，但我们已采取措施控制一般及行政费用并提升运营效率。我们的运营费用占总收入的百分比从截至2024年6月30日止六个月的2,348.9%下降至截至2025年6月30日止六个月的1,041.7%。我们简化了工作流程并优化了资源配置，以确保以具有成本效益的方式满足运营需求。此外，我们对销售及营销活动采取了聚焦式方法，以进一步提升相关费用的效率。通过将资源和精力导向基于云的部署并利用数据驱动的洞察，我们优化了销售及营销策略。因此，随着我们加快解决方案的商业化以实现收入增长，并提升销售、营销及行政活动的效率以及相关支出的管理，我们预计运营费用占收入的百分比将有所下降。
advancements and iterations of our foundation models, boosting our profitability. Our research and development expenses amounted to RMB84.4 million, RMB528.9 million, RMB2,195.4 million, RMB859.2 million and RMB1,594.7 million in 2022, 2023, 2024 and the six months ended June 30, 2024 and 2025, respectively accounting for 147.0%, 424.7%, 702.7%, 1,913.2% and 835.4% of our total revenue in the same periods, respectively. Our R&D expenses as a percentage of our total revenue increased significantly from 2022 to 2024 primarily because we expanded our R&D team and devoted significant efforts to iterate our foundation models and advance our technology infrastructure. Our R&D expenses as a percentage of our total revenue were relatively high in the six months ended June 30, 2024 and 2025 compared to that in 2022, 2023 and 2024 primarily because we strategically increased our R&D investment in early 2024 to accelerate the development and iteration of our next-generation flagship foundation models and multimodal agents, which result in a higher computing service fees. Going forward, we plan to further optimize our R&D processes and strengthen our internal R&D collaboration mechanisms to enable dynamic allocation of R&D resources. As computing service fees contribute to a substantial portion of our research and development expenses, we plan to focus on increasing the efficiency of computing resource through advancements in our research and development capabilities. As we optimize training and inference of our models and advance our self-developed infrastructure platforms, we expect to achieve more efficient consumption of computing services. As a result, our R&D expenses as a percentage of our total revenue decreased from the six months ended June 30, 2024 to the six months ended June 30, 2025. In addition, we can leverage our leading industry position and our models' broad compatibility for computing power infrastructure and seek business partners to secure better pricing on computing service. Furthermore, benefiting from our MaaS platform, we have achieved efficient mass-production of AI applications. Leveraging the increasing capabilities of our MaaS platform, we aim to generate economies of scale and shorten the time-to-market of our AI model deployment and commercialization, achieving improvement in operating leverage as our business grows.
Our general and administration expenses amounted to RMB32.3 million, RMB66.3 million, RMB133.6 million, RMB51.5 million and RMB185.2 million in 2022, 2023, 2024 and the six months ended June 30, 2024 and 2025, respectively, accounting for 56.3%, 53.2%, 42.8%, 114.6% and 97.0% of our total revenue in the same periods, respectively. Our general and administration expenses as a percentage of our total revenue decreased from 2022 to 2024 and from the six months ended June 30, 2024 to the six months ended June 30, 2025, primarily attributable to our effective cost control measures and our efforts in increasing our operational efficiency. Our general and administration expenses as a percentage of our total revenue were relatively high in the six months ended June 30, 2024 and 2025 compared to that in 2022, 2023 and 2024 primarily due to the expansion of our management team, as well as the implementation of our share incentive schemes in late 2024 and early 2025 to attract and retain our key talent. We expect our general and administration expenses to grow significantly alongside our business growth mainly due to expected increase in employee benefit expenses. We expect to continue to evaluate and monitor the effectiveness and efficiency of our general and administration expenses in order to improve our operating leverage. Except for an expected temporary increase in our general and administration expense as a percentage of our revenue in 2025 primarily due to expected significant increase in equity-settled share-based compensation expenses and expected listing expenses, we expect our general and administration expenses to remain relatively stable as percentage of our revenue in the near future with a decrease in such percentage in the long run. To improve management efficiency, enable faster decision-making and achieve higher stability of the core team, we have streamlined our internal finance, operations and project management processes. Going forward, we plan to introduce digital management tools and data-driven decision support systems to refine budget management, project monitoring and performance evaluation processes to enhance our resource allocation.
Our selling and marketing expenses amounted to RMB15.1 million, RMB101.2 million and RMB387.5 million, RMB144.2 million and RMB208.6 million in 2022, 2023, 2024 and the six months ended June 30, 2024 and 2025, respectively, accounting for 26.4%, 81.3%, 124.0%, 321.1% and 109.3% of our total revenue in the same periods, respectively. Our selling and marketing expenses as a percentage of
our total revenue increased from 2022 to 2024 primarily attributable to significant increase in our advertising and marketing expenses, as we strategically made more advertising investment in order to swiftly take advantage of emerging market opportunities. Our selling and marketing expenses as a percentage of our total revenue then decreased from the six months ended June 30, 2024 to the six months ended June 30, 2025 as we adopted a more focused and effective marketing approach. Our selling and marketing expenses as a percentage of our total revenue were relatively high in the six months ended June 30, 2024 and 2025 compared to that in 2022, 2023 and 2024 primarily driven by our increased efforts to promote the launches of our major new AI models and our expanded MaaS offerings as we invested in brand promotion, customer education, industry events and building our partner ecosystem. These initiatives improved client awareness and our market acceptance, expanded our partner network and accelerated the adoption of our new AI model and agent solutions. We expect our selling expenses to grow alongside our business growth mainly due to expected increases in employee benefit expenses. As we increase our brand awareness, we expect our selling and marketing expenses to decrease as a percentage of our revenue in the long run. To improve marketing efficiency, reduce client acquisition costs, accelerate market expansion and achieve sustained growth in brand influence, we plan to use data analytics to enable targeted client engagement and channel resource allocation, as well as strengthen brand communications and client experience management.
We have a healthy cash balance to support our operations and future business expansion. During the Track Record Period, we funded our cash requirements principally with cash generated from our operations and capital contribution from shareholders. We had cash and cash equivalents of RMB218.9 million, RMB1,249.2 million, RMB2,268.2 million and RMB2,552.0 million as of December 31, 2022, 2023 and 2024 and June 30, 2025, respectively. We believe that we possess sufficient working capital, taking into account the financial resources available to us, including (i) cash and cash equivalents; (ii) short-term investments measured at FVPL; (iii) available bank facilities; and (iv) the estimated net proceeds from the Global Offering.
We had net cash flows used in operating activities of RMB68.2 million, RMB648.0 million, RMB2,244.9 million, RMB994.7 million and RMB1,327.2 million in 2022, 2023, 2024 and the six months ended June 30, 2024 and 2025, respectively. Our operating cash flows were primarily due to (i) our net losses and (ii) our significant investments in our research and development efforts to enhance our services during the Track Record Period. We expect to improve our net operating cash outflows position by taking advantage of (i) our continuous revenue growth; (ii) our improved operating leverage; and (iii) our improved working capital. In addition, we plan to reinforce our receivable collection efforts to reduce the accounts receivable collection cycles. We plan to (i) further strengthen our client management practices, including rigorous review of payment terms at the contract stage, performing periodic review, monitoring payment behavior and implementing credit assessment procedures to ensure their financial creditworthiness and (ii) enhance collection of accounts receivable by promptly issuing invoices, regularly checking with our clients to ensure collection and implementing remedial measures when customers do not make timely payment.
We have a broad and diverse client base, which expanded rapidly over the Track Record Period. Our customers include enterprises, public sector entities and individual users. We generally enter into written agreements with our enterprise and public sector clients, the major terms and conditions of which are set out below.
• Service scope. We provide licenses of use rights to our model and/or agent solutions, maintenance and upgrade service related to our model and/or agent solutions and/or hardware embedded with our model and/or agent solutions.
Pricing. Our pricing is primarily determined by functions of solutions, scope of services, technological sophistication and advantages of our solutions, deployment method, costs of procuring hardware and components and value created for our clients. In addition, we take into consideration the prices of our competitors' offerings and overall market demand. We may grant a credit term to certain customers, which generally ranges from ten days to six months.
Payment terms. We grant credit terms to certain customers on a case-by-case basis, which generally ranges from 10 to 180 days.
Duration. The term of agreement is usually one year.
Ownership. All intellectual property rights of the model and/or agent solutions, services and technical materials provided by us under the agreement belong to us and will not change due to the transfer of product ownership. For agreements involving new R&D requirements, the ownership of intellectual property rights for such R&D results is based on commercial negotiations. No party shall conduct any reverse engineering, decompiling, disassembly or use other methods to obtain the source code and underlying algorithms of the software, hardware and related technologies provided by any other party.
Data use. When necessary, our clients may authorize us to process their data for the purposes agreed upon with them. In circumstances where the data are from or generated by our clients' users, we obtain consent from our clients for the purposes specified in our cooperation agreements. We typically undertake to comply with all applicable laws and regulations in connection with the collection of our clients' data, including but not limited to any laws in respect of intellectual property rights, privacy, data protection and image rights. See "—Data Privacy and Personal Information Protection."
Warranty. We typically provide warranty for one to three years after the delivery of the project. During the warranty period, we also provide free technical support and maintenance to customers.
Compliance. Clients certify that all solutions will be used in compliance with all applicable laws and regulations.
Confidentiality. Each party shall maintain confidentiality of information obtained in relation to the relevant agreement and its contractual terms, and not use information obtained for other purposes.
Individual users are required to agree to our terms of services before using our AI model and agent solutions. In each year/period during 2022, 2023, 2024 and the six months ended June 30, 2025, revenue from our five largest customers accounted for 55.4%, 61.5%, 45.5% and 40.0% of our total revenue, respectively. In each year/period during 2022, 2023, 2024 and the six months ended June 30, 2025, revenue from our largest customer accounted for 15.4%, 14.7%, 19.0% and 11.0% of our total revenue, respectively.
The table below sets forth the details of our five largest customers, to which we provided on-premise deployment of AI model and agent solutions along with technology support and consulting services, in each year during the Track Record Period.
| Customer | Principal Business | Number of Years of Business Relationship | Credit Term | Revenue (RMB'000) | % of Total Revenue | |---|---|---|---|---|---| | **Six months ended June 30, 2025** | | | | | | | A(1) | art-related learning services, live-streaming e-commerce, cultural tourism research and study, smart education services, and AI education | 2 | 10 working days | 20,977 | 11.0% | | B(2) | telecommunications services | 2 | 30 days | 18,980 | 9.9% | | C(3) | telecommunications infrastructure, multimedia telecommunications, and information and communications technology services | 1 | 35 days | 17,927 | 9.4% | | D(4) | information system integration services, technical services, retail of computer software, hardware and auxiliary equipment | 1 | 6 months | 9,623 | 5.0% | | E(5) | information system integration services and software development | 1 | 30 days | 9,026 | 4.7% | | **Total** | | | | **76,533** | **40.0%** | | **Year ended December 31, 2024** | | | | | | | F(6) | technology development and consulting services; computer system and data processing services; development of AI applications | 1 | 10 working days | 59,465 | 19.0% | | G(7) | scientific research and education | 4 | pay upon acceptance | 31,038 | 9.9% | | H(8) | value-added telecommunication services; operation, production and distribution of audio-visual and electronic media | 1 | 10 working days | 26,770 | 8.6% | | I(9) | technology development and technical services related to mobile communications and the internet; wholesale and leasing of mobile phones and internet equipment | 5 | 2 months | 14,297 | 4.6% | | J(10) | technology marketing and application services | 1 | 10 working days | 10,619 | 3.4% | | **Total** | | | | **142,189** | **45.5%** |
| Customer | Principal Business | Number of Years of Business Relationship | Credit Term | Revenue (RMB'000) | % of Total Revenue | |---|---|---|---|---|---| | **Year ended December 31, 2023** | | | | | | | K(11) | technology development, consulting, communication, transfer and promotion services; wholesale of computer software and hardware; information system integration services | 3 | 60 days | 18,244 | 14.6% | | L(12) | technology promotion and computer system services; software development and computer graphic design; organization of exhibition and display activities; corporate planning; business management and consulting; conference services | 1 | 10 working days | 17,536 | 14.1% | | M(13) | development of AI applications; technology development, consulting, communication, transfer and promotion services; computer system services | 2 | 20 working days | 16,809 | 13.5% | | N(14) | production of mobile terminals and communication equipment, electronic components and computer software and hardware | 1 | 15 working days | 14,875 | 11.9% | | O(15) | technology promotion services; computer system services; wholesale of communication equipment, electronic products, computers, software and auxiliary equipment | 1 | 10 working days | 9,163 | 7.4% | | **Total** | | | | **76,627** | **61.5%** | | **Year ended December 31, 2022** | | | | | | | P(16) | development, consulting, service, transfer and training of electronic information and software technology; manufacturing, wholesale and retail of computers, auxiliary equipment and software; integration of computer systems | 1 | 10 days | 8,850 | 15.4% | | Q(17) | sale of robots and industrial automatic control devices; technology development, consulting, communication, transfer and promotion services | 1 | N/A | 7,925 | 13.8% | | R(18) | implementation of national strategies, regulations and policies of science and technology development; drafting of relevant local legislative and regulatory and formulation and implementation of related policy measures | 1 | 60 working days | 6,545 | 11.4% | | S(19) | technology development, consulting, communication, transfer and promotion services; development of AI software | 1 | pay upon acceptance | 5,675 | 9.9% | | T(20) | education and training of talent | 1 | 15 working days | 2,830 | 4.9% | | **Total** | | | | **31,825** | **55.4%** |
(1) Customer A is an IT service company based in Beijing, China, with a registered capital of RMB868.3 million. Customer A is listed on the Shenzhen Stock Exchange.
(2) Customer B is the Ningxia branch of a telecommunications network operation company. Its parent company has a registered capital of RMB213.1 billion and is listed on both the Shanghai Stock Exchange and HKEx.
(3) Customer C is a company that provides telecommunication infrastructure and multimedia telecommunication services, based in Wilayah Persekutuan, Malaysia.
(4) 客户D是一家总部位于中国上海的IT及云计算服务公司，注册资本为800亿美元。 (5) 客户E是一家总部位于中国北京的软件及IT服务公司，注册资本为人民币5,000万元。 (6) 客户F是一家总部位于中国北京的IT服务公司，注册资本为人民币1亿元。客户F是一家在深圳证券交易所上市公司的全资子公司。 (7) 客户G是一家总部位于中国深圳的高等教育机构。 (8) 客户H是一家总部位于中国杭州的IT服务公司，注册资本为人民币1.395亿元。客户H在深圳证券交易所上市。 (9) 客户I是一家总部位于中国北京的电信网络运营公司的分支研究机构。其母公司注册资本为人民币532亿元，并在香港联合交易所上市。 (10) 客户J是一家总部位于中国北京的IT服务公司，注册资本为人民币7,000万元。 (11) 客户K是一家总部位于中国北京的快递服务公司，注册资本为50.5亿美元。客户K是一家在香港联合交易所上市公司的全资子公司。 (12) 客户L是一家总部位于中国北京的招聘服务公司，注册资本为人民币1,000万元。 (13) 客户M是一家总部位于中国北京的AI软件开发公司，注册资本为人民币190万元。 (14) 客户N是一家总部位于中国东莞的电子设备公司，注册资本为人民币6,500万元。 (15) 客户O是一家总部位于中国北京的IT服务公司，注册资本为人民币1,000万元。客户O是一家在上海证券交易所上市公司的全资子公司。 (16) 客户P是一家总部位于中国天津的IT服务公司，注册资本为人民币15亿元。客户P在上海证券交易所上市。 (17) 客户Q是一家总部位于中国杭州的云计算及人工智能技术公司，注册资本为人民币10亿元。 (18) 客户R是一家总部位于中国北京的政府机构，负责制定和推动该市的科技政策。 (19) 客户S是一家总部位于中国北京的IT服务公司，注册资本为人民币9,060万元。 (20) 客户T是一家总部位于中国北京的科技机构，负责为科技专业人员提供培训项目及人才服务。
于业绩记录期间及直至最后实际可行日期，本公司董事、彼等的联系人或本公司任何现有股东（据本公司董事所知，持有本公司超过5%股本的股东）均与业绩记录期间每年五大客户中的任何一家不存在根据《上市规则》须予披露的任何权益。
于业绩记录期间及直至最后实际可行日期，本公司未曾发生任何重大产品责任索赔、产品质量问题及客户投诉。
本公司的供应商主要包括：(i) 计算资源提供商，例如计算硬件及计算服务提供商；(ii) 硬件设备供应商，包括服务器、存储设备及网络设备供应商；(iii) 研发支持服务提供商，例如数据清洗及大模型评估服务提供商；以及 (iv) 营销服务提供商。
We have established a set of internal measures on selection of suppliers. We take into account various factors in selecting our suppliers, which primarily include product quality, price, service quality, qualifications and credentials. When engaging suppliers, our legal and finance department is primarily responsible for reviewing the procurement agreements in accordance with our procurement protocols. We usually enter into framework agreements with our computing resource suppliers, particularly cloud computing service providers, the major terms of which are set out below.
• Service scope. Under the framework agreements, the suppliers generally provide cloud computing service that meet our needs for computing services, along with necessary computing infrastructure and technical support services.
• Duration. The term of agreement generally range from one to four years and is renewable upon mutual agreement.
• Payment terms. We typically settle the payments in installments in accordance with the performance progress of the agreements. Generally, we may be granted a credit term of up to 40 days.
• Termination. Typically, either party may terminate the agreements with written notice and the other party's written consent or in the event of the opposing party's material violation of laws and regulations or material breach of contracts. In such cases, the terminating party is liable for any losses suffered by the other party.
During the Track Record Period, we had not experienced any significant fluctuation in prices set by our suppliers, material breach of contract on the part of our suppliers or delay in delivery of our orders from our suppliers.
In each year/period during 2022, 2023, 2024 and the six months ended June 30, 2025, purchases from our five largest suppliers accounted for 54.5%, 53.6%, 47.3% and 50.2% of our total purchases, respectively. In each year/period during 2022, 2023, 2024 and the six months ended June 30, 2025, purchases from our largest supplier accounted for 33.1%, 16.4%, 15.6% and 13.4% of our total purchases, respectively. The table below sets forth the details of our five largest suppliers in each year during the Track Record Period.
| Supplier | Principal Business | Products/Services Procured | Number of Years of Business Relationship | Credit Term | Purchase Amount (RMB'000) | % of Total Purchases | |---|---|---|---|---|---|---| | **Six months ended June 30, 2025** | | | | | | | | A(1) | Software development, technical services and development and information system integration services | Computing service | 2 | 30 days | 207,921 | 13.4% | | B(2) | Software development, technical services and development and information system integration services | Computing service | 1 | 10 working days | 157,052 | 10.1% | | C(3) | Software development; technology service and development; information system integration services | Computing service | 1 | Prepayments | 152,430 | 9.8% | | D(4) | Technical services and development and software development | Computing service | 4 | 5–15 working days | 139,442 | 9.0% |
| Supplier | Principal Business | Products/Services Procured | Number of Years of Business Relationship | Credit Term | Purchase Amount (RMB'000) | % of Total Purchases | |---|---|---|---|---|---|---|
E(5) | technical services and development and software development | Computing service | 3 | N/A | 123,474 | 7.9% |
| Supplier | Principal Business | Products/Services Procured | Number of Years of Business Relationship | Credit Term | Purchase Amount (RMB'000) | % of Total Purchases | |---|---|---|---|---|---|---| | E | technical service and development; software development | computing services | 3 | N/A | 394,771 | 15.6% | | C | software development; technology service and development; information system integration services | computing services | 1 | prepayments | 379,829 | 15.0% | | F(6) | technical service and development; sale of computer software and hardware | computing hardware | 2 | 7 days | 191,150 | 7.6% | | G(7) | technical service and development; sale of construction materials; data processing services | computing services | 1 | 10 working days | 119,834 | 4.7% | | D | technical service and development; software development | computing services | 4 | 5 - 15 working days | 112,394 | 4.4% |
| Supplier | Principal Business | Products/Services Procured | Number of Years of Business Relationship | Credit Term | Purchase Amount (RMB'000) | % of Total Purchases | |---|---|---|---|---|---|---| | E | technical service and development; software development | computing services | 3 | N/A | 158,890 | 16.4% | | H(8) | technical service and development; software development | computing services and hardware | 2 | 2 - 5 days | 124,471 | 12.8% | | I(9) | internet information services; labor dispatch services | computing hardware | 2 | prepayments | 95,841 | 9.9% | | J(10) | software and technology consulting services | computing hardware | 3 | 15 - 40 days | 83,869 | 8.6% | | K(11) | technical service and development; software development; data processing and storage support | computing hardware | 2 | prepayments | 57,345 | 5.9% |
| Supplier | Principal Business | Products/Services Procured | Number of Years of Business Relationship | Credit Term | Purchase Amount (RMB'000) | % of Total Purchases | |---|---|---|---|---|---|---| | J | software and IT services | computing hardware | 3 | 10 days | 25,204 | 33.1% | | L(12) | data processing and storage support services; technical service and development | computing services | 3 | 5 days | 5,834 | 7.7% | | D | technical service and development; software development | computing services | 4 | 5 days | 5,028 | 6.6% | | E | technical service and development; software development | computing services | 3 | N/A | 2,804 | 3.7% | | M(13) | technical service and development; software development | computing services | 3 | prepayments | 2,548 | 3.4% |
Notes: (1) Supplier A is an IT service company based in Beijing, China, with a registered capital of RMB10.0 million. (2) Supplier B is an IT service company based in Beijing, China, with a registered capital of RMB1.6 billion. (3) Supplier C is an IT service company based in Jiaxing, China, with a registered capital of RMB721.1 million. (4) Supplier D is an IT service company based in Beijing, China, with a registered capital of RMB58.2 million. (5) Supplier E is an IT service company based in Beijing, China, with a registered capital of RMB1.0 billion. (6) Supplier F is an IT service company based in Beijing, China, with a registered capital of RMB10.0 million. (7) Supplier G is an IT service and data processing service company based in Beijing, China, with a registered capital of RMB673.5 million. (8) Supplier H is an IT service company based in Beijing, China, with a registered capital of RMB47.8 million. Supplier H is listed on Shanghai Stock Exchange. (9) Supplier I is an internet information services company based in Chongqing, China, with a registered capital of RMB50.0 million. (10) Supplier J is an IT service company based in Beijing, China, with a registered capital of RMB700.0 million. Supplier J is a wholly owned subsidiary of a company listed on Shanghai Stock Exchange. (11) Supplier K is an IT service and data processing and collection company based in Beijing, China, with a registered capital of RMB14.0 million. (12) Supplier L is an IT service and data processing and collection company based in Beijing, China, with a registered capital of RMB200.0 million. (13) Supplier M is an IT service company based in Beijing, China, with a registered capital of RMB92.3 million.
During the Track Record Period and up to the Latest Practicable Date, to the best knowledge of our Directors, none of our Directors, their associates or any of our current Shareholders (who, to the knowledge of our Directors, own more than 5% of our share capital) had any interest in our five largest suppliers in each year during the Track Record Period that are required to be disclosed under the Listing Rules.
Customer A, our largest customer in the six months ended June 30, 2025, was also our supplier in 2024 and 2025. We procured database and IP authorization services from Customer A in 2024 and 2025 and provided on-premise deployment to Customer A in the same years. In 2022, 2023, 2024 and the six months ended June 30, 2025, our purchases from Customer A amounted to nil, nil, RMB12.9 million and RMB11.8 million, accounting for nil, nil, 0.5% and 0.8% of our total purchases in the same periods, respectively, and our sales to Customer A amounted to nil, nil, RMB3.2 million and RMB21.0 million, accounting for nil, nil, 1.0% and 11.0% of our revenue during the same periods, respectively.
Supplier E, our largest supplier in 2023 and 2024, was also our customer in 2024 and 2025. Supplier E is a cloud service provider, and a subsidiary of a large internet technology company in China. During the Track Record Period, we procured computing services from Supplier E. We also provided on-premise and cloud-based deployment to Supplier E in 2024. In 2022, 2023, 2024 and the six months ended June 30, 2025, our purchases from Supplier E amounted to RMB2.8 million, RMB158.9 million, RMB394.8 million and RMB123.5 million, accounting for 3.7%, 16.4%, 15.6% and 8.2% of our total purchases in the same periods, respectively, and our sales to Supplier A amounted to nil, nil, RMB2.4 million and RMB4.7 million, accounting for nil, nil, 0.8% and 2.5% of our revenue during the same periods, respectively.
客户G在业绩记录期间亦为本公司供应商。客户G为中国一所公立大学。在业绩记录期间，本公司向客户G提供本地部署服务，并就与该大学的合作向其采购研究及开发服务。于2022年、2023年、2024年及截至2025年6月30日止六个月，本公司向客户G的销售额分别为人民币160万元、人民币40万元、人民币3,100万元及零，分别占同期本公司收入的2.8%、0.3%、9.9%及零；本公司向客户G的采购额分别为人民币70万元、人民币140万元、人民币170万元及人民币160万元，分别占同期本公司总采购额的0.9%、0.1%、0.1%及0.1%。
客户I在业绩记录期间亦为本公司供应商。客户I为中国一家在线招聘平台。在业绩记录期间，本公司向客户I提供本地部署服务，并向其采购招聘服务。于2022年、2023年、2024年及截至2025年6月30日止六个月，本公司向客户I的销售额分别为零、人民币1,750万元、零及零，分别占同期本公司收入的零、14.1%、零及零；本公司向客户I的采购额分别为人民币2.4万元、人民币10万元、人民币20万元及人民币90万元，分别占同期本公司总采购额的0.03%、0.01%、0.01%及0.1%。
客户P╱供应商L亦为同年本公司五大供应商之一。客户P╱供应商L为中国一家超级计算机制造商。在业绩记录期间，本公司向客户P╱供应商L提供本地部署服务，并向其采购算力服务。于2022年、2023年、2024年及截至2025年6月30日止六个月，本公司向客户P╱供应商L的销售额分别为人民币890万元、零、零及零，分别占同期本公司收入的15.4%、零、零及零；本公司向客户P╱供应商L的采购额分别为人民币580万元、人民币520万元、零及零，分别占同期本公司总采购额的7.7%、0.5%、零及零。
客户Q在业绩记录期间亦为本公司供应商。客户Q为中国一家云计算公司。在业绩记录期间，本公司向客户Q提供本地部署服务，并向其采购算力服务。于2022年、2023年、2024年及截至2025年6月30日止六个月，本公司向客户Q的销售额分别为人民币790万元、零、零及零，分别占同期本公司收入的13.8%、零、零及零；本公司向客户Q的采购额分别为人民币98万元、人民币2,720万元、人民币7,660万元及人民币9,450万元，分别占同期本公司总采购额的1.3%、2.8%、3.0%及6.3%。
本公司董事认为，业绩记录期间与上述兼具客户及供应商身份的各方之交易均在日常业务过程中以公平交易原则按正常商业条款进行。
作为大模型公司，本公司在通常称为大语言模型（LLM）市场的领域内运营，该市场为人工智能市场的细分市场。LLM市场竞争激烈。根据Frost & Sullivan的资料，LLM供应商的竞争因素包括：（i）技术能力，如自主研发的LLM预训练框架以及模型定制与优化能力；（ii）灵活的商业模式及交付策略；（iii）生态系统构建能力；以及（iv）具有深厚技术背景及丰富经验的人才。本公司同时与中国境内及国际市场上的独立及非独立LLM供应商展开竞争。未来亦可能面临新进入者带来的竞争压力。对本公司而言，重要的主要竞争因素包括服务产品的范围、性能及安全性、用户体验、本公司的研发能力及人才储备。有关本公司所处行业竞争格局的更多详情，请参阅"行业概览"。有关本公司行业竞争力的相关风险，请参阅"风险因素——人工智能行业以持续变革为特征。若本公司无法升级、增强或创新其技术及服务，本公司的业务、经营业绩、财务状况及前景可能受到不利影响"。
The following table sets out a summary of the major awards and recognitions we had received as of the Latest Practicable Date.
| Year | Awards or Recognition | Issuing Authority | |------|-----------------------|-------------------| | 2024 | 2024 World Computing Conference Special Exhibition—Outstanding Achievement (2024世界計算大會專題展優秀成果) | World Computing Conference (世界計算大會) | | 2024 | Scientific and Technological Progress Award—First Prize (科技進步一等獎) | Chinese Institute of Electronics (中國電子學會) | | 2023 | Digital Economy Industry Innovation Achievement (數字經濟產業創新成果) | Global Digital Economy Conference | | 2023 | Artificial Intelligence Key Technology and Application Evaluation Key Laboratory (人工智能關鍵技術和應用評測重點實驗室) | China Artificial Intelligence Industry Alliance (中國人工智能產業發展聯盟) | | 2023 | The 25th China Hi-Tech Fair—Excellent Product Award (第二十五届中國國際高新技術成果交易會優秀產品獎) | China Hi-Tech Fair Organizing Committee (中國國際高新技術成果交易組委會) | | 2021 | State Scientific and Technological Progress Award—Second Prize (國家科學技術進步獎二等獎) | State Council | | 2021 | Beijing Invention Patent Award—First Prize (北京市發明專利一等獎) | The People's Government of Beijing Municipality | | 2021 | 2021 "Sci-Tech Innovation China" List—Rising Enterprise (2021年"科創中國"榜單新銳企業) | China Association for Science and Technology (中國科學技術協會) | | 2021 | Zhongguancun High-Tech Enterprise (中關村高新技術企業) | Zhongguancun Science Park Administrative Committee (中關村科技園區管理委員會) | | 2020 | SIGKDD Test of Time Award | Association for Computing Machinery | | 2020 | 2020 Outstanding Scientific and Technological Achievement Award (2020年度優秀科技成果獎) | Chinese Association for Artificial Intelligence (中國人工智能學會) | | 2020 | Zhongguancun Golden Seed Enterprise (中關村金種子企業) | Zhongguancun Science Park Administrative Committee (中關村科技園區管理委員會) |
During the Track Record Period and up to the Latest Practicable Date, we had obtained all material licenses, permits, approvals and certificates necessary to conduct our actual business operations from the
relevant government authorities, and such licenses, permits, approvals and certificates remained in full effect. The following table sets out the details of our material licenses and permits as of the Latest Practicable Date.
| License/Permit | Entity Holding the License/Permit | Issuing Authority | Grant Date | Expiration Date | |---|---|---|---|---| | Value-added Telecommunications Business Operation License (for provision of internet information services) | Our Company | MIIT | May 30, 2025 | April 16, 2029 | | Value-added Telecommunications Business Operation License (for provision of internet information services) | Beijing Knowledge Huixing | MIIT | January 26, 2025 | January 26, 2030 |
As of June 30, 2025, we had 883 employees, substantially all of whom were based in China. The following table sets forth the number of our employees by function.
| Function | Number of Employees | Percentage | |---|---|---| | Research and Development | 657 | 74.4% | | Sales and Marketing | 145 | 16.4% | | Management and administrative | 81 | 9.2% | | Total | 883 | 100.0% |
Our success depends on our ability to attract, retain and motivate qualified personnel. We adopt high standards and strict procedures in our recruitment, including campus recruitment, online recruitment, internal referral and third-party recruiters, to satisfy our demands for different types of talents. We provide regular and specialized training tailored to the needs of our employees in different departments. New employees will receive pre-job training and general training.
We believe we offer our employees competitive compensation packages. In addition, we regularly evaluate the performance of our employees and reward those who perform well with higher compensation or promotion. We enter into standard contracts and agreements regarding confidentiality, intellectual property, employment, commercial ethics and non-compete with our executive officers and full-time employees. During the Track Record Period, we made contributions to social insurance and housing provident funds in compliance with applicable PRC laws and regulations in all material respects.
None of our employees are currently represented by labor unions. We believe that we maintain a good working relationship with our employees, and we have not experienced any material labor disputes or any difficulty in recruiting staff for our operations during the Track Record Period.
We consider our insurance coverage to be adequate as we have in place all the mandatory insurance policies required by PRC laws and regulations and in accordance with the commercial practices in our
industry. In line with general market practice, we do not maintain any business interruption insurance or product liability insurance, which are not mandatory under PRC laws. We do not maintain key man life insurance, insurance policies covering damages to our network infrastructures or information technology systems or any insurance policies for our properties. See "Risk Factors—Our insurance coverage may not be sufficient to cover all losses or potential claims by our customers which would affect our business, results of operations, financial condition and prospects."
根据市场惯例，我们不持有任何营业中断保险或产品责任险，该等保险在中国法律下并非强制要求。我们不持有关键人寿保险、涵盖网络基础设施或信息技术系统损害的保险保单，亦不持有任何财产保险保单。请参阅"风险因素——我们的保险覆盖范围可能不足以覆盖所有损失或客户的潜在索赔，从而影响我们的业务、经营业绩、财务状况和前景。"
As of the Latest Practicable Date, we did not own any real property. As of June 30, 2025, we had no single property with a carrying amount of 15% or more of our total assets, and on this basis, we are not required by Rule 5.01A of the Listing Rules to include in this prospectus any valuation report. Pursuant to section 6(2) of the Companies Ordinance (Exemption of Companies and Prospectuses from Compliance with Provisions) Notice, this prospectus is exempted from compliance with the requirements of section 342(1)(b) of the Companies (Winding Up and Miscellaneous Provisions) Ordinance in relation to paragraph 34(2) of the Third Schedule to the Companies (Winding Up and Miscellaneous Provisions) Ordinance, which requires a valuation report with respect to all of our interests in land or buildings.
于最后实际可行日期，我们并不拥有任何不动产。于2025年6月30日，我们没有任何单项物业的账面价值达到或超过我们总资产的15%，因此，根据《上市规则》第5.01A条，我们无需在本招股说明书中纳入任何估值报告。根据《公司条例》（豁免公司及招股书遵从条文）公告第6(2)条，本招股说明书获豁免遵从《公司（清盘及杂项条文）条例》第342(1)(b)条的规定，该规定涉及《公司（清盘及杂项条文）条例》附表三第34(2)段，该段要求就我们在土地或建筑物中的所有权益提供估值报告。
As of the Latest Practicable Date, we leased 18 properties with a GFA of 25,860 sq.m. in the PRC as our offices from Independent Third Party. As of the Latest Practicable Date, the lessors of five out of our 18 leased premises used as our offices had not provided copies of the real property title certificates to us despite our continuous effort in requesting such documents. As advised by our PRC Legal Advisors, if the relevant lessor has no right to lease the leased property and a third party other than the parties to the relevant lease contracts have legal title to such leased property, such third party may claim that the relevant lease contracts are null and void or have no effect thereto, or request us to cease our use and move out of such leased property. However, considering that (i) as of the Latest Practicable Date, we had not received any notices requiring us to cease our use or move out of such leased property, (ii) according to Frost & Sullivan, there are abundant unoccupied properties available for lease at similar costs and we believe we would be able to relocate our facilities to a different site relatively easily if we are required by third parties, and (iii) in accordance with the relevant provisions of the PRC Civil Code, if we are unable to use or accrue proceeds from the leased property due to any claim by a third person, we may request reduction of rent or refuse to pay rent, our Directors are of the view that such incidents will not have a material adverse impact on our continuous operation, financial condition and results of operations. In the event that we are required to relocate from the leased properties, we expect that the relocation cost will be minimal for each leased property. In light of the foregoing, our Directors are of the view that if we are required to relocate, it will not have a material disruption on our business operations. As advised by our PRC Legal Advisors, as of the Latest Practicable Date we were not in violations of the fire safety requirements pursuant to the relevant laws and regulations in the PRC in all material respects. Accordingly, our Directors are of the view that our leased properties comply with the fire safety requirements in all material respects pursuant to the relevant laws and regulations. Based on the foregoing, nothing material has come to the attention of the Sole Sponsor that would cause them to cast doubt on the Directors' view relating to the fire safety issue.
于最后实际可行日期，我们向独立第三方租赁了中国境内共18处物业作为办公室，建筑面积合计25,860平方米。于最后实际可行日期，尽管我们持续索取相关文件，我们作为办公室使用的18处租赁物业中，有5处物业的出租人未向我们提供不动产权证书副本。根据我们的中国法律顾问的意见，若相关出租人无权出租该等租赁物业，且非相关租赁合同当事方的第三方对该等租赁物业拥有合法产权，则该第三方可能主张相关租赁合同无效或对其不产生效力，或要求我们停止使用并迁出该等租赁物业。然而，考虑到以下因素：(i) 于最后实际可行日期，我们尚未收到任何要求我们停止使用或迁出该等租赁物业的通知；(ii) 根据弗若斯特沙利文的资料，市场上有大量可供租赁且租金相近的空置物业，我们认为，如被第三方要求迁出，我们能够相对容易地将设施迁至其他地点；以及(iii) 根据《中华人民共和国民法典》的相关规定，若因第三方主张权利导致我们无法使用租赁物业或从中获得收益，我们可要求减少租金或拒绝支付租金，我们的董事认为，上述情况不会对我们的持续经营、财务状况及经营业绩产生重大不利影响。如我们被要求迁出租赁物业，预计每处租赁物业的迁移费用将微乎其微。有鉴于此，我们的董事认为，如我们被要求迁移，不会对我们的业务运营造成重大干扰。根据我们的中国法律顾问的意见，于最后实际可行日期，我们在所有重大方面均未违反中国相关法律法规规定的消防安全要求。因此，我们的董事认为，我们的租赁物业在所有重大方面均符合相关法律法规规定的消防安全要求。基于上述情况，独立保荐人未发现任何重大事项，足以令其对董事就消防安全问题所持立场产生怀疑。
As of the Latest Practicable Date, 14 out of our 18 leased properties used as our offices had not been registered and filed with relevant land and real estate administration bureaus in the PRC. The failure to register such lease agreements is due to factors beyond our control including, among other things, the lessors' willingness to cooperate in the registration process and provision of relevant documents for registration is necessary. As advised by our PRC Legal Advisors, failure to complete the registration and filing of lease agreements will not affect the validity of such lease agreements nor the lawful and effective use of leased properties pursuant to the lease agreements. However, the relevant authorities may require us to rectify such noncompliance within a prescribed period and we may be subject to a fine ranging from RMB1,000 to RMB10,000 for each of such properties if we fail to rectify such non-compliance within the prescribed period. During the Track Record Period, we have not been subject to any administrative penalties imposed by the competent authorities for failure to complete the registration and filing of the lease
于最后实际可行日期，我们作为办公室使用的18处租赁物业中，有14处尚未在中国相关土地及房地产管理局完成登记备案。未能完成上述租赁合同登记备案，系由于若干超出我们控制范围的因素所致，包括但不限于出租人配合登记程序的意愿，以及提供登记所需相关文件的必要性。根据我们的中国法律顾问的意见，未完成租赁合同的登记备案不会影响该等租赁合同的效力，亦不会影响依据租赁合同对租赁物业的合法有效使用。然而，相关主管部门可能要求我们在规定期限内纠正上述违规情况，若我们未能在规定期限内完成整改，每处物业可能面临人民币1,000元至人民币10,000元的罚款。在业绩记录期内，我们未曾因未完成租赁合同登记备案而遭受主管部门处以行政处罚。
agreements. Based on the above, we believe that the failure to register and file the leased properties will not have any material adverse effect on our operation and financial condition. See "Risk Factors—Our legal right to some leased properties may be challenged."
On January 16, 2025, our Company and nine of our subsidiaries (together, "Listed Entities") were added to the Entity List administered by the BIS. The addition of the Listed Entities to the Entity List (the "Entity List Addition") restricts our ability to purchase or otherwise access certain goods, software and technology (collectively, "items") that are subject to the EAR without a license from the BIS. Items subject to the EAR include, among other things, U.S.-origin items, as well as non-U.S. items that contain more than a de minimis portion of U.S.-origin controlled content, and certain items of non-U.S.-origin that are the direct product of certain U.S.-origin controlled software or technology. For further information and the potential consequences for violating U.S. export controls, see "Regulatory Overview—U.S. Export Control Laws and Regulations."
As advised by our International Sanctions Counsel, the Entity List restrictions do not apply to non-listed entities in our Group that are legally distinct from the Listed Entities (the "Non-listed Entities"). That is, BIS has explicitly advised that "the licensing and other obligations imposed on an entity by virtue of being listed on the Entity List do not per se apply to its subsidiaries, sister companies, or other legally distinct affiliates that are not listed on the Entity List." However, a Non-listed Entity (or any other person) may not act as an agent, a front, or a shell company for a Listed Entity in order to facilitate transactions that would not otherwise be permissible with the Listed Entity.
Our International Sanctions Counsel further confirmed that our suppliers can continue to provide items subject to the EAR to the Non-listed Entities that are legally distinct from the Listed Entities so long as (i) those items are not exported, reexported or transferred, directly or indirectly, to the Listed Entities; (ii) those items are not diverted to the Listed Entities by the Non-listed Entities; and (iii) the Listed Entities do not otherwise serve as the purchaser, intermediate consignee, ultimate consignee or end-user of the items. In addition, our International Sanctions Counsel also confirmed that (i) our suppliers can still continue to provide items that are not subject to the EAR to the Listed Entities; and (ii) our customers can lawfully purchase items subject to the EAR from the Listed Entities so long as those items were obtained lawfully by the Listed Entities (e.g., obtained prior to the Entity List Addition or obtained pursuant to a license from the BIS).
During the Track Record Period and prior to the Entity List Addition, we purchased certain computing hardware containing U.S.-origin AI chips subject to the EAR in compliance with the relevant export control rules and regulations and we have not procured any items subject to the EAR since August 2023. We did not rely on any specific EAR items during the Track Record Period. After the Entity List Addition, we are unable to purchase AI chips subject to the EAR and have not procured any AI chips subject to the EAR. However, this has not had an adverse impact on our business, as we did not plan to develop our own AI data center. In fact, we have not procured any AI chips, whether subject to the EAR or not, since the Entity List Addition. Instead, we procure computing resources from reputable cloud computing service providers in China, whose services were not subject to the EAR as of the Latest Practicable Date.
In light of the above, the Entity List Addition has not had, and our Directors are of the view that (assuming there is no expansion of the EAR restrictions or the scope of the Entity List Addition) it will not have in the near future, any material adverse impact on our business and financial performance. In addition, as of the Latest Practicable Date, none of our material investors, customers or suppliers had withdrawn their investment or ceased doing business with us due to the Entity List Addition.
为应对《出口管理条例》（EAR）下的相关问题，我们已采取一切合理步骤，依照商务部工业和安全局（BIS）发布的《出口合规指引——有效出口合规计划的要素》，建立了一套由专职合规资源支持的出口管制合规计划。出于审慎考虑，我们已将该出口管制合规计划的适用范围扩展至整个集团。我们的法律部门负责监督出口管制合规事宜。为确保出口管制合规计划的可持续性，我们还专门成立了出口管制合规管理委员会，成员包括我们的总经理以及来自法律部门、财务部门、信息系统管理部门和采购部门的人员，作为员工就出口管制相关问题进行咨询的中央联络点。我们的出口管制合规计划涵盖了BIS发布的有效出口合规计划的全部八项要素。
(i) 管理层意识与承诺。我们将向全体员工发布年度声明，表明我们对出口管制合规的支持、违反EAR的后果以及相关事项的指定联络人。所有员工均须签署确认书，确认其遵守我们的出口管制合规政策和程序。
b) (i) 通过获取相关出口管制信息及出口管制分类编号（"ECCN"）（如适用），对从供应商处采购的物项进行识别和分类；以及 (ii) 通过管辖权分析和ECCN分类，对我们自行创建的物项进行识别和分类；以及
c) 依据美国、英国、欧盟、联合国及澳大利亚相关制度下的限制名单和受制裁方名单，对我们的业务伙伴（包括客户和供应商）进行筛查。此外，我们的政策规定，非上市主体不得向上市主体提供EAR管制物项。
(iii) 出口授权。我们将对采购物项和可销售物项进行监控并定期审查（内部审查，必要时亦委托外部顾问进行），以确定其是否受EAR约束，如受约束，则进一步确定相关分类及许可要求。
(iv) 记录保留。我们对客户身份记录进行更新、核实和保存。此外，若对此前收集的信息存在疑问，或发现或怀疑存在任何异常活动，我们将进行核实并维护最新记录。
(v) 培训。我们将以多种形式（包括面授和在线培训）对全体员工开展出口合规培训。
(vi) 审计。我们将就出口管制合规计划的执行情况开展内部审计。
(vii) 出口违规及纠正措施。我们的出口管制合规管理政策规定了就实际或疑似违反出口管制法律法规的情况进行举报的渠道及调查程序。出口管制合规管理委员会将对任何潜在问题进行调查并采取纠正措施。所有举报均以严格保密的方式处理。
(viii) 出口管制合规管理政策。我们已采纳一项适用于集团所有实体的出口管制合规管理政策。该政策明确规定了与《出口管理条例》（EAR）相关的规则和法规。我们将定期审查和更新该政策，以确保其反映《出口管理条例》及公司运营的任何变化。
除上述出口管制合规计划外，我们将继续聘请外部法律顾问和咨询顾问，持续对我们的出口管制合规计划进行审查并提供建议。因此，我们的国际制裁法律顾问已审查了我们的出口管制合规政策，并建议：在上述合规政策得到落实和执行的前提下，我们的出口管制合规措施为我们识别和降低与实体清单相关的任何重大风险提供了合理充分且有效的内部控制框架。
基于上述内容，我们的董事认为，我们的出口管制合规政策提供了合理充分且有效的内部控制框架，以识别和降低与被列入实体清单相关的任何重大风险，且建立出口管制合规计划将有利于未来申请将被列名实体从实体清单中移除。截至最后实际可行日期，我们尚未向工业和安全局（BIS）提交移除申请。然而，我们仍致力于将被列名实体从实体清单中移除，并拟建立出口管制合规计划的合规记录。我们将在适当时机尽快推进移除申请。
基于上述内容，独家保荐人未发现任何重大事项，使其对董事就本集团出口管制合规政策在识别和降低与实体清单相关的重大风险方面的有效性和充分性所持观点产生怀疑。
自2025年2月起，美中之间的关税战升级，美国对来自中国的进口商品加征大幅额外关税，此后中国亦采取了进一步措施及反制措施。然而，由于（a）在业绩记录期间及截至最后实际可行日期，我们未曾向美国出口任何产品；（b）在业绩记录期间及截至最后实际可行日期，我们未曾直接从美国采购任何原材料或零部件；以及（c）自2025年1月被列入实体清单以来，由于未获得适当授权，我们已无法从美国进口任何物项，因此截至最后实际可行日期，美国加征关税及中国采取的反制措施对我们业务运营的影响有限。截至最后实际可行日期，我们未收到任何客户和供应商因美国近期关税上调而提出的服务取消、重大价格调整或暂停交货的通知。因此，我们未因关税上调而从客户和供应商方面遭受任何重大间接影响。
我们须遵守多项社会、健康、安全及环境法律法规，且我们的运营定期接受地方政府主管部门的检查。我们相信，我们已制定充分的政策以确保符合所有社会、健康、安全及环境保护法规。尤其是，我们认为我们的持续增长有赖于将社会价值融入我们的业务之中。我们高度重视环境、社会及管治（"ESG"）事宜，并致力于推动企业社会责任及环境保护。
Given the nature of our business, we do not operate any production facilities or otherwise impose any material threats to the environment or the climate. Therefore, we are not subject to significant environmental or climate-related risks. Nonetheless, we have made significant efforts toward environmental protection, climate change and sustainability.
Our direct energy consumption primarily consist of electricity usage in our offices and our resource consumption primarily consist of water source from municipal water supply. For years ended December 31, 2022, 2023 and 2024 and the six months ended June 30, 2025, our Company's energy and resource consumption expenses for our offices were RMB0.1 million, RMB0.4 million, RMB0.8 million and RMB0.4 million, respectively. The following table sets forth our Company's energy and resource consumption in 2024 and the six months ended June 30, 2025.
| | Year Ended December 31, 2024 | Six Months Ended June 30, 2025 | |---|---|---| | Electricity (MWh) | 82 | 44 | | Electricity intensity (MWh per million RMB of revenue) | 0.3 | 0.2 | | Water consumption (tons) | 1,456 | 1,084 | | Water consumption intensity (tons per million RMB of revenue) | 4.7 | 5.7 |
Notes: (1) We gradually relocated our offices in 2023. Prior to this relocation, we only maintained data on consumption expenses, but not on consumption volume. Given the significant differences in the office premises and associated facilities before and after the relocation, the consumption volume from our previous offices is not comparable to that after the relocation and is not relevant for setting our future targets. Therefore, we have not retrospectively collected consumption volume data for the period prior to 2024. (2) Our water consumption increased primarily due to the increase in our employees, which was in line with our business expansion.
As we optimize our models and R&D capabilities, we are able to utilize the hardware resources efficiently, thereby achieving the same results with less computing power. We also strive to empower enterprises to improve operating efficiency and achieve carbon neutrality via our services. Moreover, we believe that as AI and its application develop and mature, AI can solve problems more efficiently than traditional technologies, ultimately resulting in less energy consumption.
In addition, we make efforts to save electricity energy in our daily office life as a part of our corporate culture. We operate most of our businesses digitally and utilize cloud-based services to reduce consumption of paper and renovate our offices with environmental-friendly materials, in an effort to keep our carbon consumption low. For example, we paint our offices in light colors and make sure the curtains are open during daytime to reduce use of lighting. We also arrange our office superintendents to inspect the building regularly and turn down the lights in empty rooms. We have imposed office policies for air conditioning with considerations to season, weather and use scenario to manage the energy consumption of air conditioning and displayed notices adjacent to the air conditioners to remind our employees of the environmental impact.
We aim to reduce our Company's electricity intensity by 5% by 2027 compared to the level in the year ended December 31, 2024. We plan to achieve this goal by implementing stricter controls over the use of high-wattage appliance.
根据全球报告倡议组织（GRI）及温室气体协议（GHG Protocol）的相关标准，由于本公司不从事任何生产活动，业务运营中亦无直接排放，故不涉及范围一温室气体排放。本公司范围二温室气体排放仅来源于办公室购买电力的消耗。下表列示本公司于2024年度及截至2025年6月30日止六个月的温室气体排放情况。
| | 截至2024年12月31日止年度 | 截至2025年6月30日止六个月 | |---|---|---| | 范围二温室气体排放量（吨二氧化碳当量） | 438 | 237 | | 范围二温室气体排放强度（吨二氧化碳当量／百万元人民币收入） | 1.4 | 1.2 |
本公司目标是与截至2024年12月31日止年度的水平相比，于2027年前将范围二温室气体排放强度降低5%。本公司计划通过以能效更高的设备替换部分现有设备来实现这一目标。
本公司目前正在评估范围三排放的适用性，识别相关部门及外部利益相关方，并评估相关数据收集的可行性。展望未来，本公司致力于加深对范围三温室气体排放的了解，并按照联交所发布的相关指引作出所需披露。
由于本公司租用办公室及部分办公设备（如打印机），产生的危险废弃物通常可忽略不计。相关物业管理公司及设备提供商负责收集和处置可能产生的任何危险废弃物。本公司的无害废弃物主要由员工在办公室日常活动中产生的一般生活垃圾构成。此类废弃物的产生不可避免，且对本集团不构成任何重大直接财务或运营影响。下表列示本公司于追踪记录期间的固体废弃物排放情况。
| | 截至2022年12月31日止年度 | 截至2023年12月31日止年度 | 截至2024年12月31日止年度 | 截至2025年6月30日止六个月 | |---|---|---|---|---| | 无害废弃物（吨） | 20 | 45 | 70 | 35 | | 无害废弃物强度（吨／百万元人民币收入） | 0.3 | 0.4 | 0.2 | 0.2 |
由于本公司在租用的办公场所内运营，所有无害废弃物均由物业管理公司负责收集及处置。本公司目标是与截至2024年12月31日止年度的水平相比，于2027年前将无害废弃物强度降低2%。
本公司认为，均衡的生活方式对于保持良好的工作心态至关重要。因此，本公司鼓励员工通过参与体育及文娱活动来保持良好的身心健康。在安全政策方面，本公司要求所有员工遵守安全规定并接受安全培训，内容包括消防演练及有关疏散程序和其他消防安全措施的视频培训。
We nurtured a friendly and inspirational corporate culture that we believe is attractive to the talented scientists who are keen on our success, and we invest heavily in training and retaining them. We provide adequate resources to help them succeed, including easy access to our rich internal resources for training and studying, our invaluable industry-related insights and opportunities to work in an inclusive community with our similar-minded scientists. We continually increase our investment in employee training with the aim of enhancing both professional competence and overall capabilities. Since January 1, 2024, we have accumulated over 360 hours in new staff induction session, enabling new employees to quickly blend in the corporate culture, master essential job skills and clarify their career development paths. Meanwhile, we have organized cultural training programs for senior management, with a cumulative training duration of over 330 hours from January 1, 2023 to June 30, 2025. In addition, we continue to optimize our leadership development training program, concentrating on key topics such as objective management, performance management, empowerment and motivation, effective collaboration and case studies. During the Track Record Period, we provided nearly 1,100 hours of training sessions to our mid-level and senior managers under this program, supporting their comprehensive improvement in areas including team building, decision-making and execution capabilities.
We have also played a role in advancing social inclusion and accessibility by developing the digital sign language broadcaster, an AI-powered digital figure that delivers real-time, professional sign language interpretation of event news for people with hearing impairments, for 2022 Winter Olympics. This innovative effort not only supports equal participation in society and improves quality of life for people with disabilities, but also promotes the standardization of national sign language and sets a benchmark for the human-centered application of technology.
We have established a systematic and multi-layered governance framework to promote algorithmic transparency to regulators and responsible AI practices. At the model development and system design stage, we incorporate "explainability" as a core design principle and adopts model architectures that can provide clear decision paths so that our AI models and matching algorithms are developed in a way that allows us to trace and explain how key outputs are generated. We maintain comprehensive documentation, including system architecture descriptions, data processing flows and decision logic, so that regulators and other stakeholders, including those without a technical background, can understand the basic principles of the operations of our AI models. At the model deployment and operational stage, in compliance with relevant regulations, we have implemented transparency management measures such as periodically publishing technical reports or academic papers on our models, proactively disclosing material algorithm updates and maintaining user inquiry and feedback channels. These measures also help enhance public and user trust in the reliability and integrity of our AI models. We actively participates in open-source communities and have open-sourced most of our self-developed models, which further enhances our algorithmic transparency. In addition, as part of our data governance framework, we incorporate fairness and bias-mitigation considerations throughout the lifecycle of our AI models, including data selection, training, testing and monitoring, to reduce discriminatory outcomes and promote fairness in matching results.
As part of our efforts to promote corporate social responsibility and sustainable development, we are in the process of optimizing our corporate governance on environmental, social and corporate governance. We have established a Strategy and ESG Committee which is responsible for overseeing and guiding our ESG initiatives. We also plan to adopt a comprehensive ESG policy. In addition, we intend to set up an ESG task force led by the Strategy and ESG Committee, which would be responsible for the formulation, implementation and evaluation of our ESG initiatives and report to our Strategy and ESG committee regularly. Moreover, we plan to engage professional external ESG consultants to help us establish and improve our ESG policies and standards.
Each year, the ESG task force will perform identification and evaluation of ESG risk factors, weighing the risks of different ESG factors, including environmental, social and climate-related risks, by evaluating aspects such as affected scope, frequency of occurrence, recovery costs, predictability and social awareness, so that we can pinpoint the key ESG risk factors and formulate mitigation measures accordingly. In such assessment, we may also use metrics such as percentage of renewable energy used for our solutions, employee air travel emissions, percentage of suppliers that have an environmental program in place, potable water use per square foot and/or waste to landfill per square foot.
Worldwide environmental issues such as climate change and natural resources depletion are becoming increasingly prominent. Despite that our operations do not involve any production facilities or otherwise impose any material threats to the environment, we still make our best effort to minimize our impact on the environment. We intend to adopt a company-wide environmental management system that aligns with customary international standards. Our Strategy and ESG committee are responsible for overseeing, and the ESG task force will be responsible for implementing, the environmental management system.
During the Track Record Period and up to the Latest Practicable Date, we had not been subject to any fines or other penalties due to noncompliance in relation to health, work safety or environment regulations and had not had any incident, or received any claim for personal or property damage made by our employees, which had materially and adversely affected our financial condition or business operations.
Given that we operate our business primarily in the office, and that a majority of our operations are conducted online, we leave limited impact on the environment with a small carbon footprint. As advised by our PRC Legal Advisors, we were in compliance with laws or regulations in relation to health, work safety or environment in all material respects during the Track Record Period and up to the Latest Practicable Date. In light of such business nature, environmental-related and social-related risks and climate-related issues are not likely to have material negative impacts on our business, strategy and financial performance going forward. During the Track Record Period and up to the Latest Practicable Date, we had not incurred material capital expenditures or compliance costs related to climate and environmental protection. In 2022, 2023, 2024 and the six months ended June 30, 2025, our expenses in relation to environmental compliance matters were RMB0.04 million (万元), RMB4.5 million (万元), RMB0.02 million (万元) and RMB0.2 million (万元), respectively. We also do not anticipate to incur material capital expenditures or compliance costs related to climate in the foreseeable future.
We have established and currently maintain risk management and internal control systems consisting of policies and procedures that we consider to be appropriate for our business operations. We are dedicated to continually improving these systems. We have adopted and implemented comprehensive risk management policies in various aspects of our business operations. Our Board of Directors is responsible for the establishment and updating of our internal control systems, while our senior management monitors the daily implementation of the internal control procedures and measures with respect to each subsidiary and functional departments.
We have adopted comprehensive accounting policies in connection with our financial reporting risk management, such as financial management, budget management and financial statement preparation. We also have procedures in place to carry out such accounting policies, and our finance department reviews our management accounts in accordance with such procedures. In addition, we provide ongoing training to our finance staff to ensure that these policies are well-observed and effectively implemented.
参见"——数据隐私与个人信息保护"。
我们已设计并采用严格的内部程序，以确保我们的业务运营符合相关规则和法规，并保护我们的知识产权。我们的法律部门在签订任何合同或业务安排之前，会审查合同条款并审核与业务运营相关的所有文件，包括交易对手方或我们为履行合同义务而取得的许可证和许可，以及所有必要的基础尽职调查材料。在记录期间及截至最后实际可行日期，不存在重大且系统性的违规情况。
我们制定了详细的内部程序，以确保我们的内部法律部门在服务向公众开放之前对其进行监管合规审查。我们的法律部门还负责识别任何监管要求、获取任何必要的政府预先批准或同意，以及完成任何所需的监管备案，包括在规定的监管时限内向相关政府机构准备和提交所有必要的备案文件，并确保所有必要的商标、版权和专利注册申请、续期或备案已及时向主管机关提出。
我们已设计并采用严格的内部程序，以确保我们的业务运营符合相关规则和法规。我们维持内部程序，以确保已取得业务运营所需的所有重要许可证、许可及批准，并定期进行审查，以监控上述许可证及批准的状态和有效性。我们取得必要的政府批准或同意，包括在规定的监管时限内向相关政府机构准备和提交所有必要的备案文件。
我们已建立内部控制和风险管理政策，涵盖人力资源管理的各个方面，例如招聘、培训、职业道德和法律合规。我们在招聘方面保持高标准，制定严格程序以确保新员工的素质，并根据不同部门员工的需求提供专项培训。我们还定期对员工进行绩效考核，其薪酬以绩效为基础。我们定期监控内部风险管理政策的执行情况，以识别、管理和降低与我们集团各层级可能违反行为准则、职业道德及内部政策或违法行为相关的内部风险。
特别地，我们在公司内部制定了一套全面的反腐败和反贿赂政策，以促进和支持遵守适用的反腐败法律法规，就反腐败和反贿赂实践、举报渠道以及执行政策的责任提供指引。我们所有员工和第三方代理人均须理解并遵守上述政策，我们也会不时向员工和第三方代理人提供反腐败培训。
我们的投资部门负责投资项目的挖掘、筛选、执行和投资组合管理。该部门按照我们的投资策略寻找投资项目，并进行全面的投前尽职调查，以评估投资项目的风险、业务协同效应及潜在回报。
During the Track Record Period and up to the Latest Practicable Date, we had not been involved in any actual or pending legal, arbitration or administrative proceedings (including any bankruptcy or receivership proceedings) that we believe would have a material adverse effect on our business, results of operations, financial condition or reputation and compliance.
During the Track Record Period and up to the Latest Practicable Date, there were no material breaches or violations of laws or regulations applicable to us which are expected to have a material adverse effect on our business, financial condition or results of operations.
Our Board of Directors comprises nine Directors, including three executive Directors, three non-executive Directors and three independent non-executive Directors. The powers and duties of our Board include determining our business and investment plans, devise management and governance policies and exercising other powers, functions and duties as conferred by the Articles. The table below sets out the key information of our Directors:
| Name | Age | Date of joining our Group | Date of appointment as Director(1) | Existing position(s) in our Group | Roles and responsibilities | Relationship with other Directors, Supervisor and senior management | |---|---|---|---|---|---|---| | **Executive Directors** | | | | | | | | Dr. Liu Debing (劉德兵) | 49 | June 11, 2019 | March 26, 2025(2) | Co-founder, executive Director and chairman of the Board | Responsible for the strategic planning, business direction and overall management of our Group | Party acting in concert with Dr. Zhang and Dr. Li | | Dr. Zhang Peng (張鵬) | 46 | June 11, 2019 | March 26, 2025(2) | Co-founder, executive Director, chief executive officer and general manager | Responsible for the business development, R&D and the daily operations and management of our Group | Party acting in concert with Dr. Liu and Dr. Li | | Ms. Zhang Xiaohan (張笑涵) | 28 | July 11, 2022 | March 26, 2025(2) | Executive director | Participating in the Board as an employee director | None | | **Non-executive Directors** | | | | | | | | Dr. Li Juanzi (李涓子) | 61 | June 11, 2019 | June 28, 2025 | Co-founder and non-executive Director | Responsible for providing guidance for the R&D, strategy and business development of our Group | Party acting in concert with Dr. Liu and Dr. Zhang | | Mr. Li Jiaqing (李家慶) | 52 | February 10, 2023 | March 26, 2025(2) | Non-executive Director | Responsible for providing guidance for the strategy and business development of our Group | None | | Mr. Wang Meng (王盟) | 42 | August 19, 2024 | March 26, 2025(2) | Non-executive Director | Responsible for providing guidance for the strategy and business development of our Group | None |
| Name | Age | Date of joining our Group | Date of appointment as Director(1) | Existing position(s) in our Group | Roles and responsibilities | Relationship with other Directors, Supervisor and senior management | |---|---|---|---|---|---|---| | **Independent Non-executive Directors** | | | | | | | | Dr. Yang Qiang (楊強) | 64 | December 30, 2025 | June 28, 2025 | Independent non-executive Director | Responsible for providing independent advice to our Board | None | | Dr. Xie Deren (謝德仁) | 53 | December 30, 2025 | June 28, 2025 | Independent non-executive Director | Responsible for providing independent advice to our Board | None | | Mr. Tang Ying (唐穎) | 48 | December 30, 2025 | June 28, 2025 | Independent non-executive Director | Responsible for providing independent advice to our Board | None |
Notes: (1) The re-designation of each Director as an executive Director or a non-executive Director and the appointment of independent non-executive Directors were approved by the Shareholders on June 28, 2025. The appointment of the independent non-executive Directors took effect as of the date of this prospectus. (2) This date of the appointment as Director refers to the appointment date at the Shareholders' meeting of our Company for the conversion into a joint stock company with limited liability in March 2025. See "History, Development and Corporate Structure—Establishment and Major Shareholding Changes of Our Company—Conversion into a Joint Stock Company in 2025" for further details of the conversion.
Dr. Liu Debing (劉德兵), aged 49, is our co-founder, executive Director and chairman of the Board. He is primarily responsible for the strategic planning, business direction and overall management of our Group.
Dr. Liu has nearly 18 years of experience in the computing technology industry. From September 2007 to December 2012, he worked at the Beijing Institute of Technicolor (China) Technology Co., Ltd. (特藝 (中國) 科技有限公司) with his last position as a research engineer. Dr. Liu subsequently worked as senior engineer at Tsinghua University.
Dr. Liu obtained a bachelor's degree in computer science and technology from Beijing Jiaotong University (北京交通大學) in the PRC in July 1999 and a Ph.D. degree in computer science and technology from Institute of Computing Technology, Chinese Academy of Sciences (中國科學院計算技術研究所) in the PRC in July 2007.
In October 2013, Dr. Liu received the Science and Technology Progress Award—First Prize from the Chinese Association for Artificial Intelligence. In November 2017, he was awarded the title of Senior Engineer by the Chinese Academy of Sciences and the Beijing Science and Technology Progress Award—First Prize by the Beijing Municipal People's Government. In September 2021, Dr. Liu received the National Science and Technology Progress Award—Second Prize by the State Council. In March 2024, Dr. Liu received Science and Technology Progress—First Prize from the China Institute of Electronics.
Dr. Zhang Peng (張鵬), aged 46, is our co-founder, executive Director, chief executive officer and general manager. He is primarily responsible for the business development, R&D and the daily operations and management of our Group.
Dr. Zhang has nearly 20 years of experience in the field of computer science. He worked at Tsinghua University from August 2005 to December 2020.
Dr. Zhang obtained a bachelor's degree in computer science and technology, a master's degree in computer science and technology and a Ph.D. degree in electronics and information in July 2002, July 2005 and June 2025, respectively, from Tsinghua University in the PRC.
In November 2009, Dr. Zhang received the Fourth Wang Xuan News Science and Technology Award from the China Association of Press and Technology Professionals. In March 2024, Dr. Zhang was awarded the Science and Technology Progress Award—First Prize by the China Institute of Electronics. In April 2025, Dr. Zhang was awarded the title of Model Worker of the National Industry and Information System by the Ministry of Human Resources and Social Security and the MIIT.
Ms. Zhang Xiaohan (張笑涵), aged 28, is our executive Director. She is our employee director. Ms. Zhang joined our Group in October 2021 and has served as core manager of our Group's data labeling operations and Zhipu QingYan since July 2022.
Ms. Zhang obtained a bachelor's degree in engineering mechanics in July 2019 and a master's degree in data science and information technology from in July 2022 from Tsinghua University in the PRC.
Dr. Li Juanzi (李涓子), aged 61, is our co-founder and non-executive Director. She is primarily responsible for providing guidance for the R&D, strategy and business development of our Group.
From July 1989 to July 1996, Dr. Li was a lecturer in the Department of Computer Science and Technology at Shanxi University (山西大學). Dr. Li served as lecturer from December 2001 to December 2002, associate professor from December 2002 to December 2008 and tenured professor since December 2008 in the Department of Computer Science and Technology at Tsinghua University.
Since December 2014, Dr. Li has served as director of the Special Committee on Language and Knowledge Computing of the Chinese Information Processing Society of China. Since January 2019, Dr. Li has served as director of the Knowledge Intelligence Center at the Institute for Artificial Intelligence of Tsinghua University. Dr. Li also serves as director of the Joint Research Center between Tsinghua University's Department of Computer Science and Technology and Siemens (China) Co., Ltd for Industrial Intelligence and Internet of Things.
Dr. Li obtained a bachelor's degree and a master's degree in computer science and technology from Shanxi University in the PRC in July 1986 and July 1989, respectively. She received her Ph.D. degree in computer science and technology from Tsinghua University in the PRC in January 2000. Dr. Li also conducted postdoctoral research in the same field from December 2000 to December 2001.
In October 2013, Dr. Li received the Scientific and Technological Progress—First Prize from the Chinese Association for Artificial Intelligence. In November 2017, Dr. Li received the Beijing Scientific and Technological Progress—First Prize from the Beijing People's Municipal Government. In March 2020, Dr. Li was recognized by the China Language Resources Protection Outstanding Individual Award. In April 2021, Dr. Li received the Beijing Invention Patent Award—First Prize from the Beijing People's Municipal Government. In September 2021, Dr. Li received the National Scientific and Technological Progress—Second Prize from the State Council. In March 2024, Dr. Li received Scientific and Technological Progress—First Prize from the China Institute of Electronics. In April 2025, Dr. Li was recognized as Outstanding Mentor from Tsinghua University.
Mr. Li Jiaqing (李家慶), aged 52, is our non-executive Director. He is primarily responsible for providing guidance for the strategy and business development of our Group.
From July 1999 to January 2000, Mr. Li served as business development manager at Lenovo Group (聯想集團). Since July 2001, he has worked at and currently serves as the president of Legend Capital (君聯資本). Mr. Li served as director of Hichain Logistics Co., Ltd. (江蘇海晨物流股份有限公司), a company listed on the Shenzhen Stock Exchange (stock code: 300873), since May 2016; as director of Howbuy Wealth Management Co., Ltd. (好買財富管理股份有限公司), a company listed on the National Equities Exchange and Quotations (stock code: 834418), from November 2012 to September 2024; as non-executive director of Pharmaron Beijing Co., Ltd. (康龍化成（北京）新藥技術股份有限公司), a company listed on the Shenzhen Stock Exchange (stock code: 300759) and the Stock Exchange (stock code: 3759), since October 2016; as non-executive director of Eastern Air Logistics Co., Ltd. (東方航空物流股份有限公司), a company listed on the Shanghai Stock Exchange (stock code: 601156), from July 2017 to April 2024; and as non-executive director of UCloud Technology Co., Ltd. (優刻得科技股份有限公司), a company listed on the Shanghai Stock Exchange (stock code: 688158), from June 2020 to August 2024, in each case providing strategic and governance guidance to the respective company.
Mr. Li obtained bachelor's degrees in engineering and business administration in July 1996 and a master's degree in management science and engineering in July 1999 from Tsinghua University in the PRC and a master's degree in business administration from Collège des Ingénieurs in France in June 2001.
Mr. Li holds a fund qualification certificate (基金從業資格).
Mr. Wang Meng (王盟), aged 42, is our non-executive Director. He is primarily responsible for providing guidance for the strategy and business development of our Group.
Mr. Wang worked at China Mobile Communications Group Co., Ltd. (中國移動通信集團有限公司) as project manager of the support and maintenance center of its management information system department and as deputy general manager of its cloud computing center. From January 2018 to January 2022, Mr. Wang served as executive director and chief representative at the Beijing representative office of China Merchants Innovation Investment Management Co., Ltd. (招商局創新投資管理有限公司). Since February 2022, Mr. Wang served as director of the investment and corporate development department at Ant Group (螞蟻集團).
Mr. Wang obtained a bachelor's degree in communications engineering in July 2005 and a master's degree in communications and information system in July 2007 from Beijing Institute of Technology (北京理工大學) in the PRC.
Dr. Yang Qiang (楊強), aged 64, is our independent non-executive Director. He is primarily responsible for providing independent advice to our Board.
Dr. Yang served as assistant and associate professor with tenure at the University of Waterloo in Canada from September 1989 to August 1995. He was associate professor and later tenured professor at Simon Fraser University in Canada from September 1995 to September 2004, during which period he was also the NSERC Industry Chair Professor. From August 2001 to June 2012, Dr. Yang was associate professor and later tenured professor at the Hong Kong University of Science and Technology. He was the founding director of the Noah's Ark Laboratory of Huawei from June 2012 to October 2014. He served as dean of the Department of Computer Science and Engineering and New Bright Chair Professor of Engineering at the Hong Kong University of Science and Technology from November 2014 to February 2018.
Dr. Yang co-founded and served as director at 4Paradigm Inc. (第四範式集團股份有限公司) since November 2016, a company listed on the Stock Exchange (stock code: 6682). He served as an independent
non-executive director of WeBank Co., Ltd. (深圳前海微众银行股份有限公司) from December 2016 to April 2018 and as chief artificial intelligence officer. Since May 2018, Dr. Yang has also served as an independent non-executive director of China Mobile Limited (中国移动有限公司), a company listed on the Stock Exchange (stock code: 941) and Shanghai Stock Exchange (stock code: 600941).
Dr. Yang has been chair professor of artificial intelligence in the Department of Data Science and Artificial Intelligence and chief artificial intelligence officer at the Hong Kong Polytechnic University, a fellow of the Institute of Advanced Studies of Lingnan University in Hong Kong and Chair Professor at The Hong Kong University of Science and Technology (Guangzhou).
Dr. Yang obtained a bachelor's degree in astrophysics from Peking University in the PRC in July 1982 and master's degrees in astrophysics and computer science in July 1985 and July 1987, respectively, and a Ph.D. degree in computer science, specializing in artificial intelligence, in July 1989, from the University of Maryland in the United States.
Dr. Yang is a fellow of the Institute of Electrical and Electronics Engineers (IEEE), the American Association for the Advancement of Science (AAAS), the International Association for Pattern Recognition (IAPR), the Association for Computing Machinery (ACM) and the Chinese Association for Artificial Intelligence (CAAI). He was the first Chinese fellow of the Association for the Advancement of Artificial Intelligence (AAAI) in 2013. Dr. Yang has also been a fellow of the Royal Society of Canada (RSC) and the Canadian Academy of Engineering (CAE) in 2021.
Dr. Xie Deren (谢德仁), aged 53, is our independent non-executive Director. He is primarily responsible for providing independent advice to our Board.
Dr. Xie successively served as a lecturer and an associate professor and served as a professor since December 2005 at the School of Economics and Management, Tsinghua University (清华大学经济管理学院). Dr. Xie is now a council member of Accounting Society of China (中国会计学会) and the vice Chairman of Enterprise Accounting Standards Committee of Accounting Society of China. He became a member of the 17th Issuance Review Committee of the China Securities Regulatory Commission (中国证监会发行审核委员会) in September 2017. He had been a member of the First, Second and Third Advisory Committee for Enterprises Accounting Standards of the Ministry of Finance (财政部) from July 2016 to August 2023. Dr. Xie is a member of the Auditing Standards Committee of the Chinese Institute of Certified Public Accountants since December 2023.
Dr. Xie has been serving as an independent non-executive director and the chairman of the audit committee and remuneration committee of HengTai Securities Co., Ltd. (恒泰证券股份有限公司), a company listed on the Stock Exchange (stock code: 1476) from January 2020 to September 2023; an independent non-executive director, the chairman of audit committee and a member of remuneration committee of Xiamen Bank Co., Ltd. (厦门银行股份有限公司), a company listed on the Shanghai Stock Exchange (stock code: 601187) since March 2021; an independent non-executive director, the chairman of audit committee and a member of remuneration committee of Beijing Jingwei Hirain Technologies Co., Ltd. (北京经纬恒润科技股份有限公司), a company listed on Shanghai Stock Exchange (stock code: 688326) since October 2020; an independent non-executive director, the chairman of audit committee and a member of remuneration committee of Liaoning Chengda Co., Ltd. (辽宁成大股份有限公司), a company listed on Shanghai Stock Exchange (stock code: 600739) from August 2021 to January 2022; an independent non-executive director and the chairman of the audit committee and a member of remuneration committee of AInnovation Technology Group Co., Ltd. (创新奇智科技集团股份有限公司), a company listed on the Stock Exchange (stock code: 2121) since May 2021; and director of China Electronics Engineering Design Institute Co., Ltd. (中国电子工程设计院股份有限公司).
Dr. Xie obtained his bachelor's degree and Ph.D. degree in accounting from Xiamen University (廈門大學) in the PRC in July 1993 and July 1998, respectively.
Mr. Tang Ying (唐穎), aged 48, is our independent non-executive Director. He is primarily responsible for providing independent advice to our Board.
From August 1998, Mr. Tang served at Roland Berger International Management Consulting (Shanghai) Co., Ltd. (羅蘭貝格國際管理諮詢(上海)有限公司), where he held several senior management positions and was appointed a Global Principal in 2005, specializing in the automotive industry practice. From 2006 to 2008, he served as executive vice president for global strategy and business operations at Tiens Group (天獅集團).
In 2011, Mr. Tang founded and served as executive director at iFORCE Beijing Interactive Co., Ltd. (北京百孚思廣告有限公司), which was wholly acquired in 2015 by Zhewen Interactive Group Co., Ltd. (浙文互聯集團股份有限公司) (formerly known as KEDA Group Co., Ltd. (科達集團股份有限公司)), a company listed on the Shanghai Stock Exchange (stock code: 600986), and became the core digital marketing company within the Zhewen Interactive Group Co., Ltd. group. Mr. Tang served in Zhewen Interactive Group Co., Ltd. as vice general manager since June 2016, as general manager from December 2016 to July 2017, as director in July 2017, vice chairman from December 2018 to November 2020, and as chairman and chief executive officer from November 2020 to June 2025. Mr. Tang has continued to serve as director at Zhewen Interactive Group Co., Ltd. since June 2025. Since March 2025, Mr. Tang has also served as director at Juewei Foods Co., Ltd. (絕味食品股份有限公司), a company listed on the Shanghai Stock Exchange (stock code: 603517).
Mr. Tang obtained a bachelor's degree in economics from Shanghai Jiao Tong University in the PRC in July 1999 and an executive master's degree in business administration (EMBA) from the Cheung Kong Graduate School of Business (CKGSB) in the PRC in 2012. Mr. Tang also completed the CEO Program and Business Scholars Program (DBA) from CKGSB and the Global Finance GFD Program at Tsinghua University.
Save as disclosed above, each of our Directors has confirmed that he/she has no other relationship with any other Directors, Supervisor and senior management of our Company and none of our Directors, Supervisors or senior management has held any other directorships in listed companies during the three years immediately preceding the date of this prospectus.
Save as disclosed above, to the best knowledge of our Directors, there are no other matters relating to the appointment of our Directors and Supervisor that need to be brought to the attention of our Shareholders and there is no other information in relation to our Directors and Supervisor which is required to be disclosed pursuant to Rule 13.51(2) of the Listing Rules.
Each of our Directors has confirmed that he/she obtained legal advice on June 24, 2025 with regard to the requirements under the Listing Rules that are applicable to him/her as a director of a listed issuer and the possible consequences of making a false declaration or giving false information to the Stock Exchange as set out in Rule 3.09D of the Listing Rules and he/she understood his/her obligations as a director of a listed issuer.
Each of our independent non-executive Directors has confirmed his independence with regards to each of the factors as set out in Rule 3.13(1) to (8) of the Listing Rules and that there are no other factors that may affect his independence at the time of his appointment.
As of the Latest Practicable Date, we had one Supervisor. In accordance with the applicable PRC laws and regulation, the Supervisor will resign as Supervisor of the Company with effect from the completion of Listing, and our Company will no longer have any Supervisor. The table below sets out the key information of our Supervisor:
| Name | Age | Date of joining our Group | Date of appointment as Supervisor | Existing position(s) | Roles and responsibilities | Relationship with other Directors and senior management | |---|---|---|---|---|---|---| | Mr. Pei Bo (裴博) | 25 | July 1, 2024 | March 26, 2025 | Supervisor | Responsible for monitoring the performance of the Directors and senior management | None |
Mr. Pei Bo (裴博), aged 25, was appointed as our Supervisor in March 2025. He is primarily responsible for monitoring the performance of the Directors and senior management.
Mr. Pei obtained a bachelor's degree in law in June 2024 from Tsinghua University in the PRC. He holds a PRC Legal Professional Qualification Certificate (法律職業資格證書).
Our senior management is responsible for the day-to-day management of our business. The table below sets out the key information of our senior management:
| Name | Age | Date of joining our Group | Date of appointment as senior management | Existing position(s) in our Group | Roles and responsibilities | Relationship with other Directors, Supervisor and senior management | |---|---|---|---|---|---|---| | Dr. Zhang Peng (張鵬) | 46 | June 11, 2019 | January 1, 2024 | Co-founder, executive Director, chief executive officer and general manager | Responsible for the business development, R&D and the daily operations and management of our Group | Party acting in concert with Dr. Liu and Dr. Li | | Mr. Wang Shaolan (王紹蘭) | 52 | June 11, 2019 | July 1, 2019 | Co-founder and deputy general manager | Responsible for the business development, financing and the daily operations and management of our Group | None | | Mr. Xiao Lei (肖磊) | 37 | October 30, 2025 | December 18, 2025 | Board secretary | Responsible for the corporate governance and financing operations of our Group | None |
Dr. Zhang Peng (張鵬), aged 46, is our co-founder, executive Director, chief executive officer and general manager. See "—Directors—Executive Directors" above for his biographical details.
Mr. Wang Shaolan (王紹蘭), aged 52, is our co-founder and deputy general manager. He is mainly responsible for the business development, financing and the daily operations and management of our Group. Mr. Wang has nearly more than 20 years of experience in the technology industry. From August 2001 to May 2012, he worked at Beijing O2Micro Microelectronics Development Co., Ltd. (北京凹凸微系電子開發有限公司). From July 2012 to June 2013, he worked at Nebula Sunac (Beijing) Technology Co., Ltd. (星雲融創(北京)科技有限公司). From July 2013 to December 2017, he served as chief executive officer at Wankang Century Technology (Beijing) Co., Ltd. (萬康世紀科技(北京)有限公司). From March 2018 to June 2019, he worked at Tsinghua University.
Mr. Wang obtained a master's degree in electronics and electrical drive from Anhui University of Technology (安徽工業大學) in the PRC in June 1999.
Mr. Xiao Lei (肖磊), aged 37, is our board secretary. He is mainly responsible for the corporate governance and financing operations of our Group.
Mr. Xiao served as a legal consultant of COFCO Land Management Co., Ltd. (中糧置地管理有限公司) from August 2012 to August 2014. Mr. Xiao held several positions, including the deputy general manager and general manager of the legal department, secretary to the board, general manager of the investment banking department and deputy general manager of Shoutai Jinxin (Beijing) Equity Investment Fund Management Co., Ltd. (首泰金信(北京)股權投資基金管理股份有限公司) from May 2014 to January 2020. Mr. Xiao served as director of finance and investment, joint company secretary, board secretary, chief financial officer and vice president of AInnovation Technology Group Co., Ltd (創新奇智科技集團股份有限公司), a company listed on the Stock Exchange (stock code: 2121). He served as vice president of Shanghai Soybean Network Technology Co., Ltd. (上海黃豆網絡科技有限公司) from February 2024 to October 2025.
Mr. Xiao obtained his bachelor's degree in law from China University of Political Science and Law (中國政法大學) in the PRC in July 2009 and his master's degree in civil and commercial law from China University of Political Science and Law in June 2012.
Mr. Cheng Ching Kit (鄭程傑) was appointed as our company secretary in June 2025.
Mr. Cheng Ching Kit is an assistant vice president of SWCS Corporate Services Group (Hong Kong) Limited, a professional services provider specializing in corporate services, and has over 12 years of experience in the corporate secretarial field. He is an associate member of both The Hong Kong Chartered Governance Institute and The Chartered Governance Institute in the United Kingdom.
Mr. Cheng holds a Bachelor of Commerce degree in finance from the University of Queensland, Australia and a Master of Laws degree in Chinese law from the University of Hong Kong.
Our Board has established the Audit Committee, the Remuneration Committee, the Nomination Committee and the ESG and Strategy Committee and delegated various responsibilities to these committees, which assist our Board in discharging its duties and overseeing particular aspects of our Group's activities.
We have established the Audit Committee pursuant to Rule 3.21 of the Listing Rules with written terms of reference in compliance with paragraph D.3 of Part 2 of the Corporate Governance Code as set out in Appendix C1 to the Listing Rules (the "CG Code"). The Audit Committee consists of Dr. Xie Deren,
Dr. Yang Qiang and Dr. Li Juanzi. Dr. Xie Deren is the chairperson of the Audit Committee and has the appropriate professional qualifications or accounting or related financial management expertise as required under Rule 3.10(2) of the Listing Rules.
The primary duties of the Audit Committee are (i) reviewing annually the performance of external auditors and making recommendations to the board on their appointment, reappointment, removal, fees and terms; (ii) acting as liaison between the Company, the internal audit department and external auditors and assessing the independence of external auditors and effectiveness of audit procedures; (iii) establishing policies for engaging external auditors for non-audit services and ensuring such services do not affect independence of external auditors and that the external auditors are qualified to provide non-audit services; (iv) reviewing accounting policies, financial status, reporting procedures, internal controls and financial statements; (v) communicating with the Board, management and both internal and external auditors and meet with external auditors without executive directors present to discuss important and exceptional items in company reports and accounts at least biannually; (vi) examining the Company's financial policies, internal control systems and risk management and making recommendations for improvement; and (vii) performing other duties and responsibilities as assigned by our Board and/or required by applicable laws and regulations (including the Listing Rules) from time to time.
We have established the Remuneration Committee pursuant to Rule 3.25 of the Listing Rules with written terms of reference in compliance with paragraph E.1 of Part 2 of the CG Code. The Remuneration Committee consists of Mr. Tang Ying, Dr. Liu Debing and Dr. Xie Deren. Mr. Tang Ying is the chairperson of the Remuneration Committee.
The primary duties of the Remuneration Committee are (i) formulating remuneration policies for Directors and senior management based on their duties, time commitment, responsibilities and remuneration in comparable companies; (ii) reviewing and approving remuneration proposals based on policies and objectives set by the Board; (iii) evaluating the performance of Directors and senior management annually and submitting bonus proposals for Board approval; (iv) supervising the implementation of remuneration policies; (v) advising the Board on remuneration of Directors and senior management, including non-monetary benefits, pensions and compensation for loss or termination of office, and preparing proposals on remuneration of non-executive Directors; (vi) consulting the chairman of the Board or general manager and obtaining independent advice as needed when setting remuneration for executive Directors; (vii) reviewing and approving compensation for loss or termination of office to ensure consistency with contract terms and fairness; (viii) overseeing arrangements for compensation for loss of office resulting from misconduct or dismissal; (ix) ensuring that no Director or their associates are involved in deciding their own remuneration and that remuneration of non-executive directors who serve on the Remuneration Committee is determined by other members; (x) reviewing employee incentive schemes and Directors' service contracts; and (xi) performing other duties and responsibilities as assigned by our Board and/or required by applicable laws and regulations (including the Listing Rules) from time to time.
We have established the Nomination Committee pursuant to Rule 3.27A of the Listing Rules with written terms of reference in compliance with paragraph B.3 of Part 2 of the CG Code. The Nomination Committee consists of Dr. Yang Qiang, Mr. Tang Ying and Dr. Li Juanzi. Dr. Yang Qiang is the chairperson of the Nomination Committee.
The primary duties of the Nomination Committee are: (i) developing selection criteria and procedures for Directors and senior management based on factors such as cultural background, educational history and professional experience, and making recommendations to the Board; (ii) identifying and nominating
qualified candidates for directorships, examining the suitability of nominees and providing recommendations to the Board; (iii) searching for and assessing qualified candidates for senior management and proposing appointments to the Board; (iv) evaluating the independence of independent non-executive Directors; (v) reviewing the Board's structure, size and composition at least annually, assessing skills, knowledge and experience and recommending changes based on company strategy, and nominating members to Board committees for Board approval; (vi) advising the Board on Director appointments, reappointments and succession planning, especially chair and general manager; (vii) establishing, maintaining and updating a reserve list of candidates for Directors and senior management to ensure business continuity; (viii) evaluating the performance of Directors and based on such evaluation results, recommending changes and reappointments, including the chair and chief executive officer; (ix) formulating and reviewing the Board diversity policy, monitoring progress towards diversity objectives and disclosing the policy or its summary in the Company's annual report as appropriate; and (x) performing other duties and responsibilities as assigned by our Board and/or required by applicable laws and regulations (including the Listing Rules) from time to time.
We have established the ESG and Strategy Committee. The ESG and Strategy Committee consists of Dr. Liu Debing, Dr. Li Juanzi and Dr. Yang Qiang. Dr. Li Juanzi is the chairperson of the ESG and Strategy Committee.
The primary duties of the ESG and Strategy Committee are (i) overseeing the Company's business operations; (ii) researching and monitoring domestic and international industry developments and relevant policies; (iii) reviewing and considering the Company's strategic development plans and making recommendations to the Board; (iv) providing advice and recommendations on major capital operations, key investment and financing decisions and material business reforms; (v) conducting research and providing advice and support in relation to the Company's ESG management, including but not limited to ESG goals, objectives, policies and material ESG risks; and (vi) performing such other duties as may be required under applicable laws and regulations (including the Listing Rules) and as delegated by the Board from time to time.
Our Board has adopted a board diversity policy ("Board Diversity Policy"), which sets out the approach to achieve diversity on our Board. Our Company recognizes and embraces the benefits of having a diverse Board and sees increasing diversity at the Board level as an essential element in supporting the attainment of our Company's strategic objectives and sustainable development. Our Company seeks to achieve Board diversity through the consideration of a number of factors, including but not limited to talent, skills, gender, age, cultural and educational background, ethnicity, professional experience, independence, knowledge and length of service. We will select potential Board candidates based on merit and his/her potential contribution to our Board while taking into consideration our own business model and specific needs from time to time. All Board appointments will be based on merit and candidates will be considered against objective criteria, having due regard to the benefits of diversity on our Board.
Our Board has a balanced mix of knowledge, skills and experience, including but without limitation to computer science, artificial intelligence research, business management, investment, accounting, consulting and marketing. Members of our Board have obtained degrees in various majors including computer science, engineering, economics and business administration. We have three independent non-executive Directors from different backgrounds, including artificial intelligence research and management, accounting, consulting and marketing.
With regards to gender diversity on the Board, we recognize the particular importance of gender diversity. Our Board currently comprises two female Directors and seven male Directors and expects to
continue to maintain an appropriate gender mix in the Board upon Listing. We have taken and will continue to take steps to promote and enhance gender diversity at all levels of our Company, including but without limitation at our Board and senior management levels. Our Board Diversity Policy provides that our Board should not be a single gender board. We will also ensure that there is gender diversity when recruiting staff at mid- to senior- level so that we will have a pipeline of senior management and potential successors to our Board going forward to achieve and maintain gender diversity. It is our objective to maintain an appropriate balance of gender diversity with reference to the expectations of stakeholders and international and local recommended best practices.
Our Nomination Committee is responsible for ensuring the diversity of our Board members. After Listing, our Nomination Committee will review our Board Diversity Policy and its implementation from time to time to monitor its continued effectiveness and we will disclose the implementation of our Board Diversity Policy, including any measurable objectives set for implementing the Board Diversity Policy and the progress on achieving these objectives, in our corporate governance report on an annual basis.
Our Company aims to achieve high standards of corporate governance which are crucial to the development and safeguard the interests of our Shareholders. To accomplish this, our Company expects to comply with the CG Code and the relevant Listing Rules after the Listing.
Our Directors, Supervisor and members of our senior management receive compensation from our Group in the form of fees, salaries and other benefits and contribution to pension scheme.
The aggregate remuneration (including salaries, allowances, benefits in kind, discretionary bonuses, retirement scheme contributions and share-based payments) paid or payable to our Directors and Supervisor for the three years ended December 31, 2024 and the six months ended June 30, 2025 was approximately RMB5.0 million, RMB6.0 million, RMB11.5 million and RMB43.4 million, respectively. Save as disclosed above, no amounts have been paid or are payable by any member of our Group to our Directors or Supervisor for the three years ended December 31, 2024.
The aggregate amount of salaries, allowances, benefits in kind, discretionary bonuses, retirement scheme contributions and share-based payments paid or payable to our five highest paid individuals (excluding Directors) in respect of the three years ended December 31, 2024 and the six months ended June 30, 2025 was approximately RMB2.0 million, RMB12.4 million, RMB30.8 million and RMB55.5 million, respectively.
No remuneration was paid by us to our Directors, Supervisor or the five highest paid individuals as an inducement to join or upon joining us or as a compensation for loss of office in respect of the three years ended December 31, 2024 and the six months ended June 30, 2025. Further, none of our Directors or Supervisor had waived or agreed to waive any remuneration during the same periods.
Under the arrangement currently in force, the aggregate remuneration (including salaries, allowances, benefits in kind, discretionary bonuses, retirement scheme contributions and excluding share-based payments) of our Directors and Supervisor for the year ending December 31, 2025 is estimated to be approximately RMB9.0 million.
Our Board will review and determine the remuneration and compensation packages of our Directors, Supervisor and senior management and will, following the Listing, receive recommendation from the Remuneration Committee, which will take into account salaries paid by comparable companies, time commitment and responsibilities of our Directors and performance of our Group.
For further details of our Employee Incentive Schemes, See "Appendix VI—Statutory and General Information—D. Employee Incentive Schemes" for details.
Each of our Directors confirms that, as of the Latest Practicable Date, he/she did not have any interest in any business which competes or is likely to compete directly or indirectly with our business and requires disclosure under Rule 8.10 of the Listing Rules.
We have appointed Maxa Capital Limited as our Compliance Advisor pursuant to Rule 3A.19 of the Listing Rules. Pursuant to Rule 3A.23 of the Listing Rules, our Compliance Advisor will advise our Company in the following circumstances:
• where a transaction, which might be notifiable or connected transaction under the Listing Rules, is contemplated, including shares issues, sales and transfers of treasury shares and share repurchases;
• where our Company proposes to use the proceeds from the Global Offering in a manner different from that detailed in this prospectus or where our business activities, developments or results deviate from any forecast, estimate or other information in this prospectus; and
• where the Stock Exchange makes an inquiry of our Company regarding unusual movements in the price or trading volume of our H Shares under Rule 13.10 of the Listing Rules.
The term of the appointment shall commence on the Listing Date and end on the date on which our Company distribute our annual report in respect of our financial results for the first full financial year commencing after the Listing Date.
As of the Latest Practicable Date, Beijing Lianpai, Dr. Liu, Dr. Tang, Dr. Li, Dr. Xu, Dr. Zhang, Huihui and Zhideng, by virtue of the Amended Concert Party Agreement entered into among them, were collectively interested in approximately 33.03% of the Shares. Accordingly, they are our group of Controlling Shareholders under the Listing Rules. See "History, Development and Corporate Structure—Concert Party Arrangement and Our Controlling Shareholders" for further details.
Immediately following the completion the Global Offering, our Controlling Shareholders will in aggregate hold approximately 30.22% of the Shares (assuming the Over-allotment Option is not exercised). Accordingly, upon Listing, they will remain as our group of Controlling Shareholders as defined under the Listing Rules.
As of the Latest Practicable Date, each of our Controlling Shareholders had confirmed that none of them had any interest in any business, other than our business, which competes or is likely to compete, either directly or indirectly, with our Group's business which would require disclosure under Rule 8.10 of the Listing Rules.
We believe that we are capable of carrying on our business independently from our Controlling Shareholders and their respective close associates (other than the Group) after the Listing for the following reasons:
Our Board comprises three executive Directors, three non-executive Directors and three independent non-executive Directors. Dr. Liu is our co-founder, executive Director and the chairman of our Board responsible for the strategic planning, business direction, daily operations and overall management of our Group. Dr. Zhang is our co-founder, executive Director, chief executive officer and general manager responsible for the business development, R&D and the daily operations and management of our Group. With the support of our experienced management team, Dr. Liu and Dr. Zhang are expected to continuously devote a sufficient portion of their time to the day-to-day operations of our Group upon Listing. Dr. Li is our co-founder and a non-executive Director, primarily responsible for providing guidance for the strategy and business development of our Group. Dr. Li, as a non-executive Director, is not involved in the management of our day-to-day operations.
Each of our Directors is aware of his/her fiduciary duties as a Director, which require, among other things, that he/she acts for the benefit and in the best interests of our Company and does not allow any conflict between his/her duties as a Director and his/her personal interests. In the event that there is an actual or potential conflict of interest arising out of any transaction to be entered into between our Group and any of the Directors or their respective close associates, the interested Director(s) shall abstain from voting at the relevant Board meetings of our Company in respect of such transactions and shall not be counted in the quorum.
Our Board comprises nine Directors, including three independent non-executive Directors, which represent one-third of the members of our Board. Our independent non-executive Directors have extensive experience in corporate management and governance, and they are appointed to ensure that our Board will only make decisions after due consideration of independent and impartial opinions. Certain matters of our Company must always be referred to the independent non-executive Directors for review.
We have adopted a series of corporate governance measures to manage conflicts of interest, if any, between our Group and our Controlling Shareholders that would support our independent management. For details, see "Corporate Governance Measures" in this section.
Based on the reasons above, our Directors are of the view that our Group is capable of our business independently from our Controlling Shareholders and their respective close associates after the Listing.
We have full rights to make all decisions on, and carry on our own business operations independently from our Controlling Shareholders and their respective close associates and will continue to do so after the Listing. Our Group is able to operate without reliance on our Controlling Shareholders and their respective close associates.
We have independent access to our customers, suppliers as well as our business partners. Our customer, supplier and business partner bases are diversified and unrelated to our Controlling Shareholders and their respective close associates. We also possess the relevant licenses, certificates, facilities and intellectual property rights necessary to carry on and operate our principal businesses independently.
We have full-time management team and staff to carry out our own administration and operation independently from our Controlling Shareholders and their respective close associates. All key administrative functions (including administration, finance, internal audit, human resources, legal and compliance and company secretarial functions) have been and will be carried out by our own without reliance on our Controlling Shareholders and their respective close associates. As of the Latest Practicable Date, all of our full-time employees were independent from our Controlling Shareholders and their respective close associates and were primarily recruited through both internal referrals and external sources, such as campus recruitment, recruitment websites and third-party recruiters.
Based on the reasons above, our Directors are of the view that we have full rights to make all decisions on and to carry out our own business operations independently from our Controlling Shareholders and their respective close associates and will continue to do so after the Listing.
We have independent internal control and accounting systems. We also have an independent finance department responsible for discharging financial management, accounting, reporting, funding and treasury functions. We maintain bank accounts independently and do not share any bank account with our Controlling Shareholders. We are capable of obtaining financing from third parties, if necessary, without reliance our Controlling Shareholders and their respective close associates. We also received a series of Pre-IPO Investments from third-party investors independently in the past. See "History, Development and Corporate Structure—Pre-IPO Investments" for further details of our Pre-IPO Investments.
As of the Latest Practicable Date, we did not have any outstanding loans granted or guaranteed by any of our Controlling Shareholders or their respective close associates to us.
Based on the above, our Directors believe that we are able to maintain financial independence and would not place undue reliance on our Controlling Shareholders or their respective close associates.
Our Directors recognize the importance of good corporate governance in protecting our Shareholders' interests. We will adopt the following measures to safeguard good corporate governance standards and to avoid potential conflict of interests between our Group and our Controlling Shareholders:
(a) as part of our preparation for the Global Offering, we have amended our Articles of Association to comply with the Listing Rules which will take effect upon Listing. According to our Articles of – 236 –
Association, Directors shall avoid any conflict of interest between themselves and the Company and shall not improperly use their position to benefit themselves. In particular, our Articles of Association provide that, in reviewing a connected transaction or related party transaction, the connected or related Director shall make appropriate disclosure to the Board and shall not vote on the resolution, and the Director shall not be counted in the quorum present at the meeting;
(b) we are committed that our Board should include a balanced composition with not less than one-third of independent non-executive Directors to ensure that our Board is able to effectively exercise independent judgment in its decision-making process and provide independent advice to our Shareholders. We have appointed three independent non-executive Directors and believe our independent non-executive Directors possess sufficient experience and they are free of any business or other relationship which could interfere in any material manner with the exercise of their independent judgment and will be able to provide an impartial, external opinion to protect the interests of our Shareholders. For details of our independent non-executive Directors, see "Directors, Supervisor and Senior Management—Board of Directors—Independent non-executive Directors" in this prospectus;
(c) we have appointed Maxa Capital Limited as our Compliance Advisor pursuant to Rule 3A.19 of the Listing Rules, which will provide advice and guidance to us in respect of compliance with the applicable laws and the Listing Rules including various requirements relating to Directors' duties and corporate governance;
(d) our Company has established internal control mechanisms to identify connected transactions. Upon and after the Listing, if our Company enters into connected transactions (if any) with our Controlling Shareholders or any of their associates, our Company will comply with the applicable Listing Rules; and
(e) as required by the Listing Rules, our independent non-executive Directors shall review any continuing connected transaction annually and confirm in our annual report that such transactions have been entered into in our ordinary and usual course of business, are either on normal commercial terms or on terms no less favorable to us than those available to or from independent third parties and on terms that are fair and reasonable and in the interests of our Shareholders as a whole.
Based on the above, our Directors believe that there are sufficient and adequate corporate governance measures in place to manage any existing and potential conflicts of interest that may arise between our Group and our Controlling Shareholders, and to protect minority Shareholders' interests after the Listing.
So far as our Directors are aware, immediately following the completion of the Global Offering and assuming the Over-allotment Option is not exercised, the following persons will have interests and/or short positions in the Shares or underlying shares of our Company which would fall to be disclosed pursuant to the provisions of Divisions 2 and 3 of Part XV of the SFO, or, directly or indirectly, be interested in 10% or more of the nominal value of any class of share capital carrying the rights to vote in all circumstances at general meetings of our Company:
| Name of Shareholder | Type of Shares to be held upon Listing | Nature of Interest | As of the Latest Practicable Date taking into account the Share Subdivision — Number of Shares(1) | As of the Latest Practicable Date taking into account the Share Subdivision — Approximate percentage of shareholding in the relevant type of Shares | Immediately following the completion of the Share Subdivision and the Global Offering (assuming the Over-allotment Option is not exercised) — Number of Shares(1) | Immediately following the completion of the Share Subdivision and the Global Offering (assuming the Over-allotment Option is not exercised) — Approximate percentage in the total issued Shares | |---|---|---|---|---|---|---| | Beijing Lianpai | Unlisted Shares | Beneficial interest | 34,038,390 | 8.45% | 34,038,390 | 15.23% | | | Unlisted Shares | Interests held jointly with another person(2) | 99,005,840 | 24.58% | 89,003,983 | 39.81% | | Dr. Liu | Unlisted Shares | Beneficial interest | 925,150 | 0.23% | 925,150 | 0.41% | | | Unlisted Shares | Interests held jointly with another person(2) | 132,119,080 | 32.80% | 122,117,223 | 54.62% | | Dr. Tang | Unlisted Shares | Beneficial interest | 26,835,330 | 6.66% | 26,835,330 | 12.00% | | | Unlisted Shares | Interests held jointly with another person(2) | 106,208,900 | 26.37% | 96,207,043 | 43.03% | | Dr. Li | Unlisted Shares | Beneficial interest | 3,367,760 | 0.84% | 3,367,760 | 1.51% | | | Unlisted Shares | Interests held jointly with another person(2) | 129,676,470 | 32.19% | 119,674,613 | 53.53% | | Dr. Xu | Unlisted Shares | Beneficial interest | 799,040 | 0.20% | 799,040 | 0.36% | | | Unlisted Shares | Interests held jointly with another person(2) | 132,245,190 | 32.83% | 122,243,333 | 54.68% | | Dr. Zhang | Unlisted Shares | Beneficial interest | 399,520 | 0.10% | 399,520 | 0.18% | | | Unlisted Shares | Interests held jointly with another person(2) | 132,644,710 | 32.93% | 122,642,853 | 54.86% | | Huihui | Unlisted Shares | Beneficial interest | 39,482,710 | 9.80% | 33,560,303 | 15.01% | | | Unlisted Shares | Interests held jointly with another person(2) | 93,561,520 | 23.23% | 89,482,070 | 40.41% | | Zhideng | Unlisted Shares | Beneficial interest | 27,196,330 | 6.75% | 23,116,880 | 10.34% | | | Unlisted Shares | Interests held jointly with another person(2) | 105,847,900 | 26.28% | 99,925,493 | 45.08% |
| Name of Shareholder | Type of Shares to be held upon Listing | Nature of Interest | As of the Latest Practicable Date taking into account the Share Subdivision — Number of Shares(1) | As of the Latest Practicable Date taking into account the Share Subdivision — Approximate percentage of shareholding in the relevant type of Shares | Immediately following the completion of the Share Subdivision and the Global Offering (assuming the Over-allotment Option is not exercised) — Number of Shares(1) | Immediately following the completion of the Share Subdivision and the Global Offering (assuming the Over-allotment Option is not exercised) — Approximate percentage in the total issued Shares | |---|---|---|---|---|---|---| | Mr. Chen Hao(3) | Unlisted Shares | Interest in controlled corporation | 27,109,120 | 6.73% | — | — | | | | | — | — | 27,109,120 | — |
| Tsinghua Asset Management(6) | Unlisted Shares | Interest in controlled corporation | 15,534,390 | 3.86% | 15,534,390 | 6.95% | | Tsinghua Technology(6) | Unlisted Shares | Beneficial interest | 15,534,390 | 3.86% | 15,534,390 | 6.95% | | Xu Xin(7) | Unlisted Shares | Interest in controlled corporation | 11,349,910 | 2.82% | — | — |
| Name of Shareholder | Type of Shares to be held upon Listing | Nature of Interest | As of the Latest Practicable Date taking into account the Share Subdivision | | Immediately following the completion of the Share Subdivision and the Global Offering (assuming the Over-allotment Option is not exercised) | | |---|---|---|---|---|---|---| | | | | Number of Shares(1) | Approximate percentage of shareholding in the relevant type of Shares | Number of Shares(1) | Approximate percentage in the total issued Shares |
Notes: (1) All interests stated are long positions. (2) Pursuant to the Amended Concert Party Agreement, Beijing Lianpai, Dr. Liu, Dr. Tang, Dr. Li, Dr. Xu, Dr. Zhang, Huihui and Zhideng confirmed and agreed that, during the period in which any party directly or indirectly holds or controls any shares of the Company, they will act in concert when exercising their shareholder rights as Shareholders of the Company. Therefore, under the SFO, each of Beijing Lianpai, Dr. Liu, Dr. Tang, Dr. Li, Dr. Xu, Dr. Zhang, Huihui and Zhideng is deemed to be interested in the Shares held by each other. (3) Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)) ("Junlian Xiangdao") is owned as to approximately 1.18% by its general partner, Lhasa Junqi Enterprise Management Co., Ltd. (拉薩君祺企業管理有限公司) ("Lhasa Junqi"), which is wholly owned by Legend Capital Management Co., Ltd. (君聯資本管理股份有限公司) ("LCM"). LCM is owned as to 80.00% by Beijing Juncheng Hezhong Investment Management Partnership (Limited Partnership) (北京君誠合眾投資管理合夥企業(有限合夥)) ("Juncheng Hezhong"). Juncheng Hezhong is controlled by its general partner Beijing Junqi Jiarui Enterprise Management Co., Ltd. (北京君祺嘉睿企業管理有限公司) ("Junqi Jiarui"), which is owned by Mr. Chen Hao (陳浩) as to 40% and none of the other shareholders holds 30% or more interest therein. Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯錦帆創業投資合夥企業(有限合夥)) ("Junlian Jinfan") is owned as to approximately 1.40% by its general partner, Lhasa Junqi. Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京) 合夥企業(有限合夥)) ("Social Security Zhongguancun Innovation Fund") is owned as to approximately 1.96% by Beijing Jun Chuang Li Xin Venture Capital Partnership (Limited Partnership) (北京君創勵新創業投資合夥企業(有限合夥)) as its general partner, in which Lhasa Junqi holds 50% partnership interest as the general partner. Therefore, under the SFO, each of Mr. Chen Hao, Junqi Jiarui, Juncheng Hezhong, LCM and Lhasa Junqi is deemed to be interested in the Shares held by Junlian Xiangdao, Junlian Jinfan and Social Security Zhongguancun Innovation Fund. (4) Tianjin Sankuai Technology Co., Ltd. (天津三快科技有限公司) ("Tianjin Sankuai") is wholly owned by Meituan (美团), the Class B Shares of which are listed on the Main Board of the Stock Exchange (stock codes: 3690 (HKD counter) and 83690 (RMB counter)). Therefore, under the SFO, Meituan is deemed to be interested in the Shares held by Tianjin Sankuai. (5) Shanghai Yunya Enterprise Management Consulting Co., Ltd. (上海雲玡企業管理諮詢有限公司) ("Shanghai Yunya") is wholly owned by Ant Group Co., Ltd. (螞蟻科技集團股份有限公司) ("Ant Group"). Shanghai Feiya Technology Co., Ltd. (上海飛玡科技有限公司) ("Shanghai Feiya") is wholly owned by Ant Unicorn Fund, L.P., an exempted limited partnership registered in the Cayman Islands, and managed by its general partner, Ant Unicorn Ltd., an indirect wholly owned subsidiary of Ant Group. Therefore, under the SFO, Ant Group is deemed to be interested in the Shares held by Shanghai Yunya and Shanghai Feiya.
SUBSTANTIAL SHAREHOLDERS (6) Tsinghua Control Technology Transfer Co., Ltd. (華控技術轉移有限公司) ("Tsinghua Technology") is wholly owned by Tsinghua University Asset Management Co., Ltd. (清華大學資產管理有限公司) ("Tsinghua Asset Management"), which is in turn wholly owned by Tsinghua University. Therefore, Tsinghua Asset Management is deemed to be interested in the Shares held by Tsinghua Technology. (7) Trend Mega Limited ("Trend Mega") is owned as to approximately 99.75% by Capital Today Evergreen Fund, L.P., which is controlled by its general partner, Capital Today Evergreen GenPar Ltd., a wholly owned subsidiary of Xu Xin.
Saved as disclosed herein, our Directors are not aware of any other person who will, immediately following the completion of the Global Offering (assuming that the Over-allotment Option is not exercised) and the Conversion of Unlisted Shares into H Shares, have any interest and/or short positions in the Shares or underlying shares of our Company which would fall to be disclosed to the Company pursuant to the provisions of Divisions 2 and 3 of Part XV of the SFO, or, who is, directly or indirectly, interested in 10% or more of the nominal value of any class of our share capital carrying rights to vote in all circumstances at general meetings of our Company.
截至最后实际可行日期，本公司的注册股本为人民币40,281,069元，划分为40,281,069股非上市股份，每股面值人民币1.00元。
紧接全球发售前，本公司普通股将按一拆十的比例进行股份拆分，本公司注册股本将为人民币40,281,069元，由已发行的402,810,690股非上市股份组成，每股面值人民币0.10元。
| 股份数目 | 占已发行股本总数的大约百分比 | 股份描述 | |---|---|---| | 224,528,485 | 51.00% | 非上市股份 | | 178,282,205 | 40.50% | 由非上市股份转换而成的H股 | | 37,419,500 | 8.50% | 根据全球发售发行的H股 | | 440,230,190 | 100.00% | |
| 股份数目 | 占已发行股本总数的大约百分比 | 股份描述 | |---|---|---| | 224,528,485 | 50.36% | 非上市股份 | | 178,282,205 | 39.99% | 由非上市股份转换而成的H股 | | 43,032,400 | 9.65% | 根据全球发售发行的H股 | | 445,843,090 | 100.00% | |
全球发售及非上市股份转换为H股完成后，本公司股份将由非上市股份及H股组成。非上市股份及H股均为本公司股本中的普通股，根据章程细则被视为同一类别的股份。
除中国境内的合资格境内机构投资者、沪港通及深港通下的合资格中国投资者，以及根据中国相关法律法规或经任何主管机关批准有权持有本公司H股的其他人士（例如本公司持有非上市股份的若干现有股东）外， – 242 –
by whom will be converted into H Shares according to the filing with the CSRC), H Shares generally cannot be subscribed by or traded between legal or natural PRC persons.
Unlisted Shares and H Shares shall rank pari passu with each other in all respects and, in particular, will rank equally for dividends or distributions declared, paid or made. All dividend for H Shares will be denominated and declared in Renminbi, and paid in Hong Kong dollars or Renminbi, whereas all dividends for Unlisted Shares will be paid in Renminbi. Other than cash, dividends could also be paid in the form of shares or a combination of cash and shares.
Pursuant to the PRC Company Law and the Articles of Association, our Company may from time to time by special resolution of shareholders, among others, increase its capital or decrease its capital. For details of circumstances under which our Shareholders' general meetings are required, see "Appendix V—Summary of Articles of Association" in this prospectus.
Pursuant to the regulations prescribed by the securities regulatory authorities of the State Council, the Unlisted Shares may be converted into H Shares. Such converted Shares could be listed or traded on an overseas stock exchange, provided that prior to the conversion and trading of such converted Shares, any requisite internal approval process has been duly completed and all the filing procedures with the relevant regulatory authorities, including CSRC which requires administrative filing procedures for the conversion and trading of such converted Shares, have been obtained. In addition, such conversion and trading shall comply with the regulations, requirements and procedures prescribed by the relevant overseas stock exchange. If any of the Unlisted Shares are to be converted, listed and traded as H Shares on the Stock Exchange, such conversion, listing and trading will need to be filed with the relevant PRC regulatory authorities, including the CSRC, and the approval of the Stock Exchange.
We will perform the following procedures for the Conversion of Unlisted Shares into H Shares after receiving the approval of the Stock Exchange: (a) giving instructions to our H Share Registrar regarding relevant share certificates of the converted H Shares; and (b) enabling the converted H Shares to be accepted as eligible securities by HKSCC for deposit, clearance and settlement in the CCASS.
The Conversion of Unlisted Shares into H Shares will involve an aggregate of 178,282,205 Unlisted Shares (taking into account the Share Subdivision) held by 57 existing Shareholders, representing approximately 44.26% of total issued Shares of the Company as of the Latest Practicable Date and approximately 40.50% of total issued Shares of the Company upon completion of the Conversion of Unlisted Shares into H Shares and the Global Offering (assuming the Over-allotment Option is not exercised).
In accordance with the Guidelines for the "Full Circulation" Program for Domestic Unlisted Shares of H-share Listed Companies (《H股公司境內未上市股份申請「全流通」業務指引》) and the Overseas Listing Trial Measures announced by the CSRC, H-share listed companies which apply for the conversion of domestic shares and unlisted foreign shares into H shares for listing and circulation on the Stock Exchange shall file the application with the CSRC according to the administrative filing procedures necessary for the Overseas Listing Trial Measures. An H-share listed company may apply for a "Full Circulation" separately or when applying for refinancing overseas. An unlisted domestic joint stock company may apply for "Full Circulation" when applying for an overseas initial public offering.
We have filed with the CSRC for, and received the filing notice from the CSRC dated December 15, 2025 in relation to the Global Offering and the conversion of 178,282,205 Unlisted Shares (taking into account the Share Subdivision) into H Shares on a one-for-one basis upon Listing.
We have applied to the Stock Exchange for the approval for the granting of listing of, and permission to deal in, our H Shares to be issued pursuant to the Global Offering (including any H Shares which may be issued pursuant to the exercise of the Over-allotment Option) and the H Shares to be converted from 178,282,205 Unlisted Shares (taking into account the Share Subdivision) on the Stock Exchange, which is subject to the approval by the Stock Exchange.
The PRC Company Law provides that in relation to the public offering of a company, the shares issued prior to the public offering shall not be transferred within a period of one year from the date on which the publicly offered shares are listed on any stock exchange. Accordingly, Shares issued by our Company prior to the Global Offering shall be subject to such statutory restriction and not be transferred within a period of one year from the Listing Date. See "History, Development and Corporate Structure—Pre-IPO Investments."
For details of the lock-up undertaking given pursuant to the PRC Company Law and the Listing Rules, see "History Development and Corporate Structure—Lock-up Periods" and "Underwriting—Underwriting Arrangements and Expenses—Undertakings to the Stock Exchange pursuant to the Listing Rules."
According to the Guidelines for the "Full Circulation" Program for Domestic Unlisted Shares of H-share Listed Companies, the domestic shareholders of unlisted shares shall handle share transfer registration in accordance with the relevant rules of China Securities Depository and Clearing Corporation Limited (中國證券登記結算有限責任公司) (the "CSDC"), and H-share companies should submit relevant status reports to the CSRC within 15 days after the shares involved in the application completing the transfer registration in the CSDC.
Subject to the completion of the Global Offering, pursuant to the Shareholders resolutions of the Company, our Directors have been granted general unconditional mandates to allot, issue Shares, or sell and/or transfer Shares out of treasury that are held as treasury shares. See "Statutory and General Information—A. Further Information about Our Group—5. Resolutions of our Shareholders" in Appendix VI to this prospectus for further details.
The following discussion and analysis of our financial condition and results of operations should be read in conjunction with our audited consolidated financial statements and the accompanying notes included in the Accountants' Report set out in Appendix I to this prospectus. We have prepared our financial information in accordance with IFRS Accounting Standards as issued by the International Accounting Standards Board, which may differ in certain material aspects from generally accepted accounting principles in other jurisdictions, including the United States.
The following discussion contains forward-looking statements that reflect our current view with respect to future events and financial performance. These statements are based on our assumptions and analysis in light of our experience and perception of historical trends, current conditions and expected future developments, as well as factors that we believe are appropriate under the circumstances. However, our actual results could differ materially from those anticipated in these forward-looking statements as a result of various factors. Factors that could cause or contribute to such differences include, without limitation, those discussed in the sections headed "Forward-Looking Statements," "Risk Factors" and "Business" in this prospectus.
我们是中国领先的人工智能公司，致力于开发通用大模型。我们于2019年成立，秉承在中国探索通用人工智能（AGI）创新的宏大理想。我们在前沿研究的各个领域持续取得突破性技术成果，并稳步扩大其商业应用规模，实现了营业收入的快速增长。截至2025年6月30日，我们的模型已赋能逾八千家机构客户，截至最后实际可行日期，已赋能约8,000万台设备。根据弗若斯特沙利文的数据，按2024年收入计算，我们在中国通用大模型独立开发商中排名第一，在所有开发商中排名第二。
We are a leading AI company in China, dedicated to developing general-purpose large models. We were founded in 2019 on the bold idea of pursuing innovation toward artificial general intelligence (AGI) in China. We have solidly delivered breakthrough technology across the full spectrum of cutting-edge research and steadily scaled up its commercial application to achieve fast growth in revenue. Our models had empowered over eight thousand institutional customers as of June 30, 2025 and approximately 80 million devices as of the Latest Practicable Date. According to Frost & Sullivan, we ranked first among China's independent developers and second among all developers of general-purpose large models in terms of revenue in 2024.
As we commercialize our technology to seize the tremendous market opportunity presented by advance AI, we organize our offerings around our all-in-one MaaS platform. Through this product development and commercialization platform, we deliver intelligence to enterprise clients, developers and end customers in the most suitable, sensible and scalable way despite the great heterogeneity in computing infrastructure, devices and applications. Our MaaS platform offers flexible custom model deployment options to meet the diverse needs of businesses while maintaining efficiency, scalability and data security. We primarily offer two deployment approaches — cloud-based and on-premise deployment. We achieved significant growth in revenue during the Track Record Period. In 2022, 2023 and 2024 and for the six months ended June 30, 2024 and 2025, our revenue was RMB57.4 million (人民币5,740万元), RMB124.5 million (人民币1.245亿元), RMB312.4 million (人民币3.124亿元), RMB44.9 million (人民币4,490万元) and RMB190.9 million (人民币1.909亿元), respectively.
Our results of operations and financial conditions have been, and are expected to continue to be affected by a number of factors, including those key factors set out below:
As a leading AI company, we are committed to enhancing the capabilities of the AI models and agents in our comprehensive AI suite, innovating new ones, and improving and developing efficient and cost-effective AI services for ever more industry verticals and other use cases. To achieve this, we have prioritized in-house R&D of AI technologies since our inception. Technological leadership requires sustained substantial investment in R&D. Our research and development expenses significantly increased from RMB84.4 million (人民币8,440万元) in 2022 to RMB528.9 million (人民币5.289亿元) in 2023, and further increased to RMB2,195.4 million (人民币21.954亿元) in 2024. Our research and development expenses also increased from RMB859.2 million (人民币8.592亿元) for the six months ended June 30, 2024 to RMB1,594.7 million (人民币15.947亿元) for the six months ended June 30, 2025.
Going forward, we plan to continue to invest in advancing model performance and driving progress toward more advanced AI. We will also continue to attract and retain top-tier R&D professionals, as well as specialists from a wide range of disciplines. Specifically, we plan to invest around 70.0% of our net proceeds from the Global Offering to strengthen our research and development capabilities in general purpose large AI model. For more information, see "Future Plans and Use of Proceeds" in this prospectus. Our ability to continuously develop and introduce new technologies that meet our customers' demands is subject
In particular, computing power is essential for AI companies, as it underpins both the training, inference, and continuous development of our models and agentic tools in our research and development activities. Our computing service fees paid to third party computing power provider necessary for our R&D activities amounted to RMB14.6 million, RMB311.7 million, RMB1,552.8 million, RMB603.2 million and RMB1,145.1 million, in 2022, 2023 and 2024 and for the six months ended June 30, 2024 and 2025, respectively. Controlling such costs is key to improving our operational efficiency. According to Frost & Sullivan, the future unit price of computing power is expected to decline which will help us lower both our cost of revenue and research and development expenses, supporting our efforts to reduce costs while enhancing operational effectiveness to a number of risks and uncertainties, many of which are beyond our control. For more information, see "Risk Factors—Risks Related to Our Research and Development—We have made and expect to continue to make substantial investments in R&D. If we cannot continuously invest in our R&D activities while achieving technological innovation, our business, results of operations, financial condition and prospects may be materially and adversely affected."
During the Track Record Period, we generated revenue from the provision of large model-related services through our MaaS platform, including on-premise deployment and cloud-based deployment. These services were designed to meet the diverse needs of our customers while ensuring efficiency, scalability, and data security. Our significant revenue growth during this period was driven by our continual iteration and enhancement of our models and rapid market expansion.
As our services gained greater recognition and our brand presence expanded, we were able to develop and deliver services with stronger capabilities and additional functionalities for our customers. For instance, we released GLM-4-Plus (latest model of the GLM-4 series) in August 2024, which reached SOTA performance in various benchmark tests, and significantly outperformed comparable models in long-video processing and small action understanding. In addition, building upon GLM-4 series as our base model, we further developed the reflection model (GLM-Z1) and rumination model (GLM-Z1-Rumination). Reflection model and rumination model both spend additional time "deep thinking" before generating an answer, which makes them better for complex reasoning tasks. For more information, see "Business—Our Models" in this prospectus. We believe that these enhanced capabilities will enable us to meet the needs of an even broader range of customers across more industry verticals.
Furthermore, we are also committed to continuously expanding our service portfolio across a variety of parameter scales, ensuring compatibility with a wide range of computing power and end devices. This approach enables our models to better match end devices' computational capacity and reduce hardware limitations. At the application level, we remain focused on delivering convenient, user-friendly experiences that make large model technology more accessible, enabling companies and organizations of all sizes, as well as individual users, to unlock large models' potential. Our future success largely depends on our ability to further expand our services offerings and enhance our existing ones.
Our growth depends on our ability to expand our customer base. We have formed regional and industry-focused sales strategy, through which we penetrate, expand and maintain connections in targeted markets to serve our customers. We not only seek to work closely with industry leaders, but also to forge partnerships with entities that specialize in specific industry applications. By targeting influential and
innovative leaders in the respective industry sector, we can co-develop and deliver benchmark cases to showcase the tangible benefits and capabilities of our services. We have participated in various branding and marketing activities to reach potential customers, including in-person and online events, content marketing, partner marketing, developer outreach, search engine optimization, social media and public relations, and we adjust our marketing strategies from time to time based on our management's judgment. For the years ended December 31, 2022, 2023 and 2024 and for the six months ended June 30, 2024 and 2025, our selling and marketing expenses were RMB15.1 million, RMB101.2 million, RMB387.5 million, RMB144.2 million and RMB208.6 million, respectively. The increases during the Track Record Period were primarily due to the increased advertising activities and expansion of sales and marketing team, in line with our business growth.
We have a broad and diverse client base, which expanded rapidly over the Track Record Period. Our customers include enterprises, public sector entities and individual users. We target influential and innovative leaders in the respective industry sector and co-develop and deliver benchmark cases to showcase the tangible benefits and capabilities of our services. Through collaboration with established industry application providers, we integrate the core capabilities of our models into the solutions offered by these industry partners to help enhance their products and their services for end users. For more information, see "Business—Sales and Marketing—Sales" in this prospectus.
We believe through partnerships, we would be able to extend our reach and deliver practical model capabilities and value to users across different industries and customers. Our ability to attract new customers and deepen our relationships with existing customers is a key factor to our business growth and financial conditions.
We operate in the rapidly growing AI industry. Our business, financial performance, and future growth are affected by the development of this industry, including the general factors affecting the global and China's AI market, macroeconomic conditions, regulatory environment, as well as the market acceptance, adoption and demand of large models and related services. According to Frost & Sullivan, the size of the AI market in China expanded rapidly from RMB52.4 billion in 2020 to RMB160.7 billion in 2024, representing a CAGR of 32.3%.
Furthermore, the size of the AI market in China is expected to grow further to RMB993.0 billion in 2030, with a CAGR of 35.5%. According to Frost & Sullivan, by 2030, AI will empower at least 20% of daily business decision-making worldwide and enable at least 80% of consumer mainstream smart devices globally, achieving over US$20.0 trillion in value.
According to Frost & Sullivan, we were the largest independent vendor among all LLMs vendors in China by revenue in 2024. Competition within the AI market significantly influences our financial performance. We believe that our competitive advantage is derived from the scope, performance and safety of our service offerings, user experience, our R&D capabilities and our talent. However, intensified competition or an inability to sustain this advantage may adversely affect our operational performance. For more information, see "Risk Factors—Risks Related to Our Commercialization—We may not be able to compete effectively against current or future competitors."
Our business and operating results are also affected by general factors affecting the AI industry, which include:
• relevant laws and regulations, governmental policies and initiatives.
Any changes in any of these general factors may have a material impact on our business operations and results of operations.
We were incorporated in China on June 11, 2019 as a limited liability company. In March 2025, we were converted from a limited liability company into a joint stock limited liability company.
Our historical financial information has been prepared in accordance with all applicable IFRS Accounting Standards as issued by the International Accounting Standards Board ("IASB"). Further details of the material accounting policy information adopted are set out in Note 2 of the Accountants' Report in Appendix I to this prospectus.
The IASB has issued a number of new and revised IFRS Accounting Standards. For the purpose of preparing this historical financial information, we have consistently applied all applicable new and revised IFRS Accounting Standards throughout the Track Record Period. We have not adopted any new and revised accounting standards and interpretations issued but not yet effective for the accounting period beginning January 1, 2024 which are set out in Note 35 of the Accountants' Report in Appendix I to this prospectus.
The historical financial information also complies with the applicable disclosure provisions of the Listing Rules.
The preparation of the Historical Financial Information in conformity with IFRS Accounting Standards requires our management to make judgments, estimates and assumptions that affect the application of policies and reported amounts of assets, liabilities, income and expenses. The estimates and associated assumptions are based on historical experience and various other factors that are believed to be reasonable under the circumstances, the results of which form the basis of making the judgments about carrying values of assets and liabilities that are not readily apparent from other sources. Actual results may differ from these estimates.
Below is a summary of the accounting policies in accordance with IFRS Accounting Standards that we believe are important to the presentation of our financial results and involve the need to make estimates and judgments about the effect of matters that are inherently uncertain. We also have other policies that we consider to be key accounting policies. We do not expect such estimates and assumptions to be likely to change significantly in the future. For more information, please see Note 2 to the Accountants' Report set out in Appendix I to this prospectus.
We classify income as revenue when it arises from the sale of goods or the provision of services in the ordinary course of our business.
We recognize revenue when control over a product or service is transferred to the customer at the amount of promised consideration to which we expect to be entitled, excluding those amounts collected on behalf of third parties such as value added tax or other sales taxes.
In determining whether we act as a principal or as an agent, it considers whether it obtains control of the product or service before they are transferred to the customers. Control refers to our ability to direct the use of and obtain substantially all of the remaining benefits from the product or service.
We principally engage in the provision of large model-related services, which offers on-premise deployment and cloud-based deployment.
Our on-premise deployment consists primarily of localized deployment of large models and all necessary on-site services to facilitate such deployment. We recognize revenue at the point of time when the large model and related services are delivered to the customer's designated location and accepted by the customer.
We also provide other related services such as model training and fine-tuning. We recognize revenue from these services upon the transfer of control, either over time or at a point in time, depending on the nature of the arrangements.
Our cloud-based deployment is provided through cloud infrastructure. We recognize revenue over the contract term. For subscription-based contract, we generally recognize revenue ratably over the contract term. For usage-based contract, we recognize revenue base on the customer's utilization of the resources when the services are rendered to the customers.
We recognize dividend income in profit or loss on the date on which our right to receive payment is established.
We recognize interest income using the effective interest method. The effective interest rate that exactly discounts estimated future cash receipts through the expected life of the financial asset to the gross carrying amount of the financial asset. In calculating interest income, the effective interest rate is applied to the gross carrying amount of the asset (when the asset is not credit-impaired). However, for financial assets that have become credit-impaired subsequent to initial recognition, interest income is calculated by applying the effective interest rate to the amortized cost of the financial asset. If the asset is no longer credit-impaired, then the calculation of interest income reverts to the gross basis.
We recognize government grants that compensate us for expenses incurred are recognized in the statement of financial position as deferred income and are recognized as a reduction of the expenses related to the grants in profit or loss on a systematic basis in the same periods in which such expenses are incurred.
A contract that contains an obligation for the Company or the Group to purchase its own equity instruments for cash or another financial asset gives rise to a financial liability even if the Company's or the Group's obligation to purchase is conditional on the counterparty exercising its right to redeem.
We initially recognize the financial instruments issued to investors and subsequently measure the financial instruments issued to investors at the present value of the redemption amount, which represents the settlement that would be triggered by the event with the highest settlement outcome. We recognize changes in the carrying amounts of the financial liabilities in profit or loss as "changes in carrying amounts of financial instruments issued to investors".
The carrying amounts of the financial instruments issued to investors are reclassified to share premium upon the termination of the counterparty's redemption right.
We designated convertible bonds as financial liabilities measured at FVPL on initial recognition. Subsequent to initial recognition, the convertible bonds are carried at fair value with changes in fair value recognized in profit or loss as changes in fair value of financial instruments measured at FVPL.
• financial assets measured at amortized cost (including cash at bank and on hand, trade and other receivables, and time deposits); and
• contract assets.
ECLs are a probability-weighted estimate of credit losses. Generally, we measure credit losses as the present value of all expected cash shortfalls between the contractual and expected amounts.
The expected cash shortfalls of trade and other receivables and contract assets are discounted using the effective interest rate determined at initial recognition or an approximation thereof if the effect of discounting is material.
The maximum period considered when estimating ECLs is the maximum contractual period over which we are exposed to credit risk.
• 12-month ECLs: these are the portion of ECLs that result from default events that are possible within the 12 months after the reporting date (or a shorter period if the expected life of the instrument is less than 12 months); and
• lifetime ECLs: these are the ECLs that result from all possible default events over the expected lives of the items to which the ECL model applies.
We measure loss allowances at an amount equal to lifetime ECLs, except for the following, which are measured at 12-months ECLs:
• other financial instruments for which credit risk (i.e. the risk of default occurring over the expected life of the financial instrument) has not increased significantly since initial recognition.
Loss allowances for trade receivables and contract assets are always measured at an amount equal to lifetime ECLs.
When determining whether the credit risk of a financial instrument has increased significantly since initial recognition and when measuring ECLs, we consider reasonable and supportable information that is
relevant and available without undue cost or effort. This includes both quantitative and qualitative information and analysis, based on our historical experience and informed credit assessment, that includes forward-looking information.
We assume that the credit risk on a financial asset has increased significantly if it is past due.
ECLs are remeasured at each reporting date to reflect changes in the financial instrument's credit risk since initial recognition. Any change in the ECL amount is recognized as an impairment gain or loss in profit or loss. We recognize an impairment gain or loss for all financial instruments with a corresponding adjustment to their carrying amounts through a loss allowance account, except for investments in non-equity securities that are measured at FVOCI (recycling).
At the end of each year during the Track Record Period, we assess whether a financial asset is credit-impaired. A financial asset is credit-impaired when one or more events that have a detrimental impact on the estimated future cash flows of the financial asset have occurred.
• the disappearance of an active market for a security because of financial difficulties of the issuer.
The gross carrying amount of a financial asset or contract asset is written off to the extent that there is no realistic prospect of recovery. This is generally the case when we determine that the debtor does not have assets or sources of income that could generate sufficient cash flows to repay the amounts subject to the write-off.
We recognize subsequent recoveries of an asset that was previously written off as a reversal of impairment in profit or loss in the period in which the recovery occurs.
If circumstances indicate that the carrying amount of a non-current asset other than financial assets may not be recoverable, the asset may be considered impaired and an impairment loss may be recognized in accordance with accounting policy for impairment of non-current assets to the Accountants Report. These assets are tested for impairment periodically or whenever the events or changes in circumstances indicate that their recorded carrying amounts may not be recoverable. When such a decline has occurred, the carrying amount is reduced to recoverable amount. The recoverable amount is the greater of the fair value less costs of disposal and value in use. In determining the value in use, expected future cash flows generated by the asset are discounted to their present value, which requires significant judgment relating to the level of revenue and amount of operating costs.
In accordance with IAS 36.12, we assess at the end of each reporting period whether there is any indication that non-current assets (other than inventories, deferred tax assets and financial assets) may be impaired. If any such indication exists, we estimate the recoverable amount of the relevant assets. Impairment tests on property and equipment, right-of-use assets, intangible assets and other non-current assets were performed at the level of cash-generating units (CGUs). In performing such tests, the carrying
The recoverable amount of each CGU is compared to its recoverable amount. The recoverable amount is determined based on the estimated future cash flows, discounted to their present value using a pre-tax discount rate that reflects current market assessments of the time value of money and the risks specific to the assets. An impairment loss in respect of goodwill is not reversed.
In accordance with IAS 36, we performed impairment tests at the end of each period on non-current assets, primarily including property, plant and equipment, right-of-use assets, intangible assets and other non-current assets at the CGU level. The recoverable amounts of these assets exceeded their respective carrying amounts at the end of each period, therefore, no impairment loss was recognized during the Track Record Period.
The following table summarizes our results of operations and as percentage of our total revenue for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | | | | | |---|---|---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | Amount | % of revenue | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | Revenue | 57,409 | 100.0 | 124,538 | 100.0 | 312,414 | 100.0 | 44,909 | 100.0 | 190,877 | 100.0 | | Cost of revenue | (26,049) | (45.4) | (44,056) | (35.4) | (136,525) | (43.7) | (22,950) | (51.1) | (95,453) | (50.0) | | Gross profit | 31,360 | 54.6 | 80,482 | 64.6 | 175,889 | 56.3 | 21,959 | 48.9 | 95,424 | 50.0 | | Other income | 1,784 | 3.1 | 9,965 | 8.0 | 19,281 | 6.2 | 4,174 | 9.3 | 4,614 | 2.4 | | Selling and marketing expenses | (15,139) | (26.4) | (101,198) | (81.3) | (387,475) | (124.0) | (144,194) | (321.1) | (208,570) | (109.3) | | General and administration expenses | (32,316) | (56.3) | (66,302) | (53.2) | (133,603) | (42.8) | (51,452) | (114.6) | (185,165) | (97.0) | | Research and development expenses | (84,377) | (147.0) | (528,884) | (424.7) | (2,195,436) | (702.7) | (859,217) | (1,913.2) | (1,594,661) | (835.4) | | Impairment losses on financial assets | (10,867) | (5.7) | (98,719) | (172.0) | (625,723) | (502.4) | (2,538,352) | (812.5) | (1,029,493) | (2,292.4) | | Loss from operations | (5,694) | (9.9) | (26,332) | (21.1) | (38,321) | (12.3) | (12,212) | (27.2) | (53,270) | (27.9) | | Finance costs | (31) | (0.1) | (19,786) | (15.9) | (17,008) | (5.4) | (763) | (1.7) | — | — | | Share of profits less losses of associates | (453) | (0.4) | 21,254 | 6.8 | 324 | 0.7 | 14,147 | 7.4 | — | — | | Changes in fair value of financial instruments measured at fair value through profit or loss ("FVPL") | 5,972 | 10.4 | 26,022 | 20.9 | 66,271 | 21.2 | 7,004 | 15.6 | 9,791 | 5.1 | | Changes in the carrying amounts of financial instruments issued to investors | (468,859) | (150.1) | (201,174) | (448.0) | (429,295) | (224.9) | (45,209) | (78.7) | (161,471) | (129.7) | | Loss before taxation | (143,650) | (250.2) | (787,957) | (632.7) | (2,958,007) | (946.8) | (1,235,551) | (2,751.2) | (2,357,852) | (1,235.3) | | Income tax | — | — | — | — | — | — | — | — | — | — | | Loss for the year | (143,650) | (250.2) | (787,957) | (632.7) | (2,958,007) | (946.8) | (1,235,551) | (2,751.2) | (2,357,852) | (1,235.3) | | Loss attributable to: | | | | | | | | | | | | Equity holders of the Company | (143,374) | (249.7) | (787,960) | (632.7) | (2,956,491) | (946.3) | (1,235,551) | (2,751.2) | (2,351,173) | (1,231.8) | | Non-controlling interests | (276) | (0.5) | 3 | 0.0 | (1,516) | (0.5) | — | — | (6,679) | (3.5) |
To supplement our consolidated financial statements that are presented in accordance with IFRS Accounting Standards, we also use adjusted loss for the year (a non-IFRS measure), as an additional financial measures which is not required by or presented in accordance with IFRS Accounting Standards. We believe that this non-IFRS measure facilitates comparisons of operating performance from period to period by eliminating potential impact of certain items. We believe that this measure provides useful information to investors and others in understanding and evaluating our consolidated financial statements in the same manner as they help our management. However, our presentation of adjusted loss for the year (a non-IFRS measure) may not be comparable to similar measures presented by other companies. The use of such non-IFRS measure has limitations as an analytical tool, and you should not consider them in isolation from, or as substitute for analysis of, our consolidated financial statements or financial condition as reported under IFRS Accounting Standards. We define adjusted loss for the year (a non-IFRS measure) as loss for the year/period adjusted for adding back equity-settled share-based compensation expenses, changes in the carrying amount of financial instruments issued to investors, and listing expenses.
| | Year Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Net loss for the year/period | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Add: | | | | | | | — Equity-settled share-based compensation expenses(1) | 1,024 | 5,502 | 23,579 | 4,217 | 158,852 | | — Changes in the carrying amount of financial instruments issued to investors(2) | 45,209 | 161,471 | 468,859 | 201,174 | 429,295 | | — Listing expense(3) | — | — | — | — | 17,731 | | Adjusted net loss for the year/period (non-IFRS measure) | (97,417) | (620,984) | (2,465,569) | (1,030,160) | (1,751,974) |
(1) Equity-settled share-based compensation expenses represented share-based compensation expenses incurred in connection with our share incentive plan. Equity-settled share-based compensation expenses are not expected to result in future cash payments. The reconciling item is non-cash and does not result in cash outflow, and the adjustment has been consistently made during the Track Record Period.
(2) We adjust changes in the carrying amount of financial instruments issued to investors because it was non-cash in nature. We recognized the financial instruments at present value of financial instruments, with changes in such carrying amounts being booked in profit or loss, arising from redemption rights issued to Pre-IPO Investors. These redemption rights issued will be terminated and converted into equity upon the Global Offering.
(3) Listing expenses represent professional fees, underwriting commission and fees incurred in connection with the Listing and the Global Offering.
During the Track Record Period, we derived our revenue from providing large model services through our MaaS platform. This platform offers customers access to a matrix of models and a suite of agentic tools with flexible deployment options, including on-premise deployment and cloud-based deployment. During the Track Record Period, we generated substantially all of our revenue from mainland China.
The following table sets forth a breakdown of revenue by segment both in absolute amount and as a percentage of our total revenue for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | |---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | On-premise deployment | 54,815 | 95.5% | 112,614 | 90.4% | 263,930 | 84.5% | 26,806 | 59.7% | 161,777 | 84.8% | | Cloud-based deployment | 2,594 | 4.5% | 11,924 | 9.6% | 48,484 | 15.5% | 18,103 | 40.3% | 29,100 | 15.2% | | Total | 57,409 | 100.0% | 124,538 | 100.0% | 312,414 | 100.0% | 44,909 | 100.0% | 190,877 | 100.0% |
On-premise deployment refers to the provision of large model services according to the customers' specific instructions and needs on the customers' infrastructure. This service is designed to meet the commercial needs of customers, such as large private enterprises and public sector entities, who have stringent requirements around data security, regulatory compliance and system customization. This approach allows customers to use our AI models in their specific domains, with the peace of mind that their sensitive data remain secure and private. Unlike our on-premise deployment, where we provide localized services at the client's designated site pursuant to detailed service agreements, our cloud-based deployment services are offered on a subscription or usage basis through cloud infrastructure. When registering for our cloud-based services, users are not required to disclose their respective industries, and we do not proactively collect or retain such information in the course of service usage. As a result, we are unable to determine the industries in which users of our cloud-based services operate.
The following table sets forth a breakdown of revenue derived from on-premise deployment by industry vertical both in absolute amount and as a percentage of our total revenue derived from on-premise deployment for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | |---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | (RMB in thousands, except for percentages) | | | | | | | | | | Internet and technology | 30,258 | 55.2% | 76,177 | 67.6% | 130,956 | 49.6% | 11,598 | 43.3% | 61,924 | 38.3% | | Public services | 19,752 | 36.0% | 10,267 | 9.1% | 57,967 | 22.0% | 693 | 2.6% | 47,533 | 29.4% | | Telecommunications | 1,504 | 2.7% | 21,742 | 19.3% | 42,731 | 16.2% | 13,748 | 51.3% | 22,045 | 13.6% | | Traditional corporate(1) | 2,198 | 4.0% | 939 | 0.8% | 26,559 | 10.1% | 64 | 0.2% | 18,679 | 11.5% | | Consumer electronics | 915 | 1.7% | 241 | 0.2% | 3,405 | 1.3% | 42 | 0.2% | 9,177 | 5.7% | | Others(2) | 188 | 0.4% | 3,248 | 3.0% | 2,312 | 0.8% | 661 | 2.4% | 2,419 | 1.5% | | Total | 54,815 | 100.0% | 112,614 | 100.0% | 263,930 | 100.0% | 26,806 | 100.0% | 161,777 | 100.0% |
Notes: (1) Includes financial services, manufacturing and energy sectors. (2) Includes retail, media and consulting sectors.
Starting from 2024, our large model on-premise deployment services have generated revenue from overseas customers, primarily from customers in Southeast Asia. The following table sets forth a breakdown of revenue derived from on-premise deployment by geographical location of our customers both in absolute amount and as a percentage of our total revenue derived from on-premise deployment for the years/periods indicated.
| | Years Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands, except for percentages) | | On-premise deployment | | | | | | | Mainland China | 54,815 | 100.0% | 112,614 | 100.0% | 262,479 | 99.5% | 26,806 | 100.0% | 142,990 | 88.4% | | Southeast Asia(1) | — | — | — | — | 1,080 | 0.4% | — | — | 17,927 | 11.1% | | Hong Kong | — | — | — | — | 371 | 0.1% | — | — | — | — | | Others(2) | — | — | — | — | — | — | — | — | 860 | 0.5% | | **Total** | **54,815** | **100.0%** | **112,614** | **100.0%** | **263,930** | **100.0%** | **26,806** | **100.0%** | **161,777** | **100.0%** |
Notes: (1) Includes Malaysia and Singapore. (2) Includes the United States.
Cloud-based deployment allows customers to access our large portfolio of foundation models, agents and tools. This approach is particularly sensible for businesses seeking agility and ease of implementation. Cloud-based deployment helps accelerate our customers' AI adoption process by minimizing up-front investment and offering flexible, secure and efficient access to cutting-edge model capabilities across diverse industry use cases.
During the Track Record Period, our cost of revenue consisted of (i) payroll cost, mainly representing the wages and benefits of service personnel involved in the deployment and maintenance of our services; (ii) computing service fee paid to providers of computing power necessary for providing our services, which includes the cost of computing power used for model training and fine-tuning prior to delivery of on-premise deployment to clients, as well as the cost of computing power used in providing services to clients through cloud-based deployment. In particular, for services provided through cloud-based deployment, our clients access and invoke our proprietary large models hosted on cloud infrastructure based on their specific needs, and each such invocation consumes computing power to process and analyze the relevant inputs and generate the corresponding outputs, resulting in computing service fees primarily representing amounts paid to third-party cloud service providers for computing resources that support the training, hosting and inference of our models; (iii) provision for warranty; (iv) depreciation and amortization; (v) technical service and consultation fees as we outsourced to third-party service providers certain standard labor intensive tasks to save costs, such as data annotation; (vi) tax and surcharges; and (vii) others.
The following table sets forth a breakdown of our cost of revenue by nature both in absolute amount and as a percentage of our total cost of revenue for the years/periods indicated.
| | 2022 | | Year Ended December 31, 2023 | | 2024 | | Six Months Ended June 30, 2024 | | 2025 | | |---|---|---|---|---|---|---|---|---|---|---| | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | Payroll cost | 16,434 | 63.1% | 15,375 | 34.9% | 74,262 | 54.4% | 10,832 | 47.2% | 37,628 | 39.4% | | Computing service fee | — | — | 11,996 | 27.2% | 30,173 | 22.1% | 8,451 | 36.8% | 35,845 | 37.6% | | Provision for warranty | 2,101 | 8.1% | 4,522 | 10.3% | 12,171 | 8.9% | 1,672 | 7.3% | 6,670 | 7.0% | | Depreciation and amortization | 1,912 | 7.3% | 6,436 | 14.6% | 6,769 | 5.0% | 549 | 2.4% | 2,074 | 2.2% | | Technical service and consultation fees | 4,553 | 17.5% | 623 | 1.4% | 9,126 | 6.7% | 356 | 1.6% | 6,165 | 6.5% | | Taxes and surcharges | 57 | 0.2% | 1,713 | 3.9% | 541 | 0.4% | 361 | 1.6% | 1,109 | 1.2% | | Others | 992 | 3.8% | 3,391 | 7.7% | 3,483 | 2.5% | 729 | 3.1% | 5,962 | 6.1% | | Total | 26,049 | 100.0% | 44,056 | 100.0% | 136,525 | 100.0% | 22,950 | 100.0% | 95,453 | 100.0% |
The following table sets forth a breakdown of our cost of revenue by segment in absolute amount and as a percentage of our total cost of revenue for the years/periods indicated:
| | 2022 | | Year Ended December 31, 2023 | | 2024 | | Six Months Ended June 30, 2024 | | 2025 | | |---|---|---|---|---|---|---|---|---|---|---| | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | On-premise deployment | 25,429 | 97.6% | 35,833 | 81.3% | 89,674 | 65.7% | 10,030 | 43.7% | 66,237 | 69.4% | | Cloud-based deployment | 620 | 2.4% | 8,223 | 18.7% | 46,851 | 34.3% | 12,920 | 56.3% | 29,216 | 30.6% | | Total | 26,049 | 100.0% | 44,056 | 100.0% | 136,525 | 100.0% | 22,950 | 100.0% | 95,453 | 100.0% |
The following table sets forth a breakdown of our gross profit and gross profit margin for the years/periods indicated.
| | Year Ended December 31, | | | | Six Months Ended June 30, | | | | |---|---|---|---|---|---|---|---|---| | | 2022 | | 2023 | | 2024 | | 2024 | | 2025 | | | | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | Gross Profit | Gross Profit Margin | | | (RMB in thousands, except for percentages) | | | | | | (Unaudited) | | | | | On-premise deployment | 29,386 | 53.6% | 76,781 | 68.2% | 174,256 | 66.0% | 16,776 | 62.6% | 95,540 | 59.1% | | Cloud-based deployment | 1,974 | 76.1% | 3,701 | 31.0% | 1,633 | 3.4% | (116) | (0.4%) | 5,183 | 28.6% | | Total | 31,360 | 54.6% | 80,482 | 64.6% | 175,889 | 56.3% | 21,959 | 48.9% | 95,424 | 50.0% |
The level of our overall gross profit margin is affected by the variety of our service offerings. During the Track Record Period, our on-premise deployment had notably higher gross profit margins while our cloud-based deployment had relatively lower gross profit margins.
Specifically, on-premise deployment is typically provided to large private enterprises and public sector customers who have pressing requirements for data security, regulatory compliance, and extensive customisation. These customers are often willing and able to pay higher prices for dedicated deployment, dedicated model services and comprehensive on-site support. Cloud-based deployment enables the customers to deploy AI solutions quickly and cost effectively, without the need for costly local infrastructure.
During the Track Record Period, our other income comprised (i) interest income from bank deposits; (ii) net (loss)/gain on disposal of property and equipment and intangible assets; and (iii) others.
The following table sets forth a breakdown of other income for the years/periods indicated.
| | Year Ended December 31, | Year Ended December 31, | Year Ended December 31, | Six Months Ended June 30, | Six Months Ended June 30, | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Interest income | 1,426 | 11,236 | 13,406 | 4,163 | 4,607 | | Net (loss)/gain on disposal of property and equipment and intangible assets | — | (1,539) | 6,807 | — | (15) | | Others | 358 | 268 | (932) | 11 | 22 | | **Total** | **1,784** | **9,965** | **19,281** | **4,174** | **4,614** |
During the Track Record Period, our research and development expenses consisted of (i) computing service fees paid to third party providers for the computing power used in our research and development activities; (ii) payroll cost, mainly representing the wages and benefits of our research and development personnel, including the share-based payment expenses; (iii) depreciation and amortization; (iv) technical service and consultation fees, mainly representing fees for outsourced R&D activities, such as data cleansing, large model evaluation and other activities; and (v) others.
The following table sets forth a breakdown of research and development expenses both in absolute amount and as a percentage of our total research and development expenses for the years/periods indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands, except for percentages) | | | (Unaudited) | | | Computing service fees | 14,633 | 17.3% | 311,746 | 58.9% | 1,552,832 | 70.7% | 603,237 | 70.2% | 1,145,124 | 71.8% | | Payroll cost | 41,982 | 49.8% | 137,296 | 26.0% | 324,193 | 14.8% | 107,988 | 12.6% | 265,567 | 16.7% | | Depreciation and amortization | 14,689 | 17.4% | 49,206 | 9.3% | 259,038 | 11.8% | 118,517 | 13.8% | 123,966 | 7.8% | | Technical service and consultation fees | 11,412 | 13.5% | 25,184 | 4.8% | 45,036 | 2.1% | 24,670 | 2.9% | 52,209 | 3.3% | | Others(1) | 1,661 | 2.0% | 5,452 | 1.0% | 14,337 | 0.6% | 4,805 | 0.5% | 7,795 | 0.4% | | Total | 84,377 | 100.0% | 528,884 | 100.0% | 2,195,436 | 100.0% | 859,217 | 100.0% | 1,594,661 | 100.0% |
Note: (1) Others mainly included utilities, offices and transportation and travel expenses associated with our research and development activities.
We are committed to capitalizing on the rapid expansion of the AI industry. To achieve this goal, we have been continually updating our foundation models and investing in more advanced model training infrastructure. In parallel, we are exploring new application scenarios to increase our models' compatibility to better match end devices' computational capacity and reduce hardware limitations.
All these activities require us to run frequent and large-scale model training and retraining. We are constantly experimenting with larger models, richer datasets and more frequent iterations which need substantially more computing power from third party suppliers.
We recognize expenditure on research and development activities as research and development expenses in profit or loss when it incurred. Such expenditure is capitalized only when (i) the cost can be measured reliably, (ii) the related products and/or services are both technically and commercially feasible, (iii) there is a high probability of future economic benefits of such products and/or services, and (iv) we intend to complete the development and have sufficient resources to do so, as well as to use or sell the resulting products or services. Once capitalized, development expenditure is carried at cost, less any accumulated amortization and impairment losses. The Company did not capitalize any research and development expenses during the Track Record Period.
During the Track Record Period, our selling and marketing expenses mainly represented (i) advertising and marketing expenses; (ii) payroll cost, mainly representing the wages and benefits of our sales personnel, including the share-based payment expenses; (iii) technical service and consultation fees, mainly representing outsourcing service fees, market research expenses, translation fees, bidding fees and other fees in relation to our sales and distribution activities; (iv) computing service fees paid to third party computing power providers used in our sales and promotion activities; (v) depreciation and amortization; and (vi) others.
The following table sets forth a breakdown of selling and marketing expenses both in absolute amount and as a percentage of our total selling and marketing expenses for the years/periods indicated.
| | 2022 | | Year Ended December 31, 2023 | | 2024 | | Six Months Ended June 30, 2024 (Unaudited) | | 2025 | | |--|--|--|--|--|--|--|--|--|--|--| | | (RMB in thousands, except for percentages) | | Advertising and marketing expenses | 1,167 | 7.7% | 54,868 | 54.2% | 237,453 | 61.3% | 80,713 | 56.0% | 95,675 | 45.9% | | Payroll cost | 11,378 | 75.2% | 30,334 | 30.0% | 93,695 | 24.2% | 44,493 | 30.9% | 71,744 | 34.4% | | Technical service and consultation fees | 210 | 1.4% | 3,767 | 3.7% | 21,963 | 5.7% | 9,061 | 6.3% | 8,778 | 4.2% | | Computing service fees | — | — | 1,852 | 1.8% | 12,840 | 3.3% | 2,287 | 1.6% | 20,564 | 9.9% | | Depreciation and amortization | 1,705 | 11.3% | 5,615 | 5.5% | 6,054 | 1.6% | 2,956 | 2.1% | 4,194 | 2.0% | | Others(1) | 679 | 4.4% | 4,762 | 4.8% | 15,470 | 3.9% | 4,684 | 3.1% | 7,615 | 3.6% | | Total | 15,139 | 100.0% | 101,198 | 100.0% | 387,475 | 100.0% | 144,194 | 100.0% | 208,570 | 100.0% |
Note: (1) Others mainly included general office expenses, transportation and travel expenses in relation to selling and marketing activities, and market research expenses.
During the Track Record Period, our general and administrative expenses consisted of (i) payroll cost, mainly representing the wages and benefits of our general and administrative personnel, including the share-based payments; (ii) professional service fees in relation to our general and administrative activities; (iii) depreciation and amortization; (iv) transportation and travel expenses in relation to our general and administrative activities; and (v) others.
The following table sets forth a breakdown of general and administrative expenses both in absolute amount and as a percentage of our total general and administrative expenses for the years/periods indicated.
| | Year Ended December 31, 2022 | | Year Ended December 31, 2023 | | Year Ended December 31, 2024 | | Six Months Ended June 30, 2024 (Unaudited) | | Six Months Ended June 30, 2025 | | |--|--|--|--|--|--|--|--|--|--|--| | | (RMB in thousands, except for percentages) | | Payroll cost | 20,923 | 64.7% | 33,956 | 51.2% | 78,369 | 58.7% | 29,041 | 56.4% | 124,655 | 67.3% | | Professional service fees | 2,835 | 8.8% | 11,931 | 18.0% | 14,845 | 11.1% | 6,608 | 12.8% | 28,345 | 15.3% | | Depreciation and amortization | 1,712 | 5.3% | 7,674 | 11.6% | 8,076 | 6.0% | 5,656 | 11.0% | 8,802 | 4.8% | | Transportation and travel expenses | 2,583 | 8.0% | 3,914 | 5.9% | 5,759 | 4.3% | 2,559 | 5.0% | 3,516 | 1.9% | | Others(1) | 4,263 | 13.2% | 8,827 | 13.3% | 26,554 | 19.9% | 7,588 | 14.8% | 19,847 | 10.7% | | Total | 32,316 | 100.0% | 66,302 | 100.0% | 133,603 | 100.0% | 51,452 | 100.0% | 185,165 | 100.0% |
Note: (1) Others mainly included utilities, offices and property management expenses associated with our general and administrative activities.
Impairment losses on financial assets mainly represented provision of trade and other receivables and contract assets. We recognized impairment loss on trade and other receivables and contract assets of RMB0.03 million, RMB19.8 million, RMB17.0 million, RMB0.8 million and RMB10.9 million in 2022, 2023 and 2024 and for the six months ended June 30, 2024 and 2025, respectively.
During the Track Record Period, our finance costs consisted of (i) interest on lease liabilities; (ii) interest on bank loans; (iii) transaction costs in relation to the financial instruments issued to investors; and (iv) foreign currency exchange loss/(gain), net. The following table sets forth a breakdown of finance costs in absolute amount for the years/period indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Interest on lease liabilities | 788 | 7,498 | 30,931 | 11,840 | 17,983 | | Interest on bank loans | — | — | 1,716 | — | 1,807 | | Transaction costs in relation to the financial instruments issued to investors | 4,906 | 16,488 | 20,119 | 5,185 | 31,975 | | Foreign currency exchange loss/(gain), net | — | 2,346 | (14,445) | (4,813) | 1,505 | | Total | 5,694 | 26,332 | 38,321 | 12,212 | 53,270 |
During the Track Record Period, our share of profits less losses of associates represented our share in the profits or losses of our associates. We recognized losses of associates of RMB0.5 million in 2023, and profits of associates of RMB21.3 million in 2024. We recognized profits of associates of RMB0.3 million and RMB14.1 million for the six months ended June 30, 2024 and 2025, respectively.
During the Track Record Period, our changes in fair value of financial instruments measured at FVPL represented changes in fair value of (i) our equity investment of certain unlisted entities; and (ii) our investments of wealth management products. We recognized changes in fair value of financial instruments measured at FVPL of RMB6.0 million, RMB26.0 million, RMB66.3 million, RMB7.0 million and RMB9.8 million in 2022, 2023 and 2024, and for the six months ended June 30, 2024 and 2025, respectively.
During the Track Record Period, we recognized the financial instruments issued to Pre-IPO Investors as financial liabilities, because they were granted redemption rights to require us to redeem all of the
instruments upon certain redemption or liquidation events. For more details, please see "History, Development and Corporate Structure—Pre-IPO Investments" in this prospectus. We expect to continue to recognize such redemption rights held by Pre-IPO Investors for the period from December 31, 2024 to the date when the redemption rights held by our Pre-IPO Investors were terminated, and we do not expect to recognize changes in the carrying amount of financial instruments issued to investors thereafter. For more details, see Note 26 of the Accountants' Report set forth in Appendix I to this prospectus. We recognized losses in the changes in the carrying amount of financial instruments issued to investors of RMB45.2 million, RMB161.5 million, RMB468.9 million, RMB201.2 million and RMB429.3 million in 2022, 2023 and 2024 and for the six months ended June 30, 2024 and 2025, respectively.
Our total revenue increased by 325.0%, from RMB44.9 million for the six months ended June 30, 2024 to RMB190.9 million for the six months ended June 30, 2025, in line with our overall business expansion and increase in the sales volume of both on-premise deployment and cloud-based deployment.
Our revenue derived from on-premise deployment increased by 503.5%, from RMB26.8 million for the six months ended June 30, 2024 to RMB161.8 million for the six months ended June 30, 2025 as we continued to provide more complex large model services to our customers and to extend our large model services to additional customers.
Our revenue derived from cloud-based deployment increased by 60.7%, from RMB18.1 million for the six months ended June 30, 2024 to RMB29.1 million for the six months ended June 30, 2025, primarily resulted from an increase in the number of customers and application scenarios of our service offerings.
Our cost of revenue increased by 315.9%, from RMB23.0 million for the six months ended June 30, 2024 to RMB95.5 million for the six months ended June 30, 2025. The increase was in line with our business expansion and revenue growth. Specifically, the increase was primarily due to (i) an increase in computing service fees of RMB27.4 million, which was in line with our business expansion and revenue growth, (ii) an increase in technical service and consultation fees of RMB5.8 million, and (iii) an increase in payroll cost of RMB26.8 million, in line with our business expansion.
Our gross profit increased by 334.6%, from RMB22.0 million for the six months ended June 30, 2024 to RMB95.4 million for the six months ended June 30, 2025. Our gross profit margin increased from 48.9% for the six months ended June 30, 2024 to 50.0% for the six months ended June 30, 2025.
Our gross profit for on-premise deployment increased by 469.5% from RMB16.8 million for the six months ended June 30, 2024 to RMB95.5 million for the six months ended June 30, 2025. Our gross profit margin for on-premise deployment remained relatively stable at 62.6% for the six months ended June 30, 2024 and 59.1% for the six months ended June 30, 2025.
Our gross profit for cloud-based deployment was RMB5.2 million for the six months ended June 30, 2024, while we recorded gross loss of RMB0.1 million for the six months ended June 30, 2025. Our gross profit margin for cloud-based deployment was 28.6% for the six months ended June 30, 2024, while we recorded gross loss margin of 0.4% for the six months ended June 30, 2025, primarily due to the decrease in price of cloud-based deployment in line with the market trend, and our strategy to rapidly gain market share with competitive price.
Our other income increased from RMB4.2 million for the six months ended June 30, 2024 to RMB4.6 million for the six months ended June 30, 2025, mainly due to an increase in interest income of RMB0.4 million associated with our cash and cash equivalent balance.
Our research and development expenses increased by 85.6%, from RMB859.2 million for the six months ended June 30, 2024 to RMB1,594.7 million for the six months ended June 30, 2025, primarily due to (i) an increase in computing service fees of RMB541.9 million, paid to third party computing power providers as we devote significant efforts to iterate our foundation models, and investing in more advanced model training infrastructure; and (ii) an increase in payroll cost of RMB157.6 million, in line with our business expansion.
Our selling and marketing expenses increased by 44.6%, from RMB144.2 million for the six months ended June 30, 2024 to RMB208.6 million for the six months ended June 30, 2025, primarily due to an increase in payroll cost of RMB27.3 million, in line with our business expansion.
Our general and administrative expenses increased from RMB51.5 million for the six months ended June 30, 2024 to RMB185.2 million for the six months ended June 30, 2025, primarily due to an increase in payroll cost of RMB95.6 million, in line with our business expansion.
Our impairment losses on financial assets increased from RMB0.8 million for the six months ended June 30, 2024 to RMB10.9 million for the six months ended June 30, 2025, primarily due to an increase in trade receivables and contract asset in line with our business expansion.
Our finance costs increased from RMB12.2 million for the six months ended June 30, 2024 to RMB53.3 million for the six months ended June 30, 2025, primarily due to an increase in the transaction costs on issuance of financial instruments to investors of RMB26.8 million associated with the equity financing.
Our share of profit less losses of associates increased from RMB0.3 million for the six months ended June 30, 2024 to RMB14.1 million for the six months ended June 30, 2025, primarily due to the operations and financial performance of our associates in each period.
Changes in fair value of financial instruments measured at FVPL increased from RMB7.0 million for the six months ended June 30, 2024 to RMB9.8 million for the six months ended June 30, 2025, primarily due to an increase in the fair value of the unlisted entity.
Our changes in the carrying amounts of financial instruments issued to investors increased from RMB201.2 million for the six months ended June 30, 2024 to RMB429.3 million for the six months ended June 30, 2025. The increase was primarily due to the additional equity financing in 2025.
As a result of the foregoing, our loss for the period increased from RMB1,235.6 million for the six months ended June 30, 2024 to RMB2,357.9 million for the six months ended June 30, 2025.
Our total revenue increased by 150.9%, from RMB124.5 million in 2023 to RMB312.4 million in 2024, in line with our overall business expansion and increase in the sales volume of both on-premise deployment and cloud-based deployment.
Our revenue derived from on-premise deployment increased by 134.4%, from RMB112.6 million in 2023 to RMB263.9 million in 2024. Since 2023, there were iterations of our models which resulted in improvements in their versatility, performance and sophistication. This enabled us to provide more complex large model services to our customers and to extend our large model services to customers in a broader range of industry sectors. In addition, as an industry leader with deep expertise in the AI sector, we enjoy broad industry recognition and a solid market share. As the market expanded and demand accelerated, we believe that the advantages of our leading position became even more prominent, enabling us to secure orders more effectively.
Our revenue derived from cloud-based deployment increased by 306.6%, from RMB11.9 million in 2023 to RMB48.5 million in 2024. Our cloud-based deployment has similarly benefited from model improvement and industry expansion, which resulted in an increase in the number of customers and application scenarios of our service offerings.
Our cost of revenue increased by 209.9%, from RMB44.1 million in 2023 to RMB136.5 million in 2024. The increase was in line with our business expansion and revenue growth in 2024. Specifically, the increase was primarily due to (i) an increase in payroll cost of RMB58.9 million as we recruited more service personnel and incurred more share-based payment expenses, and (ii) an increase in computing service fees of RMB18.2 million, which was in line with our business expansion and revenue growth.
Our gross profit increased by RMB95.4 million, or 118.5%, from RMB80.5 million in 2023 to RMB175.9 million in 2024. Our gross profit margin decreased from 64.6% in 2023 to 56.3% in 2024.
本地部署的毛利润从2023年的7,680万元人民币增长127.0%至2024年的1.743亿元人民币。本地部署的毛利润率保持相对稳定，2023年为68.2%，2024年为66.0%。
云端部署的毛利润从2023年的370万元人民币下降55.9%至2024年的160万元人民币。云端部署的毛利润率从31.0%下降至3.4%，原因是我们持续战略性地下调服务价格，顺应市场趋势。
我们的其他收入从2023年的1,000万元人民币增至2024年的1,930万元人民币，主要由于处置物业、设备及无形资产的净收益增加。
研发费用增长315.1%，从2023年的5.289亿元人民币增至2024年的21.954亿元人民币。增长的主要原因如下：（i）薪酬成本增加1,869万元人民币，因为我们扩大了研发团队规模，并产生了更多股权激励支出；（ii）向第三方算力提供商支付的算力服务费增加12.411亿元人民币，因为我们投入大量资源持续迭代基础模型，并在更先进的模型训练基础设施方面加大投入；（iii）折旧及摊销增加2.098亿元人民币，主要由于用于研发用途的电子设备（如计算硬件）数量增加以及使用权资产增加所致。
销售及营销费用增长282.9%，从2023年的1.012亿元人民币增至2024年的3.875亿元人民币。增长的主要原因如下：（i）薪酬成本增加634万元人民币，因为我们扩大了销售及营销团队规模，并产生了更多股权激励支出；（ii）广告及营销费用增加1.826亿元人民币，因为我们战略性地加大广告投入，以迅速把握新兴市场机遇；（iii）技术服务及咨询费用增加182万元人民币，与业务扩张规模相符。
一般及行政费用增长101.5%，从2023年的663万元人民币增至2024年的1.336亿元人民币。增长的主要原因是薪酬成本增加444万元人民币，因为我们招募了更多一般及行政人员，并产生了更多股权激励支出。
金融资产减值损失减少14.0%，从2023年的1,980万元人民币降至2024年的1,700万元人民币，原因是2023年存在其他应收款项的一次性核销。
我们的财务费用从2023年的2,630万元人民币增至2024年的3,830万元人民币，主要由于租赁负债利息大幅增加2,340万元人民币，部分被外币汇兑收益增加1,680万元人民币所抵消。
We recorded share of profits less losses of associates of RMB0.5 million in 2023, and share of profits less losses of associates of RMB21.3 million in 2024, primarily due to the operations and financial performance of our associates in each year.
Changes in fair value of financial instruments measured at FVPL significantly increased from RMB26.0 million in 2023 to RMB66.3 million in 2024, primarily due to the increase in the fair value of some investee companies and our procurement of wealth management products.
Changes in the carrying amount of financial instruments issued to investors increased by 190.4% from RMB161.5 million in 2023 to RMB468.9 million in 2024, mainly driven by the completion of additional equity financing.
As a result of the above factors, our loss for the year increased by 275.4%, from RMB788.0 million in 2023 to RMB2,958.0 million in 2024.
Our total revenue increased by 116.9%, from RMB57.4 million in 2022 to RMB124.5 million in 2023, in line with our overall business expansion and increase in the sales volume of both on-premise deployment and cloud-based deployment of our large models.
Our revenue derived from on-premise deployment increased by 105.4%, from RMB54.8 million in 2022 to RMB112.6 million in 2023. Since 2023, there were iterations of our models which resulted in improvements in their versatility, performance and sophistication. This enabled us to provide more complex large model services to our customers and to extend our large model services to customers in a broader range of industry sectors. In addition, as the market expanded and demand accelerated, we believe that the advantages of our leading position became even more prominent, enabling us to secure orders more effectively.
Our revenue derived from cloud-based deployment increased by 359.7% from RMB2.6 million in 2022 to RMB11.9 million in 2023, primarily due to model improvement and industry expansion.
Our cost of revenue increased by 69.1%, from RMB26.0 million in 2022 to RMB44.1 million in 2023. The increase in cost of revenue was in line with business expansion and revenue growth in 2023. Specifically, the increase was primarily due to an increase in computing service fees of RMB12.0 million, as we processed more requests from a broader customer base and addressed more demands.
As a result of the foregoing, our gross profit increased by 156.6%, from RMB31.4 million in 2022 to RMB80.5 million in 2023. Our gross profit margin increased from 54.6% in 2022 to 64.6% in 2023.
Our gross profit for on-premise deployment increased by 161.3% from RMB29.4 million in 2022 to RMB76.8 million in 2023. Our gross profit margin for on-premise deployment increased from 53.6% in 2022 to 68.2% in 2023. The increase in our gross profit margin of on-premise deployment was primarily due to the enhanced sophistication and complexity of our services driven by the improved diversification and performance of our models.
Our gross profit for cloud-based deployment increased by 87.5% from RMB2.0 million in 2022 to RMB3.7 million in 2023. Our gross profit margin for cloud-based deployment decreased from 76.1% in 2022 to 31.0% in 2023. A growing number of users had been accessing our models via API interfaces. According to Frost & Sullivan, this approach had been increasingly adopted by users in recent years. However, due to the intense market competition, the prices of such services had decreased. We had also made strategic reductions to our service prices to expand our market share and attract additional customers. As a result, both the gross profit and gross profit margin of our cloud-based deployment services decreased significantly.
Our other income increased from RMB1.8 million in 2022 to RMB10.0 million in 2023, primarily due to the increase in interest income of RMB9.8 million as our cash and cash equivalent balance increased.
Research and development expenses increased by 526.8%, from RMB84.4 million in 2022 to RMB528.9 million in 2023. The increase was primarily due to (i) an increase in computing service fees of RMB297.1 million, paid to third party computing power providers as we devote significant efforts to iterate our foundation models, and investing in more advanced model training infrastructure; and (ii) an increase in payroll cost of RMB95.3 million, as we expanded our R&D team and incurred more share-based payment expenses.
Selling and marketing expenses increased by 568.5%, from RMB15.1 million in 2022 to RMB101.2 million in 2023. The increase was primarily due to (i) an increase in payroll cost of RMB19.0 million as we expanded our sales and marketing team and incurred more share-based payment expenses, and (ii) an increase in advertising and marketing expenses of RMB53.7 million, as we expanded our sales and promotion channels. We strategically made more advertising investment in order to swiftly take advantage of emerging market opportunities.
General administrative expenses increased by 105.2%, from RMB32.3 million in 2022 to RMB66.3 million in 2023. The increase was primarily due to an increase in payroll cost of RMB13.0 million as we recruited additional general and administrative personnel and incurred more share-based payment expenses.
Impairment loss on financial assets significantly increased from RMB0.03 million in 2022 to RMB19.8 million in 2023 due to a one-time write-off of other receivables in 2023.
Our finance costs increased from RMB5.7 million in 2022 to RMB26.3 million in 2023 primarily due to (i) an increase in interest on lease liabilities of RMB6.7 million as we rented additional office premises; and (ii) an increase in transaction costs in relation to the financial instruments issued to investors of RMB11.6 million due to the related financial advisor's fees in relation to financing activities.
Share of losses of associates increased from nil in 2022 to losses of RMB0.5 million in 2023, primarily due to the operations and financial performance of our associates in each year.
Fair value change of financial instruments measured at FVPL increased by 335.7%, from RMB6.0 million in 2022 to RMB26.0 million in 2023, primarily due to the increase in the fair value of investee companies and procurement of wealth management products.
Changes in the carrying amount of financial instruments issued to investors increased by 257.2% from RMB45.2 million to RMB161.5 million, mainly driven by the completion of the additional equity financing.
As a result of the above factors, our loss for the year increased by 448.5%, from RMB143.7 million in 2022 to RMB788.0 million in 2023.
The following table sets forth a summary of our consolidated statement of financial position as of the dates indicated.
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | ASSETS | | | | | | Non-current assets | | | | | | Property and equipment | 41,418 | 787,537 | 866,363 | 772,619 | | Intangible assets | 21,027 | 54,573 | 50,359 | 55,399 | | Goodwill | — | 39,379 | 39,379 | 39,379 | | Interests in associates | — | 13,047 | 201,198 | 290,345 | | Other non-current assets | — | 46 | 97,260 | 202,614 | | Time deposits | — | 102,093 | 105,343 | — | | **Total non-current assets** | **62,445** | **996,675** | **1,359,902** | **1,360,356** |
Current assets | | | | Short-term investments measured at FVPL | 31,777 | 158,904 | 42,621 | 549,364 Inventories and contract costs | 10,342 | 28,782 | 32,465 | 67,866 Trade and other receivables | 27,886 | 416,441 | 666,841 | 453,387 Contract assets | 403 | 9,960 | 4,718 | 6,654 Time deposits | 10,092 | — | — | 106,968 Cash at bank and on hand | 219,031 | 1,249,391 | 2,269,222 | 2,556,116
Current liabilities | | | | Trade and other payables | 25,834 | 288,197 | 603,488 | 838,417 Contract liabilities | 35,230 | 74,062 | 75,059 | 75,367 Bank loans | — | — | 137,246 | 137,214 Lease liabilities | 12,832 | 66,765 | 213,161 | 213,458 Financial instruments issued to investors | 457,959 | 3,179,864 | 6,676,943 | 9,564,760 Convertible bonds | — | — | 132,158 | —
Non-current liabilities | | | | Lease liabilities | — | 204,117 | 458,107 | 386,132 Deferred income | 10,309 | 29,741 | 34,752 | 36,204
The following table sets forth our current assets and current liabilities as of the dates indicated.
As of December 31, | As of June 30, | As of October 31, 2022 | 2023 | 2024 | 2025 | 2025 (RMB in thousands) | | | | (Unaudited)
Current assets | | | | | Short-term investments measured at FVPL | 31,777 | 158,904 | 42,621 | 549,364 | 228,516 Inventories and contract costs | 10,342 | 28,782 | 32,465 | 67,866 | 94,958 Trade and other receivables | 27,886 | 416,441 | 666,841 | 453,387 | 521,688 Contract assets | 403 | 9,960 | 4,718 | 6,654 | 2,972 Time deposits | 10,092 | — | — | 106,968 | — Cash at bank and on hand | 219,031 | 1,249,391 | 2,269,222 | 2,556,116 | 2,584,011
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | As of October 31, | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | 2025 | | (RMB in thousands) | | | | | (Unaudited) |
| | As of December 31, 2022 | As of December 31, 2023 | As of December 31, 2024 | As of June 30, 2025 | As of October 31, 2025 | |---|---|---|---|---|---| | Trade and other payables | 25,834 | 288,197 | 603,488 | 838,417 | 1,209,354 | | Contract liabilities | 35,230 | 74,062 | 75,059 | 75,367 | 117,451 | | Bank loans | — | — | 137,246 | 137,214 | 268,367 | | Lease liabilities | 12,832 | 66,765 | 213,161 | 213,458 | 216,879 | | Financial instruments issued to investors | 457,959 | 3,179,864 | 6,676,943 | 9,564,760 | 9,904,392 | | Convertible bonds | — | — | 132,158 | — | — | | **Total current liabilities** | **531,855** | **3,608,888** | **7,838,055** | **10,829,216** | **11,716,443** | | **Net current liabilities** | **(232,324)** | **(1,745,410)** | **(4,822,188)** | **(7,088,861)** | **(8,284,298)** |
Our net current liabilities increased slightly from RMB7,088.9 million as of June 30, 2025 to RMB8,284.3 million as of October 31, 2025, primarily due to (i) an increase in trade and other payables of RMB370.9 million, and (ii) a decrease in short-term investments measured at FVPL of RMB320.8 million. To improve our net current liabilities position as of October 31, 2025, we have secured substantial banking facilities that provide us with a strong liquidity cushion, which allows us to optimize our debt structure and prudently refinance certain short-term borrowings into longer-tenor facilities. In addition, we implement disciplined working-capital management, including close monitoring of trade receivables collection and careful management of procurement and payment schedules, which together enable a more efficient cash-flow cycle, reduce working-capital requirements and strengthen operating cash flows. Through these combined financing and operational measures, we expect to further enhance our liquidity and improve our net current liabilities position.
Our net current liabilities increased by 47.0% from RMB4,822.2 million as of December 31, 2024 to RMB7,088.9 million as of June 30, 2025, primarily due to an increase in financial instruments issued to investors of RMB2,887.8 million.
Our net current liabilities increased by 176.3% from RMB1,745.4 million as of December 31, 2023 to RMB4,822.2 million as of December 31, 2024, primarily attributed to an increase of RMB3,497.1 million in financial instruments issued to investors, partially offset by an increase of RMB1,019.8 million in cash and cash equivalents.
Our net current liabilities increased by 651.3% from RMB232.3 million as of December 31, 2022 to RMB1,745.4 million as of December 31, 2023, primarily attributed to an increase of RMB2,721.9 million in financial instruments issued to investors, partially offset by an increase of RMB1,030.4 million in cash and cash equivalents.
During the Track Record Period, our property and equipment consisted of (i) right-of-use assets, (ii) electronic equipment and others and (iii) leasehold improvement. The following table sets forth the breakdown of our property and equipment as of the dates indicated.
| | 2022 | As of December 31, 2023 | 2024 | As of June 30, 2025 | |---|---|---|---|---| | | (RMB in thousands) | | | | | **Property and equipment** | | | | | | Right-of-use assets | 11,971 | 257,582 | 623,825 | 555,302 | | Electronic equipment and others | 29,447 | 511,348 | 228,130 | 202,690 | | Leasehold improvements | — | 18,607 | 14,408 | 14,627 | | **Total** | **41,418** | **787,537** | **866,363** | **772,619** |
Our property and equipment significantly increased by 1,801.4% from RMB41.4 million as of December 31, 2022 to RMB787.5 million as of December 31, 2023, primarily due to an increase in electronic equipment of RMB481.9 million, mainly resulted from the procurement of large quantity of electronic equipment, mainly comprising computing hardware that provide us with additional computing power and an increase in right-of-use assets of RMB245.6 million mainly consisted of increase in the leased office premise and electronic equipment such as computing hardware in line with our business expansion.
Our property and equipment increased by 10.0% from RMB787.5 million as of December 31, 2023 to RMB866.4 million as of December 31, 2024, primarily due to an increase of right-of-use asset of RMB366.2 million mainly resulted from additional leases of office premise and electronic equipment, such as computing hardware, in line with our business expansion, partially offset by a decrease in electronic equipment of RMB283.2 million due to the disposal of some electronic equipment, including computing hardware. This is because we partially shifted from owning to leasing electronic equipment for operational efficiency.
Our property and equipment decreased by 10.8% from RMB866.4 million as of December 31, 2024 to RMB772.6 million as of June 30, 2025, primarily due to depreciation in property and equipment.
During the Track Record Period, our intangible assets mainly reflected our software and patents. Our intangible assets increased by 159.5% from RMB21.0 million as of December 31, 2022 to RMB54.6 million as of December 31, 2023, primarily due to an increase in patent rights of RMB31.8 million primarily resulted from the acquisition of one subsidiary in 2023, pursuant to which its patent rights were integrated into our Group. Our intangible assets remained relatively stable at RMB54.6 million as of December 31, 2023 and RMB50.4 million as of December 31, 2024, respectively. Our intangible assets increased from RMB50.4 million as of December 31, 2024 to RMB55.4 million as of June 30, 2025, primarily due to the acquisition of domain name.
Goodwill arising from the acquisition of subsidiaries represents the excess of the consideration transferred, the amount of any non-controlling interest in the acquiree and the acquisition-date fair value of any previous equity interest in the acquiree over the fair value of the identified net assets acquired. During the Track Record Period, goodwill arose from our acquisition of Beijing Lingxin Intelligent in 2023. It amounted to nil, RMB39.4 million, RMB39.4 million and RMB39.4 million as of December 31, 2022, 2023 and 2024, and June 30, 2025, respectively.
Goodwill impairment reviews are undertaken annually or more frequently if events or changes in circumstances indicate a potential impairment. Determining whether goodwill is impaired requires us to estimate the recoverable amounts of the cash generating unit ("CGU") to which we have allocated goodwill, which is the higher of fair value less costs of disposal and value in use. This recoverable amounts calculation requires us to estimate the future cash flows expected to arise from the CGU and a suitable discount rate to calculate the present value. Where the book value of the CGU exceeds its recoverable amounts, an impairment loss may arise.
The goodwill has been allocated to the CGU of Beijing Lingxin Intelligent. The recoverable amount of CGU of Beijing Lingxin Intelligent is determined based on value in use calculation, determined by discounting the future cash flows to be generated from the continuing operation of the CGU of Beijing Lingxin Intelligent with reference to valuation reports issued by an independent valuer. These calculations use cash flow projections based on financial budgets approved by management covering an eight-year period. We adopted a forecast period of longer than five years in view that the business is still under significant growth and will require additional time for the underlying technology to reach stable status. The key assumptions used in the estimation of the recoverable amounts are as follows:
| | As of December 31, | | As of June 30, | |---|---|---|---| | | 2023 | 2024 | 2025 | | Annual revenue growth rate(i) | 15%–35% | 8%–35% | 12%–35% | | Annual gross profit margin(i) | 54%–58% | 54%–58% | 55%–58% | | Growth rate beyond the forecast period(ii) | 2% | 2% | 2% | | Pre-tax discount rate(iii) | 19% | 18% | 17% |
Notes: (1) The annual revenue growth rates and gross profit margins are based on the current operational status and expectations of future changes in the industry and adjusted for other factors that are specific to the CGU. (2) The growth rate beyond the forecast period is based on relevant industry growth forecasts and does not exceed the average growth rate of the relevant industry. (3) The pre-tax discount rate reflects specific risks relating to the CGU of Beijing Lingxin Intelligent.
The headroom of CGU of Beijing Lingxin Intelligent as of December 31, 2023 and 2024 and June 30, 2025 amounted to RMB16.0 million, RMB23.0 million and RMB18.0 million, respectively. We have undertaken sensitivity analysis on the impairment test of goodwill. The following table sets out the hypothetical change to pre-tax discount rate that would have removed the remaining headroom:
| | As of December 31, | | As of June 30, | |---|---|---|---| | | 2023 | 2024 | 2025 | | Pre-tax discount rate | 2% | 3% | 2% |
As a result of the impairment tests, we believe that there was no impairment of goodwill as of December 31, 2023 and 2024 and June 30, 2025. Reasonable possible changes in key assumptions would not lead to impairment of the goodwill as of December 31, 2023 and 2024 and June 30, 2025. No impairment of goodwill was recorded during the Track Record Period. For more details, see Note 13 of the Accountants' Report set forth in Appendix I to this prospectus.
During the Track Record Period, the interests in associates primarily represented our limited partnership interest in Beijing XingLian DingSen Equity Investment Fund Partnership (Limited Partnership) ("Xinglian"), which is a limited partnership in which we were one of the limited partners. During the Track – 271 –
Record Period, our interests in associates amounted to nil, RMB13.0 million, RMB201.2 million and RMB290.3 million as of December 31, 2022, 2023 and 2024 and June 30, 2025, respectively. The increase was due to our injection of our capital contribution into Xinglian. For more information, see Note 15 of the Accountants' Report set forth in Appendix I to this prospectus.
During the Track Record Period, our other non-current assets consisted of (i) input VAT to be deducted; (ii) prepayment of computing service fees; and (iii) contract assets representing revenue recognized while the payment milestones have yet to be met. The following table sets forth the breakdown of our non-current assets as of the dates indicated.
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | | | | Input VAT to be deducted | — | — | 49,979 | 130,925 | | Prepayment of computing service fees | — | — | 39,522 | 61,123 | | Contract assets | — | 46 | 7,759 | 10,566 | | **Total** | **—** | **46** | **97,260** | **202,614** |
We recorded other non-current assets of RMB97.3 million as of December 31, 2024 as we substantially expanded our business with significant increase in related procurements that entail huge amount of input VAT to be deducted. We had prepayment of computing service fees in 2024 as we substantially increased our procurement of computing power. In addition, for some on-premise deployment solutions, as commercially agreed with our customers, certain portion of the service fee was withheld by such customers until the completion of the warranty period or other conditions. Our non-current assets increased from RMB97.3 million as of December 31, 2024 to RMB202.6 million as of June 30, 2025 primarily due to a significant increase in input VAT to be deducted of RMB80.9 million, which is associated with the increased procurement as a result of our continuous business expansion.
During the Track Record Period, our inventories and contract costs primarily consisted of (i) contract fulfillment cost, representing costs incurred to fulfill ongoing contracts. Contract fulfillment cost will be recognized as cost of revenue in the period in which revenue from the related service is recognized; and (ii) purchased hardware and components.
The table below sets forth our balance of our inventories and contract costs as of the dates indicated.
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | | | | Contract fulfillment cost | 10,939 | 30,082 | 32,979 | 70,046 | | Purchased hardware and components | — | — | 1,218 | — | | Less: write-down of inventories | (597) | (1,300) | (1,732) | (2,180) | | **Total** | **10,342** | **28,782** | **32,465** | **67,866** |
Our inventories and contract costs mainly include the contract costs incurred in fulfilling our on-premises deployment service contracts. In measuring impairment loss on contract costs, we determined the recoverable amount of these contract costs by comparing the remaining expected amount of consideration to be received in exchange for the goods or services to which the asset relates and the costs that relate directly to providing those goods or services and that have not been recognized as expenses in accordance with IFRS 15. In measuring impairment losses of inventories, we determine the carrying amount of inventories at the lower amount of cost and net realizable value. The cost of our purchased hardware and components is determined at the transaction price. Net realizable value is the estimated selling price in the ordinary course of business less the estimated costs necessary to make the sale. The impairment provision for inventories and contract costs has been assessed accordingly, specifically: (i) for inventories and contract costs related to signed contracts, the estimated selling price was determined based on binding agreements; and (ii) the estimated costs to complete and selling costs were projected with reference to the historical execution of similar projects and approved budgets submitted during the project initiation phase, which were reviewed and authorized by our management. Based on the aforementioned impairment assessment, we recognized write-down of inventories and contract costs of RMB0.6 million, RMB1.3 million, RMB1.7 million and RMB2.2 million as of December 31, 2022, 2023 and 2024 and June 30, 2025, respectively. Our Directors are of the view that impairment for inventories and contract costs was assessed appropriately and that sufficient provision was made based on the impairment assessment mentioned above.
Our inventories and contract costs increased by 178.3% from RMB10.3 million as of December 31, 2022 to RMB28.8 million as of December 31, 2023, primarily attributed to an increase in on-premise deployment services in progress. Our inventories and contract costs increased by 12.8% from RMB28.8 million as of December 31, 2023 to RMB32.5 million as of December 31, 2024, primarily affected by our expanded operations and an increase in the number of on-premise deployment services in progress, especially those requiring on-premise deployment of equipment, software and models. Our inventories and contract costs increased by 109.0% from RMB32.5 million as of December 31, 2024 to RMB67.9 million as of June 30, 2025, primarily due to an increase in contract fulfillment cost of RMB37.1 million primarily due to the completion and acceptance of certain ongoing on-premise services.
The table below sets forth our average inventory and contract costs turnover days as of the dates indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | Average inventory and contract costs turnover days(1) | 124 | 162 | 82 | 95 |
(1) Average turnover day of inventory and contract costs is calculated based on the average balance of inventory and contract costs divided by cost of revenue for the relevant year/period and multiplied by 365 days for a given year and 180 days for a six-month period. Average balance of inventory and contract costs is calculated by dividing the sum of inventory and contract costs at the beginning and the end of the year/period by two.
Our average inventory and contract costs turnover days increased from 124 days in 2022 to 162 days in 2023, primarily due to the complexity and diversification of our services, and decreased to 82 days in 2024, primarily due to the enhanced delivery efficiency of our services and the increased volume of cloud-based deployment with relatively quick service delivery. Our average inventory and contract costs turnover days remained relatively stable at 95 days for the six months ended June 30, 2025.
As of October 31, 2025, we had subsequently sold RMB10.8 million, or 15.9%, of our outstanding inventories and contract costs as of June 30, 2025.
During the Track Record Period, our trade and other receivables included (i) trade receivables primarily representing the amounts due from customers for our services provided in the ordinary course of business; (ii) other receivables primarily in relation to facilitating the purchase of hardware for certain of our customers; (iii) deposits paid to or withheld by our suppliers and landlords; (iv) receivables from disposal of investments in equity securities measured at FVPL; and (v) receivables of capital contribution from equity holders.
The table below sets forth our trade and other receivables as of the dates indicated.
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | | | | Trade receivables | 6,643 | 8,883 | 100,170 | 171,657 | | Less: credit loss allowance | (29) | (988) | (9,035) | (25,880) | | | 6,614 | 7,895 | 91,135 | 145,777 | | Other receivables | 1,833 | 117,591 | 263,805 | 95,381 | | Deposits | 3,640 | 16,584 | 67,912 | 70,257 | | Receivables from disposal of investments in equity securities measured at FVPL | — | — | 45,216 | 7,098 | | Receivables of capital contribution from equity holders | — | 120,328 | — | — | | | 5,473 | 254,503 | 376,933 | 172,736 | | Less: loss allowance | (3) | (18,779) | (27,327) | (21,172) | | | 5,470 | 235,724 | 349,606 | 151,564 | | Input VAT deductible | 4,169 | 84,527 | 98,729 | 83,139 | | Prepayments for computing service fees | 11,633 | 88,295 | 127,371 | 72,907 | | | 15,802 | 172,822 | 226,100 | 156,046 | | **Total** | **27,886** | **416,441** | **666,841** | **453,387** |
Our total trade and other receivables significantly increased by 1,393.4%, from RMB27.9 million as of December 31, 2022 to RMB416.4 million as of December 31, 2023, primarily due to (i) an increase in other receivables of RMB115.8 million as we helped facilitate purchase of hardware for certain of our customers; Our other receivables increased as we made advance payments to our suppliers for the procurement of hardware; and (ii) receivables of capital contribution from equity holders of RMB120.3 million representing investment amount from Pre-IPO Investors.
Our total trade and other receivables increased by 60.1%, from RMB416.4 million as of December 31, 2023 to RMB666.8 million as of December 31, 2024, primarily due to (i) an increase in other receivables of RMB146.2 million. We facilitated one of our customers in the technology industry for the procurement of certain hardware. Our on-premise deployment services typically do not include sales of computing hardware, which customers are generally required to procure independently. Given our expertise with the relevant hardware and our industry resources, at the customer's request, we procured specified hardware items and provided them together with our on-premise deployment service, charging the customer based on the quantity and model of the hardware. Under this transaction, we acted as solely an agent in the delivery
of the hardware. As a result, the receivables due from the customer were recorded as other receivables. As of October 31, 2025, RMB192.8 million, or 100% of other receivables in such nature were settled. There were no recoverability issue regarding payment collection for such transaction; (ii) an increase in trade receivables of RMB91.3 million, resulted from our business expansion and increase in the sales volume of our services, (iii) a significant increase in deposits, due to the increase in our procurement for business activities and additional leases; and (iv) an increase in receivables from disposal of investments in equity securities measured at FVPL of RMB45.2 million as we transferred equity interests in certain of our investee companies to Xinglian.
Our total trade and other receivables decreased by 32.0%, from RMB666.8 million as of December 31, 2024 to RMB453.4 million as of June 30, 2025, primarily due to a decrease in other receivables of RMB168.4 million as such receivables were subsequently settled in majority as of June 30, 2025, partially offset by an increase in trade receivables of RMB71.5 million in line with our business expansion.
We typically set forth the trading terms with our customers in the relevant service agreements. During the Track Record Period, for our on-premise deployment, we generally received milestone payments from our customers based on milestones as specified in the commercial agreements with such customers.
The table below sets forth an aging analysis of our trade receivables, net of loss allowance as of the dates indicated.
| | As of December 31, | | | As of June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | | (RMB in thousands) | | Within 3 months | 6,498 | 5,379 | 74,191 | 76,694 | | 3 months to 6 months | — | 1,425 | 13,804 | 28,891 | | 6 months to 1 year | 116 | 82 | 1,444 | 38,504 | | 1 year to 2 years | — | 1,009 | 1,302 | 1,358 | | More than 2 years | — | — | 394 | 330 | | **Total** | **6,614** | **7,895** | **91,135** | **145,777** |
When estimating the expected credit loss for trade receivables, we took into account all reasonable and supportable information available, including information about past events, current conditions and forecasts of future economic conditions in accordance with IFRS 9. Based on our assessment, we had recognized loss allowance of trade receivables of RMB0.03 million, RMB1.0 million, RMB9.0 million and RMB25.9 million as of December 31, 2022, 2023 and 2024 and June 30, 2025, respectively. Our Directors are of the view that impairment for trade receivables was assessed appropriately and that sufficient expected credit loss was recognized.
The table below sets forth our average trade receivables turnover days as of the dates indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | Average trade receivables turnover days(1) | 23 | 21 | 58 | 112 |
(1) Average turnover days of trade and other receivables is calculated based on the average balance of trade receivables divided by revenue for the relevant year/period and multiplied by 365 days for a given year and 180 days for a six-month period. Average balance of trade receivables is calculated by dividing the sum of trade receivables at the beginning and the end of the year/period by two.
Our average trade receivables turnover days remained relatively stable at 23 days in 2022 and 21 days in 2023. In 2023, we experienced significant growth in our customer base and sales volume. For our new customers, their payment practices varied widely. We also broadened our service portfolio, particularly through on-premise deployment that follow a milestone-based payment structure. Under these arrangements, customers make payments when specific contractual milestones are completed, generally resulting in longer payment periods than before. Consequently, our trade receivables turnover days increased notably to 58 days in 2024 and further increased to 112 days for the six months ended June 30, 2025.
We do not foresee any material recoverability issue with our trade receivables based on our evaluation of the historical credit standing and the credit records of our customers, which are generally private or state-owned enterprises with strong economic performance and credit history. We will continue to strengthen our management in trade receivables and improve the collection rate in the future, and our Directors are of the view that sufficient provision has been made to trade receivables during the Track Record Period. See "—Disclosure About Financial Risks—Credit Risks" in this section.
As of October 31, 2025, RMB86.5 million, or 50.4%, of our trade receivables as of June 30, 2025 had been settled by our customers.
During the Track Record Period, our short-term investments measured at FVPL represented (i) our equity investments in certain unlisted companies; and (ii) various wealth management products that we purchased from commercial banks in China. The wealth management products we purchased were primarily structured deposits. The interest rate of wealth management products during the Track Record Period ranged from 1.55% to 3.86% per annum.
Our short-term investments measured at FVPL significantly increased by 400.1% from RMB31.8 million as of December 31, 2022 to RMB158.9 million as of December 31, 2023, primarily due to (i) an increase in investments in equity securities of RMB47.1 million primarily because we made additional equity investments in 2023; and (ii) an increase in wealth management products of RMB80.0 million as we purchased additional structured deposits. Our short-term investments measured at FVPL significantly decreased from RMB158.9 million as of December 31, 2023 to RMB42.6 million as of December 31, 2024 due to the maturity of our wealth management products. Our short-term investments measured at FVPL increased from RMB42.6 million as of December 31, 2024 to RMB549.4 million as of June 30, 2025 due to the procurement of additional structured deposits.
We purchase wealth management products as a supplemental means to improve utilization of our cash on hand on a short-term basis. We believe that making such investments is in the best interest of the Company, and we can make better use of our cash by utilizing principal-guaranteed structured deposits, to enhance our income without interfering with our business operations or capital expenditures. The purchases of wealth management products are subject to the approval of our Board and the purchases are carefully reviewed and assessed by the staff in our finance department with financial management or accounting background. Additionally, we have established a set of risk management and capital preservation investment policy, and have implemented a series of internal control measures regarding our investment in wealth management products. These policies and measures include:
• our investment decisions are made on a case-by-case basis and after due and careful consideration of a number of factors, such as the duration of the investment and the expected returns;
• we only purchase low-risk wealth management products issued by qualified financial institutions, and in any given period, we invest in products provided by multiple issuers to mitigate concentration risks;
在完成投资后，我们定期密切监控其表现及公允价值。
未来，我们可能会根据剩余现金情况，继续购买期限较短的低风险理财产品，以最大限度提高资金使用效率。我们在理财产品方面的投资须符合《上市规则》第14章的相关规定。
于追踪记录期间，我们的贸易及其他应付款项包括：(i) 贸易应付款项，主要代表日常业务中就计算硬件、计算服务及广告应付的款项；(ii) 应付计算服务费；(iii) 应付市场推广及宣传费用；(iv) 应付员工薪酬；(v) 其他应付款项及应计费用；(vi) 股东预付待注入资本款项；(vii) 其他应付税项；以及 (viii) 保修准备金。
下表列示于所示日期我们的贸易及其他应付款项。
| | 截至12月31日止 | | | 截至2025年6月30日 | |---|---|---|---|---| | | 2022年 | 2023年 | 2024年 | | | | （人民币千元） | | | |
| 应付第三方贸易款项 | 663 | 6,754 | 58,293 | 69,403 | | 应付计算服务费 | 1,509 | 104,536 | 269,467 | 495,348 | | 应付市场推广及宣传费用 | — | 4,357 | 89,052 | 28,981 | | 应付员工薪酬 | 15,386 | 45,226 | 104,229 | 105,846 | | 其他应付款项及应计费用 | 1,987 | 12,309 | 43,767 | 109,173 | | **小计** | **19,545** | **173,182** | **564,808** | **808,751** | | 股东预付待注入资本款项 | — | 97,750 | — | — | | 其他应付税项 | 3,401 | 11,038 | 22,304 | 8,348 | | 保修准备金 | 2,888 | 6,227 | 16,376 | 21,318 | | **合计** | **25,834** | **288,197** | **603,488** | **838,417** |
我们的贸易及其他应付款项由截至2022年12月31日的人民币2,580万元增加至截至2023年12月31日的人民币28,820万元，主要原因为：(i) 与我们运营及研发活动相关的应付第三方供应商计算服务费增加人民币10,300万元；以及 (ii) 2023年出现股东预付款项人民币9,780万元，原因为我们于2023年在交易完成前提前收到若干IPO前投资者的股权融资款项。
我们的贸易及其他应付款项由截至2023年12月31日的人民币28,820万元增加至截至2024年12月31日的人民币60,350万元，主要原因为：(i) 与我们运营及研发活动相关的应付计算服务费增加人民币16,490万元；(ii) 随着业务扩张，贸易应付款项增加人民币5,150万元；以及 (iii) 由于广告活动增加，应付市场推广及宣传费用增加人民币8,470万元。
我们的贸易及其他应付款项由截至2024年12月31日的人民币60,350万元增加至截至2025年6月30日的人民币83,840万元，主要原因为与我们运营及研发活动相关的应付计算服务费大幅增加人民币22,590万元。
The following table sets forth an aging analysis of our trade payables as of the dates indicated.
| | As of December 31, | | | As of June 30, 2025 | |---|---|---|---|---| | | 2022 | 2023 | 2024 | | | | (RMB in thousands) | | | | | Within 3 months | 663 | 5,333 | 57,676 | 59,596 | | 3 months to 6 months | — | 865 | 57 | 5,848 | | 6 months to 1 year | — | 526 | 257 | 3,680 | | More than 1 year | — | 30 | 303 | 279 | | **Total** | **663** | **6,754** | **58,293** | **69,403** |
The table below sets forth the average trade payables turnover days for the years/period indicated.
| | Year Ended December 31, | | | Six Months Ended June 30, 2025 | |---|---|---|---|---| | | 2022 | 2023 | 2024 | | | Average trade payables turnover days(1) | 7 | 31 | 87 | 120 |
Note: (1) Average turnover days of trade and other payables is calculated based on the average balance of trade payables divided by cost of revenue for the relevant year/period and multiplied by 365 days for a given year and 180 days for a six-month period. Average balance of trade payables is calculated by dividing the sum of trade and other payables at the beginning and the end of the year/period by two.
Our average trade payables turnover days increased from 7 days in 2022 to 31 days in 2023, and further increased to 87 days in 2024, primarily attributable to enhanced supply chain management. Our average trade payables turnover days increased to 120 days for the six months ended June 30, 2025 primarily due to enhanced supply chain management and bargaining power.
As of October 31, 2025, we had settled RMB359.2 million, or 63.6%, of our outstanding trade payables as of June 30, 2025.
During the Track Record Period, our contract liabilities represented advance payments made by customers for provision of services. Our contract liabilities increased by 110.2% from RMB35.2 million as of December 31, 2022 to RMB74.1 million as of December 31, 2023, generally in line with the business expansion and the increase in the sales volume of our services, and remained relatively stable at RMB75.1 million as of December 31, 2024, primarily because the delivery and revenue recognition progress of our sales contracts varied, which resulted in fluctuations in our contract liabilities as of the end of the relevant years. Our contract liabilities remained relatively stable at RMB75.1 million as of December 31, 2024 and RMB75.4 million as of June 30, 2025, respectively. As of October 31, 2025, RMB22.5 million or 29.8% of our contract liabilities as of June 30, 2025 had been subsequently recognized as revenue.
Deferred income during the Track Record Period represents government grants, which were non-recurring financial support or subsidies from government primarily as incentives for research and – 278 –
development projects in relation to AI and related technology. Some governmental grants were conditioned upon the completion of the related research and development projects. Government grants are recognized as a reduction of the expenses related to the grants in profit or loss on a systematic basis in the same periods in which such expenses are incurred. We recorded deferred income of RMB10.3 million, RMB29.7 million, RMB34.8 million and RMB36.2 million as of December 31, 2022, 2023 and 2024 and June 30, 2025, respectively. Such fluctuations were primarily due to the joint impact of new grants received and the amount released to profit or loss during the relevant years. During the Track Record Period, RMB2.9 million, RMB11.2 million, RMB36.3 million and RMB5.7 million were charged to profit or loss.
Our Directors are of the opinion that, taking into account of the financial resources available to us, including (i) cash and cash equivalents; (ii) short-term investments measured at FVPL; (iii) available committed bank facilities; and (iv) the estimated net proceeds from the Global Offering, we have sufficient working capital to meet our present requirements and for the next 12 months from the date of this prospectus.
Our cash burn rate refers to the average monthly (i) net cash used in operating activities, (ii) net cash used in the purchases of property and equipment, (iii) net cash used in/generated from the purchases/disposal of wealth management products and (iv) lease payments. We consider these items to be key indicators reflecting our operational condition and efficiency, which can significantly impact our cash flow. To be specific, a major portion of our expenditure on equipment purchases and leasing is related to computing power equipment, which is essential in our day-to-day operations and research and development activities and representing significant cash outflows. In addition, we procured wealth management products as a supplemental mean to improve utilization of our cash on hand on a short-term basis, which also significantly affects our cash flow. Our historical cash burn rate was RMB3.0 million, RMB105.9 million, RMB194.5 million and RMB327.3 million for each of the years ended December 31, 2022, 2023 and 2024 and for the six months ended June 30, 2025, respectively, mainly representing our investment in R&D activities and business operations. During the Track Record Period, we recorded expenditure in purchase of property and equipment primarily due to our procurement and lease of computing hardware and offices, for the operation activities and R&D activities in line with our business expansion.
We had cash and cash equivalents, short-term investment measured at FVPL, and available committed bank facilities of RMB8,943.1 million in aggregate as of October 31, 2025. We estimate that we will receive net proceeds of RMB3,785.3 million (HK$4,173.4 million) after deducting the underwriting commissions and other estimated offering expenses payable by us in connection with the Global Offering, assuming no Over-allotment Option is exercised and based on the Offer Price of HK$116.20 per Offer Share.
Assuming that the average cash burn rate going forward will be RMB327.3 million, similar to the cash burn rate level for the six months ended June 30, 2025 based on the underlying assumptions that (i) the number of our employees will not increase significantly, particularly in the R&D department; (ii) we do not expect substantial capital investment; and (iii) we do not expect significant acquisitions of fixed assets, we estimate that our cash and cash equivalents, short term investment measured at FVPL, and available committed bank facilities as of October 31, 2025 will be able to maintain our financial viability for 27.3 months or, if we take into account 10% of the estimated net proceeds from the Listing (namely, the portion allocated for our working capital and other general corporate purposes), 28.5 months or, if we also take into account the estimated net proceeds from the Listing, 38.9 months. We will continue to monitor our cash flows from operations closely and maintain our financial viability through a variety of means, including, among others, banking facilities and external financings. In addition, we plan to reinforce our receivable collection efforts to reduce the accounts receivable collection cycles. We plan to (i) further
strengthen our client management practices, including rigorous review of payment terms at the contract stage, performing periodic review, monitoring payment behavior and implementing credit assessment procedures to ensure their financial creditworthiness and (ii) enhance collection of accounts receivable by promptly issuing invoices, regularly checking with our clients to ensure collection and implementing remedial measures when customers do not make timely payment.
The table below sets forth, for the years/periods indicated, a summary of our consolidated statements of cash flows.
| | Year Ended December 31, | Year Ended December 31, | Year Ended December 31, | Six Months Ended June 30, | Six Months Ended June 30, | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | | (Unaudited) | | | Net cash used in operating activities | (68,246) | (648,017) | (2,244,919) | (994,678) | (1,327,150) | | Net cash generated from/(used in) investing activities | 32,921 | (784,965) | (48,559) | (138,775) | (556,373) | | Net cash generated from financing activities | 191,196 | 2,463,043 | 3,312,073 | 816,341 | 2,165,110 | | Net increase/(decrease) in cash and cash equivalents | 155,871 | 1,030,061 | 1,018,595 | (317,112) | 281,587 | | Cash and cash equivalents at the beginning of the year/period | 63,057 | 218,928 | 1,249,175 | 1,249,175 | 2,268,164 | | Effect of exchange rate changes | — | 186 | 394 | 3 | 2,219 | | Cash and cash equivalents at the end of the year/period | 218,928 | 1,249,175 | 2,268,164 | 2,551,970 | 932,066 |
During the Track Record Period, our cash generated from operating activities is primarily derived from the sales of our services. Our cash outflow from operating activities comprises mainly operating expenses.
For the six months ended June 30, 2025, our net cash used in operating activities amounted to RMB1,327.2 million, primarily attributable to the loss before tax of RMB2,357.9 million, mainly offset by (i) changes in the carrying amounts of financial instruments issued to investors of RMB429.3 million and (ii) an increase in trade and other payables of RMB238.3 million.
For the year ended December 31, 2024, our net cash used in operating activities amounted to RMB2,244.9 million, primarily attributable to the loss before tax of RMB2,958.0 million, adjusted mainly by (i) increase in trade and other receivables of RMB415.0 million and (ii) changes in fair value of financial instruments measured at FVPL of RMB66.3 million. The foregoing was partially offset by (i) changes in carrying amount of financial instruments issued to investors of RMB468.9 million, (ii) depreciation on property and equipment of RMB270.3 million, and (iii) increase in trade and other payables of RMB416.0 million.
For the year ended December 31, 2023, our net cash used in operating activities amounted to RMB648.0 million, primarily attributable to the loss before tax of RMB788.0 million, adjusted mainly by increase in trade and other receivables of RMB269.5 million. The foregoing was partially offset by
(i) 贸易及其他应付款项增加人民币1.432亿元，(ii) 已发行予投资者的金融工具账面金额变动人民币1.615亿元，及(iii) 物业及设备折旧人民币6,380万元。
截至2022年12月31日止年度，我们经营活动所用净现金为人民币6,820万元，主要归因于税前亏损人民币1.437亿元，经贸易及其他应收款项增加人民币1,960万元调整。上述情况部分被以下各项所抵销：(i) 已发行予投资者的金融工具账面金额变动人民币4,520万元，(ii) 贸易及其他应付款项增加人民币1,290万元，(iii) 物业及设备折旧人民币1,660万元，及(iv) 合同负债增加人民币1,310万元。
我们投资活动所得现金主要包括出售物业及设备所得款项，以及以公允价值计量且其变动计入损益的短期投资到期所得款项。于业绩记录期间，我们投资活动所用现金主要包括购置物业及设备的款项，以及购买以公允价值计量且其变动计入损益的短期投资的款项。
截至2025年6月30日止六个月，我们投资活动所用净现金为人民币5.564亿元，主要归因于：(i) 购买理财产品的款项人民币5亿元，及(ii) 向联营公司注资的款项人民币7,500万元。上述情况部分被出售以公允价值计量且其变动计入损益的股权证券所得款项人民币4,250万元所抵销。
截至2024年12月31日止年度，我们投资活动所用净现金为人民币4,860万元，主要归因于：(i) 出售物业及设备所得款项人民币1.146亿元，(ii) 出售理财产品所得款项人民币2.005亿元，及(iii) 出售以公允价值计量且其变动计入损益的股权投资所得款项人民币1.608亿元。上述情况部分被以下各项所抵销：(i) 向联营公司注资的款项人民币1.7亿元，(ii) 购置物业及设备的款项人民币1.26亿元，(iii) 购买以公允价值计量且其变动计入损益的股权投资人民币1.187亿元，及(iv) 购买理财产品人民币1亿元。
截至2023年12月31日止年度，我们投资活动所用净现金为人民币7.85亿元，主要归因于：(i) 购买理财产品人民币11亿元，及(ii) 购置物业及设备的款项人民币5.068亿元。上述情况部分被理财产品到期所得款项人民币10.236亿元所抵销。
截至2022年12月31日止年度，我们投资活动所得净现金为人民币3,290万元，主要归因于以公允价值计量且其变动计入损益的短期投资到期所得款项人民币11.172亿元。上述情况部分被购买理财产品人民币10.42亿元所抵销。
我们融资活动所得现金主要包括向投资者发行金融工具所得款项。于业绩记录期间，我们投资活动所用现金主要包括已付租赁租金的资本部分。
截至2025年6月30日止六个月，我们融资活动所得净现金为人民币21.651亿元，主要归因于(i) 向投资者发行金融工具所得款项人民币16.25亿元，及(ii) 发行可转换债券所得款项人民币7亿元。
For the year ended December 31, 2024, our net cash generated from financing activities amounted to RMB3,312.1 million, mainly attributable to proceeds from the issuance of financial instruments to investors of RMB3,019.6 million.
For the year ended December 31, 2023, our net cash generated from financing activities amounted to RMB2,463.0 million, mainly attributable to proceeds from the issuance of financial instruments to investors of RMB2,419.0 million.
For the year ended December 31, 2022, our net cash generated from financing activities amounted to RMB191.2 million, mainly attributable to proceeds from the issuance of financial instruments to investors of RMB208.0 million.
| | 截至12月31日止年度 | | | 截至2025年6月30日止六个月 | |---|---|---|---|---| | | 2022年 | 2023年 | 2024年 | | | | （人民币千元） | | | | | 研发成本（1） | 25,894 | 247,811 | 1,425,605 | 927,001 | | 劳动力雇佣（2） | 84,825 | 181,943 | 490,639 | 339,125 | | 直接生产成本（含材料）（3） | 5,545 | 16,001 | 20,068 | 66,770 | | 产品营销（4） | 2,056 | 59,182 | 198,161 | 192,684 | | 非所得税、特许权使用费及其他政府收费（5） | 1 | 205 | 2,490 | 1,790 | | 合计 | 118,321 | 505,142 | 2,136,963 | 1,527,370 |
(1) 现金经营成本中的研发成本，指研发费用（扣除研发费用中的薪酬成本及非现金项目）经上一年末及本年末与研发活动相关的营运资金变动调整后的金额。
(2) 与劳动力雇佣相关的现金经营成本，指研发费用、行政费用、销售成本及销售与营销费用中薪酬成本的合计（扣除以股份为基础的非现金薪酬开支），经上述经营费用项下上一年末及本年末与薪酬成本相关的营运资金变动调整后的金额。
(3) 与直接生产成本（含材料）相关的现金经营成本，指收入成本（扣除合同履行成本中的薪酬成本及非现金项目），经上一年末及本年末与服务相关的营运资金变动调整后的金额。
(4) 与产品营销相关的现金经营成本，指销售与营销费用（扣除销售与营销费用中的薪酬成本及非现金项目），经上一年末及本年末与销售及分销活动相关的营运资金变动调整后的金额。
(5) 与非所得税、特许权使用费及其他政府收费相关的现金经营成本，主要指增值税、印花税、税费及附加税的缴纳。
Our indebtedness during the Track Record Period primarily consisted of bank loans and lease liabilities. The following table sets forth a breakdown of our indebtedness as of the dates indicated.
| | As of December 31, | As of December 31, | As of December 31, | As of June 30, | As of October 31, | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | 2025 | | | (RMB in thousands) | | | | (unaudited) | | **Current portion:** | | | | | | | Lease liabilities | 12,832 | 66,765 | 213,161 | 213,458 | 216,879 | | Bank loans | — | — | 137,246 | 137,214 | 268,367 | | Financial instruments issued to investors | 457,959 | 3,179,864 | 6,676,943 | 9,564,760 | 9,904,392 | | Advances received from equity shareholders to be injected as capital | — | 97,750 | — | — | — | | Convertible bonds | — | — | 132,158 | — | — | | **Non-current portion:** | | | | | | | Lease liabilities | — | 204,117 | 458,107 | 386,132 | 348,528 | | **Total** | 470,791 | 3,548,496 | 7,617,615 | 10,301,564 | 10,738,166 |
Our lease liabilities were primarily related to our office leases and computing hardware. Our lease liabilities were RMB12.8 million, RMB270.9 million, RMB671.3 million, RMB599.6 million and RMB565.4 million as of December 31, 2022, 2023 and 2024, June 30, 2025 and October 31, 2025, respectively. Our lease liabilities increased by 2,011.0% from RMB12.8 million as of December 31, 2022 to RMB270.9 million as of December 31, 2023 as we rented new premises as offices and additional computing hardware, in line with our business expansion. Our lease liabilities increased by 147.8% from RMB270.9 million as of December 31, 2023 to RMB671.3 million as of December 31, 2024, as we rented additional premises. Our lease liabilities then decreased by 10.7% to RMB599.6 million, primarily due to our payment of outstanding leases. The lease liabilities remained relatively stable at RMB565.4 million as of October 31, 2025.
As of December 31, 2022, 2023 and 2024, June 30, 2025 and October 31, 2025, other than as disclosed above, we did not have any other borrowings, charges, mortgages, debentures or debt securities issued or outstanding, or authorized or otherwise created but unissued, or other similar indebtedness, hire purchase and finance lease commitments, liabilities under acceptance, acceptance credits, any material guarantees or other material contingent liabilities.
We had one bank loan in 2024, with effective interest rate of 2.65%. Such bank loan was not secured or guaranteed, and was subsequently repaid in full in July 2025. From July 2025 to October 2025, we added ten bank loans with aggregated amount of RMB268.2 million, with effective interest rate of 2.20% and 2.11%, respectively. Such bank loans were not secured or guaranteed. As of December 31, 2024, June 30, 2025 and October 31, 2025, the outstanding balance of the bank loan (with accrued interest) amounted to RMB137.2 million, RMB137.2 million and RMB268.4 million, respectively.
我们的银行贷款包含中国商业银行贷款惯常的标准条款、条件及契诺。我们的董事确认，于追踪记录期间及直至最后实际可行日期，我们在获取银行贷款或其他借款方面未遇任何困难，亦未发生银行贷款或其他借款的付款违约或违反契诺的情况。
截至2025年10月31日，我们已承诺但未动用的银行融资额度为人民币61.305亿元。
我们向投资者发行的金融工具代表我们授予若干投资者的优先权利，该等投资者通过购买现有股份或资本投入的方式取得股权，如约定条件未获满足，彼等有权要求我们赎回其投入资本。截至2022年、2023年及2024年12月31日、2025年6月30日及2025年10月31日，我们向投资者发行的金融工具金额分别为人民币4.580亿元、人民币31.799亿元、人民币66.769亿元、人民币95.648亿元及人民币99.044亿元。上述持续增长主要由新增投资驱动。详情请参阅"历史、重组及公司架构——首次公开发售前投资"。
我们的可转换债券（"可转换债券"）代表我们于2024年及2025年向若干投资者发行的可转换债券，本金总额分别为人民币1.300亿元及人民币7.000亿元。可转换债券的年利率最高为8%。于2025年5月，债券持有人将全部未偿还可转换债券转换为我们向投资者发行的金融工具。如需进一步了解，请参阅本招股说明书附录一所载会计师报告附注27，以及本招股说明书"历史、重组及公司架构——首次公开发售前投资"。
自2025年10月31日至本招股说明书日期，由于我们的租赁负债、银行贷款、向投资者发行的金融工具及可转换债券保持相对稳定，董事确认，我们的债务及或然负债并无任何重大变化。
截至2022年、2023年及2024年12月31日及2025年6月30日，我们并无任何重大或然负债。
于追踪记录期间，本公司的研发支出主要由研发费用构成，并经以下项目调整：向第三方收购并已资本化的与研发软件相关的无形资产，以及计入研发支出的已资本化无形资产摊销费用。下表列示本公司于所示年度／期间的总研发支出。
| | 截至12月31日止年度 | | | 截至6月30日止六个月 | | |---|---|---|---|---|---| | | 2022年 | 2023年 | 2024年 | 2024年 | 2025年 | | | （人民币千元） | | | | | | **年度研发支出** | | | | | | | 研究及开发费用 | 84,377 | 528,884 | 2,195,436 | 859,217 | 1,594,661 | | 加：向第三方收购并已资本化的与研发软件相关的无形资产 | — | 38,655 | 6,125 | 6,125 | 427 | | 减：计入研发支出的已资本化无形资产摊销费用 | (3,393) | (4,874) | (9,497) | (4,621) | (5,001) | | **总研发支出** | **80,984** | **562,665** | **2,192,064** | **860,721** | **1,590,087** |
下表列示本公司于所示期间的年度及总运营支出。
| | 截至12月31日止年度 | | | 截至6月30日止六个月 | | |---|---|---|---|---|---| | | 2022年 | 2023年 | 2024年 | 2024年 | 2025年 | | | （人民币千元） | | | | | | **年度总运营支出** | | | | | | | 研究及开发费用 | 84,377 | 528,884 | 2,195,436 | 859,217 | 1,594,661 | | 销售及营销费用 | 15,139 | 101,198 | 387,475 | 144,194 | 208,570 | | 一般及行政费用 | 32,316 | 66,302 | 133,603 | 51,452 | 185,165 | | **调整：** | | | | | | | 加：向第三方收购并已资本化的与研发软件相关的无形资产 | — | 38,655 | 6,125 | 6,125 | 427 | | 减：计入研发支出的已资本化无形资产摊销费用 | (3,393) | (4,874) | (9,497) | (4,621) | (5,001) | | **总运营支出** | **128,439** | **730,165** | **2,713,142** | **1,056,367** | **1,983,822** |
The table below sets forth our annual R&D expenditure ratio and the total R&D expenditure ratio for the periods indicated.
| | For the Year Ended December 31, | | | For the Six Months Ended June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | Annual R&D expenditure ratio(1) | 63.1% | 77.1% | 80.8% | 81.5% | 80.2% | | Total R&D expenditure ratio | | | 79.4%(2) | | 79.7%(3) |
Notes: (1) Calculated by dividing annual R&D expenditure by annual total operating expenditure. (2) Calculated by dividing total R&D expenditure for the three financial years prior to listing by total operating expenditure for the three financial years prior to listing. (3) Calculated by dividing total R&D expenditure over the Track Record Period by total operating expenditure over the Track Record Period.
During the Track Record Period, our capital expenditure primarily consisted of purchases of property and equipment and intangible assets in relation to office equipment and leasehold improvements in relation to renovation.
The table below sets forth, for the periods indicated, our capital expenditures.
| | For the year ended December 31, | | | For the six months ended June 30, | | |---|---|---|---|---|---| | | 2022 | 2023 | 2024 | 2024 | 2025 | | | (RMB in thousands) | | Payment for the purchase of property and equipment | 31,651 | 506,760 | 125,958 | 98,388 | 12,740 | | Payment for the purchase of intangible assets | — | 1,809 | 6,802 | 6,150 | 11,168 | | Total | 31,651 | 508,569 | 132,760 | 104,538 | 23,908 |
We expect that our capital expenditures in 2025 will primarily consist of purchases of property and equipment and intangible assets. We intend to fund our future capital expenditures with our existing cash balance, cash generated from our operating activities and financing activities and proceeds from the Global Offering. See "Future Plans and Use of Proceeds" in this prospectus.
We did not have any material capital commitments as of December 31, 2022, 2023 and 2024, and June 30, 2025.
FINANCIAL INFORMATION KEY FINANCIAL RATIOS The following table sets forth our key financial ratios as of the dates indicated, or for the years/periods indicated:
| | As of/For the Year Ended December 31, | As of/For the Year Ended December 31, | As of/For the Year Ended December 31, | As of/For the Six Months Ended June 30, | |---|---|---|---|---| | | 2022 | 2023 | 2024 | 2025 | | Revenue growth ratio | / | 1.2 | 1.5 | 3.3 | | Current Ratio(1) | 0.6 | 0.5 | 0.4 | 0.3 | | Quick Ratio(2) | 0.5 | 0.5 | 0.4 | 0.3 | | Gearing ratio(3) | 7.1% | 27.6% | 20.4% | 12.0% |
(1) Current ratio is calculated by dividing current assets by current liabilities as of the date indicated. (2) Quick ratio is calculated by dividing current assets less inventories by current liabilities as of the date indicated. (3) Gearing ratio is calculated by dividing bank loans and lease liabilities divided by total equity as of the end of the period multiplied by 100%.
We are exposed to various types of risks, including interest rate risk, foreign currency risk, credit risk, and liquidity risk.
Our credit risk refers to the risk that a counterparty will default on its contractual obligations resulting in a financial loss to us. Our credit risk is primarily attributable to trade and other receivables and contract assets. Our exposure to credit risk arising from cash at bank is limited because the counterparties are banks and financial institutions with high credit standing, for which we consider to have low credit risk.
We do not provide any guarantees which would expose us to credit risk.
We have established a credit risk management policy under which individual credit evaluations are performed on all customers requiring credit over a certain amount. These evaluations focus on the customer's past history of making payments when due and current ability to pay, and take into account information specific to the customer as well as pertaining to the economic environment in which the customer operates. We generally require customers of on-premise deployment to settle progress billings, and cloud-based deployment customers to pay in advance under usage-based contracts and periodic billings under subscription-based contracts.
Significant concentrations of credit risk primarily arise when we have significant exposure to individual customers. As of December 31, 2022, 2023 and 2024 and June 30, 2025, 74.9%, 75.3%, 50.9% and 39.0% of the total trade receivables were due from our five largest trade debtors, respectively.
We measure loss allowances for trade receivables at an amount equal to lifetime ECLs, which is calculated using a provision matrix. As our historical credit loss experience does not indicate significantly different loss patterns for different customer segments and geographic regions, the loss allowance based on past due status is not further distinguished between our different customer bases. For more details, see Note 31(a) of the Accountants' Report set forth in Appendix I to this prospectus.
We regularly monitor current and expected liquidity requirements, to ensure that we maintain sufficient reserves of cash and adequate committed lines of funding from major financial institutions to meet its liquidity requirements in the short and longer term. For more details, see Note 31(b) of the Accountants' Report set forth in Appendix I to this prospectus.
As of the Latest Practicable Date, we did not enter into any off-balance sheet transactions or arrangements.
We did not declare or pay any dividend during the Track Record Period. We do not currently have a formal dividend policy or a fixed dividend payout ratio. We currently intend to retain all available funds and earnings, if any, to fund the development and expansion of our business and we do not anticipate paying any cash dividends in the foreseeable future. Investors should not purchase our ordinary shares with the expectation of receiving cash dividends. Any future determination to pay dividends will be made at the discretion of our Directors and may be based on a number of factors, including our future operations and earnings, capital requirements and surplus, general financial condition, contractual restrictions and other factors that our Directors may deem relevant. Regulations in the PRC currently permit payment of dividends of a PRC company only out of accumulated distributable after-tax profits less any recovery of accumulated losses and appropriations to statutory and other reserves that we are required to make, as determined in accordance with its articles of association and the accounting standards and regulations in China. As advised by our PRC Legal Adviser, taking into account the aforesaid, we may not have sufficient or any distributable profits to make dividend distributions to our Shareholders in a given year, in view of our accumulated losses, or even if we become profitable, as we will only be able to declare or pay dividends out of our distributable profits until (i) the accumulated losses are covered by our after-tax profits, and (ii) sufficient statutory and other reserves are drawn in accordance with the relevant laws, regulations and our constitutional documents. In light of our accumulated losses as disclosed in this prospectus, it is unlikely that we will be eligible to pay dividends out of our profits in the foreseeable future.
As of June 30, 2025, we did not have any distributable reserves.
We enter into transactions with our related parties from time to time. During the Track Record Period, we entered into a number of related party transactions. For more information, see Note 33 to the Accountants' Report in Appendix I to this document. Our Directors are of the view that each of our transactions with related parties during the Track Record Period were conducted in the ordinary course of business on an arm's length basis and with normal commercial terms between the relevant parties. Our Directors are also of the view that our related party transactions during the Track Record Period would not distort our results of operations or cause our historical results to become non-reflective of our future performance.
Listing expenses represent professional fees, underwriting commission and fees incurred in connection with the Listing and the Global Offering. We recorded listing expenses of nil, nil, nil and RMB17.7 million in 2022, 2023, 2024 and for the six months ended June 30, 2025, respectively. Our listing expenses are – 288 –
The listing expenses are estimated to be approximately RMB158.5 million (HK$174.7 million) (including underwriting commission), accounting for 4.0% of the gross proceeds of the Global Offering, based on the Offer Price of HK$116.20 per Share and no exercise of the Over-allotment Option. Among these, approximately RMB121.4 million (HK$133.8 million) is directly attributable to the issuance of Shares and will be charged to equity upon completion of the Listing, and approximately RMB37.1 million (HK$40.9 million) will be charged to our consolidated statements of profit or loss. The listing expenses we expect to incur would consist of approximately RMB118.3 million (HK$130.4 million) in underwriting-related expenses and fees, and approximately RMB40.2 million (HK$44.3 million) in non-underwriting-related expenses and fees, which consist of fees and expenses of legal advisors and reporting accountants of approximately RMB21.8 million (HK$24.0 million) and other fees and expenses of approximately RMB18.4 million (HK$20.3 million).
The listing expenses above are the latest practicable estimate for reference only, and the actual amount may differ from this estimate.
Our Directors have confirmed that as of the Latest Practicable Date, they were not aware of any circumstances that would give rise to a disclosure requirement under Rules 13.13 to Rules 13.19 of the Listing Rules.
The Directors confirm that, up to the date of this prospectus, there has been no material adverse change in our financial or trading position since June 30, 2025, and there is no event since June 30, 2025 which would materially affect the information shown in the Accountants' Report in Appendix I to this prospectus.
The following unaudited pro forma statement of adjusted consolidated net tangible assets of our Group is prepared in accordance with Rule 4.29 of the Listing Rules and is set out below to illustrate the effect of the Global Offering on the consolidated net tangible liabilities of the Group attributable to the equity shareholders of the Company as of June 30, 2025 as if the Global Offering had taken place on June 30, 2025.
The unaudited pro forma statement of adjusted consolidated net tangible assets has been prepared for illustrative purposes only and, because of its hypothetical nature, it may not give a true picture of the financial position of the Group had the Global Offering been completed as of June 30, 2025 or at any future date.
| | Consolidated net tangible liabilities of the Group attributable to equity shareholders of the Company as of June 30, 2025(1) | Estimated net proceeds from the Global Offering(2) | Estimated net impact upon the termination of financial instruments issued to investors(3) | Unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company | Unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company per Share | |---|---|---|---|---|---| | | RMB'000 | RMB'000 | RMB'000 | RMB'000 | RMB(4) | HK$(5) | | Based on an Offer Price of HK$116.20 per H Share | (6,240,424) | 3,804,426 | 9,564,760 | 7,128,762 | 16.19 | 17.85 |
Notes: (1) The consolidated net tangible liabilities of the Group attributable to the equity shareholders of the Company as of June 30, 2025 is arrived at after (i) deducting intangible assets of RMB55,399,000 and (ii) deducting goodwill of RMB39,379,000 from the total deficit attributable to equity shareholders of the Company of RMB6,145,646,000 which is extracted from the Accountants' Report set out in Appendix I to this prospectus.
(2) The estimated net proceeds from the Global Offering are based on the 37,419,500 H shares expected to be issued pursuant to the Global Offering and the Offer Price of HK$116.20 per H Share, after deduction of the estimated underwriting fees and other expenses relating to the Global Offering paid or payable by the Group (excluding the listing expense that have been charged to profit or loss during the Track Record Period), and does not take into account of any shares which may be issued upon the exercise of the Over-allotment Option or the share incentive plans.
The estimated net proceeds of the Global Offering have been converted to Renminbi at an exchange rate of HK$1 to RMB0.9070. No representation is made that Hong Kong dollars amount have been, could have been or may be converted to Renminbi, or vice versa, at that rate or at any other rate.
(3) As of June 30, 2025, the carrying amount of financial instruments issued to investors of the Company was RMB9,564,760,000 (as set out in Note 26 of Appendix I). Upon the Listing, the redemption rights of the financial instruments will be automatically terminated, and the financial instruments issued to investors will be reclassified from liabilities to equity accordingly.
(4) The unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company per H Share is arrived at after the adjustments referred to in the preceding paragraphs and on the basis that 440,230,190 shares were in issue (being 402,810,690 shares in issue and outstanding as of June 30, 2025 taking into account the Share Subdivision and 37,419,500 H Shares to be issued pursuant to Global Offering) and does not take into account of any shares which may be issued upon the exercise of the Over-allotment Option or the share incentive plans.
(5) The unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company per H Share amounts in RMB are converted to Hong Kong dollar with an exchange rate of RMB1 to HK$1.1016. No representation is made that Renminbi amount have been, could have been or may be converted to Hong Kong dollars, or vice versa, at that rate or at any other rate.
(6) No adjustment has been made to reflect any trading results or other transactions of the Group entered into subsequent to June 30, 2025.
我们已与下列基石投资者（"基石投资者"）订立基石投资协议（"基石投资协议"），该等基石投资者已同意按发售价认购一定数目的发售股份（向下取整至最近100股H股的整数倍），认购总金额约为2,984.2百万港元（不包括经纪佣金、证监会交易征费、会计及财务汇报局交易征费及联交所交易费）（"基石配售"）。
按每股发售股份116.20港元的发售价计算，基石投资者将认购的发售股份总数为25,681,600股H股，约占（i）假设超额配股权未获行使，全球发售项下发售股份的68.63%；（ii）假设超额配股权未获行使，全球发售完成后本公司已发行股本总额的5.83%；及（iii）假设超额配股权获全额行使，全球发售完成后本公司已发行股本总额的5.76%。
本公司认为，借助基石投资者的投资经验及市场地位，基石配售将有助提升本公司的知名度，并表明该等基石投资者对本公司业务及前景充满信心。本公司在日常运营过程中，通过本集团的商业网络或经本公司现有股东或参与全球发售的整体协调人的介绍，与各基石投资者相识。
基石配售将构成国际配售的一部分，基石投资者将不会在全球发售项下（基石投资协议所规定者除外）购买任何发售股份。基石投资者所认购的发售股份与已发行的缴足股份在各方面享有同等权利，并就《上市规则》第19A.13A条的规定计入本公司的公众持股量。紧随全球发售完成后，（i）基石投资者及其紧密联系人均不会成为本公司的主要股东；（ii）基石投资者及其紧密联系人不会仅凭其基石投资而在本公司拥有任何董事会席位；及（iii）就《上市规则》第8.08(3)条的规定而言，三名最大公众股东于本公司实益拥有的股权将少于50%。除按发售价获保证配发相关发售股份外，基石投资者在基石投资协议中并不享有较其他公众股东优先的任何权利。本公司与基石投资者之间并无任何附带安排或协议，亦无任何因上市而直接或间接给予基石投资者的利益，惟按照《上市指引》第4.15章所载原则，按最终发售价获保证配发相关发售股份除外。
据本公司所知、所信及所悉，除JSC International Investment Fund SPC（代表Qizhi SP行事）、JinYi Capital Multi-Strategy Fund SPC Ltd.（代表Structured Credit SP Fund行事）及凌雲光技術國際有限公司（Luster LightTech International Limited）（统称"相关投资者"）（其各自均为本公司现有股东的紧密联系人）外，（i）其他基石投资者及其各自的最终实益拥有人均为独立第三方；（ii）基石投资者均不惯于接受或未曾接受本公司、董事、监事、本公司行政总裁、主要股东、现有股东或其各自的附属公司或其各自的紧密联系人就收购、处置、投票或以其他方式处置发售股份作出的指示；及（iii）基石投资者认购发售股份的资金均非直接或间接由本公司、董事、监事、本公司行政总裁、主要股东、现有股东或其各自的附属公司或其各自的紧密联系人提供融资。联交所已就严格遵守相关规定的要求给予豁免
根据《上市规则》第10.04条以及《上市规则》附录F1第1C(2)段的豁免，允许将国际发售股份配售予相关投资者。详情请参阅本文件"豁免严格遵守《上市规则》相关规定——根据《上市规则》第10.04条及附录F1第1C(2)段就现有股东的紧密联系人作为基础投资者认购发售股份所获豁免及同意"一节。
就本公司所知及经各基础投资者确认，其根据基础投资认购事项项下的认购款项将由其自身内部资源、其股东资源或（就属于基金或投资管理人的基础投资者而言）其为投资者管理的资产支付，且各基础投资者均拥有足够资金以结算其根据基础投资配售项下的各自投资款项。各基础投资者已确认，就基础投资配售事项已取得一切必要审批。
基础投资者（及就任何将通过合格境内机构投资者（QDII）认购本公司发售股份的基础投资者而言，相关QDII）根据基础投资认购事项拟认购的发售股份总数，可能受到国际发售与香港公开发售之间重新分配发售股份的影响（如香港公开发售出现超额认购），如本招股章程"全球发售架构——香港公开发售——重新分配及回拨"一节所述。根据基础投资协议的条款，各基础投资者拟获配发售股份的数目，可能在考虑《上市规则》第18C.09条的规定以及保荐人总协调人（代表国际包销商）行使超额配股权的酌情权后，因应香港公开发售的公众需求而予以削减。此外，保荐人总协调人及本公司可全权酌情调整各基础投资者拟获配发售股份的数目，以遵守《上市规则》第8.08(3)条、第19A.13A条及第19A.13C条以及《上市规则》附录F1（股权证券配售指引）的规定。各基础投资者实际获分配发售股份数目的详情将于本公司于2025年1月7日或前后刊发的配发结果公告中披露。基础投资者已同意在联交所开始买卖本公司H股前，就其已认购的相关发售股份悉数付款。部分基础投资者已同意，其将认购的全部或部分发售股份的交付可延迟至上市日期之后的某一日期。此延迟交付安排旨在便利国际配售中的超额配发。如国际配售中不存在超额配发，则不会作出延迟交付安排。有关超额配股权及稳定价格代理人的稳定价格行动的详情，请参阅本招股章程"全球发售架构——超额配股权"及"全球发售架构——稳定价格措施"各节。
以下有关本公司基础投资者的资料由各基础投资者就基础投资配售事项提供。
| 基础投资者 | 投资总额（注1）（百万元） | 拟认购发售股份数目（注2） | 假设超额配股权未获行使——发售股份约占百分比 | 假设超额配股权未获行使——全球发售完成后本公司已发行总股本约占百分比 | 假设超额配股权获全额行使——发售股份约占百分比 | 假设超额配股权获全额行使——全球发售完成后本公司已发行总股本约占百分比 | |---|---|---|---|---|---|---| | JSC International Investment Fund SPC（代表Qizhi SP行事） | 美元179.0 | 11,985,900 | 32.03% | 2.72% | 27.85% | 2.69% | | JinYi Capital Multi-Strategy Fund SPC Ltd.（代表Structured Credit SP Fund行事） | 美元7.0 | 468,700 | 1.25% | 0.11% | 1.09% | 0.11% | | Perseverance Asset Management | 美元29.0 | 1,941,800 | 5.19% | 0.44% | 4.51% | 0.44% | | 上海高毅资产管理合伙企业（有限合伙）及中国国际金融（香港）证券有限公司（就高毅场外互换相关事宜） | 美元9.0 | 602,600 | 1.61% | 0.14% | 1.40% | 0.14% | | WT Asset Management | 美元44.9 | 3,006,500 | 8.03% | 0.68% | 6.99% | 0.67% | | 泰康人寿 | 美元30.0 | 2,008,800 | 5.37% | 0.46% | 4.67% | 0.45% | | 广发基金 | 美元42.0 | 2,812,300 | 7.52% | 0.64% | 6.54% | 0.63% | | 3W基金 | 美元18.0 | 1,205,200 | 3.22% | 0.27% | 2.80% | 0.27% | | RIME | 港币75.0 | 645,400 | 1.72% | 0.15% | 1.50% | 0.14% | | Optimas Capital Limited | 美元10.0 | 669,600 | 1.79% | 0.15% | 1.56% | 0.15% | | Luster LightTech International Limited | 美元5.0 | 334,800 | 0.89% | 0.08% | 0.78% | 0.08% | | **合计** | **港币2,984.2** | **25,681,600** | **68.63%** | **5.83%** | **59.68%** | **5.76%** |
注1：投资金额不包括经纪佣金、证监会交易征费、联交所交易费及会计及财务汇报局交易征费。
注2：须向下调整至最接近的整手数（每手100股H股）。按"本招股章程及全球发售相关资料——汇率换算"一节所载汇率计算。
of Shanghai Gaoyi Asset Management Co., Ltd. (上海高毅資產管理合夥企業(有限合夥)) ("Shanghai Gaoyi") and/or its managed funds (the "Gaoyi Funds"). Pursuant to the Gaoyi OTC Swaps, CICC FT will subscribe for the Offer Shares on behalf of Shanghai Gaoyi and/or the Gaoyi Funds. CICC FT's subscription of the Offer Shares under the Gaoyi OTC Swaps is for and on behalf of Shanghai Gaoyi and/or the Gaoyi Funds, and CICC FT is acting as a bare trustee and does not have any economic interest in the Offer Shares subscribed thereunder. Shanghai Gaoyi is principally engaged in asset management and its ultimate beneficial owners do not include any single entity or individual holding 30% or more interest in Shanghai Gaoyi.
JSC International Investment Fund SPC（代表Qizhi SP行事）由璟泉啟智（北京）股權投資基金合夥企業（有限合夥）（"JSC啟智（北京）"）間接全資擁有，JSC啟智（北京）的普通合夥人為璟泉善誠管理諮詢（北京）有限公司（"JSC管理諮詢（北京）"），持有其0.07%的合夥份額。JSC管理諮詢（北京）受北京金融控股集團有限公司控制，而北京金融控股集團有限公司由北京市人民政府國有資產監督管理委員會（"北京國資委"）全資擁有。北京京能綠色能源並購投資基金（有限合夥）（"北京京能"）是在中國境內設立的有限合夥企業，為JSC啟智（北京）的有限合夥人，持有其30.69%的合夥份額。北京京能同鑫投資管理有限公司（"北京京能同鑫"）為北京京能的普通合夥人，北京京能的唯一有限合夥人為北京能源集團有限責任公司（"北京能源集團"），持有其88.00%的合夥份額。北京京能同鑫及北京能源集團均由北京國資委全資擁有。除北京京能外，JSC啟智（北京）中並無其他持有30.00%或以上權益的有限合夥人。
JinYi Capital Multi-Strategy Fund SPC Ltd.（"進益資本"）是根據開曼群島《公司法》註冊成立的分類組合公司，為在開曼群島註冊成立的獲豁免有限責任公司。進益（天津）資產管理有限責任公司為進益資本的投資管理人，並最終受范翔控制。除獨立第三方范翔外，進益資本中並無其他最終實益擁有人持有30%或以上的權益。參與全球發售的JinYi Capital Multi-Strategy Fund SPC Ltd. — Structured Credit SP Fund的資金來源為清華大學教育基金會。
清華大學教育基金會成立於1994年，是經中國民政部登記註冊並批准的全國性非公募基金會，以中國教育部為其主管單位。清華大學教育基金會由清華大學發起設立，其資金主要來源於社會捐贈。根據《慈善組織保值增值投資活動管理暫行辦法》及其他適用法規，該基金會可投資於符合相關規定的金融產品，以實現資產保值增值，支持教育及慈善事業的發展。
Perseverance Asset Management International (Singapore) Pte. Ltd.（"Perseverance資產管理"）以全權委託方式擔任不超過六隻投資基金及╱或獨立管理賬戶（統稱"Perseverance基金"）的投資顧問或投資管理人。Perseverance基金中各基金均無單一最終實益擁有人持有30%或以上的權益。Perseverance資產管理為於2018年10月在新加坡註冊成立的私人有限公司，並持有新加坡金融管理局頒發的基金管理資本市場服務牌照。Perseverance資產管理由Perseverance Asset Management International全資擁有，後者主要從事投資管理及投資顧問服務，為獨立第三方。Perseverance資產管理作為投資顧問或投資管理人的若干投資基金曾以基石投資者身份投資於紫金黃金國際有限公司（股票代碼：2259.HK）、寧德時代新能源科技股份有限公司（股票代碼：3750.HK）及先瑞達醫療科技控股有限公司（股票代碼：6669.HK）。Perseverance資產管理以投資顧問或投資管理人的身份代表Perseverance基金與本公司訂立基石投資協議。
中國國際金融（香港）交易有限公司（"中金交易"）為中國國際金融股份有限公司的全資子公司，中國國際金融股份有限公司的股份分別於上海證券交易所（股票代碼：601995）及聯交所（股票代碼：3908）上市。
中金交易及中國國際金融股份有限公司（"中金公司"）將與上海高毅資產管理合夥企業（有限合夥）（"上海高毅"）及╱或其管理的基金（"高毅基金"）各自訂立一系列跨境Delta一對一場外股票掉期交易（統稱"高毅場外掉期"）。根據高毅場外掉期，中金交易將代表上海高毅及╱或高毅基金認購發售股份。中金交易根據高毅場外掉期認購發售股份係代表上海高毅及╱或高毅基金行事，中金交易作為名義受託人，對所認購的發售股份不持有任何經濟利益。上海高毅主要從事資產管理業務，其最終實益擁有人中並無任何單一實體或個人持有上海高毅30%或以上的權益。
其他最终客户（"中金财富最终客户（高毅）"），据此，中金财富将以非全权委托方式持有发售股份，以对冲高毅场外互换交易，而相关发售股份的经济风险和回报将转移至中金财富最终客户（高毅），但须缴付惯常费用及佣金。高毅场外互换交易将由中金财富最终客户（高毅）全额出资。在高毅场外互换交易存续期间，中金财富认购的发售股份的所有经济回报将通过高毅场外互换交易转移至中金财富最终客户（高毅），所有经济损失亦由中金财富最终客户（高毅）通过高毅场外互换交易承担，中金财富不参与任何与发售股份相关的经济回报或承担任何经济损失。高毅场外互换交易与发售股份的表现挂钩，中金财富最终客户（高毅）可在锁定期届满后（锁定期自中金财富与本公司订立基石协议之日起计，至上市日期后六个月届满之日止），自行决定要求提前终止高毅场外互换交易，届时中金财富可根据高毅场外互换交易的条款及条件出售发售股份并以现金结算高毅场外互换交易。尽管中金财富将自行持有相关发售股份的法定所有权，但根据其内部政策，在高毅场外互换交易存续期间，其将不行使相关发售股份所附带的投票权。据中金财富在作出一切合理查询后所知，中金财富最终客户（高毅）的每一方均为中金财富、中国国际资本股份有限公司香港证券有限公司（"中金香港证券"）及与中金香港证券属同一集团成员公司的独立第三方，且没有任何单一最终实益拥有人持有中金财富最终客户（高毅）任何一方30%或以上的权益。
中金财富最终客户（高毅）为由上海高毅资产管理合夥企业（有限合夥）（"上海高毅"）管理的投资基金。上海高毅为在中国境内成立的有限合夥企业，主要从事资产管理及投资管理业务，以二级市场投资为主。上海高毅管理的若干投资基金就紫金黄金国际有限公司（股票代码：2259）、南京维立志博生物科技股份有限公司（股票代码：9887）及宁德时代新能源科技股份有限公司（Contemporary Amperex Technology Co., Limited）（股票代码：3750）的基石投资订立了delta-one场外互换交易，并承担所有经济回报和损失。上海高毅持有由中国证券投资基金业协会认定的私募投资基金管理人资格。上海高毅的普通合夥人为上海高毅投资管理有限公司（"高毅投资"）。Perseverance Asset Management为上海高毅的关联方。经上海高毅确认，没有任何单一最终实益拥有人持有中金财富最终客户（高毅）30%或以上的权益。上海高毅及中金财富最终客户（高毅）各方均为独立第三方。
本公司已就上市规则附录F1第1C(2)段向联交所申请豁免，联交所已批准本公司将发售股份配发予中金财富。请参阅"偏离上市规则规定的严格合规要求的豁免——就若干基石投资者（其为关联客户）拟认购H股事宜的同意书"。
WT Asset Management Limited（"WT资产管理"）为一家在香港注册成立的有限责任公司，已获证监会发牌从事第9类（资产管理）受规管活动。WT资产管理由王通书先生100%实益拥有，王通书先生为独立第三方。WT资产管理已同意促使其拥有全权投资管理权的若干投资者，即WT China Fund Limited、WT China Focus Fund、WT Growth Fund及╱或独立管理账户（由WT资产管理（作为投资经理）专业管理的投资组合，投资者直接持有相关投资的所有权）（统称"基金"），认购若干数量的
Offer Shares. The Funds are managed by WT Asset Management as investment manager. The Funds pursue to achieve absolute return and long-term capital appreciation by investing primarily in the listed securities of companies which have great exposure or material impact by the PRC. Investors of the Funds include but are not limited to pension funds, fund of funds, family offices and other sophisticated institutional investors. Save for Mr. Tongshu Wang (王通書) who hold over 30% interests in WT Growth Fund and WT China Focus Fund, and the single ultimate beneficial owner of the segregated management account which is a pension fund based in North America respectively, no other single ultimate beneficial owner holds 30% or more interests in the Funds. Each of the Funds is an Independent Third Party. As of September 30, 2025, the total AUM of the Funds is approximately US$4.25 billion.
泰康人寿保险股份有限公司（"泰康人寿"，Taikang Life Insurance Co., Ltd），一家在中国注册成立的公司，是泰康保险集团股份有限公司的全资子公司。泰康保险集团股份有限公司没有持股30%或以上的股东。泰康人寿为个人及家庭提供全面的人身保障及投资理财产品与服务。所提供的产品对应客户在儿童及青少年、女性及高收入人群等细分市场的不同需求，同时满足医疗健康与意外保障、养老及财富管理等多维度需求。泰康保险集团股份有限公司是一家以保险、资产管理及健康养老为主业的保险金融服务集团。该公司总部位于北京，旗下子公司包括泰康人寿、泰康资产、泰康养老、泰康健投、泰康健康及TK.CN。其产品覆盖人寿保险、互联网金融保险、企业年金、资产管理、健康养老、健康管理及商业地产等领域。
Taikang Life Insurance Co., Ltd ("Taikang Life"), a company incorporated in China, is a wholly owned subsidiary of Taikang Insurance Group Inc. There is no shareholder holding 30% or more in Taikang Insurance Group Inc. Taikang Life provides a full range of personal security and investment and wealth management products and services for individuals and families. The products on offer correspond to the different requirements of customers in terms of market segments such as the children and teenagers, females and high-income population groups. They also meet multidimensional demands regarding health care and accident cover, pensions and wealth management, among others. Taikang Insurance Group Inc. is an insurance and financial service conglomerate focused on insurance, asset management and health and elderly care as main businesses. The Beijing-headquartered company consists of several subsidiaries including Taikang Life, Taikang AMC, Taikang Pension, Taikang Healthcare, Taikang Health, and TK.CN. Its product offering covers life insurance, internet-based financial insurance, enterprise annuity, asset management, health and elderly care, health management and commercial real estate, among others.
广发基金管理有限公司（GF Management Co., Ltd.）（"广发基金管理"）及广发国际资产管理有限公司（GF International Investment Management Limited）（"广发基金香港"，与广发基金管理合称"广发基金"）已分别与本公司签订基石投资协议。
GF Fund Management Co., Ltd. (廣發基金管理有限公司) ("GF Fund Management") and GF International Investment Management Limited (廣發國際資產管理有限公司) ("GF Fund HK", together with GF Fund Management, "GF Fund") have, respectively, entered into Cornerstone Investment Agreement with our Company.
GF Fund Management was established on August 5, 2003. GF Fund Management and its subsidiaries are licensed to conduct business as Qualified Investment Manager of Public Fund, Entrusted Domestic Investment Manager of National Social Security Fund (NSSF), qualified investment management institution of Basic Pension Insurance Funds, qualified fund management company to provide asset management services for specific clients, Qualified Domestic Institutional Investor (QDII), RMB Qualified Foreign Institutional Investor (RQFII), Qualified Foreign Institutional Investors (QFII), Qualified Domestic Limited Partner (QDLP), entrusted insurance Funds investment manager, entrusted investment manager of asset management for Insurance Security Funds and fund investment advisor, making it a large Fund management company with comprehensive asset management capabilities and experience. The controlling shareholder of GF Fund Management is GF Securities Co., Ltd. (廣發證券股份有限公司) ("GF Securities"), a limited company listed on the Stock Exchange (stock code: 1776) and Shenzhen Stock Exchange (stock code: 000776), which owns 54.53% shareholding in GF Fund Management. Apart from GF Securities, no other shareholder has a 30% or more shareholding in GF Fund Management.
GF Fund HK is a wholly owned subsidiary of GF Fund Management. GF Fund HK (central number in the Hong Kong Securities and Futures Commission license: AXL121) was incorporated in Hong Kong in December 2010. GF Fund HK is licensed by SFC to carry on Type 1 (dealing in securities), Type 4 (advising on securities) and Type 9 (asset management) regulated activities in Hong Kong. GF Fund HK serves as the global investment and business platform for its parent company, GF Fund Management. As GF Fund Management's window company overseas, GF Fund HK strategically connects China and the overseas market. GF Fund HK capitalizes the investment and research capabilities of GF Management and its competitive advantage in the overseas market to provide comprehensive quality service to its clients.
The subscription of the Offer Shares as a cornerstone investor will be made by GF Fund Management and GF Fund HK in their capacity as the discretionary investment manager of Guangfa Global Selective Equities, Guangfa Theme Investment Discretionary Account No.37, Guangfa Theme Investment Discretionary Account No.36, Guangfa Xinhui Discretionary Account No.5, Guangfa Theme Investment Discretionary Account No.53, Guangfa Organic Growth (QDII) Discretionary Account No.8 and Global Absolute Return Fund under their management. Based on the best knowledge of GF Fund Management and GF Fund HK, each fund is an Independent Third Party, and no ultimate beneficial owner holds 30% or more interest.
3W Fund Management Limited ("3W Fund") is incorporated in Hong Kong with limited liability and licensed by the Hong Kong SFC to carry out type 9 (asset management) regulated activity. 3W Fund, which is ultimately wholly owned by Mr. Weiwei WU, an Independent Third Party, has agreed to procure 3W Global Fund, over which 3W Fund has discretionary investment management power, to subscribe for such number of the Investor Shares. 3W Global Fund pursues to maximize absolute return and seek long-term capital growth primarily through fundamental investment principle with value approach. No single investor holds 30% or more interests in 3W Global Fund.
RIME Capital Limited ("RIME") is incorporated in Hong Kong with limited liability and licensed by the SFC to carry on type 1, 4, 9 regulated activities. The firm is ultimately owned by ZHUO Ying, who owns 64% shares of RIME and is an Independent Third Party. RIME is a discretionary investment manager of RIME Selected Shares LPF ("RIME LPF") and certain funds and separated managed accounts. RIME has agreed to procure RIME LPF to subscribe for the Offer Shares. RIME LPF is a limited partnership incorporated in Hong Kong and the ultimate beneficial owner of RIME LPF is SENG Iek Chon, an Independent Third Party, who owns 60% shares of the RIME LPF. Save for SENG Iek Chon, there is no other ultimate beneficial owner holding 30% or more interest in RIME LPF.
Optimas Capital Limited ("OCL") is an investment management company incorporated in Hong Kong and wholly owned by Ms. Cai Yun. OCL principally provides multi-product investment services to global investors and is licensed by the Securities and Futures Commission to carry on Type 4 (Advising on Securities) and Type 9 (Asset Management) regulated activities under Part V of the Securities and Futures Ordinance. OCL has entered into Cornerstone Investment Agreement with the Company, on behalf of Optimas GCM Fund SPC ("GCM Fund") as its sub-manager.
GCM Fund is a segregated portfolio company incorporated in the Cayman Islands and its principal business activities include investment and trading. GCM Fund seeks to generate investment returns by primarily investing on companies with high growth prospects globally. No ultimate beneficial owner holds 30% or more interest in GCM Fund.
Luster LightTech International Limited (凌雲光技術國際有限公司) is a wholly-owned subsidiary of Luster LightTech Co., Ltd. (凌雲光技術股份有限公司) ("Luster") and primarily focuses on international trade and marketing activities in the field of optoelectronic information. Luster is a company listed on the Shanghai Stock Exchange (stock code: 688400), whose business consists of machine vision and fiber optics based on optical technology innovation, and is committed to becoming a global leader in the fields of visual artificial intelligence and optoelectronic information. Approval by Luster's shareholders and Shanghai Stock Exchange is not required for Luster LightTech International Limited's subscription for the Offer Shares pursuant to the Cornerstone Investment Agreement.
Luster is one of our customers. Our Group has entered into sales contracts with Luster in connection with our provision of services to Luster in the ordinary course of business on an arm's length basis on normal commercial terms.
The obligation of each of Cornerstone Investors to acquire the Offer Shares under the respective Cornerstone Investment Agreement is subject to, among other things, the following closing conditions:
(i) the Hong Kong Underwriting Agreement and the International Underwriting Agreement being entered into and having become effective and unconditional (in accordance with their respective original terms or as subsequently waived or varied by agreement of the parties thereto) by no later than the time and date as specified in the Hong Kong Underwriting Agreement and the International Underwriting Agreement, and neither the Hong Kong Underwriting Agreement nor the International Underwriting Agreement having been terminated;
每位基石投资者均已同意，在未事先获得本公司、独家保荐人及整体协调人书面同意的情况下，其在上市日期后六个月期间（首尾两日均计算在内）（"禁售期"）内，不论直接或间接，均不会以任何方式出售其根据各自基石投资协议所认购的发售股份，但某些有限情况除外，例如转让予其任何全资附属公司（该等全资附属公司须受与基石投资者相同义务的约束，包括禁售期限制）。
(iv) 基石投资者根据各自基石投资协议作出的协议、陈述、保证、确认、承诺及确认函（于各自基石投资协议签署日期）及将会（于各自基石投资协议所界定的交割时）在各方面均属准确和真实，且不具误导性，且相关基石投资者并无重大违反各自基石投资协议。
每位基石投资者均已同意，在未事先获得本公司、独家保荐人及整体协调人书面同意的情况下，其在上市日期后六个月期间（首尾两日均计算在内）（"禁售期"）内，不论直接或间接，均不会以任何方式出售其根据各自基石投资协议所认购的发售股份，但某些有限情况除外，例如转让予其任何全资附属公司（该等全资附属公司须受与基石投资者相同义务的约束，包括禁售期限制）。
Assuming that the Over-allotment Option is not exercised, after deducting the underwriting commissions and other estimated offering expenses payable by us in connection with the Global Offering, and based on the Offer Price of HK$116.20 per Share, we estimate that we will receive net proceeds of approximately HK$4,173.4 million from the Global Offering. We intend to use the proceeds from the Global Offering for the purposes and in the amounts set forth below:
(i) Approximately 70.0% (or HK$2,921.4 million) will be used to continuously strengthen our research and development capabilities in general-purpose large AI models, of which approximately 34.5% (or HK$1,439.8 million), 26.0% (or HK$1,085.1 million) and 9.5% (or HK$396.5 million) is expected to be used in 2026, 2027 and 2028, respectively:
(a) approximately 60.0% (or HK$2,504.0 million) will be allocated to enhance our large model capabilities and optimize our training infrastructure. We are committed to consolidating our technological leadership by investing in the iterative development of our foundation models, key algorithms and large-scale training and inference infrastructures. Specifically,
• approximately 20.0% (or HK$834.7 million) will be used for the ongoing research and development of pre-trained large models, with a focus on maintaining our technological leadership and enhancing the core general capabilities of such models, which will enable our models to develop foundational capabilities and accelerate downstream learning efficiency. We plan to further invest in enhancing our computing power, attracting and retaining top-tier research and development professionals and acquiring high-quality structured data;
• approximately 20.0% (or HK$834.7 million) will be used for the ongoing research and development of deep reasoning models, with the aim of improving their self-learning capabilities through training. We will continue to invest in building versatile, powerful large models, focusing on designing new model architecture, and optimizing training infrastructure to reduce training costs and boost performance. This would include new attention and memory mechanisms, infinity context, test-time and online learning, deep reasoning algorithms, self-refinement and self-evolution paradigms. By these means, our models will be able to autonomously explore, reflect and operate in both online and real-world environments, enabling them to tackle complex tasks and continuously improve their performance through self-learning and reflection;
• approximately 10.0% (or HK$417.3 million) will be used for the ongoing research and development of AI agents, with an emphasis of enhancing the self-learning capabilities of these models through training. We plan to further upgrade our agent workspace that enables users to easily and seamlessly integrate diverse model applications and tools. This will facilitate the efficient deployment of AI agent solutions deeply tailored to specific industries, regions and scenarios, and drive advancements in intelligent automation; and
• approximately 10.0% (or HK$417.3 million) will be used to explore cutting-edge and emerging technologies, with the aim of maintaining our technological leadership in the development of advanced AI in China.
(b) approximately 10.0% (or HK$417.3 million) will be allocated to strengthen our GLM framework, optimize our data processing platforms, and build high-quality corpora. Specifically,
• approximately 5.0% (or HK$208.7 million) will be used to strengthen our GLM framework. In particular, we plan to continue investing in the research and development of our technology infrastructure to improve model performance, increase the efficiency of underlying computing resources and ensure that computing resources provided by our computing resource partners are optimally suited to our models and sufficient for our scalable training and model deployment across diverse platforms and hardware; and
• approximately 5.0% (or HK$208.7 million) will be used to optimize our data processing platform and build high-quality corpora. We plan to upgrade our data processing platforms to support high-quality storage and analysis of datasets, providing a solid foundation for further expanding the frontiers of large model capabilities.
(ii) Approximately 10.0% (or HK$417.3 million) will be used to continuously optimize our MaaS platform by offering the latest foundation models and training/inference tools and infrastructures, of which approximately 6.0% (or HK$250.4 million), 3.0% (or HK$125.2 million) and 1.0% (or HK$41.7 million) is expected to be used in 2026, 2027 and 2028, respectively:
(a) approximately 5.0% (or HK$208.7 million) will be allocated to the continuous deployment of large models with varying parameter scales, designed to be compatible with a wide range of computing resources and end devices. We plan to provide more offerings over our MaaS platform by strategically expanding our model portfolio across a spectrum of parameter scales, ensuring optimal performance on various computing resources and devices. By engineering models that adapt to differing computational capabilities, we reduce hardware barriers and democratize access to advanced AI technologies. Specifically,
• approximately 2.5% (or HK$104.3 million) will be used to strengthen the deployment of our platform portfolio by optimizing our large models and ensuring their compatibility with a wide range of computing resources and end devices, which enables our models to better match end devices' computational capacity and reduce hardware limitations. We also plan to diversify our MaaS platform service offerings, reach more end users through more IoT devices empowered by our AI models, further empowering a broad range of industries with our large model capabilities; and
• approximately 2.5% (or HK$104.3 million) will be used for the research and development of large model products that run on consumer-grade hardware, including our applications and hardware products such as consumer electronics. We remain focused on delivering convenient, user-friendly experiences that make large model technology more accessible, enabling companies and organizations of all sizes, as well as individual users, to unlock large models' potential.
(b) approximately 5.0% (or HK$208.7 million) will be allocated to strengthen our sales and marketing teams and conduct additional marketing activities to further enhance our brand name, increase our market penetration and maintain strong connections in targeted markets to serve our clients. By working closely with our client support team, our sales and marketing teams can deliver tailored solutions that address specific customer needs and maximize the value of our services. By strengthening our sales and marketing teams, we will be able to gain deeper insights into evolving market trends, which in turn supports the continuous optimization of our MaaS platform.
(iii) Approximately 10.0% (or HK$417.3 million) will be used for the development of our business partner network, as well as for strategic investments, of which approximately 4.0% (or HK$166.9 million), 4.0% (or HK$166.9 million) and 2.0% (or HK$83.5 million) is expected to be used in 2026, 2027 and 2028, respectively. Specifically, we plan to invite more computing
infrastructure partners to join our upstream value chain, further enhancing the versatility and adaptability of our models. In addition, we plan to focus on investments in and acquisitions of high-quality upstream and downstream assets within the large model industry, and on establishing industry networks and strategic alliances.
Our selection criteria for potential target companies include but are not limited to those (i) whose business is closely related to AI technologies and applications, including but not limited to large models, intelligent algorithms, AI platforms and AI-enabled applications. We will also consider companies that provide sector-specific AI solutions; (ii) have a solid user/customer base and at least two years of stable operations; (iii) possess mature products or technology platforms with continuous R&D and technological iteration capabilities; (iv) led by senior management with substantial industry experience and supported by professional R&D personnel; and (v) demonstrate solid growth potential in their respective AI fields. By investing in these types of companies, we expect to strengthen our capabilities across the AI value chain. Strategic acquisitions in targeted AI sectors are intended to accelerate our model training efficiency, and enhance our multimodal capabilities. These improvements support our overall competitiveness, enabling us to deliver more advanced, efficient and diversified AI solutions to our customers. We intend to make aforementioned investments and acquisitions mainly through equity, both controlling and non-controlling, evaluated on the case-by-case basis.
According to Frost & Sullivan, as of June 30, 2025, there were over 3,000 potential targets in the market that met our criteria. As such, our Directors are of the opinion that there is a sufficient number of suitable target companies available in the market for our investment and acquisition plans. The criteria are subject to adjustment based on changes in the market conditions and our strategic needs. However, there is no guarantee that we will be able to invest in or acquire the targets that we desire. As of the Latest Practicable Date, we had not identified any potential acquisition or investment targets and had not entered into any letters of intent or definitive agreement for the acquisition of any company.
(iv) Approximately 10.0% (or HK$417.3 million) will be used for working capital and other general corporate purposes.
The additional net proceeds that we would receive if the Over-allotment Option were exercised in full would be HK$632.6 million (based on the Offer Price of HK$116.20 per Share).
To the extent that the net proceeds from the Global Offering are either more or less than expected, we will adjust our allocation of the net proceeds for the above purposes on a pro rata basis.
To the extent that the net proceeds of the Global Offering are not immediately used for the above purposes, we will only deposit such funds into short-term interest-bearing accounts at licensed commercial banks and/or other authorized financial institutions (as defined under the Securities and Futures Ordinance or applicable laws and regulations in other jurisdictions). In such event, we will comply with the appropriate disclosure requirements under the Listing Rules. We will issue an appropriate announcement if there is any material change to the above proposed use of proceeds in accordance with the Listing Rules.
China International Capital Corporation Hong Kong Securities Limited Huatai Financial Holdings (Hong Kong) Limited BOCOM International Securities Limited Guotai Junan Securities (Hong Kong) Limited China Merchants Securities (HK) Co., Limited Futu Securities International (Hong Kong) Limited Tiger Brokers (HK) Global Limited SPDB International Capital Limited CMB International Capital Limited Zheshang International Financial Holdings Co., Limited Riche Bright Securities Limited UOB Kay Hian (Hong Kong) Limited
The Hong Kong Underwriting Agreement was entered into on December 29, 2025. Pursuant to the Hong Kong Underwriting Agreement, our Company is offering the Hong Kong Offer Shares for subscription on the terms and conditions set out in this prospectus and the Hong Kong Underwriting Agreement at the Offer Price.
Subject to (a) the Stock Exchange granting approval for the listing of, and permission to deal in, the H Shares to be issued pursuant to the Global Offering (including the H Shares which may be issued pursuant to the exercise of the Over-allotment Option and the H Shares to be converted from Unlisted Shares) as mentioned herein on the Main Board of the Stock Exchange and such approval not subsequently having been revoked and (b) certain other conditions set out in the Hong Kong Underwriting Agreement, the Hong Kong Underwriters have agreed severally but not jointly to subscribe for, or procure subscribers for, their respective applicable proportions of the Hong Kong Offer Shares being offered which are not taken up under the Hong Kong Public Offering on the terms and conditions set out in this prospectus and the Hong Kong Underwriting Agreement.
The Hong Kong Underwriting Agreement is conditional on, among other things, the International Underwriting Agreement having been executed and becoming unconditional and not having been terminated in accordance with its terms.
The Sponsor-Overall Coordinator (for itself and on behalf of the Hong Kong Underwriters) can, in its sole and absolute discretion, by a notice in writing to us, terminate the Hong Kong Underwriting Agreement with immediate effect if, at any time at or prior to 8:00 a.m. on the Listing Date:
任何新法律或法规，或任何现行法律或法规的变更或发展，或预期变更或发展，或任何可能导致现行法律或法规发生变更或发展（包括预期变更或发展）的事件或一系列事件或情况，或任何法院或香港、中国、美国或与本集团或全球发售相关的任何其他司法管辖区内任何主管机关对现行法律或法规的解释或适用（每个称为"相关司法管辖区"，统称为"相关司法管辖区"）；
监管、货币、信贷或市场状况或情绪、税收、股票证券或货币汇率或管制，或任何货币或交易结算系统，或外商投资法规（包括但不限于港元、美元或人民币对任何外币的贬值，港元与美元挂钩制度或人民币与任何外币挂钩制度的变更）或其他金融市场（包括但不限于股票及债券市场、货币及外汇市场、银行同业市场及信贷市场的状况及情绪）的变化，涉及或影响任何相关司法管辖区，或影响对发售股份的投资；
任何属不可抗力性质的事件或一系列事件或情况（包括但不限于任何政府行为、宣布地区性、全国性或国际性紧急状态或战争、灾难、危机、经济制裁、罢工、劳资纠纷、其他工业行动、停工、火灾、爆炸、洪水、海啸、地震、火山爆发、内乱、暴动、叛乱、公共秩序混乱、政府运作瘫痪、战争行为、流行病、大流行病、传染病爆发或升级、变异或加剧、意外事故或交通运输中断或延误、局部、全国性、地区性或国际性敌对行动的爆发或升级（无论战争是否已经或曾经宣布）、天灾或恐怖主义行为（无论责任是否已被认领））发生于或影响任何相关司法管辖区；
对任何相关司法管辖区内证券交易所、上海证券交易所、深圳证券交易所、纽约证券交易所或纳斯达克全球市场的股份或证券交易一般性实施或宣布任何暂停、中止或限制措施（包括但不限于实施或要求任何最低或最高价格限制或价格范围）；
除非事先获得保荐人兼总协调人的书面同意，本公司根据《公司（清盘及杂项条文）条例》或《上市规则》，或应联交所及/或证监会的任何要求或请求，就发售股份的发售及出售发出或被要求发出本招股章程或其他文件的补充或修订；
(vii) 以任何形式对本集团任何成员或任何控股股东直接或间接实施经济制裁或出口管制，或由任何相关司法管辖区或针对任何相关司法管辖区实施，或以任何形式直接或间接撤销于香港包销协议日期存在的交易特权，惟本招股章程所披露者除外；
(i) 已经、将会或可能对本集团整体的利润、亏损、经营业绩、资产、负债、总体事务、业务、管理、表现、前景、股东权益、地位或状况（财务、贸易或其他方面）产生重大不利影响或任何涉及预期重大不利影响的发展（"重大不利影响"）；
(iii) 已经、将会或可能导致香港承销协议、香港公开发售或全球发售的任何重要部分无法按预期执行或落实，或香港公开发售和/或全球发售无法进行，或无法就全球发售进行推介，或无法按发售文件（定义见其中）所拟定的条款及方式交付或分发发售股份；或
(i) 本招股说明书、正式通知、中国证监会备案文件和/或本公司就香港公开发售发出或使用的任何通知、公告、广告、通讯或其他文件（包括任何补充或修订文件）（统称"相关文件"）中（承销商信息除外，即承销商的标志、名称、地址及资质）所载的任何陈述，在发出时或其后在任何重大方面失实、不正确、不准确或具有误导性；或任何该等相关文件中所载的任何估计、预测、意见表达、意向或预期，在发出时或其后在任何方面不公平或具有误导性，或基于失实、不诚实或不合理的假设，或以不诚信方式作出；
(iii) 本公司或担保股东（定义见香港承销协议）在香港承销协议或国际承销协议中作出的任何保证（定义见香港承销协议）遭到违反，或任何事件或情况导致该等保证在任何方面失实、不正确或具有误导性，且将对全球发售产生重大不利影响；
(viii) 上市委员会对根据全球发售（包括根据行使超额配股权）已发行及将予发行的H股的上市及买卖批准被拒绝或未能在上市日期或之前获得（惯常条件除外），或获批准后，该批准被撤回、取消、附加条件（惯常条件除外）、撤销或暂缓；
with the prior written consent of the Stock Exchange, he/she/it will not, and will ensure that his/her/its associates will not, during the period commencing on the date by reference to which disclosure of his/her/its shareholding in our Company is made in this prospectus and ending on the date which is 24 months from the Listing Date (the "Key Persons Lock-up Period"):
(a) dispose of (including, but not limited to, any sale, transfer, or other disposition), or agree to dispose of, any securities in our Company beneficially owned by him/her/it as shown in this prospectus; or
(b) create or agree to create any options, rights, interests, or encumbrances (save as pursuant to a pledge or charge as security in favor of an authorized institution (as defined in the Banking Ordinance (Chapter 155 of the Laws of Hong Kong)) for a bona fide commercial loan) in respect of any securities in our Company beneficially owned by him/her/it as shown in this prospectus.
若集团任何成员公司遭遇清盘或清算申请令或呈请，或集团任何成员公司与其债权人达成任何妥协或安排，或进入任何债务重组安排，或通过任何有关清盘的决议，或集团任何成员公司全部或部分资产或业务被委任临时清盘人、接管人或管理人，或集团任何成员公司发生任何类似情况，且上述情况将对全球发售产生重大不利影响；
(xii) (A) 中国证监会就中国证监会备案发出的接受通知及/或在中国证监会网站上公布的中国证监会备案结果遭到拒绝、撤回、撤销或宣告无效；或 (B) 除获保荐人-总协调人事先书面同意外，本公司根据中国证监会规则或应中国证监会的任何要求或请求，须就中国证监会备案发出或须发出补充文件或修订文件；或 (C) 中国证监会备案不符合中国证监会规则或任何其他适用法律；或
(xiii) (i) 在簿记建档过程中落单或确认的大部分订单，或 (ii) 任何基础投资者根据与相关基础投资者签订的基础投资协议所作出的任何投资承诺，已被撤回、终止或取消，或相关订单及╱或投资承诺的付款未在规定时间及方式内收妥或结清，或存在其他情形。
根据《上市规则》第10.08条，本公司已向联交所承诺，自上市日期起六个月内（无论该等股份或证券的发行是否将于上市日期起六个月内完成），本公司将不行使权力发行任何额外股份或可转换为股份的证券（无论是否属于已上市的类别），亦不就该等发行订立任何协议，惟以下情况除外：(a) 根据全球发售（包括可能因行使超额配股权而发行的任何额外股份）；或 (b) 符合《上市规则》第10.08条规定的任何情况。
根据《上市规则》第10.07条及第18C.13条，各控股股东成员已分别向联交所及本公司承诺，在本招股说明书披露其于本公司持股量的参考日期起至上市日期起计12个月届满之日止的期间内，未经联交所事先书面同意或除非符合《上市规则》的适用规定，其不会且将促使相关登记持有人不会出售或订立任何出售协议，或就本招股说明书显示其为实益拥有人的任何本公司证券以其他方式设立任何期权、权利、权益或产权负担（根据《银行业条例》（香港法例第155章）所界定的认可机构就真诚商业贷款所作的质押或押记除外）。
根据《上市规则》第10.07(2)条附注(3)，各控股股东成员已进一步向联交所及本公司承诺，在本招股说明书披露其于本公司持股量的参考日期起至上市日期起计12个月届满之日止的期间内，其将且将促使相关登记持有人：
(ii) 当其收到质权人╱押记人的口头或书面指示，表明任何已质押╱押记的本公司证券将被出售时，立即通知本公司有关指示。
本公司在获控股股东告知上述第(i)及(ii)段所述事宜（如有）后，将立即通知联交所，并须按照《上市规则》的当时规定，以公告形式披露有关事宜。
根据《上市规则》第18C.14(1)条，各关键人士及其紧密联系人（"关键人士"），包括刘博士、唐博士、李博士、徐博士、张博士、王少兰、北京联派、知灯、汇汇、严星宇、顾小涛、张波及杜正孝，已不可撤回及无条件地向本公司及联交所承诺，除根据全球发售外，或
the Over-allotment Option, it/he/she shall not and shall procure that its/his/her respective close associates and the relevant registered Shareholder(s) controlled by it/him/her shall not, in the period commencing on the date by reference to which disclosure of its/his/her shareholdings (or its/his/her respective close associate's shareholdings, if applicable) in our Company is made in this prospectus and ending on the date which is 12 months from the Listing Date, dispose of, nor enter into any agreement to dispose of or otherwise create any options, rights, interests or encumbrances (save as (i) pursuant to a pledge or charge as security in favor of an authorized institution (as defined in the Banking Ordinance (Chapter 155 of the Laws of Hong Kong) for a bona fide commercial loan, or (ii) disposing any interest in such securities of our Company in the circumstances provided under Rule 18C.15 of the Listing Rules) in respect of, any of our securities that it/he/she (or its/his/her respective close associate, if applicable) is shown to beneficially own in this prospectus.
In accordance with Note 2 to Rule 18C.14 of the Listing Rules, each of the Key Persons has further irrevocably and unconditionally undertaken to us and the Stock Exchange, and shall procure its/his/her respective close associates, that within the period commencing on the date by reference to which disclosure of its/his/her shareholdings (or its/his/her respective close associate's shareholdings, if applicable) in our Company is made in this prospectus and ending on the date which is 12 months from the Listing Date, it/he/she will: (a) when it/he/she (or its/his/her respective close associate) pledges or charges any securities in our Company beneficially owned by it/him/her (or by its/his/her respective close associate) in favor of an authorized institution (as defined in the Banking Ordinance (Chapter 155 of the Laws of Hong Kong)), immediately inform us in writing of such pledge or charge together with the number of our securities so pledged or charged; and (b) when it/he/she (or its/his/her respective close associate) receives indications, either verbal or written, from the pledgee or chargee that any of our pledged or charged securities beneficially owned by it/him/her (or by its/his/her respective close associate) will be disposed of, immediately inform us in writing of such indications. We will also inform the Stock Exchange as soon as we have been informed of the matters mentioned in the paragraphs (a) and (b) above by any of the Key Persons and make a public disclosure in relation to such information by way of an announcement in accordance with the Listing Rules.
Pursuant to Rule 18C.14(2) of the Listing Rules, each of the Pathfinder SIIs has irrevocably and unconditionally undertaken to us and to the Stock Exchange that except pursuant to the Global Offering, or the Over-allotment Option, it shall not, and shall procure that the relevant registered holder(s) shall not, in the period commencing on the date by reference to which disclosure of its shareholdings in our Company is made in this prospectus and ending on the date which is 6 months from the Listing Date, dispose of, nor enter into any agreement to dispose of or otherwise create any options, rights, interests or encumbrances (save as (i) pursuant to a pledge or charge as security in favor of an authorized institution (as defined in the Banking Ordinance (Chapter 155 of the Laws of Hong Kong) for a bona fide commercial loan, or (ii) disposing any interest in such securities of our Company in the circumstances provided under Rule 18C.15 of the Listing Rules) in respect of, any of our securities that it is shown to beneficially own in this prospectus.
In accordance with Note 2 to Rule 18C.14 of the Listing Rules, each of the Pathfinder SIIs has further irrevocably and unconditionally undertaken to us and the Stock Exchange that, within the period commencing on the date by reference to which disclosure of its shareholdings in our Company is made in this prospectus and ending on the date which is 6 months from the Listing Date, it will: (a) when it pledges or charges any securities in our Company beneficially owned by it in favor of an authorized institution (as defined in the Banking Ordinance (Chapter 155 of the Laws of Hong Kong)), immediately inform us in writing of such pledge or charge together with the number of our securities so pledged or charged; and (b) when it receives indications, either verbal or written, from the pledgee or chargee that any of our pledged or charged securities beneficially owned by it will be disposed of, immediately inform us in writing of such indications. We will also inform the Stock Exchange as soon as we have been informed of the
matters mentioned in the paragraphs (a) and (b) above by any of the Pathfinder SIIs and make a public disclosure in relation to such information by way of an announcement in accordance with the Listing Rules.
Pursuant to the Hong Kong Underwriting Agreement, our Company has undertaken to the Sole Sponsor, the Sponsor-Overall Coordinator, the Overall Coordinators, the Hong Kong Underwriters and the Capital Market Intermediaries not to (save for the offer, allotment and issue of the Offer Shares by our Company pursuant to the Global Offering including pursuant to any exercise of the Over-allotment Option), without the prior written consent of the Sole Sponsor and Sponsor-Overall Coordinator (for itself and on behalf of the Hong Kong Underwriters) and unless in compliance with the Listing Rules, at any time during the period commencing on the date of the Hong Kong Underwriting Agreement and ending on, and including, the last date of the six months after the Listing Date (the "First Six-Month Period"):
(ii) enter into any swap or other arrangement that transfers to another, in whole or in part, any of the economic consequences of subscription or ownership (legal or beneficial) of any Locked-up Securities or any interest therein (including, without limitation, any securities convertible into or exchangeable or exercisable for, or that represent the right to receive, or any warrants or other rights to purchase, any Shares); or
(iii) enter into any transaction with the same economic effect as any transaction described in paragraphs (i) or (ii) above; or
(iv) offer to or agree to do, or announce any intention to do any such transaction described in paragraphs (i), (ii) or (iii) or announce any intention to do so,
in each case, whether any such transaction described in paragraphs (i), (ii) or (iii) above is to be settled by delivery of the Shares or other securities of our Company, in cash or otherwise (whether or not the issue of such Shares or other securities of our Company will be completed within such period of 12 months from the Listing Date).
(b) during the six months commencing on the date falling 12 months after the Listing Date, he will not, and will procure that the relevant registered holder(s), any nominee or trustee holding on trust for him and the companies controlled by him will not, enter into any of the transactions specified in paragraph (a)(i), (ii) or (iii) above or offer to or agree to or announce any intention to effect any such transaction if such transaction, offer, agreement or announcement will create a disorderly or false market in the Shares or any other equity securities of ours.
根据香港包销协议，刘博士（"承诺担保股东"）已向我们及独家保荐人、保荐人-整体协调人、整体协调人、香港包销商及资本市场中介机构各方承诺，在未经独家保荐人及保荐人-整体协调人（代表其本身及代表香港包销商）事先书面同意，且除非符合上市规则（包括上市规则第10.07(3)条及上市规则第10.07(2)条附注(2)）的情况下：
(i) 出售、要约出售、接受认购、订立合约或同意配发、发行或出售、抵押、押记、质押、抵押、借出、授予或出售任何购买期权、认股权证、合约或权利，授予或购买任何出售期权、认股权证、合约或权利，或以其他方式转让或处置或就其实益拥有的任何股份或本公司其他股权证券或其中任何权益（包括但不限于可转换为或可交换或可行使以换取，或代表获得权利的任何证券，或购买任何股份或本公司其他股权证券（视情况而定）或上述任何证券权益的任何认股权证或其他权利）（"锁定证券"）设立产权负担，或同意转让或处置或设立产权负担，无论直接或间接、有条件或无条件，或将本公司任何股份或其他证券存入存管机构以发行存托凭证；或
根据香港包销协议，刘博士（"承诺担保股东"）已向我们及独家保荐人、保荐人-整体协调人、整体协调人、香港包销商及资本市场中介机构各方承诺，在未经独家保荐人及保荐人-整体协调人（代表其本身及代表香港包销商）事先书面同意，且除非符合上市规则（包括上市规则第10.07(3)条及上市规则第10.07(2)条附注(2)）的情况下：
(i) 出售、要约出售、接受认购、订立合约或同意配发、发行或出售、抵押、押记、质押、抵押借款、借出、授予或出售任何购买期权、认股权证、合约或权利，授予或购买任何出售期权、认股权证、合约或权利，或以其他方式转让或处置或就彼实益拥有的任何股份或本公司其他股权证券或其中任何权益（包括但不限于可转换为或可交换或可行使以换取，或代表获得权利的任何证券，或购买任何股份或本公司其他股权证券（视情况而定）或上述任何前述证券权益的任何认股权证或其他权利）（"锁定证券"）设立产权负担，或同意转让或处置或设立产权负担，无论直接或间接、有条件或无条件，或将本公司任何股份或其他证券存入存管机构以发行存托凭证；或
在各种情况下，无论上文第(a)(i)、(a)(ii)或(a)(iii)段所述的任何交易是否通过交付本公司任何股份或其他股权证券、以现金或其他方式结算，以及无论任何股份或其他证券的结算或交付是否将在上市日期起12个月内完成。在上市日期后12个月期限届满之前，如其进行第(a)(i)、(a)(ii)或(a)(iii)段所述的任何交易，或要约或同意或订约或公开宣布有意进行任何该等交易，其将采取一切合理措施，确保该等处置不会在本公司证券中造成混乱或虚假市场；及
(ii) 如其收到任何质权人或押记人就任何已质押或押记的本公司股份或其他证券将予处置的口头或书面迹象，须立即通知本公司及保荐人-整体协调人有关该等迹象。
在符合上市规则适用规定的前提下，上述限制不妨碍保证股东（i）购买本公司额外股份或其他证券，并按照上市规则处置该等额外股份或证券，但任何该等购买或处置不得违反上文第(a)段所述对保证股东的锁定安排或本公司遵守最低公众持股量规定；及（ii）将其实益拥有的本公司股份或其他证券或其中任何权益用作向认可机构（定义见香港法律第155章《银行业条例》）或其他金融机构的真实商业贷款提供担保（包括押记或质押），但须按上文第(b)段所载规定提供通知。
本公司已向独家保荐人、保荐人-整体协调人、整体协调人、香港包销商及资本市场中介机构承诺，在收到保证股东的书面资料后，将尽快在上市规则、证券及期货条例及╱或任何其他适用法律规定的情况下，通知联交所及╱或其他相关机构，并以公告形式就该等资料作出公开披露。
除香港包销协议项下的责任外，截至最后实际可行日期，香港包销商于本公司并无任何股份权益，亦无任何认购或提名人士认购本公司证券的权利或期权（不论是否具有法律强制执行效力）。
全球发售完成后，香港包销商及其联属公司可能因履行其在香港包销协议项下的各自责任而持有若干比例的H股。
就国际发售而言，本公司预期将于2026年1月6日前后与国际包销商（及其他方）订立国际包销协议。根据国际包销协议并在超额配股权的规限下，国际包销商须在该协议所载若干条件的规限下，各自（而非共同）同意认购或促使认购根据国际发售初步发售的国际发售股份。预期国际包销协议可在与香港包销协议相似的理由下终止。潜在投资者须注意，如国际包销协议未能订立或遭终止，全球发售将不会进行。详情请参阅「全球发售架构——国际发售」。
在香港公开发售截止申请日后30天内，根据超额配股权，本公司可能须发行最多合共5,612,900股H股（占全球发售初步可供认购的发售股份数目的不超过15%），以发售价发行，用于弥补国际发售中的超额分配（如有）。详情请参阅"全球发售结构——超额配股权"。
包销商及资本市场中介机构将收取相当于所有发售股份（包括行使超额配股权后须发行的任何发售股份）合计发售价2%的包销佣金（"固定费用"），并从中支付任何分包销佣金及其他费用。
本公司可自行全权酌情决定向一名或多名包销商或资本市场中介机构支付相当于所有发售股份（包括行使超额配股权后须发行的任何发售股份）合计发售价最多1%的激励费（"酌情费用"）。因此，应付予所有包销商的固定费用与酌情费用之比率为67:33。
就任何未获认购而重新分配至国际发售的香港发售股份，包销佣金将不会支付予香港包销商，而是按适用于国际发售的费率支付予相关国际包销商。
本公司就全球发售应付予包销商的包销佣金总额（假设酌情激励费全额支付且超额配股权获全额行使）约为港币1.50亿元。
包销佣金及激励费总额，连同联交所上市费、会计及财务汇报局交易征费、证监会交易征费及联交所交易费、法律及其他专业费用、印刷及与全球发售相关的所有其他开支，估计约为港币1.944亿元（假设酌情激励费全额支付且超额配股权获全额行使），将由本公司支付。
本公司须向独家保荐人支付500,000美元的保荐人费用。
本公司及保证股东各成员均已同意就香港包销商可能蒙受或招致的若干损失向其作出弥偿，包括因履行香港包销协议项下义务而产生的损失，以及因本公司及保证股东任何成员违反香港包销协议而产生的损失。
香港公开发售及国际发售的包销商（统称"联合团成员"）及其联属公司可各自单独从事以下所述的各种活动，该等活动并非包销或稳定价格过程的组成部分。
联合团成员及其联属公司为多元化金融机构，在世界各地与多个国家建立有业务关系。该等实体为其自身账户及他人账户从事广泛的商业及投资银行、贷款融资、经纪、基金管理、交易、对冲、投资及其他活动。在其各项日常业务活动中，联合团成员
and their respective affiliates may purchase, sell or hold a broad array of investments and actively trade securities, derivatives, loans, commodities, currencies, credit default swaps and other financial instruments for their own account and for the accounts of their customers. Such investment and trading activities may involve or relate to assets, securities, co-investments and/or instruments of or with our Company and/or persons and entities with relationships with our Company and may also include swaps and other financial instruments entered into for hedging purposes in connection with our Company's loans and other debt.
In relation to the H Shares, the activities of the Syndicate Members and their affiliates could include acting as agent for buyers and sellers of the H Shares, entering into transactions with those buyers and sellers in a principal capacity, including as a lender to initial purchasers of the H Shares (which financing may be secured by the H Shares) in the Global Offering, proprietary trading in the H Shares, and entering into over the counter or listed derivative transactions or listed or unlisted securities transactions (including issuing securities such as derivative warrants listed on a stock exchange) which have as their underlying assets, assets including the H Shares. Such transactions may be carried out as bilateral agreements or trades with selected counterparties. Those activities may require hedging activity by those entities involving, directly or indirectly, the buying and selling of the H Shares, which may have a negative impact on the trading price of the H Shares. All such activities could occur in Hong Kong and elsewhere in the world and may result in the Syndicate Members and their affiliates holding long and/or short positions in the H Shares, in baskets of securities or indices including the H Shares, in units of funds that may purchase the H Shares, or in derivatives related to any of the foregoing.
In relation to issues by Syndicate Members or their affiliates of any listed securities having the H Shares as their underlying securities, whether on the Stock Exchange or on any other stock exchange, the rules of the stock exchange may require the issuer of those securities (or one of its affiliates or agents) to act as a market maker or liquidity provider in the security, and this will also result in hedging activity in the H Shares in most cases.
All such activities may occur both during and after the end of the stabilizing period described in the section headed "Structure of the Global Offering". Such activities may affect the market price or value of the H Shares, the liquidity or trading volume in the H Shares and the volatility of the price of the H Shares, and the extent to which this occurs from day to day cannot be estimated.
It should be noted that when engaging in any of these activities, the Syndicate Members will be subject to certain restrictions, including the following:
(a) the Syndicate Members (other than the Stabilizing Manager or its affiliates or any person acting for it) must not, in connection with the distribution of the Offer Shares, effect any transactions (including issuing or entering into any option or other derivative transactions relating to the Offer Shares), whether in the open market or otherwise, with a view to stabilizing or maintaining the market price of any of the Offer Shares at levels other than those which might otherwise prevail in the open market; and
(b) the Syndicate Members must comply with all applicable laws and regulations, including the market misconduct provisions of the SFO, including the provisions prohibiting insider dealing, false trading, price rigging and stock market manipulation.
Certain of the Syndicate Members or their respective affiliates have provided from time to time, and expect to provide in the future, investment banking, loan financing and other services to our Company and each of its affiliates for which such Syndicate Members or their respective affiliates have received or will receive customary fees and commissions.
In addition, the Syndicate Members or their respective affiliates may provide financing to investors to finance their subscriptions of Offer Shares in the Global Offering.
The Sole Sponsor satisfy the independence criteria set out in Rule 3A.07 of the Listing Rules. See "Appendix VI—Statutory and General Information—E. Other Information—3. Sole Sponsor" for further details.
This prospectus is published in connection with the Hong Kong Public Offering as part of the Global Offering. China International Capital Corporation Hong Kong Securities Limited is the Sole Sponsor and Sponsor-Overall Coordinator; and China International Capital Corporation Hong Kong Securities Limited, Huatai Financial Holdings (Hong Kong) Limited, BOCOM International Securities Limited, Guotai Junan Securities (Hong Kong) Limited and China Merchants Securities (HK) Co., Limited are the Overall Coordinators of the Global Offering.
The Listing of the H Shares on the Stock Exchange is sponsored by the Sole Sponsor. The Sole Sponsor have made an application on behalf of our Company to the Stock Exchange for the listing of, and permission to deal in, the H Shares to be issued as mentioned in this prospectus.
37,419,500 Offer Shares (subject to reallocation and the Over-allotment Option) will initially be made available under the Global Offering comprising:
(a) the Hong Kong Public Offering of initially 1,871,000 H Shares (subject to reallocation) in Hong Kong as described in "—The Hong Kong Public Offering" in this section below; and
(b) the International Offering of initially 35,548,500 H Shares (subject to reallocation and the Over-allotment Option) outside the United States (including to professional and institutional investors within Hong Kong) in offshore transactions in reliance on Regulation S to investors that are not, and not for the account or benefit of, U.S. Investors, as described in "—The International Offering" in this section below.
(ii) apply for or indicate an interest for International Offer Shares under the International Offering,
but may not do both.
The Offer Shares will represent approximately 8.50% of the enlarged issued share capital of our Company immediately following the completion of the Global Offering, assuming the Over-allotment Option is not exercised. If the Over-allotment Option is exercised in full, the Offer Shares will represent approximately 9.65% of the enlarged issued share capital of our Company immediately following the completion of the Global Offering and the issue of Offer Shares pursuant to the Over-Allotment Option.
References in this prospectus to applications, application monies or the procedure for applications relate solely to the Hong Kong Public Offering.
Subject to reallocation as mentioned below, our Company is initially offering 1,871,000 H Shares (subject to reallocation) for subscription by the public in Hong Kong at the Offer Price, representing approximately 5% of the total number of Offer Shares initially available under the Global Offering. Subject to reallocation as mentioned below, the number of Offer Shares initially offered under the Hong Kong Public Offering, subject to any reallocation of Offer Shares between the International Offering and the Hong Kong Public Offering, will represent approximately 0.43% of the enlarged issued share capital of our Company immediately following the completion of the Global Offering (assuming the Over-allotment Option is not exercised).
The Hong Kong Public Offering is open to members of the public in Hong Kong as well as to institutional and professional investors. Professional investors generally include brokers, dealers, companies (including fund managers) whose ordinary business involves dealing in shares and other securities and corporate entities that regularly invest in shares and other securities.
Completion of the Hong Kong Public Offering is subject to the conditions set out in "—Conditions of the Global Offering" in this section below.
Allocation of Offer Shares to investors under the Hong Kong Public Offering will be based solely on the level of valid applications received under the Hong Kong Public Offering. The basis of allocation may vary, depending on the number of Hong Kong Offer Shares validly applied for by applicants. Such allocation could, where appropriate, consist of balloting, which could mean that some applicants may receive a higher allocation than others who have applied for the same number of Hong Kong Offer Shares, and those applicants who are not successful in the ballot may not receive any Hong Kong Offer Shares. For allocation purposes only, the total number of Hong Kong Offer Shares available under the Hong Kong Public Offering (after taking into account any reallocation referred to below) will be divided equally (to the nearest board lot) into two pools: pool A and pool B (with any odd lot being allocated to pool A).
Pool A: The Hong Kong Offer Shares in pool A will be allocated on an equitable basis to valid applicants who have applied for Hong Kong Offer Shares with an aggregate subscription price of HK$5 million or less (excluding the brokerage, the AFRC transaction levy, the SFC transaction levy and the Stock Exchange trading fee payable).
Pool B: The Hong Kong Offer Shares in pool B will be allocated on an equitable basis to valid applicants who have applied for Hong Kong Offer Shares with an aggregate subscription price of more than HK$5 million and up to the total value in pool B (excluding the brokerage, the AFRC transaction levy, the SFC transaction levy and the Stock Exchange trading fee payable).
For the purpose of the immediately preceding paragraph only, the "price" for the Hong Kong Offer Shares means the price payable on application. See the subsection headed "—Pricing—Price Payable on Application".
Applicants should be aware that applications in pool A and applications in pool B may receive different allocation ratios. If any Hong Kong Offer Shares in one (but not both) of the pools are unsubscribed, such unsubscribed Hong Kong Offer Shares will be transferred to the other pool to satisfy demand in that other pool and be allocated accordingly.
Applicants can only receive an allocation of Hong Kong Offer Shares from either pool A or pool B and not from both pools. Multiple or suspected multiple applications under the Hong Kong Public Offering and any application for more than 935,500 Hong Kong Offer Shares (being 50% of the total number of Offer Shares initially available under the Hong Kong Public Offering) will be rejected.
The allocation of the Offer Shares between the Hong Kong Public Offering and the International Offering is subject to reallocation under the Listing Rules. Paragraph 4.2 of Practice Note 18 of the Listing Rules (as modified by Rule 18C.09 of the Listing Rules) requires a clawback mechanism to be put in place which would have the effect of increasing the number of Hong Kong Offer Shares to a certain percentage of the total number of Offer Shares offered under the Global Offering when certain prescribed total demand levels are reached under the Hong Kong Public Offering.
If the International Offering is fully subscribed or oversubscribed and the number of Offer Shares validly applied for under the Hong Kong Public Offering represents (a) 10 times or more but less than 50 times, (b) 50 times or more of the total number of Offer Shares initially available under the Hong Kong
公开发售，届时将把发售股份由国际发售重新分配至香港公开发售。由于上述重新分配，香港公开发售项下可供认购的发售股份总数将增至3,742,000股发售股份（属情况(a)）或7,483,900股发售股份（属情况(b)），分别约占全球发售（于行使超额配股权前）项下初步可供认购发售股份总数的10%及20%。在每种情况下，重新分配至香港公开发售的额外发售股份将在甲池及乙池之间分配，而国际发售项下所分配的发售股份数目将相应减少，具体方式由保荐人-整体协调人酌情决定。
除上文所述须强制重新分配外，在若干情况下，香港公开发售项下拟发售的发售股份及国际发售项下拟发售的发售股份可由保荐人-整体协调人酌情在上述两项发售之间重新分配。保荐人-整体协调人可酌情将初步分配予国际发售的发售股份重新分配至香港公开发售，以满足甲池及乙池的有效申请，具体依据《新上市申请人指引》第4.14章的指引如下：若(i)国际发售股份未获足额认购而香港发售股份已获足额认购或获超额认购；或(ii)国际发售股份已获足额认购或获超额认购，而香港发售股份获超额认购的倍数少于香港公开发售项下初步可供认购香港发售股份数目的10倍，则在发售价定为每股发售股份116.20港元的前提下，最多1,871,000股发售股份可由国际发售重新分配至香港公开发售，使香港公开发售项下于重新分配后可供认购的发售股份总数增至3,742,000股，即香港公开发售项下初步可供认购发售股份数目的两倍。
在若干情况下，香港公开发售项下拟发售的发售股份及国际发售项下拟发售的发售股份可由保荐人-整体协调人酌情在上述两项发售之间重新分配。
若香港公开发售未获足额认购，保荐人-整体协调人有权将全部或任何未获认购的香港发售股份重新分配至国际发售，具体比例由保荐人-整体协调人酌情决定。
若国际发售股份未获足额认购，且香港发售股份亦未获足额认购，则除非包销商按本招股说明书及包销协议的条款及条件认购或促使认购其各自就全球发售项下未获认购的发售股份所须承担的相应比例，否则全球发售将不会进行。
香港公开发售与国际发售之间任何发售股份重新分配的详情将于预期于2026年1月7日（星期三）晚上11时正或之前刊发的发售结果公告中披露。
香港公开发售项下的每位申请人须在其提交的申请中作出承诺及确认，说明其本人及其申请所为受益人并无申请或认购，或表示有兴趣认购，亦不会申请或认购，或表示有兴趣认购国际发售项下的任何国际发售股份。若上述承诺及╱或确认遭违反及╱或失实（视情况而定），或若该申请人已获或将获配售或分配国际发售项下的国际发售股份，则该申请人于国际发售项下的申请可能遭拒绝。
香港公开发售项下的申请人可能须于申请时（视申请渠道而定）按每股发售股份HK$116.20的发售价缴付款项，另加须就每股发售股份缴付的经纪佣金、会计及财务汇报局交易征费、证监会交易征费及联交所交易费，合计每手100股股份共须缴付HK$11,737.19。
国际发售将包括由本公司提呈发售的初步35,548,500股H股，占全球发售项下初步可供认购发售股份总数约95%（视重新分配及超额配股权而定）。国际发售项下初步提呈发售的发售股份数目（视国际发售与香港公开发售之间的任何发售股份重新分配而定），将占全球发售完成后随即已发行股份总数约8.07%（假设超额配股权未获行使）。
国际发售将包括向香港及美国以外其他司法权区的机构投资者及专业投资者以及其他预期对该等发售股份有大量需求的投资者选择性推介H股，依赖Regulation S向并非美国投资者且并非为美国投资者账户或利益而行事的投资者进行。专业投资者一般包括经纪、交易商、日常业务涉及买卖股份及其他证券的公司（包括基金经理）以及定期投资于股份及其他证券的企业实体。根据国际发售分配发售股份将按照本节下文"——定价及分配"段所述的"累计投标"程序进行，并基于若干因素，包括需求水平及时间、相关投资者在相关行业的已投资资产或股权资产总值，以及相关投资者在上市后是否预期可能进一步购买H股及╱或持有或出售其H股。该等分配旨在以有利于本公司及全体股东的方式分配H股，从而建立稳固的专业及机构股东基础。此外，根据《上市规则》第18C.08条，全球发售项下提呈发售股份总数（不包括任何因行使超额配股权而发行的股份）至少50%将由《上市规则》所界定的独立定价投资者于国际发售中认购。
保荐人-整体协调人（代表包销商）可要求任何已获配发国际发售项下发售股份并已于香港公开发售项下提出申请的投资者向保荐人-整体协调人提供充足资料，以便其识别国际发售项下的相关申请，并确保该等申请被排除于国际发售项下的任何发售股份分配之外。
根据国际发售发行的发售股份总数可能因本节上文"——香港公开发售——重新分配"段所述的回拨安排、全部或部分行使超额配股权及╱或重新分配香港公开发售项下原本所含的未获认购发售股份而有所变动。
就全球发售而言，本公司预期将向国际包销商授出超额配股权，可由保荐人-整体协调人（代表国际包销商）行使。
根据超额配售权，国际承销商将有权，由保荐人-整体协调人（代表国际承销商）行使，在国际承销协议签订之日起至香港公开发售截止申请日后30天内的任何时间，要求本公司发行最多合共5,612,900股H股，占全球发售初始可供发售股份总数不超过15%，按国际发售的发售价发行，以弥补国际发售中的超额分配（如有）。
如超额配售权获全额行使，据此发行的额外发售股份将占全球发售完成及根据超额配售权发行发售股份后本公司已发行股份总数的约1.26%。如超额配售权获行使，将作出公告。
稳定价格是承销商在某些市场中用于促进证券分销的一种惯例。为实施稳定价格，承销商可在特定期间内在二级市场上竞投或购买证券，以阻止并尽可能防止证券的初始公开市场价格跌破发售价。此类交易可在所有允许进行的司法管辖区执行，并须在每一司法管辖区遵守所有适用法律及监管规定，包括香港的相关规定。在香港，进行稳定价格行动的价格不得超过发售价。
就全球发售而言，稳定价格行动人（或其联属人或任何代其行事的人）代表承销商，可进行超额分配或实施交易，以期在上市日期后的有限期间内，将H股市价稳定或维持在其他情况下可能出现的水平之上。然而，稳定价格行动人（或其联属人或任何代其行事的人）并无责任采取任何稳定价格行动。如采取该等稳定价格行动：(a) 将完全由稳定价格行动人（或其联属人或任何代其行事的人）酌情决定，并以稳定价格行动人合理认为符合本公司最佳利益的方式进行；(b) 可随时终止；及 (c) 须于香港公开发售截止申请日后30天内终止。
根据《证券及期货条例》的《证券及期货（价格稳定）规则》，香港允许的稳定价格行动包括：(a) 为防止或尽量减少H股市价下跌而进行超额分配；(b) 出售或同意出售H股，从而建立H股的淡仓，以防止或尽量减少H股市价下跌；(c) 根据超额配售权购买或同意购买H股，以平仓上述第(a)或(b)段所建立的任何仓位；(d) 纯粹为防止或尽量减少H股市价下跌而购买或同意购买H股；(e) 出售或同意出售H股，以平仓因上述购买而建立的任何仓位；及 (f) 提出或尝试进行上述第(b)、(c)、(d)或(e)段所述的任何行动。
(d) 任何稳定价格行动对H股价格的支持期限不得超过稳定价格期，该稳定价格期将于上市日期开始，预期于香港公开发售截止申请日后第30天届满。此日期之后，将不得再采取任何稳定价格行动，H股的需求及其价格可能因此下跌；
(f) 于稳定价格行动过程中所提出的稳定价格竞投或所进行的交易，可以在等于或低于发售价的任何价格进行，因此可能以低于发售股份申请人或投资者所付价格的价格进行。
为实施稳定价格行动，稳定价格代理人将通过与已获配售国际发售股份的投资者达成延迟交付安排，安排最多合共5,612,900股H股的备兑来源，占初始发售股份的最多15%。延迟交付安排（如经某投资者明确同意）仅涉及向该投资者交付发售股份的延迟，而分配予该投资者的发售股份之发售价将于上市日期前全额缴付。
备兑来源的规模及超额配股权可行使的范围，均取决于能否与投资者达成安排，使足够数量的H股可延迟交付。若国际发售中没有任何投资者同意延迟交付安排，稳定价格代理人将不会采取任何稳定价格行动，超额配股权亦将不予行使。
本公司将确保或促使于稳定价格期届满后七日内，按照《证券及期货条例》的《证券及期货（稳定价格）规则》的规定发出公告。
就全球发售超额配售H股后，稳定价格代理人（或其联属公司或任何代其行事的人士）可透过全部或部分行使超额配股权、以不超过发售价的价格于二级市场购入H股，或综合运用上述方法，以弥补该等超额配售。
除另行公告外，发售价将为每股发售股份116.20港元，详情如下文进一步说明。
香港公开发售的申请人于申请时（视乎申请渠道）或须缴付每股发售股份116.20港元的发售价，另加1.0%经纪佣金、0.00015%会计及财务汇报局交易征费、0.0027%证监会交易征费及0.00565%联交所交易费，每手100股H股合共须缴付11,737.19港元。
of the International Offering, and with our consent, reduce the number of Offer Shares offered and/or the Offer Price below that stated in this prospectus at any time on or prior to the morning of the last day for lodging applications under the Hong Kong Public Offering. In such a case, our Company will, as soon as practicable following the decision to make such reduction, and in any event not later than the morning of the last day for lodging applications under the Hong Kong Public Offering, cause to be published on the websites of our Company and the Hong Kong Stock Exchange at www.zhipuai.cn and www.hkexnews.hk, respectively, notice of the reduction, the cancelation of the Global Offering and the relaunch of the Global Offering at the revised number of the Offer Shares and/or the revised Offer Price. This notice will also include confirmation or revision, as appropriate, of the working capital statement and the Global Offering statistics as set out in this prospectus, as well as any other financial information which may change as a result of the reduction.
We will, as soon as practicable following the decision to make the reduction, in addition to publishing the notice, issue a supplemental prospectus containing details in relation to the change in the number of Offer Shares being offered and/or the revised Offer Price. The Global Offering will be canceled and subsequently relaunched on FINI pursuant to the supplemental prospectus.
Before submitting applications for the Hong Kong Offer Shares, applicants should have regard to the possibility that any announcement of a reduction in the number of Offer Shares and/or the Offer Price may not be made until the last day for lodging applications under the Hong Kong Public Offering. In the absence of any such notice so published, the number of Offer Shares will not be reduced.
In the event of a reduction in the number of Offer Shares, the Sponsor-Overall Coordinator (for and on behalf of the Underwriters) may, at its discretion, reallocate the number of Offer Shares to be offered in the Hong Kong Public Offering and the International Offering. The Offer Shares to be offered in the Hong Kong Public Offering and the Offer Shares to be offered in the International Offering may, in certain circumstances, be reallocated between these offerings at the discretion of the Sponsor-Overall Coordinator (for and on behalf of the Underwriters).
The level of indications of interest in the International Offering, the level of applications in the Hong Kong Public Offering, the basis of allocations of the Hong Kong Offer Shares and the results of allocations in the Hong Kong Public Offering are expected to be made available through a variety of channels in the manner described in "How to Apply for Hong Kong Offer Shares – B. Publication of Results".
The Hong Kong Public Offering is fully underwritten by the Hong Kong Underwriters under the terms and conditions of the Hong Kong Underwriting Agreement and is subject to, among other things, the conditions set out in the subsection headed "—Conditions of the Global Offering".
We expect to enter into the International Underwriting Agreement relating to the International Offering on or around Tuesday, January 6, 2026.
These underwriting arrangements, including the Underwriting Agreements, are summarized in the section headed "Underwriting".
(a) the Stock Exchange granting approval for the listing of, and permission to deal in, the H Shares to be issued pursuant to the Global Offering (including the H Shares which may be issued pursuant to the
exercise of the Over-allotment Option and the H Shares to be converted from Unlisted Shares) as mentioned herein on the Main Board of the Stock Exchange and such approval not subsequently having been withdrawn, canceled or revoked;
(b) the execution and delivery of the International Underwriting Agreement on or about Tuesday, January 6, 2026; and
(c) the obligations of the Hong Kong Underwriters under the Hong Kong Underwriting Agreement and the obligations of the International Underwriters under the International Underwriting Agreement becoming and remaining unconditional and not having been terminated in accordance with the terms of the respective agreements,
in each case on or before the dates and times specified in the respective Underwriting Agreements (unless and to the extent such conditions are validly waived on or before such dates and times).
From 9:00 a.m. on Tuesday, December 30, 2025 to 11:30 a.m. on Monday, January 5, 2026. The latest time for completing full payment of application monies will be 12:00 noon on Monday, January 5, 2026.
| 申请渠道 | 平台 | 目标投资者 | 申请时间 | |---|---|---|---| | HK eIPO白表服务 | www.hkeipo.hk | 希望收到实体H股股票证书的申请人。成功申请的香港公开发售股份将以申请人本人名义配发及发行。 | 2025年12月30日（星期二）上午9时至2026年1月5日（星期一）上午11时30分。完成全额缴付申请款项的最后时间为2026年1月5日（星期一）中午12时正。 | | HKSCC电子IPO渠道 | 申请人的经纪商或托管人（须为香港结算参与者）将根据申请人的指示，通过香港结算的FINI系统代表申请人提交电子申请指示。 | 不希望收到实体H股股票证书的申请人。成功申请的香港公开发售股份将以香港结算代理人名义配发及发行，直接存入中央结算系统，并存入申请人指定的香港结算参与者的股票账户。 | 请联系您的经纪商或托管人了解提交指示的最早及最迟时间，因不同经纪商或托管人的时间可能有所不同。 |
You may only apply through the HKSCC EIPO channel if you are applying through a broker or custodian who is a HKSCC Participant. You may not apply for Hong Kong Offer Shares in any other way.
Your application must be for a minimum of 100 Hong Kong Offer Shares. Each application or electronic application instruction must be in one of the numbers set out in the table in the Application Results announcement.
You may apply for 100 Hong Kong Offer Shares or a higher number of Hong Kong Offer Shares in multiples of 100 (i.e., 100, 200, 300, 400, 500 and so on).
Your application monies (including brokerage, SFC transaction levy and Stock Exchange trading fee) should be paid in Hong Kong dollars. The maximum total application monies (including brokerage, SFC transaction levy and Stock Exchange trading fee) are HK$[●] per board lot of 100 Shares, and the amount payable per application for the number of Shares applied for will be set out in the Application Results announcement.
(ii) pay the full application monies, brokerage, SFC transaction levy and Stock Exchange trading fee.
(i) instruct your broker or custodian to submit electronic application instructions through HKSCC's FINI system on your behalf; and
(ii) ensure that you have made sufficient funds available in your account with your broker or custodian to pay the full amount of application monies, brokerage, SFC transaction levy and Stock Exchange trading fee.
By instructing your broker or custodian to apply for the Hong Kong Offer Shares on your behalf through the HKSCC EIPO channel, you (and, if you are joint applicants, each of you jointly and severally) are deemed to have instructed and authorized HKSCC to cause HKSCC Nominees (acting as nominee for the relevant HKSCC Participants) to apply for Hong Kong Offer Shares on your behalf and to do on your behalf all the things stated in this prospectus and any supplement to it.
For those applying through HKSCC EIPO channel, an actual application will be deemed to have been made for any application instructions given by you or for your benefit to HKSCC (in which case an application will be made by HKSCC Nominees on your behalf) provided such application instruction has not been withdrawn or otherwise invalidated before the closing time of the Hong Kong Public Offering.
HKSCC Nominees will only be acting as a nominee for you and neither HKSCC nor HKSCC Nominees shall be liable to you or any other person in respect of any actions taken by HKSCC or HKSCC Nominees on your behalf to apply for Hong Kong Offer Shares or for any breach of the terms and conditions of this prospectus.
Only one application may be made for the benefit of any person. If you are suspected of making more than one application through the HK eIPO White Form service or any other channel, all of your applications are liable to be rejected.
| For Individual / Joint Applicants | For Corporate Applicants | |---|---| | • Full name(s)² as shown on your identity document | • Full name(s)² as shown on your identity document | | • Identity document's issuing country or jurisdiction | • Identity document's issuing country or jurisdiction | | • Identity document type, with order of priority: | • Identity document type, with order of priority: | | i. HKID card; or | i. LEI registration document; or | | ii. National identification document; or | ii. Certificate of incorporation; or | | iii. Passport; and | iii. Business registration certificate; or | | • Identity document number | iv. Other equivalent document; and | | | • Identity document number |
Notes: 1. If you are applying through the HK eIPO White Form service, you are required to provide a valid e-mail address, a contact telephone number and a Hong Kong Address. You are also required to declare that the identity information provided by you follows the requirements as described in Note 2 below. In particular, where you cannot provide a HKID number, you must confirm that you do not hold a HKID card. The number of joint applicants may not exceed four. If you are a firm, the applicant must be in the individual members' names.
2. The applicant's full name as shown on their identity document must be used and the surname, given name, middle and other names (if any) must be input in the same order as shown on the identity document. If an applicant's identity document contains
both an English and Chinese name, both English and Chinese names must be used. Otherwise, either English or Chinese names will be accepted. The order of priority of the applicant's identity document type must be strictly followed and where an individual applicant has a valid HKID card (including both Hong Kong Residents and Hong Kong Permanent Residents), the HKID number must be used when making an application to subscribe for Hong Kong Offer Shares. Similarly for corporate applicants, a LEI number must be used if an entity has a LEI certificate.
3. If the applicant is a trustee, the client identification data ("CID") of the trustee, as set out above, will be required. If the applicant is an investment fund (i.e. a collective investment scheme, or CIS), the CID of the asset management company or the individual fund, as appropriate, which has opened a trading account with the broker will be required, as above.
4. The maximum number of joint account holders on FINI is capped at 4 in accordance with market practice.
5.
You must not make more than one application, whether individually or jointly. If you apply individually, you must not also apply jointly. The following individuals/entities are considered to be the same applicant and jointly liable for making multiple applications:
If you are applying through the HK eIPO White Form service you may refer to the table below for the amount payable for the number of Shares you have selected. You must pay the respective amount payable on application in full upon application for Hong Kong Offer Shares.
| No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | No. of Hong Kong Offer Shares applied for | Amount payable(2) on application/ successful allotment HK$ | |---|---|---|---|---|---|---|---| | 100 | 11,737.19 | 2,500 | 293,429.69 | 30,000 | 3,521,156.31 | 600,000 | 70,423,126.20 | | 200 | 23,474.37 | 3,000 | 352,115.63 | 40,000 | 4,694,875.08 | 700,000 | 82,160,313.90 | | 300 | 35,211.56 | 3,500 | 410,801.57 | 50,000 | 5,868,593.86 | 800,000 | 93,897,501.60 | | 400 | 46,948.75 | 4,000 | 469,487.51 | 60,000 | 7,042,312.62 | 935,500(1) | 109,801,390.94 | | 500 | 58,685.94 | 4,500 | 528,173.44 | 70,000 | 8,216,031.39 | | | | 600 | 70,423.12 | 5,000 | 586,859.39 | 80,000 | 9,389,750.15 | | | | 700 | 82,160.32 | 6,000 | 704,231.26 | 90,000 | 10,563,468.94 | | | | 800 | 93,897.50 | 7,000 | 821,603.14 | 100,000 | 11,737,187.70 | | | | 900 | 105,634.69 | 8,000 | 938,975.01 | 200,000 | 23,474,375.40 | | | | 1,000 | 117,371.88 | 9,000 | 1,056,346.90 | 300,000 | 35,211,563.10 | | | | 1,500 | 176,057.82 | 10,000 | 1,173,718.76 | 400,000 | 46,948,750.80 | | | | 2,000 | 234,743.75 | 20,000 | 2,347,437.55 | 500,000 | 58,685,938.50 | | |
(1) Maximum number of Hong Kong Offer Shares you may apply for and this is 50% of the Hong Kong Offer Shares initially offered.
(2) The amount payable is inclusive of brokerage, SFC transaction levy, the Stock Exchange trading fee and AFRC transaction levy. If your application is successful, brokerage will be paid to the Exchange Participants (as defined in the Listing Rules) or to the HK eIPO White Form Service Provider (for applications made through the application channel of the HK eIPO White Form service) while the SFC transaction levy, the Stock Exchange trading fee and the AFRC transaction levy will be paid to the SFC, the Stock Exchange and the AFRC, respectively.
No application for any other number of the Hong Kong Offer Shares will be considered and any such application is liable to be rejected.
agree that all applications, acceptances and contracts will be subject to the laws of Hong Kong, and in case of any conflict, the laws of Hong Kong shall prevail;
(m).
represent, warrant and undertake that: (i) you understand that the H Shares have not been and will not be registered under the U.S. Securities Act and may not be offered or sold in the United States; (ii) you and any person for whose benefit you are making the application are outside the United States; (iii) you and any person for whose benefit you are making the application are not a U.S. Person and will not be offering or selling the H Shares to, or for the account or benefit of, any U.S. Person or any other person located in the United States; and (iv) you and any person for whose benefit you are making the application are not a "distributor" (as defined in Regulation S under the U.S. Securities Act) or an affiliate of any such distributor;
(n).
agree that our Company's decision on the allocation of Hong Kong Offer Shares is final and binding on all applicants;
(o).
undertake and confirm that you or the person(s) for whose benefit you have made the application have not applied for or taken up, or indicated an interest for, and will not apply for or take up, or indicate an interest for, any International Offer Shares;
(p).
agree to accept the Hong Kong Offer Shares applied for, or any lesser number allocated to you under the application;
(q).
authorize our Company to place your name(s) or the name of the HKSCC Nominees, on our Company's H Share register of members as the holder(s) of any Hong Kong Offer Shares allocated to you, and our Company and/or its agents to send any H Share certificate(s) and/or any refund check(s) to you or the first-named applicant for joint application by ordinary post at your own risk to the address stated on the application form, unless you are eligible to collect the H Share certificates in person;
(r).
declare and represent that this is the only application made and the only application intended by you to be made to benefit you or the person for whose benefit the application is made and that you are aware of the fact that if you or the person for whose benefit you have made the application have made any other application for Hong Kong Offer Shares, all such applications will be rejected;
(s).
(if the application is made for your own benefit) warrant that no other application has been or will be made for your benefit on a WHITE or YELLOW Application Form or by giving electronic application instructions to HKSCC or through the HK eIPO White Form service by any one as nominee on your behalf;
(t).
(if you are an agent for another person) warrant that: (i) you have due authority to sign the application form or give electronic application instructions on behalf of that other person; and (ii) that other person has authorized you to apply for the Hong Kong Offer Shares; and (iii) your application is the only application made for the benefit of that person;
(u).
understand that our Company and the Overall Coordinators will rely on your declarations and representations in deciding whether or not to make any allotment of any of the Hong Kong Offer Shares to you and that you may be prosecuted for making a false declaration;
(v).
(if you are applying through the HK eIPO White Form service) authorize the designated HK eIPO White Form service provider as agent and attorney of each applicant, and/or to make payment to our Company by means of the payment method as described in the designated website.
Our Company expects to announce the results of allocations, including the Hong Kong Identity Card numbers, passport numbers or Hong Kong business registration numbers of successful applicants and the number of Hong Kong Offer Shares successfully applied for (where applicable) under the application reference numbers on the following date:
The results will be available through a variety of channels in the following manner, where the announcement on the Stock Exchange's website will be in a searchable format on the date of announcement:
(ii). in the announcement to be posted on our Company's website at [●].
Applicants who have applied through the HK eIPO White Form service may also check the results of their applications through the HK eIPO White Form service designated website at www.hkeipo.hk.
Individual applicants who are unsuccessful in their applications will not receive any notification letter.
You should note that your application is liable to be rejected in any of the circumstances set out below:
(a). you (or as the case may be, your designated HKSCC participant in the case of electronic application instructions given to HKSCC) have not paid the application monies in full;
(b). the application does not contain the applicant's full name, Hong Kong Identity Card number or passport number (or for companies, the Hong Kong business registration number or certificate of incorporation number) or the information provided is incorrect;
(c). you have applied for more than one application or have caused or permitted more than one application to be made for your benefit (and such applications will also be rejected);
(d). our Company or our agents have reason to believe that by accepting your application, we would violate the applicable securities or other laws or regulations of any jurisdiction, country or city outside Hong Kong;
(e). your application has not been duly completed and signed in accordance with the instructions set out in the application form and this prospectus;
(f). our Company or our agents believe that you are not eligible to apply for the Hong Kong Offer Shares;
(h). our Company, the Overall Coordinators or any of their respective agents do not receive your application monies in the designated account;
(i). the Underwriting Agreement does not become unconditional or is terminated in accordance with its terms.
agree to comply with the Companies Ordinance, the Companies (Winding Up and Miscellaneous Provisions) Ordinance, the Articles of Association and laws of any place outside Hong Kong that apply to your application and that neither we nor the Relevant Persons will breach any law inside and/or outside Hong Kong as a result of the acceptance of your offer to purchase, or any action arising from your rights and obligations under the terms and conditions contained in this prospectus;
(m).
confirm that (a) your application or HKSCC Nominees' application on your behalf is not financed directly or indirectly by our Company, any of the directors, chief executives, substantial Shareholder(s) or existing shareholder(s) of our Company or any of their respective close associates; and (b) you are not accustomed or will not be accustomed to taking instructions from our Company, any of the directors, chief executives, substantial shareholder(s) or existing shareholder(s) of our Company or any of its subsidiaries or any of their respective close associates in relation to the acquisition, disposal, voting or other disposition of the Shares registered in your name or otherwise held by you;
(n).
(o).
confirm that you understand that we and the Overall Coordinators will rely on your declarations and representations in deciding whether or not to allocate any Hong Kong Offer Shares to you and that you may be prosecuted for making a false declaration;
(p).
agree to accept Hong Kong Offer Shares applied for or any lesser number allocated to you under the application;
(q).
declare and represent that this is the only application made and the only application intended by you to be made to benefit you or the person for whose benefit you are applying;
(r).
represent, warrant and undertake that (i) you understand that the Hong Kong Offer Shares have not been and will not be registered under the U.S. Securities Act; and (ii) you and any person for whose benefit you are applying for the Hong Kong Offer Shares are outside the United States (as defined in Regulation S) or are a person described in paragraph (h)(3) of Rule 902 of Regulation S and are not a U.S. Investor;
(s).
undertake and confirm that you or the person(s) for whose benefit you have made the application have not applied for or taken up, or indicated an interest for, and will not apply for or take up, or indicate an interest for, any International Offer Shares nor have participated in the International Offering;
(t).
(u).
(if the application is made for your own benefit) warrant that no other application has been or will be made for your benefit by giving electronic application instructions to HKSCC directly or indirectly or through the application channel of the HK eIPO White Form service or by any one as your agent or by any other person; and – 327 –
B.
(v).
(if you are making the application as an agent for the benefit of another person) warrant that (i) no other application has been or will be made by you as agent for or for the benefit of that person or by that person or by any other person as agent for that person by giving application instructions to HKSCC and the HK eIPO White Form Service Provider and (ii) you have due authority to give electronic application instructions on behalf of that other person as its agent; and
(w).
if the laws of any place outside Hong Kong apply to your application, agree and warrant that you have complied with all these laws and none of us nor any Relevant Person will breach any of these laws as a result of the acceptance of your offer to purchase, or any action arising from your rights and obligations under the terms and conditions contained in this prospectus.
Results of Allocation You can check whether you are successfully allocated any Hong Kong Offer Shares through:
| Date/Time | Platform | |---|---| | Applying through the HK eIPO White Form service or HKSCC EIPO channel: | | | 24 hours, from 11:00 p.m. Wednesday, January 7, 2026 to 12:00 midnight on Tuesday, January 13, 2026 (Hong Kong time) | Website: the "Allotment Results" page on the designated results of allocations website at www.tricor.com.hk/ipo/result or www.hkeipo.hk/IPOResult | | | The full list of (i) wholly or partially successful applicants using the HK eIPO White Form service and HKSCC EIPO channel, and (ii) the number of Hong Kong Offer Shares conditionally allotted to them, among other things, will be displayed at www.hkeipo.hk/IPOResult or www.tricor.com.hk/ipo/result | | No later than 11:00 p.m. on Wednesday, January 7, 2026 (Hong Kong time). | The Stock Exchange's website at www.hkexnews.hk and our website at www.zhipuai.cn which will provide links to the above mentioned websites of the H Share Registrar. | | between 9:00 a.m. and 6:00 p.m., from Thursday, January 8, 2026 to Tuesday, January 13, 2026 (Hong Kong time) on a Business Day | Telephone: +852 3691 8488 – the allocation results telephone enquiry line provided by the H Share Registrar |
For those applying through HKSCC EIPO channel, you may also check with your broker or custodian from 6:00 p.m. on Tuesday, January 6, 2026 (Hong Kong time).
HKSCC Participants can log into FINI and review the allotment result from 6:00 p.m. on Tuesday, January 6, 2026 (Hong Kong time) on a 24-hour basis and should report any discrepancies on allotments to HKSCC as soon as practicable.
We expect to announce the results of the level of indications of interest in the Global Offering, the level of applications in the Hong Kong Public Offering and the basis of allocations of Hong Kong Offer Shares on the Stock Exchange's website at www.hkexnews.hk and our website at www.zhipuai.cn by no later than 11:00 p.m. on Wednesday, January 7, 2026 (Hong Kong time).
You should note the following situations in which Hong Kong Offer Shares will not be allocated to you or the person(s) for whose benefit you are applying for:
Your application or the application made by HKSCC Nominees on your behalf may be revoked pursuant to Section 44A(6) of the Companies (Winding Up and Miscellaneous Provisions) Ordinance.
We, the Overall Coordinators, the H Share Registrar and their respective agents and nominees have full discretion to reject or accept any application, or to accept only part of any application, without giving any reasons.
The allocation of Hong Kong Offer Shares will be void if the Stock Exchange does not grant permission to list the H Shares either:
• within a longer period of up to six weeks if the Stock Exchange notifies us of that longer period within three weeks of the closing date of the application lists.
• you make multiple applications or suspected multiple applications. You may refer to the paragraph headed "—A. Applications for Hong Kong Offer Shares—5. Multiple Applications Prohibited" in this section on what constitutes multiple applications;
• we or the Overall Coordinators believe that by accepting your application, it or we would violate applicable securities or other laws, rules or regulations.
Based on the arrangements between HKSCC Participants and HKSCC, HKSCC Participants will be required to hold sufficient application funds on deposit with their designated bank before balloting. After balloting of Hong Kong Offer Shares, the Receiving Bank will collect the portion of these funds required to settle each HKSCC Participant's actual Hong Kong Offer Share allotment from their designated bank.
There is a risk of money settlement failure. In the extreme event of money settlement failure by a HKSCC Participant (or its designated bank), who is acting on your behalf in settling payment for your allotted shares, HKSCC will contact the defaulting HKSCC Participant and its designated bank to determine the cause of failure and request such defaulting HKSCC Participant to rectify or procure to rectify the failure.
However, if it is determined that such settlement obligation cannot be met, the affected Hong Kong Offer Shares will be reallocated to the International Offering. Hong Kong Offer Shares applied for by you through the broker or custodian may be affected to the extent of the settlement failure. In the extreme case, you will not be allocated any Hong Kong Offer Shares due to the money settlement failure by such HKSCC Participant. None of us, the Relevant Persons, the H Share Registrar and HKSCC is or will be liable if Hong Kong Offer Shares are not allocated to you due to the money settlement failure.
You will receive one H Share certificate for all Hong Kong Offer Shares allotted to you under the Hong Kong Public Offering (except pursuant to applications made through the HKSCC EIPO channel where the H Share certificates will be deposited into CCASS as described below).
No temporary document of title will be issued in respect of the Shares. No receipt will be issued for sums paid on application.
H Share certificates will only become valid evidence of title at 8:00 a.m. on Thursday, January 8, 2026 (Hong Kong time), provided that the Global Offering has become unconditional and the right of termination described in the section headed "Underwriting" has not been exercised. Investors who trade H Shares prior to the receipt of H Share certificates or the H Share certificates becoming valid evidence of title do so entirely at their own risk.
The right is reserved to retain any H Share certificate(s) and (if applicable) any surplus application monies pending clearance of application monies.
| | HK eIPO White Form service | HKSCC EIPO channel | |---|---|---| | **Despatch/collection of H Share certificate** | | | | For application of 500,000 Hong Kong Offer Shares or more | Collection in person at Tricor Investor Services Limited, at 17/F, Far East Finance Centre, 16 Harcourt Road, Hong Kong | H Share certificate(s) will be issued in the name of HKSCC Nominees, deposited into CCASS and credited to your designated HKSCC Participant's stock account | | | Time: from 9:00 a.m. to 1:00 p.m. on Thursday, January 8, 2026 (Hong Kong time) | No action by you is required |
如果您是个人申请人，您不得授权任何其他人代您领取。如果您是企业申请人，您的授权代表必须持有贴有贵公司印章的授权书。个人申请人及授权代表在领取时均须向H股过户登记处出示可接受的身份证明文件。
注意：如果您未在上述时间内亲自领取您的H股股票，股票将按普通邮递方式寄往您申请指示中指定的地址，邮递风险由您自行承担。
| | 日期：2026年1月8日（星期四） | 视您与您的经纪商或托管人之间的安排而定 | |---|---|---| | **负责方** | H股过户登记处 | 您的经纪商或托管人 | | **通过单一银行账户支付的申请款项** | 任何退款将以HK eIPO白表电子自动退款付款指示的形式退至银行账户 | 您的经纪商或托管人将按照与您之间的安排，将退款安排退至您指定的银行账户 | | **通过多个银行账户支付的申请款项** | 退款支票将按普通邮递方式寄往您申请指示中指定的地址，邮递风险由您自行承担 | |
除非香港在上市日期前一个营业日发出下列严重恶劣天气警告（定义见下文），导致相关股票无法及时派送至香港中央结算有限公司，本公司将促使H股过户登记处按照双方商定的应急安排，安排递送支持文件及股票。您可参阅本节"—E. 严重恶劣天气安排"。
申请名单将改为在下一个于上午9时至中午12时之间没有严重恶劣天气警告的营业日上午11时45分至中午12时之间开放，及/或于中午12时关闭。
潜在投资者须注意，申请名单开放／关闭日期的延迟可能导致上市日期延迟。如本招股说明书"预期时间表"一节所述日期有任何变更，将于香港联合交易所网站www.hkexnews.hk及本公司网站www.zhipuai.cn刊登公告及公布修订后的时间表。
如果在2026年1月7日（星期三）发出任何严重恶劣天气警告，H股过户登记处将就向中央结算系统存管处服务柜台递送H股股票作出适当安排，以确保股票可于2026年1月8日（星期四）开始买卖。
如果在2026年1月7日（星期三）发出任何严重恶劣天气警告，对于申请少于500,000股香港发售股份的申请，实物H股股票将于任何上述警告解除或取消后（例如2026年1月7日（星期三）下午或2026年1月8日（星期四））邮局重开时以普通邮递方式寄出。
如果在2026年1月8日（星期四）发出任何严重恶劣天气警告，对于申请500,000股或以上香港发售股份的申请，实物H股股票将于任何上述警告解除或取消后（例如2026年1月8日（星期四）下午或2026年1月9日（星期五））在H股过户登记处办公室供领取。
潜在投资者须注意，如选择以本身名义收取实物H股股票，收取H股股票或有延迟。
如联交所批准H股在联交所上市及买卖，且本公司符合香港中央结算有限公司的股份接纳规定，H股将于H股开始买卖之日或香港中央结算有限公司选择的其他日期起，获香港中央结算有限公司接纳为合资格证券，可在中央结算系统进行存管、结算及交收。交易所参与者之间的交易须于任何交易日后第二个交收日在中央结算系统完成交收。
中央结算系统的所有活动均受香港中央结算有限公司不时生效的《香港中央结算有限公司一般规则》及《香港中央结算有限公司运作程序》所规限。
已作出一切必要安排，使股份可获纳入中央结算系统。
您应向您的经纪商或其他专业顾问寻求有关交收安排的详情，因为该等安排可能影响您的权利及权益。
以下个人资料收集声明适用于本公司、H股过户登记处、收款银行及相关人士就您所收集及持有的任何个人资料，其适用方式与适用于香港中央结算（代理人）有限公司以外的申请人个人资料的方式相同。该等个人资料可能包括客户识别码及您的身份识别信息。通过向香港中央结算有限公司发出申请指示，您确认您已阅读、理解并同意以下个人资料收集声明的所有条款。
This Personal Information Collection Statement informs the applicant for, and holder of, Hong Kong Offer Shares, of the policies and practices of our Company and the H Share Registrar in relation to personal data and the Personal Data (Privacy) Ordinance (Chapter 486 of the Laws of Hong Kong).
It is necessary for applicants and registered holders of Hong Kong Offer Shares to ensure that personal data supplied to our Company or its agents and the H Share Registrar is accurate and up-to-date when applying for Hong Kong Offer Shares or transferring Hong Kong Offer Shares into or out of their names or in procuring the services of the H Share Registrar.
Failure to supply the requested data or supplying inaccurate data may result in your application for Hong Kong Offer Shares being rejected, or in the delay or the inability of our Company or the H Share Registrar to effect transfers or otherwise render their services. It may also prevent or delay registration or transfers of Hong Kong Offer Shares which you have successfully applied for and/or the despatch of Share certificate(s) to which you are entitled.
It is important that applicants for and holders of Hong Kong Offer Shares inform our Company and the H Share Registrar immediately of any inaccuracies in the personal data supplied.
Your personal data may be used, held, processed, and/or stored (by whatever means) for the following purposes:
• processing your application and refund check and HK eIPO White Form e-Auto Refund payment instruction(s), where applicable, verification of compliance with the terms and application procedures set out in this prospectus and announcing results of allocation of Hong Kong Offer Shares;
• registering new issues or transfers into or out of the names of the holders of the Shares including, where applicable, HKSCC Nominees;
• verifying identities of applicants for and holders of the Shares and identifying any duplicate applications for the Shares;
• establishing benefit entitlements of holders of the Shares, such as dividends, rights issues, bonus issues, etc.;
• any other incidental or associated purposes relating to the above and/or to enable our Company and the H Share Registrar to discharge their obligations to applicants and holders of the Shares and/or regulators and/or any other purposes to which applicants and holders of the Shares may from time to time agree.
Personal data held by our Company and the H Share Registrar relating to the applicants for and holders of Hong Kong Offer Shares will be kept confidential but our Company and the H Share Registrar may, to the extent necessary for achieving any of the above purposes, disclose, obtain or transfer (whether within or outside Hong Kong) the personal data to, from or with any of the following:
• our Company's appointed agents such as financial advisers, receiving banks and overseas principal share registrar;
• HKSCC or HKSCC Nominees, who will use the personal data and may transfer the personal data to the H Share Registrar, in each case for the purposes of providing its services or facilities or performing its functions in accordance with its rules or procedures and operating FINI and CCASS (including where applicants for the Hong Kong Offer Shares request a deposit into CCASS);
• any agents, contractors or third-party service providers who offer administrative, telecommunications, computer, payment or other services to our Company or the H Share Registrar in connection with their respective business operation;
• the Stock Exchange, the SFC and any other statutory regulatory or governmental bodies or otherwise as required by laws, rules or regulations, including for the purpose of the Stock Exchange's administration of the Listing Rules and the SFC's performance of its statutory functions; and
• any persons or institutions with which the holders of Hong Kong Offer Shares have or propose to have dealings, such as their bankers, solicitors, accountants or brokers etc.
Our Company and the H Share Registrar will keep the personal data of the applicants and holders of Hong Kong Offer Shares for as long as necessary to fulfill the purposes for which the personal data were collected. Personal data which is no longer required will be destroyed or dealt with in accordance with the Personal Data (Privacy) Ordinance (Chapter 486 of the Laws of Hong Kong).
Applicants for and holders of Hong Kong Offer Shares have the right to ascertain whether our Company or the H Share Registrar hold their personal data, to obtain a copy of that data, and to correct any data that is inaccurate. Our Company and the H Share Registrar have the right to charge a reasonable fee for the processing of such requests. All requests for access to data or correction of data should be addressed to our Company and the H Share Registrar, at their registered address disclosed in the section headed "Corporate information" in this prospectus or as notified from time to time, for the attention of our Company secretary, or the H Share Registrar for the attention of the privacy compliance officer. – 334 –
The following is the text of a report set out on pages I-1 to I-85, received from the Company's reporting accountants, KPMG, Certified Public Accountants, Hong Kong, for the purpose of incorporation in this prospectus.
ACCOUNTANTS' REPORT ON HISTORICAL FINANCIAL INFORMATION TO THE DIRECTORS OF KNOWLEDGE ATLAS TECHNOLOGY JOINT STOCK COMPANY LIMITED AND CHINA INTERNATIONAL CAPITAL CORPORATION HONG KONG SECURITIES LIMITED
我们就知识图谱科技股份有限公司（「本公司」）及其附属公司（统称「本集团」）的历史财务资料作出报告，该等历史财务资料载于第I-3至I-85页，包括本集团于2022年、2023年及2024年12月31日及2025年6月30日的综合财务状况表及本公司财务状况表，以及截至2022年、2023年及2024年12月31日止各年度及截至2025年6月30日止六个月期间（「往绩记录期间」）的综合损益及其他全面收益表、综合权益变动表及综合现金流量表，以及重大会计政策资料及其他说明资料（统称「历史财务资料」）。载于第I-3至I-85页的历史财务资料构成本报告的组成部分，本报告乃为纳入本公司日期为2025年12月30日的招股章程（「招股章程」）而编制，与本公司股份于香港联合交易所有限公司主板首次上市有关。
本公司董事负责编制根据历史财务资料附注1所载编制及呈列基准呈真实公平反映的历史财务资料，以及本公司董事认为为使历史财务资料的编制不含因欺诈或错误而产生的重大错误陈述所必要的内部控制。
我们的责任是就历史财务资料发表意见并向你们报告我们的意见。我们按照香港会计师公会（「香港会计师公会」）颁布的香港投资通函申报业务准则第200号「投资通函中历史财务资料的会计师报告」开展工作。该准则规定我们须遵守道德准则，并规划和执行工作，以就历史财务资料是否不含重大错误陈述获取合理保证。
我们的工作包括执行程序以就历史财务资料中的金额及披露事项取得证据。所选择的程序取决于申报会计师的判断，包括评估历史财务资料因欺诈或错误而产生重大错误陈述的风险。在作出该等风险评估时，申报会计师会考虑与实体编制根据历史财务资料附注1所载编制及呈列基准呈真实公平反映的历史财务资料相关的内部控制，以设计在该等情况下适当的程序，但并非为就实体内部控制的有效性发表意见。我们的工作亦包括评估董事所采用会计政策的适当性及所作会计估计的合理性，以及评估历史财务资料的整体呈列。
我们认为，我们所获取的审计证据已足够且适当，足以为我们发表意见提供基础。
在我们看来，历史财务资料就会计师报告之目的而言，按照历史财务资料附注1所载的编制及呈列基准，真实公平地反映了本集团及本公司于2022年、2023年及2024年12月31日及2025年6月30日的财务状况，以及本集团在记录期间的财务表现及现金流量，并符合香港财务报告准则。
我们已审阅本集团的存根期对应财务资料，该资料包括截至2024年6月30日止六个月的综合损益及其他全面收益表、综合权益变动表及综合现金流量表及其他说明资料（"存根期对应财务资料"）。本公司董事负责按照历史财务资料附注1所载的编制及呈列基准编制及呈列存根期对应财务资料。我们的责任是根据我们的审阅就存根期对应财务资料发表结论。我们按照香港会计师公会颁布的《香港审阅委聘准则第2410号——由实体独立核数师执行的中期财务资料审阅》进行审阅。审阅工作主要包括向负责财务及会计事务的人员查询，以及应用分析和其他审阅程序。审阅工作的范围远少于根据香港审计准则进行的审计，因此无法使我们获得保证，令我们能察觉审计中可能识别的所有重大事项。因此，我们不发表审计意见。根据我们的审阅，我们并没有注意到任何令我们相信存根期对应财务资料就会计师报告之目的而言，在所有重大方面并非按照历史财务资料附注1所载的编制及呈列基准编制的事项。
在编制历史财务资料时，未对第I-3页所界定的基础财务报表作出任何调整。
我们参阅历史财务资料附注30(g)，该附注说明本公司在记录期间内并未派付任何股息。
HISTORICAL FINANCIAL INFORMATION Set out below is the Historical Financial Information which forms an integral part of this accountants' report. The consolidated financial statements of the Group for the Track Record Period, on which the Historical Financial Information is based, were audited by KPMG under separate terms of engagement with the Company in accordance with Hong Kong Standards on Auditing issued by the HKICPA (the "Underlying Financial Statements").
Consolidated statements of profit or loss and other comprehensive income (Expressed in Renminbi ("RMB"))
| | Note | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | | | | | | (unaudited) | | | Revenue | 4 | 57,409 | 124,538 | 312,414 | 44,909 | 190,877 | | Cost of revenue | | (26,049) | (44,056) | (136,525) | (22,950) | (95,453) | | Gross profit | | 31,360 | 80,482 | 175,889 | 21,959 | 95,424 | | Other income | 4(b) | 1,784 | 9,965 | 19,281 | 4,174 | 4,614 | | Selling and marketing expenses | 5 | (15,139) | (101,198) | (387,475) | (144,194) | (208,570) | | General and administration expenses | | (32,316) | (66,302) | (133,603) | (51,452) | (185,165) | | Research and development expenses | | (84,377) | (528,884) | (2,195,436) | (98,719) | (625,723) | | Impairment losses on financial assets | 31(a) | (31) | (19,786) | (17,008) | (5,694) | (26,332) | | | 15 | - | (453) | 21,254 | 324 | 14,147 | | | | 5,972 | 26,022 | 66,271 | 7,004 | 9,791 | | | 26 | (45,209) | (161,471) | (468,859) | (201,174) | (429,295) | | Loss from operations | | | | | | | | Finance costs | 6 | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Share of profits less losses of associates | 7 | - | - | - | - | - | | Changes in fair value of financial instruments measured at fair value through profit or loss ("FVPL") | | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Changes in the carrying amounts of financial instruments issued to investors | | (763) | (10,867) | (2,538,352) | (859,217) | (1,594,661) | | Loss before taxation | | | | | (38,321) | (12,212) | | Income tax | | | | | (53,270) | | | Loss for the year/period | | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) |
The accompanying notes form part of this Historical Financial Information.
Consolidated statements of profit or loss and other comprehensive income (continued) (Expressed in RMB)
| | Note | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | | | | | | (unaudited) | | | Loss for the year/period | | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Other comprehensive income for the year/period (after tax): | | | | | | | | Items that may be reclassified to profit or loss: | | | | | | | | – Exchange differences on translation of financial statements into presentation currency | | - | - | (79) | (1,516) | (1,516) | | | | | | | - | - | | | | | | | 414 | (6,679) | | | | | | | (6,679) | | | Total comprehensive income for the year/period | | (143,650) | (787,957) | (2,958,086) | (1,235,534) | (2,357,438) | | Loss attributable to: | | | | | | | | Equity shareholders of the Company | | (143,374) | (787,960) | (2,956,491) | (1,235,551) | (2,351,173) | | Non-controlling interests | | (276) | 3 | (1,516) | - | (6,679) | | | | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Total comprehensive income attributable to: | | | | | | | | Equity shareholders of the Company | | (143,374) | (787,960) | (2,956,570) | (1,235,534) | (2,350,759) | | Non-controlling interests | | (276) | 3 | (1,516) | - | (6,679) | | | | (143,650) | (787,957) | (2,958,086) | (1,235,534) | (2,357,438) | | Loss per share | | | | | | | | Basic and diluted (RMB) | 10 | (7.80) | (29.46) | (87.20) | (36.67) | (62.27) |
| | Note | As at 31 December | | | As at 30 June | |---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Non-current assets | | | | | | | Property and equipment | 11(a) | 41,418 | 787,537 | 866,363 | 772,619 | | Intangible assets | 12(a) | 21,027 | 54,573 | 50,359 | 55,399 | | Goodwill | 13 | - | 39,379 | 39,379 | 39,379 | | Interests in associates | 15 | - | 13,047 | 201,198 | 290,345 | | Other non-current assets | 17(a) | - | 46 | 97,260 | 202,614 | | Time deposits | 21(b) | - | 102,093 | 105,343 | - | | | | 62,445 | 996,675 | 1,359,902 | 1,360,356 | | Current assets | | | | | | | Short-term investments measured at FVPL | 16 | 31,777 | - | - | - | | Inventories and contract costs | 18(a) | 10,342 | 158,904 | 42,621 | 549,364 | | Trade and other receivables | 19(a) | 27,886 | 28,782 | 32,465 | 67,866 | | Contract assets | 20 | 403 | 416,441 | 666,841 | 453,387 | | Time deposits | 21(b) | 10,092 | 9,960 | 4,718 | 6,654 | | Cash at bank and on hand | 21(a) | 219,031 | 1,249,391 | 2,269,222 | 2,556,116 | | | | 299,531 | 1,863,478 | 3,015,867 | 3,740,355 | | Current liabilities | | | | | | | Trade and other payables | 22(a) | 25,834 | 288,197 | 603,488 | 838,417 | | Contract liabilities | 23(a) | 35,230 | 74,062 | 75,059 | 75,367 | | Bank loans | 24 | 12,832 | 66,765 | 137,246 | 137,214 | | Lease liabilities | 25(a) | - | - | 213,161 | 213,458 | | Financial instruments issued to investors | 26 | 457,959 | 3,179,864 | 6,676,943 | 9,564,760 | | Convertible bonds | 27 | - | - | 132,158 | - | | | | 531,855 | 3,608,888 | 7,838,055 | 10,829,216 | | Net current liabilities | | (232,324) | (1,745,410) | (4,822,188) | (7,088,861) | | Total assets less current liabilities | | (169,879) | (748,735) | (3,462,286) | (5,728,505) | | Non-current liabilities | | | | | | | Lease liabilities | 25(a) | 10,309 | 204,117 | 458,107 | 386,132 | | Deferred income | | - | 29,741 | 34,752 | 36,204 | | | | 10,309 | 233,858 | 492,859 | 422,336 | | NET LIABILITIES | | (180,188) | (982,593) | (3,955,145) | (6,150,841) |
The accompanying notes form part of this Historical Financial Information.
| | Note | As at 31 December | | | As at 30 June | |---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | CAPITAL AND RESERVES | | | | | | | Paid-in capital/share capital | 30 | 14,807 | 28,478 | 36,224 | 40,281 | | Reserves | | (196,160) | (1,011,071) | (3,992,853) | (6,185,927) | | Total equity - deficit attributable to equity shareholders of the Company | | (181,353) | (982,593) | (3,956,629) | (6,145,646) | | Non-controlling interests | | 1,165 | - | 1,484 | (5,195) | | TOTAL EQUITY DEFICIT | | (180,188) | (982,593) | (3,955,145) | (6,150,841) |
The accompanying notes form part of this Historical Financial Information.
Non-current assets Property and equipment | 11(b) | 41,040 | 786,127 | 844,023 | 742,140 Intangible assets | 12(b) | 19,033 | 17,710 | 18,726 | 17,483 Investments in subsidiaries | 14 | 46,440 | 94,328 | 709,950 | 1,194,062 Interests in associates | 15 | - | 13,047 | 201,198 | 290,345 Other non-current assets | 17(b) | - | 46 | 92,920 | 202,614 Time deposits | 21(b) | - | 102,093 | 105,343 | -
Current assets Short-term investments measured at FVPL | | 31,777 | 158,904 | 42,621 | 549,364 Inventories and contract costs | | 10,277 | 28,692 | 31,172 | 64,389 Trade and other receivables | | 27,893 | 419,444 | 824,474 | 750,090 Contract assets | | 403 | 9,960 | 4,718 | 6,654 Time deposits | 21(b) | 10,092 | - | - | 106,968 Cash at bank and on hand | 21(a) | 184,742 | 1,237,835 | 1,608,972 | 1,469,365
Current liabilities Trade and other payables | 22(b) | 27,045 | 292,219 | 658,985 | 918,558 Contract liabilities | 23(b) | 34,200 | 74,012 | 68,761 | 62,546 Bank loans | 24 | 12,832 | 66,421 | 137,246 | 137,214 Lease liabilities | 25(b) | - | - | 208,021 | 206,195 Financial instruments issued to investors | 26 | 457,959 | 3,179,864 | 6,676,943 | 9,564,760 Convertible bonds | 27 | - | - | 132,158 | -
Non-current liabilities Lease liabilities | 25(b) | 10,309 | 203,696 | 447,856 | 372,274 Deferred income | 16 | - | 26,741 | 24,431 | 25,883
CAPITAL AND RESERVES Paid-in capital/share capital | 18(b) | 14,807 | 28,478 | 36,224 | 40,281 Reserves | 19(b) | (185,455) | (1,003,245) | (3,906,508) | (5,934,237) TOTAL EQUITY - DEFICIT | 30 | (170,648) | (974,767) | (3,870,284) | (5,893,956)
| Note | Paid-in capital RMB'000 (Note 30(b)) | Capital reserve RMB'000 (Note 30(c)) | Other reserve RMB'000 (Note 30(d)) | Share-based payments reserve RMB'000 (Note 30(e)) | Exchange reserve RMB'000 (Note 30(f)) | Accumulated losses RMB'000 | Total RMB'000 | Non-controlling interests RMB'000 | Total equity - deficit RMB'000 |
Balance at 1 January 2022 | | 12,812 | 1,024 | - | - | - | (209,879) | (196,043) | 1,165 | (194,878) |
Changes in equity for the year ended 31 December 2022: Total comprehensive income for the year | | - | - | - | - | (276) | - | (276) | - | (276) |
Capital contributions from equity shareholders | 30(b) | - | - | - | - | - | - | - | - | - | Increase in paid-in capital through transfer from capital reserve | 30(b) | 1,324 | (1,324) | - | - | - | - | - | - | - | Equity settled share-based transactions | 29 | - | - | - | 300 | - | - | 300 | - | 300 |
| | - | - | - | - | - | (143,374) | (143,374) | (276) | (143,650) | | | - | - | - | - | - | (66,505) | (66,505) | - | (66,505) |
Balance at 31 December 2022 | | 14,807 | - | (534) | - | - | (185,455) | (181,353) | 1,165 | (180,188) |
Wait — let me reproduce the tables more carefully based on the source data.
Balance at 1 January 2022 | | 12,812 | 1,024 | - | - | - | (209,879) | (181,353) | 1,165 | (180,188) |
Changes in equity for the year ended 31 December 2022: Total comprehensive income for the year | | - | - | - | - | (276) | (143,374) | (143,650) | 1,441 | (142,209) |
Capital contributions from equity shareholders | 30(b) | - | 58,242 | - | - | - | - | 58,242 | - | 58,242 | Increase in paid-in capital through transfer from capital reserve | 30(b) | 1,324 | (45,847) | (534) | - | - | - | (45,057) | - | (45,057) | Equity settled share-based transactions | 29 | - | - | - | 300 | - | - | 300 | - | 300 |
Balance at 31 December 2022 | | 14,807 | - | (534) | - | (276) | (209,879) | (181,353) | 1,165 | (180,188) |
The accompanying notes form part of this Historical Financial Information.
| Note | Paid-in capital RMB'000 (Note 30(b)) | Capital reserve RMB'000 (Note 30(c)) | Other reserve RMB'000 (Note 30(d)) | Share-based payments reserve RMB'000 (Note 30(e)) | Exchange reserve RMB'000 (Note 30(f)) | Accumulated losses RMB'000 | Total RMB'000 | Non-controlling interests RMB'000 | Total equity - deficit RMB'000 |
Balance at 1 January 2023 | | 14,807 | - | (534) | - | (276) | (209,879) | (180,188) | 1,165 | (180,188) |
Changes in equity for the year ended 31 December 2023: Total comprehensive income for the year | | - | - | - | - | (18,782) | (787,960) | (806,742) | 3 | (806,739) |
Capital contributions from equity shareholders | 30(b) | - | 54,409 | - | - | - | - | 54,409 | - | 54,409 | Increase in paid-in capital through transfer from capital reserve | 30(b) | 9,838 | (74,467) | - | - | - | - | (64,629) | - | (64,629) | Distributions to non-controlling equity shareholders on liquidation of a subsidiary | | - | - | - | - | - | - | - | (1,168) | (1,168) | Equity settled share-based transactions | 29 | 3,833 | (3,833) | - | 28,478 | - | - | 28,478 | - | 28,478 |
Balance at 31 December 2023 | | 28,478 | 5,502 | (534) | 1,324 | (18,782) | (997,839) | (982,593) | - | (982,593) |
The accompanying notes form part of this Historical Financial Information.
| Note | Paid-in capital RMB'000 (Note 30(b)) | Capital reserve RMB'000 (Note 30(c)) | Other reserve RMB'000 (Note 30(d)) | Share-based payments reserve RMB'000 (Note 30(e)) | Exchange reserve RMB'000 (Note 30(f)) | Accumulated losses RMB'000 | Total RMB'000 | Non-controlling interests RMB'000 | Total equity - deficit RMB'000 |
Balance at 1 January 2024 | | 28,478 | 5,502 | (534) | 1,324 | (18,782) | (997,839) | (982,593) | - | (982,593) |
Changes in equity for the year ended 31 December 2024: Loss for the year | | - | - | - | - | - | (2,956,491) | (2,956,491) | (1,516) | (2,958,007) | Other comprehensive income for the year | | - | - | - | - | (79) | - | (79) | - | (79) |
Total comprehensive income for the year | | - | - | - | - | (79) | (2,956,491) | (2,956,570) | (1,516) | (2,958,086) |
Capital contributions from equity shareholders | 30(b) | - | 52,427 | - | - | - | - | (46,809) | - | (46,809) | Increase in paid-in capital through transfer from capital reserve | 30(b) | 5,764 | (1,982) | - | - | - | - | (121,276) | - | (121,276) | Capital contribution from non-controlling interests | | - | - | - | - | - | - | - | 3,000 | 3,000 | Equity settled share-based transactions | 29 | 1,982 | (46,809) | - | 36,224 | - | - | - | - | - |
Balance at 31 December 2024 | | 30,405 | 23,579 | (79) | - | - | (3,954,330) | (3,956,629) | 1,484 | (3,955,145) |
The accompanying notes form part of this Historical Financial Information.
| | | Attributable to equity shareholders of the Company | | | | | | | |---|---|---|---|---|---|---|---|---| | | Paid-in capital / share capital RMB'000 (Note 30(b)) | Capital reserve RMB'000 (Note 30(c)) | Other reserve RMB'000 (Note 30(d)) | Share-based payments reserve RMB'000 (Note 30(e)) | Exchange reserve RMB'000 (Note 30(f)) | Accumulated losses RMB'000 | Total RMB'000 | Non-controlling interests RMB'000 | Total equity - deficit RMB'000 |
Balance at 1 January 2025 | 36,224 | (3,954,330) | (2,351,173) | (79) | 414 | (2,062,211) | (3,956,629) | 1,484 | (3,955,145) |
Total comprehensive income for the period | - | - | - | - | - | (4,243,292) | (2,350,759) | - | (2,357,438) |
Capital contributions from equity shareholders | 30(b) | 4,243,292 | - | - | - | - | 4,243,292 | - | - |
| | 40,281 | (4,243,292) | - | (121,276) | - | - | (4,190,865) | (125,333) | - | - | (4,057) | - | (4,057) | - | - | 30,405 | 192,147 | 161,742 | 161,742 |
The accompanying notes form part of this Historical Financial Information.
| | | Attributable to equity shareholders of the Company | | | | | | | |---|---|---|---|---|---|---|---|---| | Note | Paid-in capital RMB'000 (Note 30(b)) | Capital reserve RMB'000 (Note 30(c)) | Other reserve RMB'000 (Note 30(d)) | Share-based payments reserve RMB'000 (Note 30(e)) | Exchange reserve RMB'000 (Note 30(f)) | Accumulated losses RMB'000 | Total RMB'000 | Non-controlling interests RMB'000 | Total equity - deficit RMB'000 |
Balance at 1 January 2024 | | 17 | (997,839) | (1,235,551) | - | 17 | (1,235,551) | (982,593) | - | (982,593) |
Total comprehensive income for the period (unaudited) | | - | - | (1,235,551) | - | 17 | (1,235,534) | - | (1,235,534) |
Capital contributions from equity shareholders (unaudited) | 30(b) | - | - | - | - | - | - | - | - |
Equity settled share-based transactions (unaudited) | | - | - | - | 4,217 | - | - | 4,217 | - | 4,217 |
Balance at 30 June 2024 (unaudited) | | 17 | (2,233,390) | - | - | (1,235,551) | (2,247,895) | (29,768) | (33,985) | (2,247,895) |
| | 2,092 | 29 | 30,570 | 2,092 | 54,409 | - | - | 54,409 | - | 28,478 | (110,544) | (36,077) | (36,077) | - | - | (74,467) | 11,043 | 4,217 | 4,217 | - | - | - | 6,826 |
The accompanying notes form part of this Historical Financial Information.
| | Note | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 |
| Loss before taxation | | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | **Adjustments for:** | | | | | | | | Depreciation on property and equipment | 6(c) | 16,567 | 63,822 | 270,252 | 131,105 | 132,908 | | Amortisation of intangible assets | 6(c) | 3,451 | 5,109 | 9,685 | 4,510 | 6,128 | | Net loss/(gain) on disposal of property and equipment and intangible assets | 5 | - | 1,539 | (6,807) | - | 15 | | Equity settled share-based transactions | 6(b) | 1,024 | 5,502 | 23,579 | 4,217 | 158,852 | | Changes in fair value of financial instruments measured at FVPL | | (5,972) | (26,022) | (66,271) | (7,004) | (9,791) | | Changes in the carrying amounts of financial instruments issued to investors | 26 | 45,209 | 161,471 | 468,859 | 201,174 | 429,295 | | Interest income from time deposits | | (462) | (2,176) | (3,250) | (1,625) | (1,625) | | Finance costs | 6(a) | 5,694 | 26,332 | 38,321 | 12,212 | 53,270 | | Share of profits less losses of associates | 15 | - | 453 | (21,254) | (324) | (14,147) | | **Changes in working capital:** | | | | | | | | Increase in inventories and contract costs | | (2,918) | (18,369) | (3,683) | (24,905) | (29,200) | | (Increase)/decrease in trade and other receivables | | (19,601) | (269,462) | (415,011) | (279,763) | 72,790 | | (Increase)/decrease in contract assets | | (403) | (9,603) | (2,471) | 8,730 | (4,743) | | Increase in restricted cash | | (103) | (113) | (842) | - | (3,088) | | Increase in trade and other payables | | 12,893 | 143,193 | 415,973 | 179,911 | 238,278 | | Increase in contract liabilities | | 38,832 | 415,973 | 997 | 14,073 | 308 | | Increase/(decrease) in deferred income | | 6,888 | 19,432 | 5,011 | (1,438) | 1,452 |
| Cash used in operations | | (68,246) | (648,017) | (2,244,919) | (994,678) | (1,327,150) | | Income tax paid | | - | - | - | - | - | | **Net cash used in operating activities** | | **(68,246)** | **(648,017)** | **(2,244,919)** | **(994,678)** | **(1,327,150)** |
The accompanying notes form part of this Historical Financial Information.
| | Note | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 |
Investing activities Payments for the purchases of property and equipment (31,651) (506,760) (125,958) (98,388) (12,740) Payments for the purchases of intangible assets (1,809) (6,802) (6,150) (11,168) Proceeds from disposal of property and equipment and intangible assets 7 - 114,603 5 Payments for the purchases of wealth management product (1,041,999) (1,100,000) (100,000) (100,000) (500,000) Proceeds from disposal of wealth management products 1,117,232 1,023,594 200,534 180,412 Placement of time deposits - (110,000) Proceeds from maturity of time deposits 370 20,175 Payments for the purchases of investments in equity securities measured at FVPL (11,038) (35,579) (118,747) (111,649) Proceeds from disposal of investments in equity securities measured at FVPL - 160,811 42,530 Payments for capital injections to associates (13,500) (170,000) - (75,000) Payment for the acquisition of a subsidiary, net of cash acquired 32 (61,086) (3,000) (3,000) Net cash generated from/(used in) investing activities
The accompanying notes form part of this Historical Financial Information.
Financing activities Proceeds from the issuance of financial instruments to investors 21(c) Payments of transaction costs for the issuance of financial instruments to investors 21(c) Advanced capital injections received from equity shareholders 21(c) Proceeds from the issuance of convertible bonds 21(c) Proceeds from bank loans 21(c) Capital element of lease rentals paid 21(c) Interest element of lease rentals paid 21(c) Capital contribution from non-controlling interests Distribution to non-controlling equity holders on liquidation of a subsidiary Proceeds from financing of property and equipment Payments of interests on bank loans 21(c) Payment of costs in connection with the proposed listing of the Company's shares
Effect of foreign exchange rate changes Cash and cash equivalents at the end of the year/period 21(a)(i)
Knowledge Atlas Technology Joint Stock Company Limited (the "Company", formerly known as Beijing Zhipu Huazhang Technology Limited) was established in the People's Republic of China (the "PRC") on 11 June 2019 as a limited liability company. The Company was converted from a limited liability company into a joint stock limited liability company on 26 March 2025.
The Company and its subsidiaries (together, the "Group") are principally engaged in the provision of large model-related services in the PRC.
| Name of Company | Place and date of establishment/ incorporation | Particulars of issued/ registered and paid-up capital | Held by the Group | Held by the Company | Held by a subsidiary | Principal activities | Auditors | |---|---|---|---|---|---|---|---| | Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司) (formerly known as Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司)) (i) | The PRC 19 April 2013 | RMB1,100,000 | 100% | 100% | - | Provision of large model-related services | 2022: RSM ChinaCPA LLP (容誠會計師事務所（特殊普通合夥）) (i) 2023-2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) | | Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司) (formerly known as Shenzhen Knowledge Future Co., Ltd. (深圳智譜未來科技有限公司)) (i) | The PRC 21 July 2021 | RMB5,000,000 | 100% | 100% | - | Provision of large model-related services | 2022: RSM ChinaCPA LLP (容誠會計師事務所（特殊普通合夥）) (i) 2023-2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) | | Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司) ("Beijing Lingxin Intelligent") (See Note 32) (i) | The PRC 19 November 2021 | RMB5,052,631 | 100% | 100% | - | Provision of large model-related services | 2023-2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) | | Jincheng Yaoda Technology Limited (ii) | The British Virgin Islands 6 January 2022 | USD100 | 100% | 100% | - | Investment holding | | | Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司) (i) | The PRC 20 September 2023 | RMB10,000,000 | 100% | 100% | - | Provision of large model-related services | 2023-2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) | | Hong Kong Xiangtai Ruifeng Technology Limited (ii) | Hong Kong 21 September 2023 | 100,000 shares | 100% | - | 100% | Investment holding | 2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) |
Place and date of Particulars of Held by Held by establishment/ issued/ registered the the Held by a incorporation and paid-up capital Group Company subsidiary Principal activities
| Name of Company | Place and date of establishment/ incorporation | Particulars of issued/ registered and paid-up capital | Held by the Group | Held by the Company | Held by a subsidiary | Principal activities | Auditors | |---|---|---|---|---|---|---|---| | Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司) (i) | The PRC 18 December 2024 | RMB15,000,000 | 100% | 100% | - | Provision of large model-related services | Not applicable | | Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司) (i) | The PRC 27 December 2024 | RMB300,000,000/ RMB150,000,000 | 100% | 100% | - | Provision of large model-related services | 2024: Beijing Chengju Certified Public Accountants (北京誠炬會計師事務所(普通合夥)) (i) | | Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司) (i) | The PRC 24 February 2025 | RMB450,000,000/ RMB200,000,000 | 100% | 100% | - | Provision of large model-related services | Not applicable | | Beijing Knowledge Qingying Technology Culture Media Co., Ltd. (北京智譜清影科技文化傳媒有限公司) (i) | The PRC 13 March 2025 | RMB5,000,000/ nil | 100% | 100% | - | Provision of large model-related services | Not applicable | | Corethinks Technology SDN. BHD. (ii) | Malaysia 30 June 2025 | MYR1 | 100% | - | 100% | Provision of large model-related services | Not applicable |
Notes: (i) These companies are limited liability companies established in the Chinese Mainland. The English translations of the names are for reference only. The official names of these entities are in Chinese.
(ii) These companies are limited liability companies established outside of the Chinese Mainland.
All companies comprising the Group have adopted 31 December as their financial year end date.
The Historical Financial Information has been prepared in accordance with IFRS Accounting Standards issued by the International Accounting Standards Board (the "IASB"). Further details of the material accounting policy information adopted are set out in Note 2.
The IASB has issued a number of new and revised IFRS Accounting Standards. For the purpose of preparing this Historical Financial Information, the Group has consistently applied all applicable new and revised IFRS Accounting Standards throughout the Track Record Period. The Group has not adopted any new and revised accounting standards and interpretations issued but not yet effective for the Track Record Period which are set out in Note 35.
The Historical Financial Information also complies with the applicable disclosure provisions of the Rules Governing the Listing of Securities on The Stock Exchange of Hong Kong Limited (the "Stock Exchange").
For the six months ended 30 June 2025, the Group incurred net loss of RMB2,357,852,000 and as at 30 June 2025, the Group recorded net liabilities of RMB6,150,841,000 and net current liabilities of RMB7,088,861,000, which included financial instruments issued to investors amounted to RMB9,564,760,000. The directors of the Company are of the opinion that no payment is expected for the settlement of the liabilities arising from financial instruments issued to investors within twelve months from the date of this report and the related redemption options would be terminated and the financial instruments issued to investors would be converted into equity upon the listing of the Company's shares on the Stock
Associates are entities over which the Group has significant influence but not control or joint control.
Investments in associates are accounted for using the equity method. Under the equity method, the investment is initially recognised at cost and the carrying amount is increased or decreased to recognise the investor's share of the profit or loss and other comprehensive income of the investee after the date of acquisition.
When the Group's share of losses of an associate equals or exceeds the Group's interest in the associate (which includes any long-term interests that, in substance, form part of the Group's net investment in the associate), the Group does not recognise further losses, unless it has incurred legal or constructive obligations or made payments on behalf of the associate.
An investment in an associate is accounted for using the equity method from the date that significant influence commences until the date that significant influence ceases. When the Group's interest in an associate decreases but significant influence is retained, only a proportionate share of the amounts previously recognised in other comprehensive income are reclassified to profit or loss where appropriate.
Gains and losses resulting from transactions between the Group and its associates are recognised only to the extent of unrelated investors' interests in the associates. Unrealised losses are eliminated unless the transaction provides evidence of an impairment of the asset transferred.
When an investment in an associate is classified as held for sale, it is accounted for in accordance with IFRS 5 Non-current Assets Held for Sale and Discontinued Operations.
Goodwill represents the excess of the cost of an acquisition over the fair value of the Group's share of the net identifiable assets of the acquired subsidiary/associate at the date of acquisition.
Goodwill is tested annually for impairment and carried at cost less accumulated impairment losses. Impairment losses on goodwill are not reversed. Gains and losses on the disposal of an entity include the carrying amount of goodwill relating to the entity sold.
Goodwill is allocated to cash-generating units for the purpose of impairment testing. The allocation is made to those cash-generating units or groups of cash-generating units that are expected to benefit from the business combination in which the goodwill arose.
— those to be measured subsequently at fair value (either through other comprehensive income ("OCI"), or through profit or loss); and
— those to be measured at amortised cost.
The classification depends on the Group's business model for managing the financial assets and the contractual terms of the cash flows.
For assets measured at fair value, gains and losses will either be recorded in profit or loss or OCI. For investments in equity instruments that are not held for trading, this will depend on whether the Group has made an irrevocable election at the time of initial recognition to account for the equity investment at fair value through other comprehensive income ("FVOCI").
The Group reclassifies debt investments when and only when its business model for managing those assets changes.
At initial recognition, the Group measures a financial asset at its fair value plus, in the case of a financial asset not at fair value through profit or loss ("FVPL"), transaction costs that are directly attributable to the acquisition of the financial asset. Transaction costs of financial assets carried at FVPL are expensed in profit or loss.
Financial assets with embedded derivatives are considered in their entirety when determining whether their cash flows are solely payments of principal and interest.
Subsequent measurement of debt instruments depends on the Group's business model for managing the asset and the cash flow characteristics of the asset. There are three measurement categories into which the Group classifies its debt instruments:
— Amortised cost: Assets that are held for collection of contractual cash flows where those cash flows represent solely payments of principal and interest are measured at amortised cost. Interest income from these financial assets is included in finance income using the effective interest rate method. Any gain or loss arising on derecognition is recognised directly in profit or loss and presented in other gains/(losses), together with foreign exchange gains and losses. Impairment losses are presented as a separate line item in the statement of profit or loss.
— FVOCI: Assets that are held for collection of contractual cash flows and for selling the financial assets, where the assets' cash flows represent solely payments of principal and interest, are measured at FVOCI. Movements in the carrying amount are taken through OCI, except for the recognition of impairment gains or losses, interest income and foreign exchange gains and losses which are recognised in profit or loss. When the financial asset is derecognised, the cumulative gain or loss previously recognised in OCI is reclassified from equity to profit or loss and recognised in other gains/(losses). Interest income from these financial assets is included in finance income using the effective interest rate method. Foreign exchange gains and losses are presented in other gains/(losses) and impairment expenses are presented as a separate line item in the statement of profit or loss.
— FVPL: Assets that do not meet the criteria for amortised cost or FVOCI are measured at FVPL. A gain or loss on a debt investment that is subsequently measured at FVPL is recognised in profit or loss and presented net within other gains/(losses) in the period in which it arises.
The Group subsequently measures all equity investments at fair value. Where the Group's management has elected to present fair value gains and losses on equity investments in OCI, there is no subsequent reclassification of fair value gains and losses to profit or loss following the derecognition of the investment. Dividends from such investments continue to be recognised in profit or loss as other income when the Group's right to receive payments is established.
Changes in the fair value of financial assets at FVPL are recognised in other gains/(losses) in the statement of profit or loss as applicable. Impairment losses (and reversal of impairment losses) on equity investments measured at FVOCI are not reported separately from other changes in fair value.
At inception of a contract, the Group assesses whether a contract is, or contains, a lease. A lease is defined as a contract that conveys the right to control the use of an identified asset for a period of time in exchange for consideration. To assess whether a contract conveys the right to control the use of an identified asset, the Group uses the definition of a lease in HKFRS 16.
This policy is applied to contracts entered into, or changed, on or after 1 January 2019.
At commencement or on modification of a contract that contains a lease component, the Group allocates the consideration in the contract to each lease component on the basis of its relative stand-alone prices. However, for the leases of property the Group has elected not to separate non-lease components and account for the lease and associated non-lease components as a single lease component.
The Group recognises a right-of-use asset and a lease liability at the lease commencement date. The right-of-use asset is initially measured at cost, which comprises the initial amount of the lease liability adjusted for any lease payments made at or before the commencement date, plus any initial direct costs incurred and an estimate of costs to dismantle and remove the underlying asset or to restore the underlying asset or the site on which it is located, less any lease incentives received.
The right-of-use asset is subsequently depreciated using the straight-line method from the commencement date to the end of the lease term, unless the lease transfers ownership of the underlying asset to the Group by the end of the lease term or the cost of the right-of-use asset reflects that the Group will exercise a purchase option. In that case the right-of-use asset will be depreciated over the useful life of the underlying asset, which is determined on the same basis as those of property and equipment. In addition, the right-of-use asset is periodically reduced by impairment losses, if any, and adjusted for certain remeasurements of the lease liability.
The lease liability is initially measured at the present value of the lease payments that are not paid at the commencement date, discounted using the interest rate implicit in the lease or, if that rate cannot be readily determined, the Group's incremental borrowing rate. Generally, the Group uses its incremental borrowing rate as the discount rate.
The Group determines its incremental borrowing rate by obtaining interest rates from various external financing sources and makes certain adjustments to reflect the terms of the lease and type of the asset leased.
Lease payments included in the measurement of the lease liability comprise the following: – fixed payments, including in-substance fixed payments; – variable lease payments that depend on an index or a rate, initially measured using the index or rate as at the commencement date; – amounts expected to be payable under a residual value guarantee; and – the exercise price under a purchase option that the Group is reasonably certain to exercise, lease payments in an optional renewal period if the Group is reasonably certain to exercise an extension option, and penalties for early termination of a lease unless the Group is reasonably certain not to terminate early.
The lease liability is measured at amortised cost using the effective interest method. It is remeasured when there is a change in future lease payments arising from a change in an index or rate, if there is a change in the Group's estimate of the amount expected to be payable under a residual value guarantee, if the Group changes its assessment of whether it will exercise a purchase, extension or termination option or if there is a revised in-substance fixed lease payment.
When the lease liability is remeasured in this way, a corresponding adjustment is made to the carrying amount of the right-of-use asset, or is recorded in profit or loss if the carrying amount of the right-of-use asset has been reduced to zero.
The Group presents right-of-use assets that do not meet the definition of investment property in 'property and equipment' and lease liabilities in 'lease liabilities' in the statement of financial position.
The Group has elected not to recognise right-of-use assets and lease liabilities for leases of low-value assets and short-term leases, including IT equipment. The Group recognises the lease payments associated with these leases as an expense on a straight-line basis over the lease term.
The Group recognises loss allowances for expected credit losses ("ECL") on: – trade and other receivables; – contract assets; and – amounts due from related parties.
Financial assets measured at FVOCI – recycling are subject to the ECL model, but the loss allowance is recognised in OCI instead of reducing the carrying amount of the financial asset.
ECLs are a probability-weighted estimate of credit losses. Credit losses are measured as the present value of all cash shortfalls (i.e. the difference between the cash flows due to the entity in accordance with the contract and the cash flows that the Group expects to receive).
The maximum period considered when estimating ECLs is the maximum contractual period over which the Group is exposed to credit risk.
When measuring ECL, the Group and the Company take into account reasonable and supportable information that is available without undue cost or effort. This includes both quantitative and qualitative information and analysis, based on the Group's historical experience and informed credit assessment and including forward-looking information.
The Group measures loss allowances at an amount equal to lifetime ECL, except for the following, which are measured at 12-month ECL: – debt securities that are determined to have low credit risk at the reporting date; and – other debt securities and bank balances for which credit risk (i.e. the risk of default occurring over the expected life of the financial instrument) has not increased significantly since initial recognition.
Loss allowances for trade receivables and contract assets are always measured at an amount equal to lifetime ECL.
When determining whether the credit risk of a financial asset has increased significantly since initial recognition and when estimating ECL, the Group considers reasonable and supportable information that is relevant and available without undue cost or effort. This includes both quantitative and qualitative information and analysis, based on the Group's historical experience and informed credit assessment and including forward-looking information.
The Group assumes that the credit risk on a financial asset has increased significantly if it is more than 30 days past due.
The Group considers a financial asset to be in default when: – the borrower is unlikely to pay its credit obligations to the Group in full, without recourse by the Group to actions such as realising security (if any is held); or – the financial asset is more than 90 days past due.
Lifetime ECLs are the ECLs that result from all possible default events over the expected life of a financial instrument.
12-month ECLs are the portion of ECLs that result from default events that are possible within the 12 months after the reporting date (or a shorter period if the expected life of the instrument is less than 12 months).
The maximum period considered when estimating ECLs is the maximum contractual period over which the Group is exposed to credit risk.
At each reporting date, the Group assesses whether financial assets carried at amortised cost and debt securities at FVOCI – recycling are credit-impaired. A financial asset is 'credit-impaired' when one or more events that have a detrimental impact on the estimated future cash flows of the financial asset have occurred.
The gross carrying amount of a financial asset is written off when the Group has no reasonable expectations of recovering a financial asset in its entirety or a portion thereof. The Group individually makes an assessment with respect to the timing and amount of write-off based on whether there is a reasonable expectation of recovery. The Group expects no significant recovery from the amount written off. However, financial assets that are written off could still be subject to enforcement activities in order to comply with the Group's procedures for recovery of amounts due.
Internal and external sources of information are reviewed at each reporting date to identify indications that the following assets may be impaired or, except in the case of goodwill, an impairment loss previously recognised no longer exists or may have decreased: – property and equipment; – right-of-use assets; – intangible assets; and – goodwill.
If any such indication exists, the asset's recoverable amount is estimated. In addition, for goodwill, intangible assets that are not yet available for use and intangible assets that have indefinite useful life, the recoverable amount is estimated annually whether or not there is any indication of impairment.
The recoverable amount of an asset is the greater of its fair value less costs of disposal and value in use. In assessing value in use, the estimated future cash flows are discounted to their present value using a pre-tax discount rate that reflects current market assessments of the time value of money and the risks specific to the asset. Where an asset does not generate cash inflows largely independent of those from other assets, the recoverable amount is determined for the smallest group of assets that generates cash inflows independently (i.e. a cash-generating unit).
An impairment loss is recognised in profit or loss if the carrying amount of an asset, or the cash-generating unit to which it belongs, exceeds its recoverable amount. Impairment losses recognised in respect of cash-generating units are allocated first to reduce the carrying amount of any goodwill allocated to the cash-generating unit (or group of units) and then, to reduce the carrying amounts of the other assets in the unit (or group of units) on a pro rata basis.
In respect of assets other than goodwill, an impairment loss is reversed if there has been a favourable change in the estimates used to determine the recoverable amount. An impairment loss in respect of goodwill is not reversed.
A reversal of an impairment loss is limited to the asset's carrying amount that would have been determined had no impairment loss been recognised in prior years. Reversals of impairment losses are credited to profit or loss in the year in which the reversals are recognised.
At inception of a contract, the Group assesses whether the contract is, or contains, a lease. This is the case if the contract conveys the right to control the use of an identified asset for a period of time in exchange for consideration. Control is conveyed where the customer has both the right to direct the use of the identified asset and to obtain substantially all of the economic benefits from that use.
Where the contract contains lease component(s) and non-lease component(s), the Group has elected not to separate non-lease components and accounts for each lease component and any associated non-lease components as a single lease component for all leases.
At the lease commencement date, the Group recognises a right-of-use asset and a lease liability, except for leases that have a short lease term of 12 months or less, and leases of low-value items. When the Group enters into a lease in respect of a low-value item, the Group decides whether to capitalise the lease on a lease-by-lease basis. If not capitalised, the associated lease payments are recognised in profit or loss on a systematic basis over the lease term.
Where the lease is capitalised, the lease liability is initially recognised at the present value of the lease payments payable over the lease term, discounted using the interest rate implicit in the lease or, if that rate cannot be readily determined, using a relevant incremental borrowing rate. After initial recognition, the lease liability is measured at amortised cost and interest expense is recognised using the effective interest method. Variable lease payments that do not depend on an index or rate are not included in the measurement of the lease liability, and are charged to profit or loss as incurred.
The right-of-use asset recognised when a lease is capitalised is initially measured at cost, which comprises the initial amount of the lease liability adjusted for any lease payments made at or before the commencement date, plus any initial direct costs incurred and an estimate of costs to dismantle and remove the underlying asset or to restore the underlying asset or the site on which it is located, less any lease incentives received. The right-of-use asset is subsequently stated at cost less accumulated depreciation and impairment losses (see Note 2(j)(ii)). Depreciation is calculated using the straight-line method over the lease term.
The lease liability is remeasured when there is a change in future lease payments arising from a change in an index or rate, if there is a change in the Group's estimate of the amount expected to be payable under a residual value guarantee, or if the Group changes its assessment of whether it will exercise a purchase, extension or termination option. When the lease liability is remeasured in this way, a corresponding adjustment is made to the carrying amount of the right-of-use asset, or is recorded in profit or loss if the carrying amount of the right-of-use asset has been reduced to zero.
The lease liability is also remeasured when there is a lease modification, which means a change in the scope of a lease or the consideration for a lease that is not originally provided for in the lease contract, if such modification is not accounted for as a separate lease. In this case, the lease liability is remeasured based on the revised lease payments and lease term using a revised discount rate at the effective date of the modification.
In the consolidated statement of financial position, the current portion of long-term lease liabilities is determined as the present value of contractual payments that are due to be settled within twelve months after the reporting period.
(i) Credit losses from financial instruments and contract assets.
• financial assets measured at amortised cost (including cash at bank and on hand, trade and other receivables, and time deposits); and
• contract assets (see Note 2(l)).
ECLs are a probability-weighted estimate of credit losses. Generally, credit losses are measured as the present value of all expected cash shortfalls between the contractual and expected amounts.
The expected cash shortfalls of trade and other receivables and contract assets are discounted using the effective interest rate determined at initial recognition or an approximation thereof if the effect of discounting is material.
The maximum period considered when estimating ECLs is the maximum contractual period over which the Group is exposed to credit risk.
• 12-month ECLs: these are the portion of ECLs that result from default events that are possible within the 12 months after the reporting date (or a shorter period if the expected life of the instrument is less than 12 months); and
• lifetime ECLs: these are the ECLs that result from all possible default events over the expected lives of the items to which the ECL model applies.
The Group measures loss allowances at an amount equal to lifetime ECLs, except for the following, which are measured at 12-months ECLs:
• other financial instruments for which credit risk (i.e. the risk of default occurring over the expected life of the financial instrument) has not increased significantly since initial recognition.
Loss allowances for trade receivables and contract assets are always measured at an amount equal to lifetime ECLs.
A contract asset is the right to consideration in exchange for goods or services transferred to the customer. If the Group performs by transferring goods or services to a customer before the customer pays consideration or before payment is due, a contract asset is recognised for the earned consideration that is conditional. A contract liability is the obligation to transfer goods or services to a customer for which the Group has received consideration (or an amount of consideration is due) from the customer. If a customer pays consideration before the Group transfers goods or services to the customer, a contract liability is recognised when the payment is made or the payment is due (whichever is earlier). Contract liabilities are recognised as revenue when the Group performs under the contract (see Note 2(v)(i)). (m)
Cash and cash equivalents comprise cash at bank and on hand, demand deposits with banks and other financial institutions, and short-term, highly liquid investments that are readily convertible into known amounts of cash and which are subject to an insignificant risk of changes in value, and have a maturity of three months or less when acquired. Bank overdrafts that are repayable on demand and form an integral part of the Group's cash management are included as a component of cash and cash equivalents for the purpose of the statement of cash flows.
Trade and other payables are initially recognised at fair value and subsequently stated at amortised cost unless the effect of discounting would be immaterial, in which case they are stated at cost.
Interest-bearing borrowings are recognised initially at fair value less attributable transaction costs. Subsequent to initial recognition, interest-bearing borrowings are stated at amortised cost with any difference between the amount initially recognised and redemption value being recognised in profit or loss over the period of the borrowings, together with any interest and fees payable, using the effective interest method.
(i) Short-term employee benefits and contributions to defined contribution retirement plans Short-term employee benefits are expensed as the related service is provided. A liability is recognised for the amount expected to be paid if the Group has a present legal or constructive obligation to pay this amount as a result of past service provided by the employee and the obligation can be estimated reliably. Obligations for contributions to defined contribution retirement plans are expensed as the related service is provided.
(ii) Share-based payments The grant-date fair value of equity-settled share-based payment arrangements granted to employees is recognised as an employee expense, with a corresponding increase in equity, over the vesting period of the awards. The amount recognised as an expense is adjusted to reflect the number of awards for which the related service conditions are expected to be met, such that the amount ultimately recognised is based on the number of awards that meet the related service conditions at the vesting date.
信用风险的重大增加 在确定金融工具自初始确认以来信用风险是否已显著增加以及计量预期信用损失时，本集团考虑在不需要付出过多成本或精力的情况下可获得的相关合理且有依据的信息。这包括定量和定性信息及分析，基于本集团的历史经验和知情信用评估，包含前瞻性信息。
本集团假定，若金融资产已逾期，则该金融资产的信用风险自初始确认以来已显著增加。
预期信用损失于每个报告日重新计量，以反映自初始确认以来金融工具信用风险的变化。预期信用损失金额的任何变动均在损益中确认为减值收益或损失。本集团对所有金融工具确认减值收益或损失，并通过损失准备账户对其账面金额进行相应调整，但以公允价值计量且其变动计入其他综合收益（可回收）的非权益证券投资除外。
于每个报告日，本集团评估金融资产是否已发生信用减值。当一项或多项对金融资产估计未来现金流量产生不利影响的事件已经发生时，该金融资产即发生信用减值。
• 由于发行人财务困难导致证券活跃市场消失。
核销政策 在没有切实可行的回收前景的范围内，金融资产或合同资产的账面总额予以核销。当本集团确定债务人没有可产生足够现金流量以偿还核销金额的资产或收入来源时，通常即属此情形。
此前已核销资产的后续回收，在回收发生的期间于损益中确认为减值转回。
(ii) 其他非流动资产的减值 于每个报告日，本集团审查其非金融资产（存货及其他合同成本、合同资产和递延税项资产除外）的账面金额，以确定是否存在任何减值迹象。若存在任何此类迹象，则估计该资产的可收回金额。商誉每年进行减值测试。
就减值测试而言，资产被归入能从持续使用中产生现金流入且该现金流入在很大程度上独立于其他资产或现金产生单元（"现金产生单元"）现金流入的最小资产组合。
资产或现金产生单元的可收回金额为其使用价值与公允价值减处置成本两者中的较高者。使用价值基于估计的未来现金流量，使用反映当前市场对货币时间价值及该资产或现金产生单元特定风险评估的税前折现率折现至现值。
若资产或现金产生单元的账面金额超过其可收回金额，则确认减值损失。
减值损失在损益中确认。减值损失首先用于减少分配至该现金产生单元的任何商誉的账面金额，然后按比例减少该现金产生单元中其他资产的账面金额。
商誉的减值损失不予转回。对于其他资产，减值损失的转回仅以转回后的账面金额不超过若未确认减值损失时经折旧或摊销后所确定的账面金额为限。
(i) 存货 存货按成本与可变现净值两者中的较低者计量。购入存货的成本按交易价格确定。可变现净值为正常经营过程中的估计售价减去完成销售所需的估计成本。
(ii) 合同成本 合同成本为取得客户合同的增量成本，或履行客户合同的成本（该成本未资本化为存货，参见注释2(k)(i)）。
若取得合同的增量成本与将在未来报告期间确认的收入相关，且预期可收回，则予以资本化。取得合同的其他成本于发生时费用化。
若履行合同的成本直接与现有合同或具体可识别的预期合同相关；能够产生或增强将用于未来提供商品或服务的资源；且预期可收回，则予以资本化。否则，未予资本化的履行合同成本于发生时费用化。
已资本化的合同成本按成本减累计摊销及减值损失列示。已资本化合同成本的摊销于与该资产相关的收入被确认时在损益中确认（参见注释2(v)(i)）。
APPENDIX I ACCOUNTANTS' REPORT
合同资产在集团确认收入（见注释2(v)(i)）但尚未根据合同条款无条件获得对价权利时予以确认。合同资产须进行预期信用损失评估（见注释2(j)(i)），并在对价收取权利变为无条件时重新分类为应收款项（见注释2(m)）。
合同负债在客户于集团确认相关收入（见注释2(v)(i)）之前支付不可退还对价时予以确认。若集团在确认相关收入之前已拥有收取不可退还对价的无条件权利，亦应确认合同负债。在后一种情况下，还应同时确认相应的应收款项（见注释2(m)）。
APPENDIX I ACCOUNTANTS' REPORT
当集团拥有收取对价的无条件权利，且仅需等待时间流逝即可收到相应款项时，应确认应收款项。
不含重大融资成分的贸易应收款项按其交易价格进行初始计量。含有重大融资成分的贸易应收款项及其他应收款项按公允价值加交易成本进行初始计量。所有应收款项其后均以摊余成本列示（见注释2(j)(i)）。
现金及现金等价物包括银行及手头现金、银行及其他金融机构的活期存款，以及可随时变现为已知金额现金、价值变动风险极低且于购入时距到期日在三个月以内的短期高流动性投资。现金及现金等价物须进行预期信用损失评估（见注释2(j)(i)）。
贸易及其他应付款项初始按公允价值确认。初始确认后，贸易及其他应付款项采用实际利率法以摊余成本列示，但若折现影响不重大，则按发票金额列示。
计息借款初始按公允价值减交易成本计量。其后，该等借款采用实际利率法以摊余成本列示。借款成本于发生当期计入损益。
若合同载有本公司或集团以现金或其他金融资产回购自身权益工具的义务，即产生金融负债，即使本公司或集团的回购义务须待交易对手行使赎回权后方可触发亦然。
向投资者发行的金融工具初始确认后，持续按赎回金额的现值计量，该现值代表由产生最高结算金额的事件所触发的结算金额。金融负债账面金额的变动在损益中确认为"向投资者发行金融工具账面金额的变动"。
向投资者发行金融工具的账面金额于交易对手赎回权终止时重新分类至股份溢价。
集团于初始确认时将可转换债券指定为以公允价值计量且其变动计入损益的金融负债。初始确认后，可转换债券按公允价值列示，公允价值变动在损益中确认为"以公允价值计量且其变动计入损益的金融工具公允价值变动"。
APPENDIX I ACCOUNTANTS' REPORT
短期雇员福利于相关服务提供时计入损益。若集团因雇员过去提供的服务而承担现时法定或推定义务须支付相关金额，且该义务能够可靠估计，则应就预期应付金额确认负债。
向设定缴款退休计划缴款的义务于相关服务提供时计入损益。
授予雇员的以权益结算的以股份为基础付款，其授予日公允价值采用二项式模型计量。该金额通常在权益奖励的归属期内确认为费用，并相应增加权益。确认为费用的金额经调整以反映预计将满足相关服务条件的权益奖励，使最终确认金额以归属日满足相关服务条件的权益奖励为基础。权益金额在以股份为基础付款储备中确认，直至权益奖励被行使（届时计入实收资本所确认金额）或权益奖励届期失效（届时直接转入留存收益）。
如对以权益结算的以股份为基础付款的条款进行修改，首先按条款未经修改的情况确认费用。此外，对于增加以股份为基础付款总公允价值或在修改日计量对雇员有利的任何修改，均应确认相应费用。
离职福利在集团不再能够撤回该等福利的要约与集团确认重组成本两者中较早发生时计入损益。
所得税费用包括当期税和递延税。除与企业合并相关或直接在权益或其他综合收益中确认的项目外，所得税费用在损益中确认。
当期税包括对本年度应税收入或亏损估计应缴纳或应退还的税款，以及对以前年度应缴或应退税款的任何调整。应缴或应退当期税款的金额是预期将缴纳或收到的税款金额的最佳估计，反映了与所得税相关的任何不确定性。当期税按报告日已颁布或实质性颁布的税率计量。当期税还包括因股息产生的任何税款。
仅在符合特定条件时，当期税资产和负债才可相互抵销。
• 初始确认商誉时产生的应税暂时性差异。
本集团就其租赁负债和使用权资产分别确认递延税资产和递延税负债。
在未来很可能获得足够应税利润以供抵扣的范围内，对未使用的税务亏损、未使用的税收抵免及可抵扣暂时性差异确认递延税资产。未来应税利润根据相关应税暂时性差异的转回确定。若应税暂时性差异金额不足以全额确认递延税资产，则根据本集团各子公司的业务计划，考虑经现有暂时性差异转回调整后的未来应税利润。递延税资产于每个报告日进行审查，并在相关税收利益不再很可能实现时予以减少；当未来应税利润的可能性改善时，上述减少额予以转回。
仅在符合特定条件时，递延税资产和负债才可相互抵销。
一般而言，准备金通过按税前折现率对预期未来现金流量进行折现来确定，该折现率反映当前市场对货币时间价值的评估及该负债特有的风险。
保修准备金在相关产品或服务销售时予以确认，依据历史保修数据及各可能结果与相关概率的加权计算确定。
亏损合同准备金按终止合同的预期成本与继续履行合同的预期净成本两者中较低者的现值计量，该成本根据履行该合同义务的增量成本及直接与履行该合同相关的其他成本的分摊确定。在确立准备金之前，本集团就与该合同相关的资产确认任何减值损失（见注2(j)(ii)）。
在经济利益流出不太可能发生或金额无法可靠估计的情况下，该义务作为或有负债披露，除非经济利益流出的可能性极小。可能的义务，其存在仅能通过一项或多项未来事项的发生或不发生加以证实，亦作为或有负债披露，除非经济利益流出的可能性极小。
当收入来源于本集团日常业务中商品销售或服务提供时，本集团将其归类为收入。
收入于产品或服务的控制权转移给客户时确认，金额为本集团预期有权收取的已承诺代价，不包括代第三方收取的款项（如增值税或其他销售税）。
在判断本集团是作为委托人还是代理人时，本集团考虑其是否在产品或服务转移给客户之前已取得对该产品或服务的控制权。控制权是指本集团主导该产品或服务的使用并从中获取实质上全部剩余利益的能力。
本集团主要从事大模型相关服务的提供，包括本地化部署和云端部署。
(a) 本地部署 本集团的本地部署主要包括大型模型的本地化部署及促进此类部署所需的所有现场服务。收入于大型模型及相关服务交付至客户指定地点并经客户验收时按时点确认。
本集团亦提供其他相关服务，如模型训练及微调。来自该等服务的收入于控制权转移时确认，视安排的性质，可按时间段或时点确认。
(b) 云端部署 本集团的云端部署通过云基础设施提供。收入在合同期内确认。对于基于订阅的合同，收入一般在合同期内按比例确认。对于基于使用量的合同，收入根据客户在向其提供服务时对资源的实际使用量确认。
(a) 股息 股息收入于本集团收取款项的权利确立之日在损益中确认。
(b) 利息收入 利息收入采用实际利率法确认。实际利率是将金融资产预期存续期内的预计未来现金收入精确折现至金融资产账面总值的利率。在计算利息收入时，实际利率适用于资产的账面总值（当资产未出现信用减值时）。然而，对于初始确认后出现信用减值的金融资产，利息收入通过将实际利率适用于金融资产的摊余成本计算得出。若资产不再出现信用减值，则利息收入的计算恢复为总额基础。
(c) 政府补助 补偿本集团已发生费用的政府补助在财务状况表中确认为递延收益，并在发生相关费用的同一期间内，以系统性方式在损益中作为相关费用的减项确认。
以外币进行的交易按交易日的汇率折算为集团各公司各自的功能货币。
以外币计量的货币性资产及负债按报告日的汇率折算为功能货币。以外币按公允价值计量的非货币性资产及负债按确定公允价值时的汇率折算为功能货币。以外币按历史成本计量的非货币性资产及负债按交易日的汇率折算。外币差异一般在损益中确认。
APPENDIX I ACCOUNTANTS' REPORT
境外经营的资产及负债按报告日的汇率折算为本集团的列报货币人民币。境外经营的收入及费用按交易日的汇率折算为人民币。外币差异在其他综合收益中确认并累计于汇兑储备，但分配至非控制性权益的折算差异除外。
(iii) 为本集团或本集团母公司关键管理人员的成员。
(i) 该实体与本集团同属一个集团的成员。
(ii) 一方实体是另一方实体的联营公司或合营企业（或该另一方实体所属集团某成员的联营公司或合营企业）。
(iii) 两个实体均为同一第三方的合营企业。
(iv) 一方实体是某第三方实体的合营企业，而另一方实体是该第三方实体的联营公司。
(v) 该实体是为本集团或与本集团存在关联关系的实体的雇员设立的离职后福利计划。
(vi) 该实体受(a)项所述人员控制或共同控制。
(vii) (a)(i)项所述人员对该实体具有重大影响，或为该实体（或该实体母公司）关键管理人员的成员。
(viii) 该实体或其所属集团的任何成员向本集团或本集团母公司提供关键管理人员服务。
某人家庭的近亲成员，是指在与实体打交道时可能影响该人或受该人影响的家庭成员。
历史财务信息中各经营分部及各分部项目的金额，均从定期提供给本集团最高级别行政管理层的财务信息中识别，用于向本集团各业务线及各地理位置分配资源及评估其业绩。
除非各分部具有相似的经济特征，且在产品及服务的性质、客户的类型或级别、产品分销或服务提供所采用的方式以及监管环境的性质方面相似，否则各重大经营分部不予合并列报。非单独重大的经营分部如符合上述大部分标准，则可予以合并。
APPENDIX I ACCOUNTANTS' REPORT
Notes 13, 27, 29 and 31(c) contain information about the assumptions and their risk factors relating to goodwill impairment, fair value of convertible bonds, share-based payments and financial instruments. Other significant sources of estimation uncertainty are as follows:
If circumstances indicate that the carrying amount of a non-current non-financial asset may not be recoverable, the asset may be considered "impaired", and an impairment loss may be recognised in accordance with accounting policy for impairment of non-current assets as described in Note 2(j)(ii). These assets are tested for impairment periodically or whenever the events or changes in circumstances indicate that their recorded carrying amounts may not be recoverable. When such a decline has occurred, the carrying amount is reduced to recoverable amount. The recoverable amount is the greater of the fair value less costs of disposal and value in use. In determining the value in use, expected future cash flows generated by the asset are discounted to their present value, which requires significant judgement relating to the level of revenue and amount of operating costs. The Group uses all readily available information in determining an amount that is a reasonable approximation of the recoverable amount, including estimates based on reasonable and supportable assumptions and projections of the level of revenue and amount of operating costs. Changes in these estimates could have a significant impact on the recoverable amount of the assets and could result in additional impairment charge or reversal of impairment in future periods.
The principal activities of the Group are the provision of large model-related services in the PRC.
Disaggregation of revenue from contracts with customers by major service types and timing of revenue recognition are as follows:
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | **Revenue from contracts with customers within the scope of IFRS 15** | | | | | | | Disaggregated by major service types: | | | | | | | - On-premise deployment | 54,815 | 112,614 | 263,930 | 26,806 | 161,777 | | - Cloud-based deployment | 2,594 | 11,924 | 48,484 | 18,103 | 29,100 | | **Total** | **57,409** | **124,538** | **312,414** | **44,909** | **190,877** |
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | **Timing of revenue recognition** | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | At a point in time | 53,423 | 88,937 | 250,521 | 13,397 | 155,636 | | Over time | 3,986 | 35,601 | 61,893 | 31,512 | 35,241 | | **Total** | **57,409** | **124,538** | **312,414** | **44,909** | **190,877** |
During the Track Record Period, the Group's customers with whom transactions have exceeded 10% of the Group's revenue in the respective years/periods are as follows:
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Customer A | * | 18,244 | 59,465 | 13,409 | 20,977 | | Customer B | * | 17,536 | * | * | - | | Customer C | * | 16,810 | | | | | Customer D | 8,850 | 14,875 | | | | | Customer E | 7,925 | * | | | | | Customer F | 6,545 | * | | | | | Customer G | - | | | | | | Customer H | | | | | | | Customer I | | | | | | | Customer J | | | | | |
Details of concentrations of credit risk are set out in Note 31(a).
**(ii) Revenue expected to be recognised in the future arising from contracts with customers in existence at the reporting date**
The Group applies the practical expedient in paragraph 121(a) of IFRS 15 of not disclosing the transaction price allocated to the remaining performance obligation as the original expected duration of the Group's contracts are one year or less.
The Group manages its businesses by service types. In a manner consistent with the way in which information is reported internally to the Group's chief operating decision maker for the purposes of resource allocation and performance assessment, the Group has presented the following two reportable segments. No operating segments have been aggregated to form the following reportable segments.
- **On-premise deployment:** this segment develops and provides customised large model-related services according to the customers' specific instructions and needs at the customers' infrastructure.
- **Cloud-based deployment:** this segment develops and provides cloud-based large model-related services to customers through cloud infrastructure.
The Group participates in defined contribution retirement benefit schemes for all qualified employees in the PRC. During the Track Record Period, the contributions made by the Group to the defined contribution retirement benefit schemes are charged to profit or loss as incurred.
For the purposes of assessing segment performance and allocating resources, the Group's most senior executive management monitors the results attributable to each reportable segment on the following bases:
Revenue and expenses are allocated to the reportable segments with reference to revenue generated by those segments and direct expenses incurred by those segments. The measure used for reporting segment result is gross profit. No inter-segment sales have occurred during the Track Record Period. Assistance provided by one segment to another, including sharing of assets and technical know-how, is not measured.
The Group's other operating income and expenses, such as other income, selling and marketing expenses, general and administrative expenses, research and development expenses, impairment losses on financial assets, finance costs, and assets and liabilities are not measured under individual segments. Accordingly, neither information on segment assets and liabilities nor information concerning capital expenditure, interest income and interest expenses is presented.
Information regarding the Group's reportable segments as provided to the Group's most senior executive management for the purposes of resource allocation and assessment of segment performance during the Track Record Period is set out below.
| | On-premise deployment RMB'000 | Cloud-based deployment RMB'000 | Total RMB'000 | |---|---|---|---| | Segment revenue derived from external customers | 54,815 | 2,594 | 57,409 | | Segment gross profit | 29,386 | 1,974 | 31,360 |
| | On-premise deployment RMB'000 | Cloud-based deployment RMB'000 | Total RMB'000 | |---|---|---|---| | Segment revenue derived from external customers | 112,614 | 11,924 | 124,538 | | Segment gross profit | 76,781 | 3,701 | 80,482 |
| | On-premise deployment RMB'000 | Cloud-based deployment RMB'000 | Total RMB'000 | |---|---|---|---| | Segment revenue derived from external customers | 263,930 | 48,484 | 312,414 | | Segment gross profit | 174,256 | 1,633 | 175,889 |
| | On-premise deployment RMB'000 | Cloud-based deployment RMB'000 | Total RMB'000 | |---|---|---|---| | Segment revenue derived from external customers | 161,777 | 29,100 | 190,877 | | Segment gross profit/(loss) | 95,540 | (116) | 95,424 |
| | On-premise deployment RMB'000 | Cloud-based deployment RMB'000 | Total RMB'000 | |---|---|---|---| | Segment revenue derived from external customers | 26,806 | 18,103 | 44,909 | | Segment gross profit | 16,776 | 5,183 | 21,959 |
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Reportable segment gross profit | 31,360 | 80,482 | 175,889 | 21,959 | 95,424 | | Other income | 1,784 | 9,965 | 19,281 | 4,174 | 4,614 | | Selling and marketing expenses | (15,139) | (101,198) | (387,475) | (144,194) | (208,570) | | General and administration expenses | (32,316) | (66,302) | (133,603) | (51,452) | (185,165) | | Research and development expenses | (84,377) | (528,884) | (2,195,436) | (859,217) | (1,594,661) | | Impairment losses on financial assets | (31) | (19,786) | (17,008) | (763) | (10,867) | | Finance costs | (5,694) | (26,332) | (38,321) | (12,212) | (53,270) | | Share of profits less losses of associates | - | (453) | 21,254 | 324 | 14,147 | | Changes in fair value of financial instruments measured at FVPL | 5,972 | 26,022 | 66,271 | 7,004 | 9,791 | | Changes in the carrying amounts of financial instruments issued to investors | (45,209) | (161,471) | (468,859) | (201,174) | (429,295) | | Consolidated loss before taxation | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) |
The Group's revenue is substantially derived from customers located in the PRC, and all of the Group's non-current assets are located or allocated to operations located in the PRC. Accordingly, no segment analysis based on geographical locations is provided.
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Interest income | 1,426 | 11,236 | 13,406 | 4,163 | 4,607 | | Net (loss)/gain on disposal of property and equipment and intangible assets | 358 | (1,539) | 6,807 | 11 | (15) | | Others | - | 268 | (932) | - | 22 | | | 1,784 | 9,965 | 19,281 | 4,174 | 4,614 |
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Interest on lease liabilities (Note 11(c)) | 788 | 7,498 | 30,931 | 11,840 | 17,983 | | Interest on bank loans | - | - | 1,716 | - | 1,807 | | Transaction costs on issuance of financial instruments to investors | 4,906 | 16,488 | 20,119 | 5,185 | 31,975 | | Foreign currency exchange loss/(gain), net | - | 2,346 | (14,445) | (4,813) | 1,505 | | | 5,694 | 26,332 | 38,321 | 12,212 | 53,270 |
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Salaries, wages and other benefits | 82,198 | 195,117 | 512,176 | 171,663 | 318,624 | | Contributions to defined contribution retirement benefit schemes (i) | 7,499 | 16,666 | 37,466 | 16,474 | 22,118 | | Equity-settled share-based compensation expenses (excluding expenses of nil, nil, nil, nil and RMB2,890,000 capitalised as inventories respectively) (Note 29) | 1,024 | 5,502 | 23,579 | 4,217 | 158,852 | | | 90,721 | 217,285 | 573,221 | 192,354 | 499,594 |
Note: (i) The Group participates in defined contribution retirement benefit schemes for all qualified employees in the PRC. During the Track Record Period, the contributions made by the Group to the defined contribution retirement benefit schemes are charged to profit or loss as incurred.
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Auditors' remuneration: | | | | | | | – Audit services | 900 | 1,600 | 2,200 | 1,200 | 1,200 | | – Non-audit services | 350 | 350 | 350 | 350 | 350 | | Depreciation of property and equipment (Note 10(a)) | 4,101 | 9,768 | 29,009 | 10,498 | 23,469 | | Depreciation of right-of-use assets (Note 11(a)) | 4,016 | 18,396 | 59,455 | 20,944 | 48,101 | | Amortisation of intangible assets (Note 12) | 1,697 | 11,188 | 45,124 | 20,040 | 25,380 |
The employees of the Company and its subsidiaries established in the Chinese Mainland participate in defined contribution retirement benefit schemes managed by the respective local governments, whereby the Company and these subsidiaries are required to contribute to the schemes at specified percentages of the employees' average salaries during the Track Record Period. Employees of the Company and these subsidiaries are entitled to receive retirement benefits, calculated based on a percentage of the average salaries level in the Chinese Mainland, from the above-mentioned retirement schemes at their normal retirement age. The Group also operates a Mandatory Provident Fund Scheme (the "MPF Scheme") under the Hong Kong Mandatory Provident Fund Scheme Ordinance for employees under the jurisdiction of the Hong Kong Employment Ordinance. The MPF Scheme is a defined contribution retirement plan administered by an independent trustee. Under the MPF Scheme, the employer and its employees are each required to make contributions to the plan at 5% of the employees' relevant salaries, subject to a cap of monthly relevant salaries of HK$30,000. Contributions to the MPF Scheme vest immediately. The Group has no further obligation for payment of other retirement benefits beyond the above contributions.
| | Note | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---|---| | | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 | 2025 RMB'000 (unaudited) | | Depreciation on property and equipment (excluding expenses of nil, nil, nil, nil and RMB3,311,000 capitalised as inventories respectively) | 11(a) | 16,567 | 63,822 | 270,252 | 131,105 | 132,908 | | Amortisation of intangible assets | 12(a) | 3,451 | 5,109 | 9,685 | 4,510 | 6,128 | | Listing expense | | - | - | - | - | 17,731 |
Income tax in the consolidated statements of profit or loss and other comprehensive income Reconciliations between income tax expenses and accounting losses at applicable tax rates:
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 | 2025 RMB'000 (unaudited) | | Loss before taxation | (143,650) | (787,957) | (2,958,007) | (1,235,551) | (2,357,852) | | Tax on loss before taxation, calculated at the rates applicable to profits in the jurisdictions concerned (i) | (35,912) | (196,989) | (739,502) | (308,888) | (589,463) | | Tax rates differentials (ii) | 13,507 | 79,736 | 290,253 | 122,511 | 221,902 | | Tax effect of additional deduction on research and development expenses (iii) | (8,406) | (41,757) | (305,408) | (119,645) | (239,946) | | Tax effect of non-deductible expenses (iv) | 7,200 | 31,978 | 83,782 | 34,123 | 96,960 | | Tax effect of unrecognised unused tax losses and deductible temporary differences | 23,611 | 127,032 | 670,875 | 271,899 | 510,547 | | | - | - | - | - | - |
Entities of the Group established in the Chinese Mainland were subject to the PRC Corporate Income Tax rate of 25% during the Track Record Period. Taxation for subsidiaries incorporated in other jurisdictions is calculated at the applicable income tax rates in the relevant jurisdictions.
Certain subsidiaries of the Group obtained the certificates of "High and New Technology Enterprise" ("HNTE") from the tax authorities and were subject to a preferential tax rate of 15% during the Track Record Period.
An additional 75% of qualified research and development expenses incurred is allowed to be deducted from taxable income under the PRC Corporate Income Tax laws and regulations before 1 October 2022. An additional 100% of qualified research and development expenses incurred is allowed to be deducted from taxable income under the PRC Corporate Income Tax laws and regulations after 1 October 2022.
Tax effect of non-deductible expenses mainly represented the changes in the carrying amounts of financial instruments issued to investors and share-based payments expenses, which are not deductible in accordance with the relevant tax regulations in the PRC.
Directors' and supervisors' emoluments Directors' and supervisors' emoluments during the Track Record Period are as follows:
| | Directors' fee RMB'000 | Salaries, allowances and other benefits RMB'000 | Discretionary bonuses RMB'000 | Retirement scheme contributions RMB'000 | Sub-total RMB'000 | Share-based payments RMB'000 (Note (i)) | Total RMB'000 | |---|---|---|---|---|---|---|---| | **Directors** | | | | | | | | | Dr. Liu Debing | - | 662 | 1,259 | 58 | 1,979 | 426 | 2,405 | | Dr. Zhang Peng | - | 895 | 257 | 56 | 1,208 | 353 | 1,561 | | Dr. Li Juanzi | - | 94 | 334 | 58 | 94 | - | 94 | | Dr. Tang Jie | - | 574 | - | - | 966 | - | 966 | | Mr. Wang Shaolan | - | - | - | - | - | - | - | | Mr. Li Jiaqing (appointed on 21 January 2022) | - | - | - | - | - | - | - | | Mr. Xiang Xiaobo | - | - | - | - | - | - | - | | Dr. Xu Bin (redesignated in on 15 February 2022) | - | - | - | - | - | - | - | | Mr. Mi Lei | - | - | - | - | - | - | - | | Mr. Zhou Zhifeng (appointed on 15 February 2022) | - | - | - | - | - | - | - | | Mr. Wu Xi | - | - | - | - | - | - | - | | **Supervisor** | | | | | | | | | Dr. Xu Bin (redesignated out on 15 February 2022) | - | 2,225 | 1,850 | 172 | 4,247 | 779 | 5,026 |
| | Directors' fee RMB'000 | Salaries, allowances and other benefits RMB'000 | Discretionary bonuses RMB'000 | Retirement scheme contributions RMB'000 | Sub-total RMB'000 | Share-based payments RMB'000 (Note (i)) | Total RMB'000 | |---|---|---|---|---|---|---|---| | **Directors** | | | | | | | | | Dr. Liu Debing | - | 694 | 545 | 63 | 1,302 | 1,261 | 2,563 | | Dr. Zhang Peng | - | 891 | 220 | 63 | 1,174 | 388 | 1,562 | | Dr. Li Juanzi | - | 10 | 161 | 63 | 10 | - | 10 | | Dr. Tang Jie | - | 579 | - | - | 803 | - | 803 | | Mr. Wang Shaolan | - | - | - | - | - | - | - | | Mr. Li Jiaqing | - | - | - | - | - | - | - | | Mr. Xiang Xiaobo | - | - | - | - | - | - | - | | Mr. Xu Bin (resigned on 8 March 2023) | - | 1 | - | - | 1 | - | 1 | | Mr. Mi Lei | - | - | - | - | - | - | - | | Mr. Zhou Zhifeng (resigned on 8 March 2023) | - | - | - | - | - | - | - | | Mr. Wu Xi (resigned on 8 March 2023) | - | - | - | - | - | - | - | | Mr. Zhang Haifeng (appointed on 8 March 2023) | - | - | - | - | - | - | - | | Mr. Wang Meng (appointed on 4 August 2023) | - | - | - | - | - | - | - | | **Supervisor** | | | | | | | | | Mr. Yan Xingyu (appointed on 8 March 2023) | - | 749 | 248 | 63 | 1,060 | - | 1,060 | | | - | 2,924 | 1,174 | 252 | 4,350 | 1,649 | 5,999 |
Directors Dr. Liu Debing Dr. Zhang Peng Dr. Li Juanzi Dr. Tang Jie Mr. Wang Shaolan Mr. Li Jiaqing Mr. Xiang Xiaobo Mr. Zhang Haifeng Mr. Wang Meng
Six months ended 30 June 2025 Salaries, allowances Directors' and other Discretionary fee benefits bonuses RMB'000 RMB'000 RMB'000
Retirement scheme Share-based contributions Sub-total payments Total RMB'000 RMB'000 RMB'000 RMB'000 (Note (i))
Directors Dr. Liu Debing Dr. Zhang Peng Dr. Li Juanzi (resigned on 26 March 2025 and reappointed on 28 June 2025) Dr. Tang Jie (resigned on 28 June 2025) Mr. Wang Shaolan (resigned on 26 March 2025) Ms. Zhang Xiaohan (appointed on 26 March 2025) Mr. Li Jiaqing Mr. Xiang Xiaobo (resigned on 26 March 2025) Mr. Zhang Haifeng (resigned on 26 March 2025) Mr. Wang Meng Supervisors Mr. Pei Bo (appointed on 26 March 2025) Mr. Yan Xingyu (resigned on 26 March 2025)
Directors Dr. Liu Debing Dr. Zhang Peng Dr. Li Juanzi Dr. Tang Jie Mr. Wang Shaolan Mr. Li Jiaqing Mr. Xiang Xiaobo Mr. Zhang Haifeng Mr. Wang Meng
These represent the estimated value of equity awards granted to the directors under the Group's equity award schemes. The value of these equity awards is measured according to the Group's accounting policies for share-based payment transactions. The details of these benefits in kind, including the principal terms and equity awards granted, are disclosed in Note 29.
The number of directors and non-directors included in the five highest paid individuals during the Track Record Period are set out below: Years ended 31 December
The emoluments of the directors are disclosed in Note 8. The emoluments of the individuals who are not directors and who are amongst the five highest paid individuals of the Group are set out below: Years ended 31 December
Salaries, wages and other benefits Discretionary bonuses Contributions to defined contribution retirement plans Equity-settled share-based compensation expenses
The number of the individuals who are not directors and who are amongst the five highest paid individuals of the Group are within the following bands: Years ended 31 December
Hong Kong Dollars ("HK$") HK$1,000,001 - HK$1,500,000 HK$1,500,001 - HK$2,000,000 HK$2,000,001 - HK$2,500,000 HK$2,500,001 - HK$3,000,000 HK$3,000,001 - HK$3,500,000 HK$3,500,001 - HK$4,000,000 HK$4,000,001 - HK$4,500,000 HK$4,500,001 - HK$5,000,000 HK$5,000,001 - HK$5,500,000 HK$8,500,001 - HK$9,000,000 HK$9,500,001 - HK$10,000,000 HK$12,500,001 - HK$13,000,000 HK$14,000,001 - HK$14,500,000 HK$33,500,001 - HK$34,000,000
Basic loss per share is calculated by dividing the loss attributable to ordinary equity shareholders of the Company by the weighted average number of ordinary shares in issue or deemed to be in issue during the Track Record Period. As described in Note 30(b), the Company was converted into a joint stock company with limited liability on 26 March 2025. The Company's paid-in capital of RMB36,224,375 was converted into 36,224,375 shares of RMB1.00 each accordingly. For the purpose of computing basic and diluted loss per share, the weighted average number of ordinary shares deemed to be in issue before the Company's conversion into a joint stock company was determined assuming the conversion into joint stock company
had occurred since 1 January 2022, at the exchange ratio established in the conversion in March 2025 and does not take into account the effect of the share subdivision plan detailed in Note 30(b). Years ended 31 December
Loss for the year/period attributable to ordinary equity shareholders of the Company (RMB'000) (Note 10(a))
Loss for the year/period attributable to ordinary equity shareholders of the Company Years ended 31 December 2022 RMB'000
Loss for the year/period attributable to all equity shareholders of the Company Allocation of loss for the year/period attributable to financial instruments issued to investors Loss for the year/period attributable to ordinary equity shareholders of the Company
Ordinary shares deemed to be in issue at 1 January Effect of ordinary shares deemed to be in issue Effect of increase in paid-in capital through transfer from capital reserve (Note 30) Effect of the financial instruments issued to investors (Note 26)
During the Track Record Period, financial instruments issued to investors (Note 26) and convertible bonds (Note 27) were not included in the calculation of diluted loss per share as their inclusion would have been anti-dilutive. Accordingly, diluted loss per share for the years ended 31 December 2022, 2023 and 2024 and the six months ended 30 June 2024 and 2025 are the same as basic loss per share for the respective years/periods. 11
Accumulated depreciation: At 1 January 2024 Charge for the year Transfer (Note) Written back on disposals
Note: The Group entered into a financing arrangement with a third party, under which the Group transferred the legal ownership of the electronic equipment to this third party while continuing to use these equipment under lease for a term of four years.
Accumulated depreciation: At 1 January 2024 Charge for the year Transfer (Note) Written back on disposals
| | Electronic equipment and others RMB'000 | Leasehold improvements RMB'000 | Right-of-use assets RMB'000 | Total RMB'000 | |---|---|---|---|---| | **Cost:** | | | | | | At 1 January 2025 | 286,937 | 19,497 | 819,455 | 1,125,889 | | Additions | 4,332 | 385 | 24,908 | 29,625 | | Disposals | (32) | - | - | (32) | | At 30 June 2025 | 291,237 | 19,882 | 844,363 | 1,155,482 | | **Accumulated depreciation:** | | | | | | At 1 January 2025 | (66,552) | (5,285) | (210,029) | (281,866) | | Charge for the period | (29,867) | (2,236) | (99,385) | (131,488) | | Written back on disposals | 12 | - | - | 12 | | At 30 June 2025 | (96,407) | (7,521) | (309,414) | (413,342) | | **Carrying amount:** | | | | | | At 30 June 2025 | 194,830 | 12,361 | 534,949 | 742,140 |
Note: The Group entered into a financing arrangement with a third party, under which the Group transferred the legal ownership of the electronic equipment to this third party while continuing to use these equipment under lease for a term of four years.
The analyses of the carrying amounts of right-of-use assets by class of underlying assets are as follows:
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Assets leased for own use, carried at depreciated cost: | | | | | | - Office premises | 11,971 | 172,967 | 140,088 | 151,536 | | - Electronic equipment | - | 84,615 | 483,737 | 403,766 | | | 11,971 | 257,582 | 623,825 | 555,302 |
The analyses of expense items in relation to leases recognised in the Group's consolidated statements of profit or loss are as follows:
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Depreciation on right-of-use assets has been charged to the consolidated statements of profit or loss as follows: | | | | | | | - Office premises | 11,971 | 41,026 | 42,808 | 21,238 | 19,868 | | - Electronic equipment | - | 3,542 | 133,444 | 53,032 | 79,970 | | Depreciation charge of right-of-use assets (Note 11(a)) | 11,971 | 44,568 | 176,252 | 74,270 | 99,838 | | Interest on lease liabilities (Note 6(a)) | 788 | 7,498 | 30,931 | 11,840 | 17,983 | | Expenses relating to short-term leases | 2,383 | 5,537 | 9,070 | 4,211 | 5,909 |
The Group leases office premises and electronic equipment under leases expiring from 1 to 5 years. Some leases include an option to renew when all terms are renegotiated. None of the leases includes variable lease payments.
The total cash outflow for leases and the maturity analyses of lease liabilities are set out in Note 21(d) and Note 25, respectively.
In accordance with IAS 36, the Group performed impairment tests at the end of each reporting period on non-current assets, primarily including property, plant and equipment, right-of-use assets, intangible assets and other non-current assets at the CGU level. The recoverable amounts of these assets exceeded their respective carrying amounts at the end of each reporting period, therefore, no impairment loss was recognized during the Track Record Period.
| | Patents RMB'000 | Software RMB'000 | Total RMB'000 | |---|---|---|---| | **Cost:** | | | | | At 1 January 2022 and 31 December 2022 | 27,139 | 269 | 27,408 | | **Accumulated amortisation:** | | | | | At 1 January 2022 | (2,884) | (46) | (2,930) | | Charge for the year | (3,392) | (59) | (3,451) | | At 31 December 2022 | (6,276) | (105) | (6,381) | | **Carrying amount:** | | | | | At 31 December 2022 | 20,863 | 164 | 21,027 |
| | Patents RMB'000 | Software RMB'000 | Total RMB'000 | |---|---|---|---| | **Cost:** | | | | | At 1 January 2023 | 27,139 | 269 | 27,408 | | Additions | - | 1,809 | 1,809 | | Acquisitions through business combinations (Note 32) | 36,846 | - | 36,846 | | At 31 December 2023 | 63,985 | 2,078 | 66,063 | | **Accumulated amortisation:** | | | | | At 1 January 2023 | (6,276) | (105) | (6,381) | | Charge for the year | (5,023) | (86) | (5,109) | | At 31 December 2023 | (11,299) | (191) | (11,490) | | **Carrying amount:** | | | | | At 31 December 2023 | 52,686 | 1,887 | 54,573 | | **Cost:** | | | | | At 1 January 2024 | 63,985 | 2,078 | 66,063 | | Additions | - | 6,802 | 6,802 | | Disposals | - | (1,699) | (1,699) | | At 31 December 2024 | 63,985 | 7,181 | 71,166 | | **Accumulated amortisation:** | | | | | At 1 January 2024 | (11,299) | (191) | (11,490) | | Charge for the year | (8,275) | (1,410) | (9,685) | | Written back on disposals | - | 368 | 368 | | At 31 December 2024 | (19,574) | (1,233) | (20,807) | | **Carrying amount:** | | | | | At 31 December 2024 | 44,411 | 5,948 | 50,359 | | **Cost:** | | | | | At 1 January 2025 | 63,985 | 7,181 | 71,166 | | Additions | 10,049 | 1,119 | 11,168 | | At 30 June 2025 | 74,034 | 8,300 | 82,334 | | **Accumulated amortisation:** | | | | | At 1 January 2025 | (19,574) | (1,233) | (20,807) | | Charge for the period | (5,296) | (832) | (6,128) | | At 30 June 2025 | (24,870) | (2,065) | (26,935) | | **Carrying amount:** | | | | | At 30 June 2025 | 49,164 | 6,235 | 55,399 |
| | Patents RMB'000 | Software RMB'000 | Total RMB'000 | |---|---|---|---| | **Cost:** | | | | | At 1 January 2022 and 31 December 2022 | 24,365 | 269 | 24,634 | | **Accumulated amortisation:** | | | | | At 1 January 2022 | (2,451) | (46) | (2,497) | | Charge for the year | (3,046) | (58) | (3,104) | | At 31 December 2022 | (5,497) | (104) | (5,601) | | **Carrying amount:** | | | | | At 31 December 2022 | 18,868 | 165 | 19,033 | | **Cost:** | | | | | At 1 January 2023 | 24,365 | 269 | 24,634 | | Additions | - | 1,809 | 1,809 | | At 31 December 2023 | 24,365 | 2,078 | 26,443 | | **Accumulated amortisation:** | | | | | At 1 January 2023 | (5,497) | (104) | (5,601) | | Charge for the year | (3,046) | (86) | (3,132) | | At 31 December 2023 | (8,543) | (190) | (8,733) | | **Carrying amount:** | | | | | At 31 December 2023 | 15,822 | 1,888 | 17,710 | | **Cost:** | | | | | At 1 January 2024 | 24,365 | 2,078 | 26,443 | | Additions | - | 6,802 | 6,802 | | Disposal | - | (1,699) | (1,699) | | At 31 December 2024 | 24,365 | 7,181 | 31,546 | | **Accumulated amortisation:** | | | | | At 1 January 2024 | (8,543) | (190) | (8,733) | | Charge for the year | (3,046) | (1,409) | (4,455) | | Written back on disposals | - | 368 | 368 | | At 31 December 2024 | (11,589) | (1,231) | (12,820) | | **Carrying amount:** | | | | | At 31 December 2024 | 12,776 | 5,950 | — |
Goodwill arose from the Group's acquisition of Beijing Lingxin Intelligent in 2023 (see Note 32). The goodwill has been allocated to the Beijing Lingxin Intelligent CGU. The recoverable amount of the Beijing Lingxin Intelligent CGU has been determined based on value in use calculation, determined by discounting the future cash flows to be generated from the continuing operation of the Beijing Lingxin Intelligent CGU with reference to valuation reports issued by an independent valuer. These calculations use cash flow projections based on financial budgets approved by management covering an eight-year period. Management adopted a forecast period of longer than five years in view that the business is still under significant growth and will require additional time for the underlying technology to reach stable status. The key assumptions used in the estimation of the recoverable amounts are as follows: As at 31 December
Annual revenue growth rate (i) Annual gross profit margin (i) Growth rate beyond the forecast period (ii) Pre-tax discount rate (iii) – I-51 –
The annual revenue growth rates and gross profit margins are based on the current operational status and expectations of future changes in the industry and adjusted for other factors that are specific to the CGU.
The growth rate beyond the forecast period is based on relevant industry growth forecasts and does not exceed the average growth rate of the relevant industry.
The pre-tax discount rate reflects specific risks relating to the Beijing Lingxin Intelligent CGU.
The headroom of the Beijing Lingxin Intelligent CGU as at 31 December 2023 and 2024 and 30 June 2025 amounted to RMB16,021,000, RMB22,973,000 and RMB18,011,000, respectively. Management have undertaken sensitivity analysis on the impairment test of goodwill. The following tables set out the hypothetical change to pre-tax discount rate that would have removed the remaining headroom: As at 31 December
As a result of the impairment tests, the Group is of the view that there was no impairment of goodwill as at 31 December 2023 and 2024 and 30 June 2025. Reasonable possible changes in key assumptions would not lead to impairment of the goodwill as at 31 December 2023 and 2024 and 30 June 2025. No impairment of goodwill was recorded during the Track Record Period. 14
Details of the Group's subsidiaries are set out in Note 1. The Group does not have any subsidiary with material NCI during the Track Record Period. 15
At the beginning of the year/period Additions Share of profits less losses of associates Disposal of an associate
The following list contains the particulars of the Group's material associates, all of which are unlisted entities: The Group's effective interest
Beijing Shudao Intelligent Computing Technology Co., Ltd. (北京數道智算 科技有限公司) ("Beijing Shudao")(i)(ii) Beijing Xinglian Dingsen Equity Investment Fund Partnership (Limited Partnership) (北京星連鼎 森股權投資基金合夥企 業(有限合夥)) ("Beijing Xinglian")(i) Beijing Doushen Zhichuang Technology Co., Ltd. (北京豆神智創 科技有限公司)(i) Guangdong Juzhen Zhituo Technology Co., Ltd. (廣東矩陣智拓科技有限 公司)(i)
These entities' official names are in Chinese. The English translations of these entities' names are for identification only.
Although the Group's effective interest in Beijing Shudao is less than 20%, the Group is able to exercise significant influence over Beijing Shudao through representation on its board of directors and participation in its policy-making processes. Therefore, the Company has classified Beijing Shudao as an associate.
Summarised financial information of Beijing Xinglian, which is a material associate of the Group established in 2024 and adjusted for any differences in accounting policies, are disclosed below: As at 31 December 2024
Aggregate information of associates that are not individually material: | | Years ended 31 December | | | Six months ended 30 June 2025 | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | RMB'000 | | Aggregate carrying amount of individually immaterial associates in the Historical Financial Information | - | 13,047 | 38,846 | 47,955 | | Aggregate amounts of the Group's share of those associates' net loss | - | (453) | (1,098) | (5,892) |
The associates of the Group have been accounted for using the equity method based on the financial information of the associates prepared under the accounting policies consistent with the Group.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Investments in equity securities (i) | 11,740 | 58,830 | 42,621 | 49,343 | | Wealth management products (ii) | 20,037 | 100,074 | - | 500,021 | | | 31,777 | 158,904 | 42,621 | 549,364 |
During the Track Record Period, the Company has made investments in unlisted companies specialised in the AI related industries. The directors of the Company consider that the Group has neither significant influence nor control over these investments and designated the investments as equity securities measured at FVPL.
The Group's wealth management products are issued by financial institutions in the PRC with expected rates of return ranging from 1.55% to 3.86%.
For information about the methods and assumptions used in determining the fair value of (i) and (ii) above, see Note 31(c).
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Input VAT deductible | 4,169 | 84,527 | 148,708 | 214,064 | | Prepayments for computing service fee and others | 11,633 | 88,295 | 166,893 | 134,030 | | Contract assets | 403 | 10,006 | 12,477 | 17,220 | | | 16,205 | 182,828 | 328,078 | 365,314 | | Less: current portion | | | | | | - Input VAT deductible (Note 19(a)) | 4,169 | 84,527 | 98,729 | 83,139 | | - Prepayments for computing service fee and others (Note 19(a)) | 11,633 | 88,295 | 127,371 | 72,907 | | - Contract assets (Note 20) | 403 | 9,960 | 4,718 | 6,654 | | | 16,205 | 182,782 | 230,818 | 162,700 | | | - | 46 | 97,260 | 202,614 |
The prepayments for computing service fees represented payments made to suppliers to secure these suppliers' computing services for a certain period, and will be deducted by subsequent utilisation of the computing services from these suppliers.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Input VAT deductible | 4,169 | 84,527 | 140,147 | 199,023 | | Prepayments for computing service fee and others | 11,633 | 88,295 | 161,421 | 133,765 | | Contract assets | 403 | 10,006 | 12,477 | 17,220 | | | 16,205 | 182,828 | 314,045 | 350,008 | | Less: current portion | | | | | | - Input VAT deductible (Note 19(b)) | 4,169 | 84,527 | 90,168 | 68,098 | | - Prepayments for computing service fee and others (Note 19(b)) | 11,633 | 88,295 | 126,239 | 72,642 | | - Contract assets (Note 20) | 403 | 9,960 | 4,718 | 6,654 | | | 16,205 | 182,782 | 221,125 | 147,394 | | | - | 46 | 92,920 | 202,614 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Purchased hardware and components | - | - | 1,218 | 70,046 | | Contract fulfilment costs | 10,939 | 30,082 | 32,979 | - | | | 10,939 | 30,082 | 34,197 | 70,046 | | Less: write-down of inventories | (597) | (1,300) | (1,732) | (2,180) | | | 10,342 | 28,782 | 32,465 | 67,866 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Purchased hardware and components | - | - | 1,218 | 66,478 | | Contract fulfilment costs | 10,874 | 29,992 | 31,601 | - | | | 10,874 | 29,992 | 32,819 | 66,478 | | Less: write-down of inventories | (597) | (1,300) | (1,647) | (2,089) | | | 10,277 | 28,692 | 31,172 | 64,389 |
| | 2022 RMB'000 | As at 31 December 2023 RMB'000 | 2024 RMB'000 | As at 30 June 2025 RMB'000 | |---|---|---|---|---| | Trade receivables | 6,643 | 8,883 | 100,170 | 171,657 | | Less: loss allowance (Note 31(a)) | (29) | (988) | (9,035) | (25,880) | | | 6,614 | 7,895 | 91,135 | 145,777 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Deposits | 3,640 | 16,584 | 67,912 | 70,257 | | Receivables from disposal of investments in equity securities measured at FVPL (i) | - | - | 45,216 | 7,098 | | Receivables of capital contributions from equity shareholders (ii) | 1,833 | 120,328 | - | - | | Other receivables | - | 117,591 | 263,805 | 95,381 | | | 5,473 | 254,503 | 376,933 | 172,736 | | Less: loss allowance (Note 31(a)) | (3) | (18,779) | (27,327) | (21,172) | | | 5,470 | 235,724 | 349,606 | 151,564 | | Financial assets measured at amortised cost | 12,084 | 243,619 | 440,741 | 297,341 | | Input VAT deductible (Note 17(a)) | 4,169 | 84,527 | 98,729 | 83,139 | | Prepayments for computing service fees and others (Note 17(a)) | 11,633 | 88,295 | 127,371 | 72,907 | | | 15,802 | 172,822 | 226,100 | 156,046 | | | 27,886 | 416,441 | 666,841 | 453,387 |
In November 2024, the Group entered into a series of equity transfer agreements with Beijing Xinglian, an associate of the Group, pursuant to which, the Group divested certain unlisted equity investments measured at FVPL to Beijing Xinglian at a total consideration of RMB202,528,000. The consideration was determined based on arm's length negotiation and was with reference to the most recent transaction price of the equity interests in those unlisted entities.
Receivables of capital contributions from equity shareholders had been received by the Group in 2024.
I'll translate the Chinese IPO prospectus text. However, I notice that the text you've provided appears to already be in English. This chunk (21/30) contains English financial statements and tables. There is no Chinese text to translate in this chunk.
The content appears to be from an accountants' report appendix containing: - Trade receivables aging analysis - Contract assets details - Cash at bank information - Time deposits - Reconciliation of financing liabilities
If you have Chinese text that needs to be translated, please provide the Chinese version of this chunk. If this is the correct chunk and you need it reproduced as-is, here it is formatted:
All of the trade and other receivables are expected to be recovered or recognised as expenses within one year.
At the end of each reporting period, the ageing analyses of trade receivable (net of loss allowance), based on the invoice date, is as follows:
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Within 3 months | 6,498 | 5,379 | 74,191 | 76,694 | | 3 months to 6 months | 116 | 1,425 | 13,804 | 28,891 | | 6 months to 1 year | - | 82 | 1,444 | 38,504 | | 1 year to 2 years | - | 1,009 | 1,302 | 1,358 | | 2 years to 3 years | - | - | 394 | 330 | | | 6,614 | 7,895 | 91,135 | 145,777 |
Further details on the Group's credit policy and credit risk are set out in Note 31(a).
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Trade receivables due from: | | | | | | - third parties | 6,643 | 8,883 | 98,159 | 145,198 | | - subsidiaries | - | - | 11,412 | 38,964 | | | 6,643 | 8,883 | 109,571 | 184,162 | | Less: loss allowance | (29) | (988) | (8,911) | (23,681) | | | 6,614 | 7,895 | 100,660 | 160,481 | | Deposits | 3,640 | 16,471 | 66,721 | 67,201 | | Receivables from disposal of investments in equity securities measured at FVPL | - | - | 38,118 | - | | Receivables of capital contributions from equity holders | - | - | - | 120,328 | | Other receivables due from: | | | | | | - third parties | 1,833 | 117,591 | 263,805 | 59,715 | | - subsidiaries | 7 | 3,116 | 166,090 | 341,961 | | | 5,480 | 257,506 | 534,734 | 468,877 | | Less: loss allowance | (3) | (18,779) | (27,327) | (20,008) | | | 5,477 | 238,727 | 507,407 | 448,869 | | Financial assets measured at amortised cost | 12,091 | 246,622 | 608,067 | 609,350 | | Input VAT deductible (Note 17(b)) | 4,169 | 84,527 | 90,168 | 68,098 | | Prepayments for computing service fees and others (Note 17(b)) | 11,633 | 88,295 | 126,239 | 72,642 | | | 15,802 | 172,822 | 216,407 | 140,740 | | | 27,893 | 419,444 | 824,474 | 750,090 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Contract assets arising from performance under on-premise deployment contracts | 405 | 10,059 | 12,943 | 17,863 | | Less: loss allowance (Note 31(a)) | (2) | (53) | (466) | (643) | | | 403 | 10,006 | 12,477 | 17,220 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Represented by: | | | | | | - current portion (Notes 17(a) and 17(b)) | 403 | 9,960 | 4,718 | 6,654 | | - non-current portion | - | 46 | 7,759 | 10,566 | | | 403 | 10,006 | 12,477 | 17,220 |
Contract assets are revenue recognised by the Group while the payment milestones have yet to be met.
The current portion of contract assets is expected to be recovered within one year.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Cash at bank and on hand in the consolidated statements of financial position | 219,031 | 1,249,391 | 2,269,222 | 2,556,116 | | Less: restricted cash (Note) | (103) | (216) | (1,058) | (4,146) | | Cash and cash equivalents in the consolidated statements of cash flows | 218,928 | 1,249,175 | 2,268,164 | 2,551,970 |
Note: As at 31 December 2022, 2023 and 2024 and 30 June 2025, restricted cash mainly represented deposits for bidding and performance guarantee.
Remittance of funds out of the PRC is subject to relevant rules and regulations of foreign exchange control.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Cash and cash equivalents | 184,742 | 1,237,619 | 1,607,914 | 1,465,519 | | Restricted cash | - | 216 | 1,058 | 3,846 | | Cash at bank and on hand | 184,742 | 1,237,835 | 1,608,972 | 1,469,365 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Time deposits | 10,092 | 102,093 | 105,343 | 106,968 | | Represented by: | | | | | | - current portion | 10,092 | 102,093 | 105,343 | 106,968 | | - non-current portion | - | - | - | - | | | 10,092 | 102,093 | 105,343 | 106,968 |
Time deposits represented deposits placed at financial institutions in the PRC with original maturity dates over one year, and they bear interest ranged from 3.00% to 3.70% during the Track Record Period.
The tables below detail changes in the Group's liabilities from financing activities, including both cash and non-cash changes. Liabilities arising from financing activities are liabilities for which cash flows were, or future cash flows will be, classified in the Group's consolidated statements of cash flows as cash flows from financing activities.
| Note | Bank loans RMB'000 (Note 24) | Financial instruments issued to investors RMB'000 (Note 26) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---| | At 1 January 2022 | - | 162,650 | - | - | - | - | 162,650 | | **Changes from financing cash flows:** | | | | | | | | | Proceeds from the issuance of financial instruments to investors | - | 208,000 | - | - | - | - | 208,000 | | Payments of transaction costs for the issuance of financial instruments to investors | - | - | - | (4,906) | - | - | (4,906) | | Capital element of lease rentals paid | - | - | - | - | - | (11,110) | (11,110) | | Interest element of lease rentals paid | - | - | - | - | - | (788) | (788) | | Total changes from financing cash flows | - | 208,000 | - | (4,906) | - | (11,898) | 191,196 |
| Note | Bank loans RMB'000 (Note 24) | Financial instruments issued to investors RMB'000 (Note 26) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---| | **Other changes:** | | | | | | | | | Net increase in lease liabilities | - | - | - | 4,906 | - | 23,942 | 23,942 | | Finance costs | 6(a) | - | - | - | - | 788 | 5,694 | | Issuance of financial instruments to investors included in other reserve | - | 42,100 | - | - | - | - | 42,100 | | Changes in the carrying amounts of financial instruments issued to investors | 26 | - | 45,209 | - | - | - | - |
| Note | Bank loans RMB'000 (Note 24) | Financial instruments issued to investors RMB'000 (Note 26) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---|
| - | 161,471 | - | - | - | - | 161,471 | |---|---|---|---|---|---|---|---| | - | 2,286 | - | - | - | - | 2,286 |
| 135,530 | 3,019,587 | - | (8,406) | 130,000 | (178,318) | 3,098,393 | |---|---|---|---|---|---|---|---|
| 1,716 | - | - | 20,119 | - | 547,773 | 547,773 | |---|---|---|---|---|---|---|---| | - | - | - | - | - | 30,931 | 52,766 |
| Note | Bank loans RMB'000 (Note 24) | Financial instruments issued to investors RMB'000 (Note 26) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---|
| - | 468,859 | - | - | - | - | 468,859 | |---|---|---|---|---|---|---|---| | - | (9,834) | - | - | - | - | (9,834) |
| 1,716 | 477,492 | 120,328 | (77,631) | 2,158 | 578,704 | 1,102,767 | |---|---|---|---|---|---|---|---|
| 137,246 | 6,676,943 | - | 14,951 | 132,158 | 671,268 | 7,632,566 | |---|---|---|---|---|---|---|---|
| (1,839) | 1,625,000 | - | (33,926) | 700,000 | (124,058) | 2,165,177 | |---|---|---|---|---|---|---|---|
| 1,807 | - | - | 31,975 | - | 34,397 | 34,397 | |---|---|---|---|---|---|---|---| | - | - | - | - | - | 17,983 | 51,765 |
| Note | Bank loans RMB'000 (Note 24) | Financial instruments issued to investors RMB'000 (Note 26) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---|
| | Financial instruments issued to investors RMB'000 (Note 26) | Bank loans RMB'000 (Note 24) | Other receivables RMB'000 (Note 19) | Other payables RMB'000 (Note 22) | Convertible bonds RMB'000 (Note 27) | Lease liabilities RMB'000 (Note 25) | Total RMB'000 | |---|---|---|---|---|---|---|---| | - | 201,174 | - | - | - | - | 201,174 |
Total cash outflow for leases Amounts included in the consolidated statements of cash flows for lease rentals paid are as follows:
| | Years ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Within operating cash flows | 2,383 | 5,537 | 9,070 | 4,211 | 5,909 | | Within financing cash flows | 11,898 | 39,327 | 178,318 | 73,611 | 124,058 | | | 14,281 | 44,864 | 187,388 | 77,822 | 129,967 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Trade payables due to third parties | 663 | 6,754 | 58,293 | 69,403 | | Payables for computing service fees | 1,509 | 104,536 | 269,467 | 495,348 | | Payables for marketing and promotion services | 15,386 | 4,357 | 89,052 | 28,981 | | Payables of staff costs | 1,987 | 45,226 | 104,229 | 105,846 | | Other payables and accruals | | 12,309 | 43,767 | 109,173 | | Financial liabilities measured at amortised cost | 19,545 | 173,182 | 564,808 | 808,751 | | Advances from equity shareholders to be injected as capital | 3,401 | 97,750 | | | | Other taxes payables | 2,888 | 11,038 | 22,304 | 8,348 | | Provisions for warranties | | 6,227 | 16,376 | 21,318 | | | 25,834 | 288,197 | 603,488 | 838,417 |
As at the end of each reporting period, the ageing analyses of trade payables, based on the invoice date, are as follows:
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Within 3 months | 663 | 5,333 | 57,676 | 59,596 | | 3 months to 6 months | - | 865 | 57 | 5,848 | | 6 months to 1 year | - | 526 | 257 | 3,680 | | More than 1 year | - | 30 | 303 | 279 | | | 663 | 6,754 | 58,293 | 69,403 |
All of the trade and other payables are expected to be settled within one year or are repayable on demand.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Trade payables due to third parties | 411 | 6,621 | 32,891 | 52,529 | | Trade payable due to subsidiaries | 2,034 | 6,010 | 167,432 | 163,326 | | Payables for computing service fees | 2,445 | 12,631 | 200,323 | 215,855 | | Payables for marketing and promotion services | 1,509 | 104,536 | 218,397 | 480,471 | | Payables of staff costs | 14,836 | 4,313 | 89,052 | 28,981 | | Other payables and accruals | 1,979 | 43,655 | 89,515 | 85,855 | | | | 12,301 | 24,707 | 80,569 | | Financial liabilities measured at amortised cost | 20,769 | 177,446 | 621,994 | 891,731 | | Advances from equity shareholders to be injected as capital | 3,388 | 97,750 | | | | Other taxes payables | 2,888 | 10,796 | 20,615 | 5,609 | | Provisions for warranties | | 6,227 | 16,376 | 21,218 | | | 27,045 | 292,219 | 658,985 | 918,558 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Receipts in advance from customers | 35,230 | 74,062 | 75,059 | 75,367 |
| | Years ended 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Balance at 1 January | 22,093 | 35,230 | 74,062 | 75,059 | | Increase in contract liabilities as a result of receipts in advance | 23,598 | 48,973 | 46,930 | 20,895 | | Decrease in contract liabilities as a result of recognising revenue during the year/period | (10,461) | (10,141) | (45,933) | (20,587) | | Balance at 31 December/30 June | 35,230 | 74,062 | 75,059 | 75,367 |
The contract liabilities are expected to be recognised as revenue within one year.
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Receipts in advance from customers | 34,200 | 74,012 | 68,761 | 62,546 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Unguaranteed and unsecured bank loan repayable within one year | - | - | 137,246 | 137,214 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Within 1 year | 12,832 | 66,765 | 213,161 | 213,458 | | After 1 year but within 2 years | - | 47,367 | 205,437 | 174,836 | | After 2 years but within 5 years | - | 156,750 | 252,670 | 211,296 | | | - | 204,117 | 458,107 | 386,132 | | | 12,832 | 270,882 | 671,268 | 599,590 |
| | As at 31 December | | | As at 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Within 1 year | 12,832 | 66,421 | 208,021 | 206,195 | | After 1 year but within 2 years | - | 47,367 | 176,548 | 168,905 | | After 2 years but within 5 years | - | 156,329 | 271,308 | 203,369 | | | - | 203,696 | 447,856 | 372,274 | | | 12,832 | 270,117 | 655,877 | 578,469 |
The movements of the financial instruments issued to investors during the Track Record Period are set out below:
| | Years ended 31 December | | | Six months ended 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | At 1 January | 162,650 | 457,959 | 3,179,864 | 6,676,943 | | Additions | 250,100 | 2,560,434 | 3,028,220 | 2,458,522 | | Changes in the carrying amounts of financial instruments issued to investors | 45,209 | 161,471 | 468,859 | 429,295 | | At 31 December/30 June | 457,959 | | | |
自2019年起及于往绩记录期间，本集团通过本公司向多名投资者进行了数轮融资，以支持本集团业务发展。本公司与上述投资者订立了投资协议，据此，该等投资者将通过收购现有股东的现有股权权益及注入新资本认购额外股权权益的方式进行投资。同时，本公司与上述投资者亦订立了股东协议，据此，本公司授予该等投资者在若干触发事件未能达成时（包括本公司股份于2028年2月前未能完成首次公开发售）要求本公司赎回投资款项的权利（即赎回权）。
From 2019 onwards and during the Track Record Period, the Group, via the Company, has conducted several rounds of financing from various investors to support the development of the Group's business. The Company and these investors have entered into investment agreements pursuant to which these investors were to acquire both existing equity interests from existing equity shareholders and additional equity interests by injections of new capital. At the same time, the Company and these investors have also entered into equity holders' agreements pursuant to which the Company has granted rights for these investors to require the Company to redeem the invested amounts if certain triggering events cannot be met, including an initial public offering of the Company's shares by February 2028 (i.e. the redemption rights).
根据本集团采纳的会计政策（见附注2(q)），本集团于综合财务状况表中将该等金融负债确认为"发行予投资者的金融工具"，往绩记录期间该等金融工具账面金额的变动已计入综合损益表。本集团已评估赎回权可能产生的最高结果，即原始本金投资加上按年利率12%计算的应计利息。赎回权将于本公司股份在任何认可证券交易所上市时终止。
In accordance with the accounting policies adopted by the Group (see Note 2(q)), the Group recognised these financial liabilities as "financial instruments issued to investors" in the consolidated statements of financial position, where changes in the carrying amounts of these financial instruments were charged to the consolidated statements of profit or loss during the Track Record Period. The Group has assessed the highest possible outcome arising from the redemption rights would be the original principal investments plus accrued interests at 12% per annum. The redemption rights will be terminated upon the listing of the Company's shares at any recognised stock exchanges.
除所确认的金融负债外，本集团已将投资者所认购额外股权权益的名义价值贷记至本公司的实收资本账户，并将相同金额加上与投资者所收购的现有股权权益相关的投资金额借记至其他储备账户。其他储备中累计的金额将于本公司股份在任何认可证券交易所成功上市时转入本公司的股份溢价账户。
In addition to the financial liabilities recognised, the Group has credited the Company's paid-in capital account with the nominal value of the additional equity interests acquired by the investors, and debited the other reserve account with the same amount plus the investment amounts related to the existing equity interests acquired by the investors. The amounts accumulated in other reserve will be transferred to the Company's share premium account upon the successful listing of the Company's shares at any recognised stock exchanges.
| | RMB'000 | |---|---| | At 1 January 2022, 31 December 2022, 31 December 2023 and 1 January 2024 | - | | Issuance of convertible bonds | 130,000 | | Fair value change of convertible bonds | 2,158 | | **At 31 December 2024 and 1 January 2025** | **132,158** | | Issuance of convertible bonds | 700,000 | | Fair value change of convertible bonds | 1,364 | | Converted into financial instruments issued to investors | (833,522) | | **At 30 June 2025** | **-** |
于2024年及2025年，本集团通过本公司订立了一系列可转换债券协议，合计本金总额为人民币830,000,000元。该等债券可由债券持有人选择转换，并按年利率0%至8%计息。
In 2024 and 2025, the Group, via the Company, entered into a series of convertible bonds agreements with total aggregate principal amount of RMB830,000,000. The bonds may be converted at the option of the bond holder and bear interest rate ranging from 0%–8% per annum.
The Group had designated the convertible bonds to be measured at FVPL, and has engaged an independent valuer to determine the fair value. The equity allocation model was adopted to determine the fair value of the convertible bonds. Details of the assumptions used are as follows:
于2025年5月，债券持有人将该等可转换债券转换为本集团发行予投资者的金融工具，其条款及条件与附注26所述者相近。
In May 2025, the bond holders converted these convertible bonds into the Group's financial instruments issued to investors with terms and conditions similar to those mentioned in Note 26.
The components of deferred tax assets/(liabilities) recognised in the consolidated statements of financial position and the movements during the Track Record Period are as follows:
| Deferred tax arising from: | Right-of-use assets RMB'000 | Lease liabilities RMB'000 | Fair value adjustments on financial instruments measured at FVPL RMB'000 | Fair value adjustments arising from business combination RMB'000 | Unused tax losses RMB'000 | Total RMB'000 | |---|---|---|---|---|---|---| | At 1 January 2022 | - | - | - | - | - | - | | (Credited)/charged to the consolidated statement of profit or loss | (1,796) | 1,889 | (93) | - | - | - | | At 31 December 2022 and 1 January 2023 | (1,796) | 1,889 | (93) | - | - | - | | Additions from business combination (Note 32) | - | - | - | (5,304) | 5,304 | - |
| | Right-of-use assets RMB'000 | Lease liabilities RMB'000 | Fair value adjustments on financial instruments measured at FVPL RMB'000 | Fair value adjustments arising from business combination RMB'000 | Unused tax losses RMB'000 | Total RMB'000 | |---|---|---|---|---|---|---| | (Credited)/charged to the consolidated statement of profit or loss | (36,953) | 36,779 | 86 | 221 | (133) | - | | At 31 December 2023 and 1 January 2024 | (38,749) | 38,668 | (7) | (5,083) | 5,171 | - | | (Credited)/charged to the consolidated statement of profit or loss | (56,264) | 56,462 | (198) | 663 | (663) | - | | At 31 December 2024 and 1 January 2025 | (95,013) | 95,130 | (205) | (4,420) | 4,508 | - | | Charged/(credited) to the consolidated statement of profit or loss | 9,907 | (7,095) | (2,812) | 332 | (332) | - | | At 30 June 2025 | (85,106) | 88,035 | (3,017) | (4,088) | 4,176 | - |
| | As at 31 December 2022 RMB'000 | As at 31 December 2023 RMB'000 | As at 31 December 2024 RMB'000 | As at 30 June 2025 RMB'000 | |---|---|---|---|---| | Net deferred tax assets in the consolidated statements of financial position | 1,889 | 43,839 | 99,638 | 92,211 | | Net deferred tax liabilities in the consolidated statements of financial position | (1,889) | (43,839) | (99,638) | (92,211) | | | - | - | - | - |
In accordance with the accounting policy set out in Note 2(t), the Group has not recognised deferred tax assets in respect of cumulative tax losses and deductible temporary difference of RMB197,398,000, RMB988,829,000, RMB5,191,680,000 and RMB8,343,905,000 as at 31 December 2022, 2023 and 2024 and 30 June 2025 respectively, as it is not probable that future taxable profits against which the losses can be utilised will be available in the relevant tax jurisdictions and entities. The unused tax losses are allowed to be carried forward for a five-year period and a ten-year period for HNTE entities.
自2021年起，集团已向集团董事及雇员，以及对集团整体业务表现及可持续发展作出直接贡献的人士授予股权激励。
股权激励附有服务条件及业绩条件。
由于本公司于2025年3月26日转制为股份有限公司，本公司在转制前为授予股权激励之目的，任意将每股权单位定价为人民币1元。
| | 截至12月31日止年度 | | | | 截至6月30日止六个月 | | |---|---|---|---|---|---|---| | | **2022年** | | **2023年** | | **2024年** | | **2025年** | | | 加权平均行使价 RMB | 股权激励数量 '000 | 加权平均行使价 RMB | 股权激励数量 '000 | 加权平均行使价 RMB | 股权激励数量 '000 | 加权平均行使价 RMB | 股权激励数量 '000 | | 于年初／期初未行使 | 6.62 | 437 | 6.11 | 444 | 3.60 | 1,161 | 2.96 | 1,895 | | 于年内／期内授予 | 1.20 | 43 | 2.04 | 739 | 1.97 | 1,005 | 1.92 | 250 | | 于年内／期内没收 | 6.40 | (36) | 2.07 | (22) | 2.00 | (271) | 1.88 | (129) | | 于期内以限制性股份替换 | - | - | - | - | - | - | 2.90 | (2,016) | | 于年末／期末未行使 | 6.11 | 444 | 3.60 | 1,161 | 2.96 | 1,895 | - | - |
于追踪记录期间，概无股权激励获行使。
| | 截至12月31日止年度 | | | 截至6月30日止六个月 | |---|---|---|---|---| | | **2022年** | **2023年** | **2024年** | **2025年** | | 计量日每股权单位公平值 | 人民币28.12元 | 人民币41.16元至人民币141.29元 | 人民币144.58元至人民币252.01元 | 人民币279.45元至人民币299.70元 | | 行使价 | 人民币1.00元 | 人民币1.00元至人民币2.00元 | 人民币1.00元至人民币2.00元 | 人民币1.00元至人民币2.00元 | | 预期波动率 | 51.92% | 56.36%至62.21% | 50.08%至61.83% | 53.25%至54.62% | | 预期股息 | 2.38% | 1.93% | 1.37%至2.58% | 1.42%至1.62% | | 无风险利率 | 2.62% | 2.15% | — | — |
预期波动率乃根据可比上市公司的历史波动率估算，并就公开资料所显示的未来波动率预期变动作出调整。
预期股息乃根据历史股息估算。
无风险利率乃基于中国人民银行制定的金融机构存款基准利率。
二项式模型用于估算股权激励之公平值。计算股权激励公平值所采用的变量及假设乃基于集团的最佳估算。若采用不同变量及假设，股权激励之公平值将有所不同。
于2025年6月，集团以限制性股份替换全部股权激励，旨在为集团合资格雇员提供激励。此次替换构成对股权激励的修改。限制性股份亦附有服务条件及业绩条件。
| | 截至2025年6月30日止六个月 | |---|---| | | 限制性股份数量 '000 | | 于期初未行使 | — | | 因替换股权激励而发行限制性股份 | 2,016 | | 于期内授予 | 4,604 | | 于期末未行使 | 6,620 |
计算限制性股份公平值所采用的变量及假设乃基于集团的最佳估算。若采用不同变量及假设，限制性股份之公平值将有所不同。
| 注释 | 实收资本／股本 人民币千元（注30(b)） | 资本储备 人民币千元（注30(c)） | 其他储备 人民币千元（注30(d)） | 股份支付储备 人民币千元（注30(e)） | 累积亏损 人民币千元 | 股东权益总额（赤字） 人民币千元 | |---|---|---|---|---|---|---| | 于2022年1月1日结余 | 12,812 | 58,776 | (2,286) | 300 | (64,710) | 4,892 | | **截至2022年12月31日止年度权益变动** | | | | | | | | 年度全面收益总额 | — | — | — | — | (134,464) | (134,464) | | 股权股东之资本注入 | 30(b) | 1,461 | — | (43,561) | — | — | (42,100) |
Increase in paid-in capital through transfer from capital reserve Equity settled share-based transactions Balance at 31 December 2022 and 1 January 2023
Change in equity for the year ended 31 December 2023: Total comprehensive income for the year Capital contributions from equity shareholders 30(b) Increase in paid-in capital through transfer from capital reserve 30(b) Equity settled share-based transactions 29 Balance at 31 December 2023 and 1 January 2024 Change in equity for the year ended 31 December 2024: Total comprehensive income for the year Capital contributions from equity shareholders 30(b) Increase in paid-in capital through transfer from capital reserve 30(b) Equity settled share-based transactions 29 Balance at 31 December 2024 and 1 January 2025 Change in equity for the six months ended 30 June 2025: Total comprehensive income for the period Capital contributions from equity shareholders 30(b) Conversion into a joint stock limited liability company 30(b) Equity settled share-based transactions 29
Change in equity for the six months ended 30 June 2024: Total comprehensive income for the period (unaudited) Capital contributions from equity shareholders (unaudited) Equity settled share-based transactions (unaudited) Balance at 30 June 2024
For the purpose of the Historical Financial Information, the paid-in capital of the Group represents the paid-in capital of the Company before it was converted into a joint stock company with limited liability.
| | Paid-in capital RMB'000 | |---|---| | At 1 January 2022 | 12,812 | | Capital contributions from investors through issuance of financial instruments (see Note 26) | 1,461 | | Increase in paid-in capital through transfer from capital reserve | 534 | | At 31 December 2022 and 1 January 2023 | 14,807 | | Capital contributions from investors through issuance of financial instruments (see Note 26) | 9,838 | | Increase in paid-in capital through transfer from capital reserve | 3,833 | | At 31 December 2023 and 1 January 2024 | 28,478 | | Capital contributions from investors through issuance of financial instruments (see Note 26) | 5,764 | | Increase in paid-in capital through transfer from capital reserve | 1,982 | | At 31 December 2024 and 1 January 2025 | 36,224 | | Conversion into a joint stock limited liability company (see Note 30(b)(ii)) | (36,224) | | At 30 June 2025 | - |
| | Number of shares '000 | Share capital RMB'000 | |---|---|---| | At 1 January 2025 | - | - | | Issuance of ordinary shares upon conversion into a joint stock company | 36,224 | 36,224 | | Capital contributions from investors through issuance of financial instruments (see Note 26) | 4,057 | 4,057 | | At 30 June 2025 | 40,281 | 40,281 |
On 26 March 2025, the Company was converted into a joint stock limited liability company and the registered capital of the Company of RMB36,224,375 was divided into 36,224,375 ordinary shares with nominal value of RMB1 each.
Immediately prior to the Global Offering, the ordinary shares of the Company will be split on a one for ten basis, and the registered share capital of our Company will be RMB40,281,069, comprising 402,810,690 Unlisted Shares in issue of nominal value RMB0.10 each.
The capital reserve comprises: (i) the differences between the net considerations received and the nominal amount of paid-in capital/share capital issued by the Company; and (ii) the differences between the net assets received and the total amount of the par value of shares issued in relation to the conversion into a joint stock company as disclosed in Note 30(b)(ii).
The other reserve represents amounts in connection with the issuance of financial instruments issued to investors as set out in Note 26.
The share-based payments reserve comprises the Company's equity settled share-based payments (see Note 29). The reserve is dealt with in accordance with the accounting policies set out in Note 2(s)(ii).
The exchange reserve comprises all foreign exchange differences arising from the translation of the financial statements of foreign operations. The reserve is dealt with in accordance with the accounting policies set out in Note 2(w).
No dividends had been declared by the Group during the Track Record Period.
The directors of the Company consider that the distributions/dividends policy during the Track Record Period is not indicative of the future dividend policy of the Company.
The Group's primary objectives when managing capital are to safeguard the Group's ability to continue as a going concern, so that it can continue to provide returns for shareholders and benefits for other stakeholders, by pricing products and services commensurately with the level of risk and by securing access to finance at a reasonable cost.
The Group actively and regularly reviews and manages its capital structure to maintain a balance between the higher shareholder returns that might be possible with higher levels of borrowings and the advantages and security afforded by a sound capital position, and makes adjustments to the capital structure in light of changes in economic conditions.
Neither the Company nor any of its subsidiaries are subject to externally imposed capital requirements.
Exposure to credit, liquidity, interest rate and currency risks arises in the normal course of the Group's business.
The Group's exposure to these risks and the financial risk management policies and practices used by the Group to manage these risks are described below.
Credit risk refers to the risk that a counterparty will default on its contractual obligations resulting in a financial loss to the Group. The Group's credit risk is primarily attributable to trade and other receivables and contract assets. The Group's exposure to credit risk arising from cash at bank is limited because the counterparties are banks and financial institutions with high credit standing, for which the Group considers to have low credit risk.
The Group does not provide any guarantees which would expose the Group to credit risk.
The Group has established a credit risk management policy under which individual credit evaluations are performed on all customers requiring credit over a certain amount. These evaluations focus on the customer's past history of making payments when due and current ability to pay and take into account information specific to the customer as well as pertaining to the economic environment in which the customer operates. The Group generally requires customers of on-premise deployment to settle progress billings, and cloud-based deployment customers to pay in advance under usage-based contracts and periodic billings under subscription-based contracts.
Significant concentrations of credit risk primarily arise when the Group has significant exposure to individual customers. As at 31 December 2022, 2023 and 2024 and 30 June 2025, 75%, 75%, 51% and 39%, respectively, of the total trade receivables was due from the Group's five largest trade debtors.
The Group measures loss allowances for trade receivables at an amount equal to lifetime ECLs, which is calculated using a provision matrix. As the Group's historical credit loss experience does not indicate significantly different loss patterns for different customer segments and geographic regions, the loss allowance based on past due status is not further distinguished between the Group's different customer bases.
The following tables provide information about the Group's exposure to credit risk and ECLs for trade receivables as at 31 December 2022, 2023 and 2024 and 30 June 2025:
As at 31 December 2022 | | Expected loss rate | Gross carrying amount RMB'000 | Loss allowance RMB'000 | |---|---|---|---| | Within 3 months | 0.38% | 6,523 | 25 | | 6 months to 1 year | 3.15% | 120 | 4 | | | | 6,643 | 29 |
As at 31 December 2023 | | Expected loss rate | Gross carrying amount RMB'000 | Loss allowance RMB'000 | |---|---|---|---| | Within 3 months | 0.53% | 5,409 | 30 | | 3 months to 6 months | 3.54% | 1,477 | 52 | | 6 months to 1 year | 3.53% | 85 | 3 | | 1 year to 2 years | 47.23% | 1,912 | 903 | | | | 8,883 | 988 |
As at 31 December 2024 | | Expected loss rate | Gross carrying amount RMB'000 | Loss allowance RMB'000 | |---|---|---|---| | Within 3 months | 3.60% | 76,958 | 2,767 | | 3 months to 6 months | 16.09% | 16,451 | 2,647 | | 6 months to 1 year | 22.07% | 1,853 | 409 | | 1 year to 2 years | 56.54% | 2,996 | 1,694 | | 2 years to 3 years | 79.39% | 1,912 | 1,518 | | | | 100,170 | 9,035 |
As at 30 June 2025 | | Expected loss rate | Gross carrying amount RMB'000 | Loss allowance RMB'000 | |---|---|---|---| | Within 3 months | 3.70% | 79,644 | 2,950 | | 3 months to 6 months | 16.57% | 34,630 | 5,739 | | 6 months to 1 year | 24.54% | 51,029 | 12,525 | | 1 year to 2 years | 68.84% | 4,358 | 3,000 | | 2 years to 3 years | 83.47% | 1,996 | 1,666 | | | | 171,657 | 25,880 |
As at 31 December 2022, 2023 and 2024 and 30 June 2025, the expected loss rates for contract assets are from 0.38% to 3.70%.
The Group has assessed the ECLs of other receivables based on the debtor's payment history and other information related to the debtors. The Group has identified certain debtors were in financial difficulties and the related credit risks have increased significantly. Accordingly, loss allowance of RMB3,000, RMB18,779,000, RMB27,327,000 and RMB21,172,000 as at 31 December 2022, 2023 and 2024 and 30 June 2025, respectively, have been recognised.
Movements in the loss allowance account in respect of trade and other receivables and contract assets during the Track Record Period are as follows:
| | Years ended 31 December | | | Six months ended 30 June | |---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2025 RMB'000 | | Balance at 1 January | 3 | 34 | 19,820 | 36,828 | | Impairment losses recognised | 31 | 19,786 | 17,008 | 10,867 | | Balance at 31 December/30 June | 34 | 19,820 | 36,828 | 47,695 |
The Group's policy is to regularly monitor current and expected liquidity requirements, to ensure that it maintains sufficient reserves of cash and adequate committed lines of funding from major financial institutions to meet its liquidity requirements in the short and longer term.
The following tables show the remaining contractual maturities at the end of each reporting period of the Group's financial liabilities, which are based on contractual undiscounted cash flows (including interest payments computed using contractual rates or, if floating, based on rates current at the end of the reporting period) and the earliest date the Group can be required to pay:
| | Within 1 year or on demand RMB'000 | More than 1 year but less than 2 years RMB'000 | More than 2 years but less than 5 years RMB'000 | Total RMB'000 | Carrying amounts in the consolidated statements of financial position RMB'000 | |---|---|---|---|---|---| | Trade and other payables measured at amortised cost (Note 22) | 19,545 | - | - | 19,545 | 19,545 | | Lease liabilities (Note 25) | 12,832 | - | - | 12,832 | 12,832 | | | 32,377 | - | - | 32,377 | 32,377 |
| | Within 1 year or on demand RMB'000 | More than 1 year but less than 2 years RMB'000 | More than 2 years but less than 5 years RMB'000 | Total RMB'000 | Carrying amounts in the consolidated statements of financial position RMB'000 | |---|---|---|---|---|---| | Trade and other payables measured at amortised cost (Note 22) | 173,182 | - | - | 173,182 | 173,182 | | Lease liabilities (Note 25) | 78,192 | 54,029 | 164,903 | 297,124 | 270,882 | | | 251,374 | 54,029 | 164,903 | 470,306 | 444,064 |
| | Within 1 year or on demand RMB'000 | More than 1 year but less than 2 years RMB'000 | More than 2 years but less than 5 years RMB'000 | Total RMB'000 | Carrying amounts in the consolidated statements of financial position RMB'000 | |---|---|---|---|---|---| | Trade and other payables measured at amortised cost (Note 22) | 564,808 | - | - | 564,808 | 564,808 | | Lease liabilities (Note 25) | 246,149 | 226,760 | 265,255 | 738,164 | 671,268 | | Bank loans (Note 24) | 140,881 | - | - | 140,881 | 137,246 | | | 951,838 | 226,760 | 265,255 | 1,443,853 | 1,373,322 |
| | Within 1 year or on demand RMB'000 | More than 1 year but less than 2 years RMB'000 | More than 2 years but less than 5 years RMB'000 | Total RMB'000 | Carrying amounts in the consolidated statements of financial position RMB'000 | |---|---|---|---|---|---| | Trade and other payables measured at amortised cost (Note 22) | 808,751 | - | - | 808,751 | 808,751 | | Lease liabilities (Note 25) | 240,307 | 191,287 | 219,448 | 651,042 | 599,590 | | Bank loans (Note 24) | 140,850 | - | - | 140,850 | 137,214 | | | 1,189,908 | 191,287 | 219,448 | 1,600,643 | 1,545,555 |
In addition to the above, the Group was also exposed to liquidity risk arising from financial instruments issued to investors at 31 December 2022, 2023 and 2024 and 30 June 2025, and convertible bonds as at 31 December 2024, the payment terms of which are further disclosed in Note 26 and Note 27 respectively.
Fair values are categorised into the three-level fair value hierarchy as defined in IFRS 13, Fair value measurement. The level into which a fair value measurement is classified is determined with reference to the observability and significance of the inputs used in the valuation technique as follows:
• Level 1 valuations: Fair value measured using only Level 1 inputs, i.e. unadjusted quoted prices in active markets for identical assets or liabilities at the measurement date.
• Level 2 valuations: Fair value measured using Level 2 inputs, i.e. observable inputs which fail to meet Level 1, and not using significant unobservable inputs. Unobservable inputs are inputs for which market data are not available.
• Level 3 valuations: Fair value measured using significant unobservable inputs.
The following tables present the fair value of the Group's financial instruments measured at the end of the reporting period on a recurring basis, categorised into the three-level fair value hierarchy.
| | RMB'000 | Fair value measurements as at 31 December 2022 categorised into Level 3 RMB'000 | |---|---|---| | Assets | | | | Investments in equity securities | 11,740 | 11,740 | | Wealth management products | 20,037 | 20,037 |
| | RMB'000 | Fair value measurements as at 31 December 2023 categorised into Level 3 RMB'000 | |---|---|---| | Assets | | | | Investments in equity securities | 58,830 | 58,830 | | Wealth management products | 100,074 | 100,074 |
| | RMB'000 | Fair value measurements as at 31 December 2024 categorised into Level 3 RMB'000 | |---|---|---| | Assets | | | | Investments in equity securities | 42,621 | 42,621 | | Liabilities | | | | Convertible Bonds | 132,158 | 132,158 |
| | RMB'000 | Fair value measurements as at 30 June 2025 categorised into Level 3 RMB'000 | |---|---|---| | Assets | | | | Investments in equity securities | 49,343 | 49,343 | | Wealth management products | 500,021 | 500,021 |
During the Track Record Period, there were no transfers between Level 1 and Level 2, or transfers into or out of Level 3. The Group's policy is to recognise transfers between levels of fair value hierarchy as at the end of the reporting period in which they occur.
Below is a summary of the valuation techniques and significant unobservable inputs to the valuation of these financial assets at 31 December 2022, 2023 and 2024 and 30 June 2025.
| | Valuation techniques | Significant unobservable inputs | |---|---|---| | Investments in equity securities | Market approach | Discount for lack of marketability | | Wealth management products | Discounted cash flow method | Expected rate of return | | Convertible bonds | Equity allocation model | Expected volatility |
During the Track Record Period, the Group's investments in equity securities measured at FVPL are investments in non-listed entities of which fair values were substantially determined based on either the latest round of equity financing obtained by these entities or based on market approach. Given the discount for lack of marketability was not developed by the Group, the management of the Group did not carry out nor present any information on sensitivity analysis.
理财产品的公允价值受预期收益率变动的影响。若在其他变量保持不变的情况下，预期收益率上升/下降1%，则截至2022年12月31日及2023年12月31日止年度以及截至2025年6月30日止六个月的税前亏损将分别减少/增加约人民币200,000元、人民币1,001,000元及人民币5,000,000元。
Fair value of the wealth management products is affected by changes in the expected rate of return. If the expected rate of return had increased/decreased by 1% with all other variables held constant, the loss before tax for the year ended 31 December 2022 and 2023 and the six months ended 30 June 2025 would have been decreased/increased by approximately RMB200,000 and RMB1,001,000 and RMB5,000,000 respectively.
The carrying amounts of the Group's financial instruments carried at amortised cost are not materially different from their fair values as at 31 December 2022, 2023 and 2024 and 30 June 2025.
Interest rate risk is the risk that the fair value or future cash flows of a financial instrument will fluctuate because of changes in market interest rates. The Group's interest rate risk arises primarily from lease liabilities and bank loans, which expose the Group to cash flow interest rate risk and fair value interest rate risk, respectively.
The following tables detail the interest rate profile of the Group's borrowings at the end of each reporting period.
| | At 31 December 2022 | | At 31 December 2023 | | At 31 December 2024 | | At 30 June 2025 | | |---|---|---|---|---|---|---|---|---| | | Effective interest rate | RMB'000 | Effective interest rate | RMB'000 | Effective interest rate | RMB'000 | Effective interest rate | RMB'000 | | Fixed rate borrowings: | | | | | | | | | | Lease liabilities (Note 25) | 4.75% | 12,832 | 4.75% | 270,882 | 4.75% | 671,268 | 4.75% | 599,590 | | Bank loans (Note 24) | N/A | - | N/A | - | 2.65% | 137,246 | 2.65% | 137,214 | | | | 12,832 | | 270,882 | | 808,514 | | 736,804 |
Prior to 7 September 2023, the Company held 5.02% equity interests in Beijing Lingxin Intelligent (北京灵心智能), which the Group accounted for as financial assets measured at FVPL. On 7 September 2023, the Company entered into a series of share purchase agreements with various equity holders of Beijing Lingxin Intelligent (北京灵心智能) to acquire an aggregate of 94.98% interest in Beijing Lingxin Intelligent (北京灵心智能) with a total cash consideration of RMB73,005,000. Beijing Lingxin Intelligent (北京灵心智能) is mainly engaged in providing artificial intelligence services. The acquisition was made as part of the Group's strategy to expand its market share of artificial intelligence in the PRC.
The acquisition was completed on 7 September 2023, and after completion of the acquisition, Beijing Lingxin Intelligent (北京灵心智能) became a wholly-owned subsidiary of the Company and the financial results of Beijing Lingxin Intelligent (北京灵心智能) was consolidated into the consolidated financial statements of the Group.
APPENDIX I ACCOUNTANTS' REPORT
The fair value of the identifiable assets and liabilities of Beijing Lingxin Intelligent (北京灵心智能) as at the date of business acquisition is set out as follows:
| | Note | On acquisition date RMB'000 | |---|---|---| | Identifiable assets and liabilities: | | | | Property and equipment | 11(a) | 184 | | Intangible assets | 12(a) | 36,846 | | Deferred tax assets | | 5,304 | | Inventories and contract costs | | 71 | | Trade and other receivables | | 213 | | Cash at bank and on hand | | 8,919 | | Trade and other payables | | (1,727) | | Deferred tax liabilities | 28(a) | (5,304) | | Fair value of net identifiable assets acquired | | 44,506 | | Cash consideration | | 73,005 | | Add: previously held equity interests in Beijing Lingxin Intelligent (北京灵心智能) | | 10,880 | | Total acquisition price | | 83,885 | | Goodwill | 13 | 39,379 |
Analysis of net cash outflow of cash and cash equivalents in respect of acquisition of Beijing Lingxin Intelligent (北京灵心智能) as at the date of acquisition:
| | RMB'000 | |---|---| | Total cash consideration | 73,005 | | Less: cash acquired as at the date of acquisition | (8,919) | | Net cash outflow for the acquisition of a subsidiary included in the consolidated statements of cash flows from investing activities | 64,086 |
Included in the consideration was RMB3,000,000 which was settled in 2024.
For the year ended 31 December 2023, Beijing Lingxin Intelligent (北京灵心智能) contributed revenue of RMB307,000 and loss of RMB5,903,000 to the Group's results. If the acquisition had occurred on 1 January 2023, management estimates that consolidated revenue would have increased by RMB3,889,000, and consolidated loss for the year would have increased by RMB13,910,000.
APPENDIX I ACCOUNTANTS' REPORT
The Group entered into the following significant related party transactions during the Track Record Period:
Remuneration for key management personnel of the Group, including amounts paid to the Company's directors as disclosed in Note 8 and certain of the highest paid employees as disclosed in Note 9:
| | Year ended 31 December | | | Six months ended 30 June | | |---|---|---|---|---|---| | | 2022 RMB'000 | 2023 RMB'000 | 2024 RMB'000 | 2024 RMB'000 (unaudited) | 2025 RMB'000 | | Salaries, wages and other benefits | 2,225 | 2,924 | 4,127 | 1,844 | 2,323 | | Discretionary bonuses | 1,850 | 1,174 | 3,069 | 1,357 | 1,084 | | Contributions to defined contribution retirement plan | 172 | 252 | 264 | 132 | 120 | | Equity-settled share-based compensation expenses | 779 | 1,649 | 4,050 | 861 | 39,878 | | Total | 5,026 | 5,999 | 11,510 | 4,194 | 43,405 |
Total remuneration was included in "staff costs" (see Note 6(b)).
(b) During the Track Record Period, transactions with the following parties are considered as related party transactions:
相关方名称： Trend Mega Limited 北京星联招集企业管理合伙企业（有限合伙） 无锡云汇数字经济投资管理合伙企业（有限合伙） 北京星联 北京抖闪智创科技有限公司 北京三快在线科技有限公司
| | 截至12月31日 | | | 截至6月30日 | |---|---|---|---|---| | | 2022年 人民币千元 | 2023年 人民币千元 | 2024年 人民币千元 | 2025年 人民币千元 | | 贸易性质 | | | | | | 贸易及其他应收款 | 952 | 2,339 | 44,694 | - | | 贸易及其他应付款 | - | 36,022 | 71,372 | 120,328 | | 非贸易性质 | | | | | | 应收股权持有人缴纳的出资款 | - | - | - | 97,750 | | 已收待注入资本的股东预付款 | - | - | - | - |
| | 截至12月31日止年度 | | | 截至6月30日止六个月 | | |---|---|---|---|---|---| | | 2022年 人民币千元 | 2023年 人民币千元 | 2024年 人民币千元 | 2024年 人民币千元（未经审计） | 2025年 人民币千元 | | 贸易性质 | | | | | | | 服务销售收入 | 1,604 | 18,244 | 20,977 | 37,863 | 29,717 | | 服务采购支出 | | | | | | | 非贸易性质 | | | | | | | 出售若干非上市公司的股权权益 | - | - | 206,027 | - | - |
上述与服务销售及服务采购相关的关联方交易将在本公司股份于联交所建议上市后继续进行，但该等交易并不构成《上市规则》第14A章项下的关连交易。
本公司董事认为，截至2025年6月30日，本公司的直接母公司及最终控股方为北京联湃科技发展中心（有限合伙）、刘德兵博士、唐杰博士、李娟子博士、徐斌博士、张鹏博士、珠海横琴汇汇企业管理合伙企业（有限合伙）及珠海横琴智灯企业管理合伙企业（有限合伙）。
| | 生效会计期间（始于） | |---|---| | 《国际财务报告准则第9号》及《国际财务报告准则第7号》修订——参照自然依存型电力的合同 | 2026年1月1日 | | 《国际财务报告准则第9号》及《国际财务报告准则第7号》修订——金融工具分类和计量修订 | 2026年1月1日 | | 国际财务报告准则会计准则年度改进——第11卷 | 2026年1月1日 | | 《国际财务报告准则第18号》——财务报表的列报与披露 | 2027年1月1日 | | 《国际财务报告准则第19号》——不具公众责任的子公司：披露 | 2027年1月1日 | | 《国际财务报告准则第10号》及《国际会计准则第28号》修订——投资者与其联营公司或合营企业之间的资产出售或注入 | 待定 |
本集团正在评估其他进展在首次应用时预期产生的影响。迄今为止，本集团得出结论，采纳上述准则不太可能对历史财务资料产生重大影响，但以下内容除外。
《国际财务报告准则第18号》将取代《国际会计准则第1号——财务报表的列报》，旨在提高实体财务报表相关信息的透明度和可比性。《国际财务报告准则第18号》对2027年1月1日或之后开始的年度报告期生效，并须追溯应用。
除其他变化外，根据《国际财务报告准则第18号》，实体须在损益表中将所有收入和费用划分为五个类别，即经营类别、投资类别、融资类别、终止经营类别及所得税类别。实体还须在财务报表的单一附注中就管理层定义的绩效指标提供具体披露。
本集团不计划提前采纳《国际财务报告准则第18号》。《国际财务报告准则第18号》将影响财务报表的列报，预计不会对本集团的财务表现及状况产生重大影响。
自2025年6月30日至本报告日期，并无重大后续事项。
本公司及其附属公司并未就2025年6月30日后的任何期间编制经审计财务报表。
以下附录中所载信息不构成香港注册会计师事务所毕马威会计师事务所（本公司申报会计师）出具的会计师报告的组成部分（该报告载于本招股说明书附录I），仅供参考。
未经审计的备考财务信息应与本招股说明书"财务信息"一节及本招股说明书附录I所载会计师报告一并阅读。
以下本集团未经审计的备考经调整综合有形净资产声明依据《上市规则》第4.29条编制，旨在说明全球发售对本公司股东于2025年6月30日应占本集团综合有形净负债的影响，假设全球发售于2025年6月30日完成。
未经审计的备考经调整综合有形净资产声明仅供说明用途，由于其假设性质，可能无法如实反映全球发售于2025年6月30日或其后任何日期完成时本集团的财务状况。
| | 本公司股东于2025年6月30日应占本集团综合有形净负债(1) | 终止向投资者发行的金融工具后的估计影响(3) | 全球发售估计净收益(2) | 本公司股东应占备考经调整综合有形净资产 | 每股备考经调整综合有形净资产 | | |---|---|---|---|---|---|---| | | 人民币千元 | 人民币千元 | 人民币千元 | 人民币千元 | 人民币(4) | 港元(5) | | 按每股H股发售价116.20港元 | (6,240,424) | 3,804,426 | 9,564,760 | 7,128,762 | 16.19 | 17.85 |
(1) 于2025年6月30日，本公司股东应占本集团综合有形净负债，是在本公司股东应占总亏损人民币6,145,646,000元的基础上，(i) 扣除无形资产人民币55,399,000元，及 (ii) 扣除商誉人民币39,379,000元后得出，上述数据摘自本招股说明书附录I所载会计师报告。
(2) 全球发售的估计净收益，以根据全球发售预期发行的37,419,500股H股及每股H股116.20港元的指示性发售价为基础，扣除本集团已付或应付的估计承销费用及与全球发售相关的其他开支（不包括于追踪记录期间已计入损益的上市费用），且不考虑行使超额配股权或股份激励计划可能导致发行的任何股份。
全球发售的估计净收益已按港元1元兑换人民币0.9070元的汇率换算为人民币。并无声明港元金额已经、可能或将能按上述汇率或任何其他汇率兑换为人民币，或反之亦然。
(3) 于2025年6月30日，本公司向投资者发行的金融工具账面金额为人民币9,564,760,000元（详见附录I注26）。上市后，金融工具的赎回权将自动终止，向投资者发行的金融工具将相应由负债重新分类至权益。
(4) The unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company per H Share is arrived at after the adjustments referred to in the preceding paragraphs and on the basis that 440,230,190 shares were in issue (being 402,810,690 shares in issue and outstanding as of 30 June 2025 taking into account the Share Subdivision and 37,419,500 H Shares to be issued pursuant to Global Offering) and does not take into account of any shares which may be issued upon the exercise of the Over-allotment Option or the share incentive plans.
(5) The unaudited pro forma adjusted consolidated net tangible assets attributable to equity shareholders of the Company per H Share amounts in RMB are converted to Hong Kong dollar with an exchange rate of RMB1 to HK$1.1025. No representation is made that Renminbi amount have been, could have been or may be converted to Hong Kong dollars, or vice versa, at that rate or at any other rate.
(6) No adjustment has been made to reflect any trading results or other transactions of the Group entered into subsequent to 30 June 2025.
APPENDIX II B.
The following is the text of a report received from the reporting accountants, KPMG, Certified Public Accountants, Hong Kong, in respect of the Group's pro forma financial information for the purpose in this Prospectus.
We have completed our assurance engagement to report on the compilation of pro forma financial information of Knowledge Atlas Technology Joint Stock Company Limited (the "Company") and its subsidiaries (collectively the "Group") by the directors of the Company (the "Directors") for illustrative purposes only. The unaudited pro forma financial information consists of the unaudited pro forma statement of adjusted consolidated net tangible assets as at 30 June 2025 and related notes as set out in Part A of Appendix II to the prospectus dated 30 December 2025 (the "Prospectus") issued by the Company. The applicable criteria on the basis of which the Directors have compiled the pro forma financial information are described in Part A of Appendix II to the Prospectus.
The pro forma financial information has been compiled by the Directors to illustrate the impact of the proposed offering of the H shares of the Company (the "Global Offering") on the Group's financial position as at 30 June 2025 as if the Global Offering had taken place at 30 June 2025. As part of this process, information about the Group's financial position as at 30 June 2025 has been extracted by the Directors from the Group's historical financial information included in the Accountants' Report as set out in Appendix I to the Prospectus.
The Directors are responsible for compiling the pro forma financial information in accordance with paragraph 4.29 of the Rules Governing the Listing of Securities on The Stock Exchange of Hong Kong Limited (the "Listing Rules") and with reference to Accounting Guideline 7 "Preparation of Pro Forma Financial Information for Inclusion in Investment Circulars" ("AG 7") issued by the Hong Kong Institute of Certified Public Accountants ("HKICPA").
We have complied with the independence and other ethical requirements of the Code of Ethics for Professional Accountants issued by the HKICPA, which is founded on fundamental principles of integrity, objectivity, professional competence and due care, confidentiality and professional behaviour.
Our firm applies Hong Kong Standard on Quality Management 1 "Quality Management for Firms that Perform Audits or Reviews of Financial Statements, or Other Assurance or Related Services Engagements", which requires the firm to design, implement and operate a system of quality management including policies or procedures regarding compliance with ethical requirements, professional standards and applicable legal and regulatory requirements.
Our responsibility is to express an opinion, as required by paragraph 4.29(7) of the Listing Rules, on the pro forma financial information and to report our opinion to you. We do not accept any responsibility
for any reports previously given by us on any financial information used in the compilation of the pro forma financial information beyond that owed to those to whom those reports were addressed by us at the dates of their issue.
We conducted our engagement in accordance with Hong Kong Standard on Assurance Engagements ("HKSAE") 3420 "Assurance Engagements to Report on the Compilation of Pro Forma Financial Information Included in a Prospectus" issued by the HKICPA. This standard requires that the reporting accountants plan and perform procedures to obtain reasonable assurance about whether the Directors have compiled the pro forma financial information in accordance with paragraph 4.29 of the Listing Rules, and with reference to AG 7 issued by the HKICPA.
For purpose of this engagement, we are not responsible for updating or reissuing any reports or opinions on any historical financial information used in compiling the pro forma financial information, nor have we, in the course of this engagement, performed an audit or review of the financial information used in compiling the pro forma financial information.
The purpose of pro forma financial information included in an investment circular is solely to illustrate the impact of a significant event or transaction on unadjusted financial information of the Group as if the event had occurred or the transaction had been undertaken at an earlier date selected for purposes of the illustration. Accordingly, we do not provide any assurance that the actual outcome of events or transactions as 30 June 2025 would have been as presented.
A reasonable assurance engagement to report on whether the pro forma financial information has been properly compiled on the basis of the applicable criteria involves performing procedures to assess whether the applicable criteria used by the Directors in the compilation of the pro forma financial information provide a reasonable basis for presenting the significant effects directly attributable to the event or transaction, and to obtain sufficient appropriate evidence about whether:
• the pro forma financial information reflects the proper application of those adjustments to the unadjusted financial information.
The procedures selected depend on the reporting accountants' judgement, having regard to the reporting accountants' understanding of the nature of the Group, the event or transaction in respect of which the pro forma financial information has been compiled, and other relevant engagement circumstances.
The engagement also involves evaluating the overall presentation of the pro forma financial information.
We believe that the evidence we have obtained is sufficient and appropriate to provide a basis for our opinion.
Our procedures on the pro forma financial information have not been carried out in accordance with attestation standards or other standards and practices generally accepted in the United States of America, auditing standards of the Public Company Accounting Oversight Board (United States) or any overseas standards and accordingly should not be relied upon as if they had been carried out in accordance with those standards and practices.
We make no comments regarding the reasonableness of the amount of net proceeds from the issuance of the Company's shares, the application of those net proceeds, or whether such use will actually take place as described in the section headed "Future Plans and Use of Proceeds" in the Prospectus.
PRC TAXATION Taxation of Security Holders The taxation of income and capital gains of holders of H Shares is subject to the laws and practices of the PRC and of jurisdictions in which holders of H Shares are resident or otherwise subject to tax. The following summary of certain relevant taxation provisions is based on current effective PRC laws and practices and no predictions are made about changes or adjustments to relevant laws or policies, and no comments or suggestions will be made accordingly. The discussion does not deal with all possible tax consequences relating to an investment in the H Shares, nor does it take into account the specific circumstances of any particular investor, some of which may be subject to special regulations. Accordingly, you should consult your own tax adviser regarding the tax consequences of an investment in H Shares. The discussion is based upon PRC laws and relevant interpretations in effect as at the date of this prospectus, all of which are subject to change. Prospective investors are urged to consult their financial adviser regarding the PRC and other tax consequences of owning and disposing of H Shares.
Pursuant to the Individual Income Tax Law of the PRC (《中華人民共和國個人所得稅法》), which was last amended on August 31, 2018 by the SCNPC and came into effect on January 1, 2019, and the Regulations on Implementation of the Individual Income Tax Law of the PRC (《中華人民共和國個人所得稅法實施條例》), which were last amended on December 18, 2018 by the State Council and came into effect on January 1, 2019, dividends paid by PRC enterprises are subject to individual income tax levied at a flat rate of 20%. For a foreign individual who is not a resident of the PRC, the receipt of dividends from a PRC enterprise in the PRC is normally subject to individual income tax of 20% unless specifically exempted by the tax authority of the State Council or reduced by an applicable tax treaty.
Pursuant to the Notice of the State Taxation Administration (the "STA") on Issues Concerning the Levy and Administration of Individual Income Tax After the Repeal of Guo Shui Fa [1993] No. 45) (Guo Shui Han [2011] No. 348) (《國家稅務總局關于國稅發[1993]045號文件廢止後有關個人所得稅徵管問題的通知 (國稅函[2011]348號)》) issued by the STA on June 28, 2011, which came into effect on the same day, domestic non-foreign-invested enterprises issuing shares in Hong Kong may, when distributing dividends, withhold individual income tax at the rate of 10%. For the individual holders of H Shares receiving dividends who are citizens of countries that have entered into a tax treaty with the PRC with tax rate of lower than 10%, non-foreign-invested enterprises listed in Hong Kong may apply on behalf of such holders for enjoying the lower preferential tax treatments, and, upon approval by the tax authorities, the excessive withholding amount will be refunded. For the individual holders of H Shares receiving dividends who are citizens of countries that have entered into a tax treaty with the PRC with tax rate of higher than 10% but lower than 20%, the non-foreign-invested enterprise is required to withhold the tax at the agreed rate under the treaties, and no application procedures will be necessary. For the individual holders of H Shares receiving dividends who are citizens of countries without taxation treaties with the PRC or are under other situations, the non-foreign-invested enterprise is required to withhold the tax at a rate of 20%.
According to the Enterprise Income Tax Law of the PRC (《中華人民共和國企業所得稅法》), which was latest amended by the SCNPC and implemented on December 29, 2018, and the Implementation Rules for the Enterprise Income Tax Law of the PRC (《中華人民共和國企業所得稅法實施條例》), which was last amended on December 6, 2024 by the State Council, a non-resident enterprise is generally subject to a 10% enterprise income tax on PRC-sourced income (including dividends received from a PRC resident enterprise that issues shares in Hong Kong), if it does not have an establishment or premise in the PRC or has an
在中国境内设有机构或场所，但其来源于中国境内的所得与该机构或场所没有实际联系的非居民企业，就其上述应缴纳的企业所得税，实行源泉扣缴，以支付人为扣缴义务人。
国家税务总局于2008年11月6日发布并实施的《关于中国居民企业向境外H股非居民企业股东派发股息代扣代缴企业所得税有关问题的通知（国税函〔2008〕897号）》进一步明确，中国居民企业向境外H股非居民企业股东派发2008年及以后年度股息，应以10%的税率代扣代缴企业所得税。此外，国家税务总局于2009年7月24日发布并生效的《关于非居民企业取得B股等股票股息征收企业所得税问题的批复（国税函〔2009〕394号）》进一步规定，在境外证券交易所上市的中国居民企业，应以10%的税率对其向非居民企业派发的2008年及以后年度股息代扣代缴企业所得税。上述税率可依据中国与相关国家或地区签订的税收协定或安排作进一步调整。
根据国家税务总局与香港政府于2006年8月21日签订的《内地和香港特别行政区关于对所得避免双重征税和防止偷漏税的安排》（"安排"），中国政府就中国公司向香港居民（包括居民个人及居民实体）派发的股息征收的税款，不得超过股息总额的10%；但若香港居民直接持有该中国公司25%或以上股权，则税款不得超过股息总额的5%。于2019年12月6日生效的《〈内地和香港特别行政区关于对所得避免双重征税和防止偷漏税的安排〉第五议定书》新增了享受协定优惠待遇的资格标准。尽管安排中可能存在其他条款，但在综合考虑所有相关事实和情况后，如相关收益被合理认定为将带来本安排项下任何直接或间接利益的安排或交易的主要目的之一，则不应给予上述标准下的协定优惠，除非在该情形下给予优惠与本安排的相关目标和宗旨相符。税收协定股息条款的适用须符合中国税法及法规的要求，例如《国家税务总局关于执行税收协定股息条款有关问题的通知（国税函〔2009〕81号）》。
居住在与中国签订了避免双重征税协定或安排的司法管辖区的非居民投资者，可能有权就其从中国企业取得的股息享受中国企业所得税的减免优惠。中国目前已与多个国家和地区签订了避免双重征税协定或安排，包括香港、澳门、澳大利亚、加拿大、法国、德国、日本、荷兰、新加坡、英国及美国。依据相关税收协定或安排享有优惠税率的非中国居民企业，须向中国税务机关申请退还超过约定税率部分的企业所得税，且退税申请须经中国税务机关批准。
根据《关于全面推开营业税改征增值税试点的通知（财税[2016]36号）》（分别于2017年7月11日及2019年3月20日部分废止），在中国境内从事销售服务的单位和个人须缴纳增值税，而在中国境内销售服务是指应税服务的销售方或购买方在中国境内的情形。该通知还规定，转让金融商品（包括转让有价证券所有权）须由一般纳税人或境外纳税人按应税收入（即销售价格扣除购买价格后的余额）缴纳6%的增值税。然而，个人转让金融商品免征增值税。
增值税纳税人还须缴纳城市维护建设税、教育费附加及地方教育附加（统称"地方附加税"），通常为应缴增值税的12%（如有）。然而，根据自2021年9月1日起生效的《中华人民共和国城市维护建设税法》，境外单位和个人在中国境内销售劳务、服务及无形资产所缴纳的增值税或消费税，不征收城市维护建设税。同时，根据《关于城市维护建设税计税依据确定办法等事项的公告》，自2021年9月1日起，教育费附加和地方教育附加的计税依据与城市维护建设税的计税依据保持一致。综上所述，自2021年9月1日起，境外单位和个人在中国境内销售无形资产所缴纳的增值税，不再征收城市维护建设税、教育费附加及地方教育附加。
然而，非中国居民企业在实际处置H股时是否须缴纳中国增值税，目前仍存在不确定性。若未来相关税费被征收，此类H股持有人的投资价值可能受到重大不利影响。
根据《中华人民共和国个人所得税法》及其实施细则，出售中国居民企业股权所取得的收益须按20%的税率缴纳所得税。
根据国家税务总局于1998年3月30日颁布并生效的《关于个人转让股票所得继续暂免征收个人所得税的通知（财税字[1998]61号）》，自1997年1月1日起，个人转让上市公司股票所得继续暂免征收个人所得税。在新修订的《中华人民共和国个人所得税法》中，国家税务总局未明确规定是否继续对个人转让上市公司股票所得免税。
此外，财政部（"财政部"）、国家税务总局及中国证券监督管理委员会（"证监会"）联合颁布并于2009年12月31日起实施的《关于个人转让上市公司限售股所得征收个人所得税有关问题的通知（财税[2009]167号）》规定，个人从公开发行和转让市场取得的上市公司限售股转让所得……
on the Shanghai Stock Exchange and the Shenzhen Stock Exchange shall continue to be exempted from the individual income tax, provided that it excludes the relevant restricted shares as defined in the Supplementary Notice Concerning the Levy of Individual Income Tax on Incomes from the Transfer of Restricted Shares of Listed Companies (Cai Shui [2010] No. 70) (《關於個人轉讓上市公司限售股所得徵收個人所得稅有關問題的補充通知 (財稅[2010]70號)》) jointly issued by these departments and implemented on November 10, 2010. As at the Latest Practicable Date, the aforementioned provisions did not specify whether to impose the individual income tax on the income from the transfer of shares of PRC-resident enterprise listed on overseas stock exchanges by non-PRC resident individuals.
In accordance with the EIT Law and its implementation rules, a non-resident enterprise that has not established an establishment or premises in the PRC or it has established an establishment and premises but the income received has no actual connection with the establishment and premises, it shall pay an enterprise income tax at a rate of 10% for the income arising within the PRC (including the income from sale of equity interests of PRC-resident enterprise). The aforesaid income tax payable for non-resident enterprises are deducted at source, where the payer of the income is required to withhold the income tax from the amount to be paid to the non-resident enterprise on each payment or when it is payable on due date. The withholding tax may be reduced pursuant to applicable treaties or agreements on avoidance of double taxation.
In accordance with the Stamp Tax Law of the PRC (《中華人民共和國印花稅法》) promulgated by the Standing Committee of the NPC on June 10, 2021 and came into effect on July 1, 2022, entities and individuals that issue taxable certificates and conduct securities transactions within the territory of PRC, or entities and individuals who issue taxable certificates and conduct securities transactions outside the territory of PRC to be used within the territory of the PRC shall subject to stamp duty.
As at the Latest Practicable Date, no estate duty is levied within the PRC.
Please refer to the section headed "Regulatory Overview" of this document.
The lawful currency of the PRC is RMB, which is currently subject to foreign exchange regulation according to relevant laws and regulations. SAFE, with the authorization of the PBOC, is empowered with the functions of administering all matters relating to foreign exchange, including the enforcement of foreign exchange regulatory regulations.
On January 29, 1996, the State Council promulgated the Regulations of the PRC for Foreign Exchange Control (《中華人民共和國外匯管理條例》) (the "Foreign Exchange Control Regulations") which became effective on April 1, 1996. The Foreign Exchange Control Regulations classify all international payments and transfers into current items and capital items. Most of the current items are no longer subject to SAFE's approval, while capital items remain unchanged. The Foreign Exchange Control Regulations were subsequently amended on January 14, 1997 and August 5, 2008. The latest amendment to the Foreign Exchange Control Regulations clearly states that no restriction will be imposed on international current payments and transfers.
On June 20, 1996, the People's Bank of China (the "PBOC") promulgated the Regulations for the Administration of Settlement, Sale and Payment of Foreign Exchange (Yin Fa [1996] No. 210) (《結匯、售匯及付匯管理規定 (銀發[1996]210號)》), which abolished the remaining restrictions on convertibility of
foreign exchange under current items, while retaining the existing restrictions on foreign exchange transactions under capital items accounts.
According to the Announcement on Improving the Reform of the RMB (the PBOC Announcement [2005] No. 16) (《关于完善人民币汇率形成机制改革的公告（中国人民银行公告[2005]第16号）》), issued by the PBOC on July 21, 2005 and effective on the same date, the PRC began to implement a managed floating exchange rate system in which the exchange rate would be determined based on market supply and demand and adjusted with reference to a basket of currencies. The RMB exchange rate was no longer pegged to the USD. The PBOC would publish the closing price of the exchange rate of the RMB against trading currencies such as the USD in the interbank foreign exchange market after the closing of the market on each working day, as the central parity of the currency against RMB transactions on the following working day. On July 1, 2014, the PBOC further improved the formation mechanism of the RMB exchange rate by authorizing the China Foreign Exchange Trade System to make inquiries with the market makers before the interbank foreign exchange market opens every day for their offered quotations which are used as samples to calculate the central parity of the RMB against the USD on that day using the weighted average of the remaining market makers' offered quotations after excluding the highest and lowest quotations, and announce the central parity of the RMB against currencies such as the USD at 9:15 a.m. on each working day. On August 11, 2015, the PBOC announced to improve the central parity quotations of RMB against the USD by authorizing market makers to provide central parity quotations to the China Foreign Exchange Trading System before the interbank foreign exchange market opens every day with reference to the interbank foreign exchange market closing rate of the previous day, the supply and demand for foreign exchange as well as changes in major international currency exchange rates.
On August 5, 2008, the State Council promulgated the revised Foreign Exchange Control Regulations of the PRC, which have made substantial changes to the foreign exchange supervision system of the PRC. First, the regulations have adopted an approach of balancing the inflow and outflow of foreign exchange. Foreign exchange income received overseas can be repatriated or deposited overseas, and foreign exchange and settlement funds under the capital account are required to be used only for purposes as approved by the competent authorities and foreign exchange administrative authorities; second, the regulations have improved the RMB exchange rate floating system based on market supply and demand under management; third, in the event that international balance of payment suffer or may suffer a material misbalance, or the national economy encounters or may encounter a severe crisis, the State may adopt necessary safeguard or control measures against international balance of payment; fourth, the regulations have enhanced the supervision and administration of foreign exchange transactions and grant extensive authorities to SAFE to enhance its supervisory and administrative powers.
According to the relevant laws and regulations in the PRC, PRC enterprises which need foreign exchange for current item transactions may, without the approval of the foreign exchange administrative authorities, effect payment through foreign exchange accounts opened at designated banks that carry foreign exchange business, on the strength of valid receipts and proof. Foreign investment enterprises which need foreign exchange for the distribution of profits to their shareholders and PRC enterprises which, in accordance with regulations, are required to pay dividends to their shareholders in foreign exchange may, after paying taxes in according to the law, on the strength of resolutions of the board of directors on the distribution of profits, effect payment from foreign exchange accounts opened at designated banks that carry foreign exchange business, or effect exchange and payment at designated banks.
The Decisions on Matters including Canceling and Adjusting a Batch of Administrative Approval Items (Guo Fa [2014] No. 50) (《关于取消和调整一批行政审批项目等事项的决定（国发[2014]50号）》) promulgated by the State Council and came into effect on October 23, 2014 provide to cancel the approval requirement of SAFE and its branches for the remittance and settlement of the proceeds raised from the overseas listing of the foreign shares into RMB domestic accounts.
根据国家外汇管理局颁布并于2014年12月26日生效的《关于境外上市外汇管理有关问题的通知（汇发[2014]54号）》，境内公司应在境外上市发行结束之日起15个工作日内，向注册地所在地的国家外汇管理局分支机构办理境外上市登记手续；境内公司境外上市所得资金可以调回境内，也可以存放境外，但资金用途应与募集说明书或其他公开披露文件的内容相符。境内公司（银行类金融机构除外）应持境外上市证明文件，在境内银行开立专用外汇账户，用于首次公开发行（或后续发行）及回购业务，办理相关业务的结汇、汇兑及资金划转。
根据国家外汇管理局于2015年2月13日颁布、2015年6月1日起生效并于2019年12月30日部分废止的《关于进一步简化和改进直接投资外汇管理政策的通知（汇发[2015]13号）》，境内直接投资外汇登记确认及境外直接投资外汇登记确认均由银行直接审核办理。国家外汇管理局及其分支机构通过银行对直接投资外汇登记实施间接管理。
根据国家外汇管理局颁布并于2023年12月4日最后一次修订的《关于改革和规范资本项目结匯管理政策的通知（汇发[2016]16号）》，已明确实施意愿结汇政策的资本项下外汇收入（包括境外上市募集资金的调回），境内机构可根据实际业务需要在银行办理结汇。境内机构资本项下外汇收入意愿结汇比例暂定为100%，国家外汇管理局可根据国际收支状况适时调整。
根据国家外汇管理局颁布并于2020年4月10日起生效的《国家外汇管理局关于优化外汇管理支持涉外业务发展的通知（汇发[2020]8号）》，符合条件的企业使用资本金、外债资金及境外上市资本项下收入在境内进行支付时，无需事先逐笔向银行提交真实性审核材料，但须确保所用资金真实合规，并符合现行资本项下收入使用相关管理规定。相关银行应按照审慎经营原则管控相关业务风险，并依据相关要求事后开展抽查。地方外汇管理部门应加强监测分析，强化事中事后监管。
This Appendix sets out summaries of certain aspects of PRC laws and regulations, which are relevant to the Company's operations and business. Laws and regulations relating to taxation in the PRC are discussed separately in "Appendix III — Taxation and Foreign Exchange" to this prospectus. The principal objective of this summary is to provide potential investors with an overview of the principal PRC legal and regulatory provisions applicable to the Company. This summary is not intended to include all the information which may be important to potential investors. For more details on laws and regulations which are relevant to our business, please refer to the section headed "Regulatory Overview" in this prospectus.
中国法律制度以《中华人民共和国宪法》（以下简称"宪法"）为基础，由成文法律、行政法规、地方性法规、自治条例、单行条例、国务院部门规章、地方政府规章、特别行政区法律、中国政府签署的国际条约及其他规范性文件共同构成。法院判决不构成具有法律约束力的判例，但可作为司法参考与指导。
根据《中华人民共和国宪法》及《中华人民共和国立法法》，全国人民代表大会（"全国人大"）及其常务委员会被授权行使国家立法权。全国人大有权制定和修改有关国家机构以及民事、刑事和其他事项的基本法律。全国人民代表大会常务委员会（"全国人大常委会"）被授权制定和修改除应由全国人大制定的法律以外的其他法律，并在全国人大闭会期间对全国人大制定的法律进行补充和修改，但上述补充和修改不得与该等法律的基本原则相抵触。
国务院是最高国家行政机关，有权根据宪法和法律制定行政法规。
省、自治区、直辖市的人民代表大会及其常务委员会可根据本行政区域的具体情况和实际需要制定地方性法规，但上述地方性法规不得与宪法、法律或行政法规的任何规定相抵触。设区的市的人民代表大会及其常务委员会可根据本市的具体情况和实际需要，就城乡建设与管理、生态文明建设、历史文化保护等方面制定地方性法规，报所在省、自治区人民代表大会常务委员会批准后施行，但上述地方性法规不得与宪法、法律、行政法规及所在省、自治区的地方性法规的任何规定相抵触。
国务院各部、委员会、中国人民银行、审计署和具有行政管理职能的直属机构，可在本部门的权限范围内，根据法律和行政法规以及国务院的决定、命令制定规章。省、自治区、直辖市和设区的市的人民政府，可根据法律、行政法规和本省、自治区、直辖市的地方性法规制定规章。
宪法具有最高法律效力，任何法律、行政法规、地方性法规、自治条例和单行条例均不得与宪法相抵触。
法律的效力高于行政法规、地方性法规和规章。行政法规的效力高于地方性法规和规章。省或自治区的人民政府制定的规章的效力高于该省和自治区内设区的市及自治州的人民政府制定的规章。
全国人民代表大会有权改变或撤销全国人民代表大会常务委员会制定的不适当的法律，以及撤销经其常务委员会批准但违反《中华人民共和国宪法》或《中华人民共和国立法法》的自治条例和单行条例。全国人民代表大会常务委员会有权撤销违反《中华人民共和国宪法》和法律的行政法规，撤销违反《中华人民共和国宪法》、法律或行政法规的地方性法规，以及撤销经有关省、自治区或直辖市人民代表大会常务委员会批准但违反《中华人民共和国宪法》和《中华人民共和国立法法》的自治条例和地方性法规。国务院有权改变或撤销不适当的部门规章和地方政府规章。省、自治区或直辖市的人民代表大会有权改变或撤销其各自常务委员会制定或批准的不适当的地方性法规。地方人民代表大会常务委员会有权撤销同级人民政府制定的不适当的规章。省和自治区的人民政府有权改变或撤销下级人民政府制定的不适当的规章。
根据《中华人民共和国宪法》，法律解释权归属于全国人民代表大会常务委员会。依据1981年6月10日通过的《全国人民代表大会常务委员会关于加强法律解释工作的决议》，凡关于法律、法令条文本身需要进一步明确界限或作补充规定的问题，由全国人民代表大会常务委员会进行解释或规定；凡属于法院审判工作中具体应用法律、法令的问题，由最高人民法院进行解释；凡属于检察院检察工作中具体应用法律、法令的问题，由最高人民检察院进行解释；凡属于上述以外的其他法律问题，由国务院及主管部门进行解释。最高人民法院和最高人民检察院的解释如果有原则性分歧，应报请全国人民代表大会常务委员会解释或裁定。国务院及其各部委亦有权对其颁布的行政法规和部门规章进行解释。在地方层面，地方法律的解释权归属于颁布该等法律的地方立法和行政机关。
根据《中华人民共和国宪法》及最近于2018年10月26日修订、2019年1月1日起施行的《中华人民共和国人民法院组织法》，人民法院分为最高人民法院、地方各级人民法院及其他专门人民法院。地方各级人民法院分为三个层级，即基层人民法院、中级人民法院和高级人民法院。基层人民法院下设民事、刑事和经济审判庭。中级人民法院设有与基层人民法院类似的机构及其他专门法庭，如知识产权法院、军事法院和海事法院。上述两级人民法院均受上级人民法院监督。最高人民检察院有权对各级人民法院已依法生效的判决和裁定实施监督，上级人民检察院有权对下级人民法院已依法生效的判决和裁定实施监督。最高人民法院是中华人民共和国最高审判机关，负责监督各级人民法院的审判工作。
The Guidance for Articles of Association of Companies Seeking Listing Overseas (到境外上市公司章 程必備條款) (the "Guidance for Articles of Association") which was revised on September 28, 2023 and came into effect on the same date.
The PRC Company Law (中華人民共和國公司法) was promulgated on December 29, 2023 and took effect on July 1, 2024. The PRC Company Law sets out the rules and regulations governing the establishment and the corporate governance of companies and the protection of the lawful rights and interests of companies, shareholders and creditors in the PRC.
The PRC Company Law provides that companies established within the PRC are either limited liability companies or joint stock limited companies. A limited liability company is established when all of its shareholders contribute their respective capital to the company in accordance with the relevant provisions of the PRC Company Law. A joint stock limited company may be established by way of promotion or by way of share subscription, with its registered capital being divided into shares of equal value. The total amount of shares subscribed to by its promoters or the total amount of shares which has been paid up constitutes the company's registered capital.
As a general rule, shareholders of a limited liability company are liable to the company only to the extent of their respective capital contributions, and shareholders of a joint stock limited company are liable to the company only to the extent of the shares held by them. A company is liable for its debts to the extent of all its assets.
The PRC Company Law provides that the valid period for subscription of shares and the dates of payment in the case of a joint stock limited company established by way of share subscription shall be specified in the prospectus. If the shares are not fully subscribed within the valid period specified, or if the subscribers fail to pay up the price for the subscribed shares within the stipulated dates of payment, the company shall not be established. Before the establishment of a company, no public offering of shares shall be made in the name of the company to be established. After a joint stock limited company is established, it shall not issue new shares below their par value.
The PRC Company Law provides that the articles of association of a company shall include information such as the name and domicile of the company, the scope of its business, its registered capital, the name of shareholders and the percentage of capital contribution of each shareholder (in the case of a limited liability company) or the total number of shares, par value per share and the classes of shares (in the case of a joint stock limited company), the structure of the company's organs and the rules of procedure and voting procedures thereof, the company's legal representative, and any other matters required by the law and administrative regulations.
The PRC Company Law provides that a joint stock limited company shall have a board of directors consisting of five to nineteen members. The board of directors of a joint stock limited company shall be accountable to the shareholders' general meeting. Among other things, the board of directors shall be responsible for convening shareholders' general meetings and reporting to the shareholders' general meetings; implementing resolutions of the shareholders' general meetings; deciding on the company's business operation plans and investment plans; formulating the company's annual financial budget and final accounts; formulating the company's profit distribution plans and plans for making up losses; formulating plans for the increase or reduction of the company's registered capital and for the issuance of corporate bonds; and formulating plans for significant acquisitions, repurchases of the company's shares, merger, division, dissolution or change in corporate form of the company.
The PRC Company Law provides that a joint stock limited company shall have a manager who is appointed or removed by the board of directors. The manager is accountable to the board of directors and shall exercise the following functions and powers, among others: to be in charge of the production, operation and management of the company, and to organize the implementation of resolutions of the board of directors; to organize the implementation of the company's annual business operation plans and investment plans; to draft the company's basic management system; and to formulate the specific rules and regulations of the company.
The PRC Company Law provides that a joint stock limited company shall have a board of supervisors consisting of no fewer than three members. The board of supervisors shall include representatives of shareholders and an appropriate proportion of representatives of the company's employees, with the proportion of employee representatives not less than one-third of the total, as stipulated by the articles of association. The board of supervisors shall be accountable to the shareholders' general meeting. Among other things, the board of supervisors shall be responsible for examining the company's financial affairs; supervising the acts of the directors and senior management in performing their duties and making proposals for the removal of directors and senior management members who violate the law, administrative regulations, the articles of association or resolutions of the shareholders' general meetings; requiring directors and senior management to rectify any act that damages the interests of the company; and proposing the convening of extraordinary shareholders' general meetings.
Alternatively, a joint stock limited company may have an audit committee under the board of directors in lieu of a board of supervisors, with functions and powers as specified in the PRC Company Law.
The PRC Company Law provides that a joint stock limited company shall hold an annual shareholders' general meeting once a year. Extraordinary shareholders' general meetings shall be convened within two months after the occurrence of any of the following: (i) the number of directors falls below the number required by the PRC Company Law or two-thirds of the number required by the articles of association; (ii) the company's losses that are not made up reach one-third of the company's total paid-up share capital; (iii) shareholders holding ten percent or more of the company's issued and outstanding voting shares request the convening of such a meeting; (iv) the board of directors deems it necessary; (v) the board of supervisors (or, if no board of supervisors has been established, the audit committee) proposes to convene; or (vi) any other circumstances as required by the articles of association.
The PRC Company Law provides that a shareholder of a joint stock limited company may transfer his/her/its shares in accordance with the law. The transfer of shares listed on a stock exchange shall be carried out in accordance with the relevant rules of the stock exchange concerned.
The PRC Company Law provides that after a company has paid income tax in accordance with the provisions of the law, its after-tax profits shall be distributed in the following order: (i) making up the company's losses; (ii) allocating to the company's statutory reserve fund (no less than 10% of after-tax profit, until the aggregate amount of the fund reaches 50% or more of the company's registered capital); (iii) allocating to the company's discretionary reserve fund (if the shareholders' general meeting resolves to do so); and (iv) distributing dividends to shareholders.
The PRC Company Law provides that where a company merges with or separates from another company, it shall prepare a balance sheet and an inventory of assets. The company shall notify its creditors within ten days from the date of the resolution made by the shareholders' general meeting on the merger or separation, and shall make a public announcement in a newspaper within thirty days. Creditors may, within thirty days from the date of receipt of the notice or, in the case of creditors who did not receive the notice, within forty-five days from the date of the public announcement, demand that the company settle the debts or provide corresponding guarantees.
The Trial Administrative Measures of Overseas Securities Offering and Listing by Domestic Companies (境內企業境外發行證券和上市管理試行辦法) (the "Overseas Listing Trial Measures") and five relevant guidelines were promulgated by the CSRC on February 17, 2023 pursuant to the PRC Securities Law and are applicable to the direct and indirect overseas share offering or listing of domestic companies.
The Overseas Listing Trial Measures provide that domestic companies conducting overseas offering and listing activities, whether direct or indirect, shall complete the filing procedure with the CSRC. A "direct overseas offering and listing" refers to offering and listing of the stock and other equity securities by a joint stock limited company incorporated in the PRC directly overseas. An "indirect overseas offering and listing" refers to overseas offering and listing by an enterprise where the main business operations are undertaken by a domestic company, with such overseas offering and listing being conducted in the name of an overseas entity and using the overseas entity's securities. The Overseas Listing Trial Measures set out the specific criteria for determining whether an overseas offering and listing is classified as direct or indirect. Among other things, where the revenues, total profits, total assets or net assets of domestic enterprises contribute more than 50% of the corresponding items in the audited consolidated financial statements of the overseas listed entity in the most recent fiscal year, such overseas listing or offering may be deemed as indirect overseas offering and listing.
The Overseas Listing Trial Measures set out the specific requirements for filing and information disclosure, including but not limited to: (i) in respect of a filing for initial public offering and listing, the relevant domestic company shall submit its filing materials with the CSRC within three business days after its application for overseas listing is accepted by the overseas regulatory authority or the stock exchange; (ii) in respect of a filing for subsequent overseas securities offering, the relevant domestic company shall submit its filing materials with the CSRC within three business days after the completion of each overseas securities offering; and (iii) in respect of material events, the relevant domestic company shall make an immediate report to the CSRC if there are any material events occurring after the initial public offering and before the listing of the company's securities overseas, or for a listed company, if there are any material events as specified under the Overseas Listing Trial Measures.
The Overseas Listing Trial Measures provide that the CSRC will not grant approval to an overseas offering and listing in any of the following circumstances: (i) the relevant laws and regulations of the PRC explicitly prohibit overseas listing; (ii) the applicant poses serious national security concerns which have been confirmed by the competent national security authority; (iii) in the past three years, the applicant has committed any serious violation of the PRC laws or regulations, or has committed any serious breach of integrity; (iv) in the past three years, the applicant's directors, supervisors or senior management members have been convicted of corruption, bribery, misappropriation of property, theft or any other criminal offence that disrupts the social and economic order, or have been subject to administrative penalties for any serious illegal activities; or (v) any other circumstances as recognized by the State Council.
The Guidance for Articles of Association of Companies Seeking Listing Overseas (到境外上市公司章程必備條款) (the "Guidance for Articles of Association") was revised on September 28, 2023 and came into effect on the same date. The Guidance for Articles of Association provides the mandatory provisions which are required to be incorporated into the articles of association of a company seeking an overseas listing, covering mainly the following areas: (i) general provisions (including the company's name, domicile, scope of business and registered capital); (ii) shares (including share issuances and transfers); (iii) shareholders and shareholders' general meetings (including shareholders' rights, convening and procedures of shareholders' general meetings and powers and functions of shareholders' general meetings); (iv) directors, board of directors, supervisors, board of supervisors (or audit committee) and senior management; (v) financial affairs and accounting; (vi) notifications and announcements; and (vii) merger, division, dissolution and liquidation.
《上市公司章程指引》（Guidelines for Articles of Association of Listed Companies）（以下简称"章程指引"）于2025年3月28日经中国证监会最新修订。公司章程参照章程指引制定，其摘要载于本招股说明书"附录五——公司章程摘要"一节。
以下为目前有效的《中华人民共和国公司法》、《境外上市试行办法》及《章程指引》中适用于本公司的主要条款摘要。
股份有限公司是指依照《中华人民共和国公司法》在中国设立的企业法人，其注册资本划分为股份。公司的全部股份应当依据公司章程的规定采用有面值股份或无面值股份。采用有面值股份的，每股金额相等。公司以其全部资产对公司的债务承担责任，股东以其认购的股份为限对公司承担责任。
公司应当依照法律、行政法规开展经营活动。公司可以向其他有限责任公司及股份有限公司投资，并以其出资额为限对所投资公司承担责任。除法律另有规定外，公司不得成为对所投资公司债务承担连带责任的出资人。
公司可以采取发起设立或募集设立的方式设立。设立公司的发起人最少为一人，最多不超过二百人，且发起人中须有半数以上在中国境内有住所。发起设立是指由发起人认购公司应发行的全部股份而设立公司。发起设立的公司在注册资本未缴足前，不得向他人发行股份。法律、行政法规及国务院决定对实缴注册资本及最低注册资本另有规定的，从其规定。
采取发起设立方式设立公司的，发起人应当书面认购其所认购的股份，并依照章程规定缴纳出资。
协会。如以非货币资产出资，应依法办理财产权的转移手续。发起人未按照前述规定缴纳出资的，应按照发起人协议中的约定承担违约责任。发起人确认章程规定的出资后，须选举董事会和监事会（法律法规另有规定的除外），由董事会授权的代表向公司登记机关申请设立登记，并提交章程及法律或行政法规要求的其他文件。
以募集方式设立公司的，发起人认购的股份不得少于公司股份总数的35%，但法律、行政法规另有规定的除外。发起人应当在股款缴足之日起三十日内主持召开创立大会。创立大会由发起人和认股人组成。招股说明书规定的认股期限届满，所发行的股份未能全部认足的，或者发起人在股款缴足之日起三十日内未能召开创立大会的，认股人可以要求发起人返还所缴股款，并按照银行同期存款利率加算利息。创立大会结束后三十日内，由董事会授权的代表向登记机关申请公司设立登记。公司在完成相关市场监督管理部门登记并取得营业执照后，正式成立并取得法人资格。
发起人可以货币出资，也可以用实物、知识产权、土地使用权等可以用货币估价并可以依法转让的非货币财产作价出资；但是，法律、行政法规规定不得作为出资的财产除外。以非货币财产出资的，应当依照法律、行政法规的规定进行评估作价，不得高估或者低估。
《中华人民共和国公司法》对单个股东持有公司股份的比例不设上限。公司股份以股票为表现形式。股票是公司签发的证明股东所持股份的凭证。公司签发的股票应采用记名股票形式。
股份的发行，应当公平、公正。同次发行的同种类股票，每股的发行条件和价格应当相同；任何单位或者个人所认购的股份，每股应当支付相同价额。股票发行价格可以等于票面金额，也可以超过票面金额，但不得低于票面金额。
根据《境外上市试点措施》，境内企业在境外发行股票的，可以以外币或人民币募集资金及进行股息分配。
(iv) 各股东取得股票的日期。
鉴于其经营和发展需要，并依照法律法规，公司可在股东大会通过决议的前提下，通过以下任何方式增加股本：(i) 公开发行股份；(ii) 非公开发行股份；(iii) 向现有股东发放红股；(iv) 将公积金转换为股份；及 (v) 法律、行政法规规定并经中国证监会批准的其他方式。
根据《中华人民共和国公司法》，公司可依据其章程发行以下与普通股权利不同的分类股份：(i) 在利润分配或剩余财产分配方面享有优先权或劣后权的股份；(ii) 每股表决权多于或少于普通股的股份；(iii) 转让须经公司同意或受其他限制的股份；(iv) 国务院规定的其他分类股份。公司公开发行股份时，不得发行上述第(ii)项和第(iii)项所规定的分类股份，但公开发行前已发行的除外。公司发行新股时，须依据章程在股东大会上就新股的种类和数量、新股的发行价格、新股发行的起止日期，以及拟向现有股东发行新股时该等新股的种类和数量作出决议。
境内公司向境外发行股份的，应当在向境外提交发行上市申请文件后三个工作日内，将发行上市的申请文件向中国证监会报备。
(v) 公司须向有关公司登记机关申请办理变更登记及注册资本减少登记。
(vi) 上市公司为维护公司价值及股东权益所必要时。
to review and approve the guarantee matters of the company as required by laws and regulations or the articles of association;
to review and approve the significant acquisition, disposal or write-off of the company's assets as required by laws and regulations or the articles of association; and
any other matters that shall be decided by the shareholders' meeting as required by laws and regulations or the articles of association.
The shareholders' meeting shall be convened by the board of directors and presided over by the chairman of the board. Where the chairman of the board is unable or fails to perform his/her duties, the meeting shall be presided over by the vice chairman. Where the vice chairman is also unable or fails to perform his/her duties, a director jointly elected by more than half of the directors shall preside over the meeting.
The shareholders' meeting shall be called and held at least once a year. An extraordinary shareholders' meeting shall be convened within two months when any of the following circumstances occur:
(i) the number of directors is less than the number required by the PRC Company Law or less than two thirds of the number specified in the articles of association;
(iii) shareholders individually or in aggregate holding 10% or more of the company's shares with voting rights request the convening of an extraordinary general meeting;
(vi) any other circumstances as required by laws, administrative regulations or the articles of association.
行使章程规定的其他职权。
股东大会可授权董事会就发行公司债券作出决议。
章程规定的其他情形。
股东大会由董事会召集，董事长主持。董事长不能履行职务或不履行职务的，由副董事长主持。副董事长不能履行职务或不履行职务的，由半数以上董事共同推举一名董事主持。董事会不能履行或不履行召集股东大会职责的，监事会应及时召集和主持。监事会不召集和主持的，连续九十天以上单独或合计持有公司百分之十以上股份的股东可自行召集和主持。单独或合计持有公司百分之十以上股份的股东请求召开临时股东大会的，董事会及监事会须于收到请求后十日内决定是否召开临时股东大会，并以书面形式答复股东。
根据中国《公司法》，股东大会通知须在会议召开二十日前发给全体股东，说明会议日期、地点及议题。临时股东大会通知须在会议召开十五日前发给全体股东。
中国《公司法》对股东大会出席人数构成法定人数并无专项规定。
根据中国《公司法》，出席股东大会的股东（不包括分类股东）每持有一股享有一票表决权，但公司持有的本公司股份不享有表决权。
根据公司章程的规定或股东大会的决议，股东大会在选举董事和监事时可以实行累积投票制。在累积投票制下，每一股份拥有与股东大会所选举的董事或监事人数相当的投票权，股东在投票时可以将其所有投票权集中投票给一名或数名董事或监事候选人。
根据《中华人民共和国公司法》，股东大会的决议须经出席会议的股东所持表决权过半数通过，但涉及公司合并、分立、解散、增减注册资本、变更公司形式或修改公司章程的决议除外，上述事项须经出席会议的股东所持表决权的三分之二以上通过。在《中华人民共和国公司法》及公司章程规定公司转让或购置重大资产或对外提供担保须经股东大会决议批准的情况下，董事会应及时召开股东大会对该等事项进行表决。
股东可以委托代理人出席股东大会，股东应明确代理事项、权限及期限。代理人应向公司出示股东授权委托书，并在授权范围内行使表决权。
股东大会审议的事项应当制作会议记录，出席会议的董事长和董事应在会议记录上签字确认。会议主持人及出席会议的董事应在会议记录上签字确认。会议记录应当与股东出席登记册及授权委托书一并妥善保管。
股份有限公司应设立董事会，董事会成员不得少于三人，但规模较小或股东人数较少的股份有限公司可设一名董事而不设立董事会。对于员工人数在三百人及以上的公司，除非已设立监事会并已包含职工代表监事，否则董事会应当包括职工代表。董事的任期应在公司章程中予以规定，但每届任期不得超过三年。董事经选举可连任。若任期届满未能及时改选，或因董事辞职导致董事会成员人数低于法定人数，董事应依照法律、行政法规及公司章程的规定，继续履行其董事职责，直至新当选的董事就任为止。
(x) 行使公司章程规定的或股东大会授权的其他职权。
章程中规定的对董事会权力的任何限制，不得对抗任何善意第三方。
董事会会议每年至少召开两次。会议通知应在会议召开前十日送达全体董事和监事。持有10%或以上表决权的股东、三分之一或以上的董事或监事会可提议召开临时董事会会议。董事长应在收到提议后十日内召集会议并主持会议。董事会可另行确定召开临时董事会会议的方式及通知期限。董事会会议须有过半数董事出席方可举行。董事会决议须经全体董事过半数通过。每位董事就须经董事会批准的决议享有一票表决权。董事应亲自出席董事会会议。如董事因故不能出席，可以书面委托书委托其他董事代为出席会议，委托书须载明授权范围。董事会应就会议所议事项的决定制作会议记录，出席会议的董事应在会议记录上签字。
如董事会决议违反任何法律、行政法规或章程或股东会决议，并因此致使公司遭受严重损失，参与该决议的董事应承担赔偿责任。但如能证明某董事在表决时明确表示反对该决议，且该反对意见已记录于会议记录中，则该董事可免除相应责任。
(v) 因未偿还大量到期未清偿债务而被人民法院列为失信被执行人的人。
Where a company elects or appoints a director to which any of the above circumstances applies, such election or appointment shall be null and void. A director to which any of the above circumstances applies during his/her term of office shall be released of his/her duties by the company.
Pursuant to the PRC Company Law, the board of directors shall appoint a chairman and may appoint a vice chairman. The chairman and the vice chairman shall be elected with approval of more than half of all the directors. The chairman shall convene and preside over board meetings and review the implementation of board resolutions. The vice chairman shall assist the chairman to perform his/her duties. Where the chairman is incapable of performing or is not performing his/her duties, the duties shall be performed by the vice chairman. Where the vice chairman is incapable of performing or is not performing his/her duties, a director elected by more than half of the directors shall perform his/her duties.
The board of supervisors shall consist of representatives of the shareholders and an appropriate proportion of representatives of the company's staff, among which the proportion of representatives of the company's staff shall not be less than one-third, and the actual proportion shall be determined in the articles of association. Representatives of the company's staff at the board of supervisors shall be democratically elected by the company's staff at the staff representative assembly, general staff meeting or otherwise. The board of supervisors shall appoint a chairman and may appoint a vice chairman. The chairman and the vice chairman of the board of supervisors shall be elected by more than half of the supervisors. Directors and senior management shall not act concurrently as supervisors.
The chairman of the board of supervisors shall convene and preside over board of supervisors meetings. Where the chairman of the board of supervisors is incapable of performing or is not performing his/her duties, the vice chairman of the board of supervisors shall convene and preside over supervisory board meetings. Where the vice chairman of the board of supervisors is incapable of performing or is not performing his/her duties, a supervisor nominated by more than half of the supervisors shall convene and preside over meetings of the board of supervisors.
Each term of office of a supervisor is three years and he/she may serve consecutive terms if re-elected. A supervisor shall continue to perform his/her duties as a supervisor in accordance with the laws, administrative regulations and the articles of association until a duly re-elected supervisor takes office, if re-election is not conducted in a timely manner upon the expiry of his/her term of office or if the resignation of supervisors results in the number of supervisors being less than the quorum.
(ii) to supervise the directors and senior management in their performance of their duties and to propose the removal of directors and senior management who have violated laws, regulations, the articles of association or shareholders' resolutions;
(iii) when the acts of directors or senior management are detrimental to the company's interests, to require the director and senior management to correct these acts;
(iv) to propose the convening of extraordinary shareholders' meetings and to convene and preside over shareholders' meetings when the board fails to perform the duty of convening and presiding over shareholders' meetings under the PRC Company Law;
(vi) to bring actions against directors and senior management pursuant to the relevant provisions of the PRC Company Law; and
(vii) to exercise any other authority stipulated in the articles of association.
Supervisors may be present at board meetings and make inquiries or proposals in respect of the resolutions of the board. The board of supervisors may investigate any irregularities identified in the operation of the company and, when necessary, may engage an accounting firm to assist its work at the cost of the company.
Under PRC Company Law, a joint stock limited company may establish an audit committee composed of directors within its board of directors pursuant to the provisions of its articles of association to exercise the functions and powers of a supervisory committee as prescribed by PRC Company Law, in lieu of establishing a supervisory committee or supervisor.
The audit committee shall comprise at least three members, with a majority not holding any position in the company other than that of director and having no relationship with the company that may affect their independent and objective judgment. Employee representatives serving on the board of directors may be appointed as audit committee members.
For listed companies with audit committees, the following matters shall require approval by a majority of all audit committee members before being resolved by the board of directors:
(iv) other matters specified by the securities regulatory authority under the State Council.
审计委员会的决议须经其成员多数批准方可通过。
根据中国《公司法》，公司应设经理一名，由董事会聘任或解聘。经理应依照公司章程的规定或董事会的授权行使职权。
公司章程中关于经理职权的其他规定亦须遵守。
经理应出席董事会会议，但除非经理同时担任董事，否则在董事会会议上无表决权。
根据中国《公司法》，高级管理人员是指上市公司的经理、副经理、财务负责人、董事会秘书及公司章程规定的其他人员。
根据中国《公司法》，董事、监事及高级管理人员须遵守相关法律、法规及公司章程，并对公司负有忠实义务和勤勉义务。控股股东或实际控制人未担任公司董事但实际执行公司事务的，亦须遵守上述规定。
董事、监事及管理人员不得滥用职权收受贿赂或取得其他非法收入，不得侵占公司财产。
(vi) 其他违反对公司忠实义务的行为。
董事、监事及高级管理人员直接或间接与公司订立合同或进行交易的，须向董事会或股东大会报告，并依照公司章程取得董事会或股东大会的决议批准。董事、监事及高级管理人员的近亲属，或由其近亲属直接或间接控制的企业，以及与董事、监事及高级管理人员存在其他关联关系的人员订立合同或进行交易的，亦须适用上述规定。
董事、监事及高级管理人员不得利用职务便利为自己或他人谋取本应由公司获得的商业机会，但该行为已依照公司章程向董事会或股东大会报告并获批准，或公司依照适用法律、行政法规及公司章程无法取得该商业机会的除外。
董事、监事及高级管理人员不得为自己或他人经营与公司同类的业务，但该行为已依照公司章程向董事会或股东大会报告并获批准的除外。
董事或高级管理人员违反上述规定所获得的收入，应归还公司。
董事、监事或高级管理人员在执行职务时违反法律、法规或公司章程，给公司造成损失的，应当对公司承担赔偿责任。
《公司章程指引》规定，公司董事及高级管理人员对公司负有勤勉义务，例如，董事在行使公司授予的权力时应当谨慎、认真、勤勉，确保公司的业务活动符合国家法律、行政法规及各项经济政策的要求，且业务活动不超出公司营业执照所载明的经营范围；董事应当平等对待所有股东；股东应当随时了解公司的经营管理状况；董事及高级管理人员均应签署书面声明，确认公司的定期报告，并确保公司披露的信息真实、准确、完整；董事及高级管理人员均应向审计委员会提供准确的信息和材料，不得干预审计委员会履行职责；董事及高级管理人员还负有法律、行政法规、部门规章及公司章程规定的其他勤勉义务。
根据《中华人民共和国公司法》，公司应当依照法律、行政法规和国务院财政主管部门的规定，建立本公司的财务、会计制度。公司在每一会计年度终了时，应当依法编制财务会计报告，并依法经会计师事务所审计。财务会计报告应当依照法律、行政法规及国务院财政主管部门的规定编制。
公司的财务会计报告应当在召开年度股东大会的二十日前置备于公司，供股东查阅。向社会公开发行股票的股份有限公司，应当公告其财务会计报告。
公司在分配每年税后利润时，应当提取税后利润的10%列入公司法定公积金，直至法定公积金达到公司注册资本的50%为止。公司法定公积金不足以弥补以前年度亏损的，在依照规定提取法定公积金之前，应当先用当年利润弥补亏损。公司从税后利润中提取法定公积金后，经股东大会决议，还可以从税后利润中提取任意公积金。公司弥补亏损并提取上述公积金后，剩余的税后利润应当按照股东持有的股份比例进行分配，但公司章程规定不按持股比例分配的除外。
违反上述规定分配给股东的利润，必须退还公司。公司不得就其所持有的本公司股份分配利润。
公司发行股份所得超过票面价值的溢价，以及相关政府主管部门规定应当列入资本公积金的其他收入，均应计入资本公积金。公司的公积金用于弥补公司亏损、扩大公司业务经营或增加公司资本。公司需以公积金弥补亏损时，应当先使用任意公积金和法定公积金，仍不足的，可依照适用法规的规定使用资本公积金。
(i) the term of its operation set out in the articles of association has expired or other events of dissolution specified in the articles of association have occurred;
(iv) the business license of the company is revoked or the company is ordered to close down or to be dissolved in accordance with the laws; or
(v) the company is dissolved by a people's court in response to the request of shareholders holding shares that represent 10% or more of the voting rights of all shareholders of the company, on the grounds that the operation and management of the company has suffered serious difficulties that cannot be resolved through other means, rendering ongoing existence of the company a cause for significant losses to the shareholders' interests.
发生上述情形，公司应当在十日内通过国家企业信用信息公示系统予以公告。
发生第（一）、（二）项情形，公司通过修改公司章程可以继续存续的，在依法赔偿损失之前，不得对股东分配财产。依照前款规定修改公司章程，需要出席股东会议的股东所持表决权三分之二以上通过。
公司依照前款第（一）、（二）、（四）、（五）项规定解散的，应当在解散事由出现之日起十五日内成立清算组，开始清算。清算组由董事或者股东大会确定的人员组成。逾期不成立清算组进行清算的，债权人可以申请人民法院指定有关人员组成清算组进行清算。人民法院应当受理该申请，并及时组织清算组进行清算。
(vii) 代表公司参与民事诉讼活动。
清算委员会应自成立之日起十日内通知公司债权人，并于六十日内在报纸上或国家企业信用信息公示系统上发布公告。
债权人应在收到通知后三十日内，或未收到通知者应在公告发布之日起四十五日内，向清算委员会申报债权。
债权人在申报债权时，应陈述债权相关事项并提供相关证据。清算委员会应对申报的债权进行登记。在申报债权期间，清算委员会不得对债权人进行清偿。
处置公司财产并编制资产负债表及财产清单后，清算委员会应制定清算方案，并提交股东大会或人民法院确认。公司财产在支付清算费用、员工工资、社会保险费用及法定补偿、未缴税款及公司债务后，剩余财产按照股东持股比例分配给股东。
公司在清算期间继续存续，但不得开展与清算无关的经营活动。在按上述规定清偿完毕之前，公司财产不得向股东分配。
处置公司财产并编制资产负债表及财产清单后，若清算委员会发现公司财产不足以清偿其债务，则必须依法向人民法院申请宣告破产。人民法院宣告破产后，清算委员会应将管理事务移交给人民法院指定的破产管理人。
清算完成后，清算委员会应编制清算报告，提交股东大会或人民法院确认，并向公司登记机关申请注销公司登记，同时发布终止公告。清算委员会成员须忠实履行职责，遵守相关法律。清算委员会成员不得利用职权收受贿赂或其他非法收入，不得侵占公司财产。清算委员会成员因故意或重大过失给公司及其债权人造成损失的，应承担赔偿责任。
依法宣告破产的公司的清算，依照企业破产相关法律处理。
根据《境外上市试点办法》，在境外市场首次公开发行股票或上市的，应在向境外提交相关申请后三个工作日内向中国证监会备案。发行人在其已发行并上市的同一境外市场进行后续证券发行的，应在发行完成后三个工作日内向中国证监会备案。此外，若备案文件齐全且符合规定要求，中国证监会将在收到备案文件后二十个工作日内完成备案程序，并在中国证监会网站上公布备案结果。若备案文件不齐全或不符合规定要求，中国证监会应在收到备案文件后五个工作日内要求补充修改。发行人应在三十个工作日内完成补充修改。
持有记名股票的股东，如其股票证书遭到盗窃、遗失或毁损，可依照《中华人民共和国民事诉讼法》规定的公示催告程序，向人民法院申请宣告该股票证书无效。在获得该宣告后，股东可向公司申请补发股票证书。
公司合并可采取吸收合并或新设合并两种方式。采用吸收合并方式的，被吸收的公司解散；采用新设合并方式的，合并各方均告解散。
合并各方须订立合并协议，并编制资产负债表及财产清单。公司应自作出合并决议之日起十日内通知债权人，并于三十日内在报纸上或国家企业信用信息公示系统上刊登公告。债权人自接到通知之日起三十日内，未接到通知的自公告之日起四十五日内，有权要求公司清偿债务或提供相应担保；公司合并时，合并各方的债权、债务由存续公司或新设公司承继。
公司与持有其不低于90%股份的公司合并时，被合并公司无需经股东大会决议批准，但须通知其他股东，其他股东有权要求公司以合理价格收购其股权或股份。若公司合并所支付的价款不超过公司净资产的10%，除公司章程另有规定外，可无需经股东大会决议批准。在上述两种豁免股东大会批准的情形下，公司合并须经董事会决议批准。
公司分立时，须对公司资产进行分割，并编制资产负债表及财产清单。公司应自作出分立决议之日起十日内通知债权人，并于三十日内在报纸上或国家企业信用信息公示系统上刊登公告。公司分立前已发生的债务，由分立后的公司承担连带责任，但公司与债权人就债务清偿事项在分立前达成书面协议另有约定的除外。
中国已颁布一系列与股份发行、交易及信息披露相关的法规。1992年10月，国务院设立证券委员会及中国证券监督管理委员会（"中国证监会"）。证券委员会负责统筹协调起草证券法规、制定证券相关政策、规划证券市场发展，并对中国境内所有证券相关机构进行指导、协调和监督，同时对中国证监会实施管理。中国证监会是证券委员会的监管执行机构，负责起草证券市场监管规定、监督管理证券公司、监管中国公司在境内外的证券公开发行、监管证券交易活动，以及汇编证券相关统计数据并开展相关研究与分析。1998年4月，国务院将证券委员会与中国证监会合并，并对中国证监会进行改革重组。
《中华人民共和国证券法》（中華人民共和國證券法）是中国第一部全国性证券法律，监管事项包括证券的发行与交易、上市公司收购、信息披露、证券交易所、证券公司及证券监管机构的义务与责任等。《中华人民共和国证券法》对中国证券市场的活动进行了全面规范。
根据《中华人民共和国证券法》，境内企业直接或间接在境外发行证券或将其证券在境外上市交易，应当符合国务院的有关规定。目前，境外发行股份的发行与交易主要受国务院及中国证监会发布的法规和规章监管。
《中华人民共和国仲裁法》（中華人民共和國仲裁法）由全国人民代表大会常务委员会于1994年8月31日颁布，于1995年9月1日起施行，并于2017年9月1日最后修订。《中华人民共和国仲裁法》规定，仲裁委员会可在中国仲裁协会颁布仲裁规则之前，依照《中华人民共和国仲裁法》及《中华人民共和国民事诉讼法》制定临时仲裁规则。在当事人已同意通过仲裁方式解决争议的情况下，人民法院将拒绝受理一方当事人向其提起的诉讼，但仲裁协议无效的除外。
根据《中华人民共和国仲裁法》及《中华人民共和国民事诉讼法》，仲裁裁决为终局裁决，对仲裁当事各方均具有约束力。如任何一方不履行仲裁裁决，裁决的另一方当事人可向人民法院申请强制执行。
（六）仲裁员在仲裁该案时有贪污受贿、徇私舞弊、枉法裁决行为的。
人民法院认定执行该裁决违背社会公共利益的，裁决将不予执行。
任何一方当事人如寻求执行中国涉外仲裁机构针对不在中国境内或其财产不在中国境内的一方所作出的仲裁裁决，可向对该案有管辖权的外国法院申请承认和执行该裁决。同样，外国仲裁机构作出的仲裁裁决，可由中国法院依据互惠原则或中国缔结或参加的国际条约予以承认和执行。
This appendix contains a summary of the main provisions of the Articles of Association of the Company adopted on June 28, 2025, which will take effect from the date of listing of H shares on the Hong Kong Stock Exchange. The main purpose of this appendix is to provide potential investors with an overview of the Articles of Association of the Company, so it may not contain all the information that is important to potential investors.
The shares of the Company shall be issued in the form of share certificates.
All shares issued by the Company shall be shares with par value, and each share shall have a par value of RMB0.10.
The Company shall issue shares in an open, fair and just manner, and each share of the same class issued at the same time shall be issued under the same conditions and at the same price per share; subscribers shall pay the same price per share for the shares they subscribe for.
The ordinary shares issued by the Company comprise domestic shares and overseas listed shares (H shares). Domestic shares and overseas listed shares have the same rights in respect of any distribution in the form of dividends (including distributions in cash and in kind) or otherwise.
The Company may, based on its business and development needs and in accordance with the laws and regulations, increase its capital in the following manners upon resolutions being adopted by the shareholders' meetings:
(v) other means required by the laws, administrative regulations and approved by CSRC and Hong Kong Stock Exchange.
Subject to the provisions of laws, regulations and the securities regulatory rules of the jurisdiction where the Company's shares are listed, the shareholders' meeting may authorize the board of directors to decide, within three years, the issuance of shares not exceeding fifty percent of the already issued shares. If the board of directors decides to issue shares in accordance with the aforementioned provisions, resulting in changes to the Company's registered capital or the number of issued shares, the amendment to the relevant provisions of the articles of association regarding such matters shall not require further approval by the shareholders' meeting.
The Company may reduce its registered capital. The Company shall reduce its registered capital in accordance with the procedures stipulated in the Company Law, other relevant regulations of the PRC, the Listing Rules, other relevant regulations and the Articles of Association.
(iv) where the shareholders, who disagree with the resolution in relation to merger or division of the Company made at the shareholders' meeting, require the Company to repurchase the shares held by such shareholders;
(v) to use the shares for conversion of corporate bonds issued by the Company which are convertible into shares;
(vi) to safeguard the value of the Company and the interests of the shareholders when necessary.
(vii) other circumstances approved by the laws, administrative regulations, the CSRC or Hong Kong securities regulatory authorities.
The Company may repurchase its shares by open centralized transaction method or other method approved by laws, administrative regulations, the Listing Rules and the CSRC as well as the securities regulatory authorities of the place where the shares of the Company are listed.
The repurchase of the Company's shares under the circumstances of utilizing its shares in employee stock ownership plans or share incentive, using the shares for conversion of corporate bonds issued by the Company which are convertible into shares and safeguarding the value of the Company and the interests of the shareholders when necessary, subject to compliance with the requirements of the Listing Rules and other securities regulatory rules of the place where the Company's shares are listed, shall be conducted by way of open and centralized trading.
When the Company repurchases its own shares under any of the circumstances specified in item (i) or (ii) mentioned above, a resolution adopted by shareholders' meeting is required. Where the Company repurchases its own shares pursuant to the provisions of items (iii), (v) and (vi), it shall be resolved by a resolution of a meeting of Board of Directors attended by at least two-thirds of the directors.
In the event that the Company repurchases its shares under the circumstances set out in item (i) thereof, the shares shall be cancelled within 10 days after the date of repurchase; where it falls under the circumstances set out in item (ii) or (iv) thereof, the shares shall be transferred or cancelled within 6 months; where it falls under the circumstances set out in items (iii), (v) and (vi) thereof, the total number of shares of the Company held by the Company shall not exceed 10% of the total number of shares issued by the Company and shall be transferred or cancelled within three years.
In the event of a repurchase of the Company's H Shares, the Company may elect either to cancel such shares forthwith or to hold them as treasury shares in compliance with the Listing Rules. Should the Board of Directors fail to designate the repurchased shares as treasury shares, such shares shall be cancelled.
All treasury shares shall be deposited in a segregated account within the Central Clearing and Settlement System which shall be clearly identified as containing treasury shares. The Company shall not exercise any rights attaching to treasury shares, including but not limited to voting rights, nor shall it declare or pay any dividends or other distributions in respect thereof. Subject to the provisions of these Articles of Association and the Listing Rules, the Company may dispose of any treasury shares on such terms and conditions as the Board of Directors may determine.
Where the Company cancels repurchased shares pursuant to these Articles of Association, it shall promptly apply to the relevant company registry for registration of the corresponding reduction in registered capital. The nominal value of all cancelled shares shall be deducted from the Company's registered capital.
TRANSFER OF SHARES Shares of the Company shall be transferred in accordance with the law. The shares of the Company issued before public offering shall not be transferred for one year from the date on which the Company's shares are listed and traded on a stock exchange. The Directors and senior officers of the Company shall report to the Company their shareholdings and changes thereof and shall not transfer more than 25% of the total number of their shares of the same class in the Company per annum during their terms of office as determined when they take office. The shares of the Company they hold shall not be transferred within one year from the date when the Company's shares are listed and traded on the stock exchange. The aforesaid persons shall not transfer their shares in the Company within half a year after they terminate service with the Company. If there are other provisions on the transfer restrictions of shares in laws, regulations or rules of CSRC, Hong Kong Stock Exchange or other securities regulatory authorities, such provisions shall prevail.
SHAREHOLDERS The Company shall establish a register of members in accordance with certificates from the securities registration and clearing organization. The register of members of the company is sufficient evidence of the shareholders' shareholdings in the Company. A shareholder shall enjoy the relevant rights and assume the relevant obligations in accordance with the class of shares he/she holds. Shareholders holding the same class of shares shall enjoy the same rights and assume the same obligations. The transfer and assignment of shares shall be registered in the register of members. In respect of the register of shareholders of overseas-listed H shares, the original register of members of shares listed in the Hong Kong Stock Exchange shall be maintained in Hong Kong.
(ii) the right to request, convene, preside over, attend or appoint a proxy to attend shareholders' meetings and to exercise the corresponding rights to speak and vote in accordance with the law;
(iii) the right to supervise the Company's business operations, to present proposals and to raise enquiries;
(iv) the right to transfer, give as a gift or pledge shares held by them in accordance with laws, administrative regulations and the Articles of Association;
(v) the right to inspect and duplicate the Articles of Association, the register of members, minutes of shareholders' meetings, resolutions of Board meetings and financial accounting reports. Qualified shareholders shall have the right to inspect the company's accounting books and accounting vouchers;
(vi) in the event of the termination or liquidation of the Company, the right to participate in the distribution of remaining assets of the Company in proportion to the shareholdings;
(vii) for shareholders who vote against any resolution adopted at the shareholders' meeting on the merger or division of the Company, the right to demand the Company to buy back their shares;
(viii) other rights under laws, administrative regulations, departmental rules, normative documents, listing rules of the places where the shares of the Company are listed and the Articles of Association.
Shareholders of the Company who abuse their shareholders' rights and cause losses to the Company or other shareholders shall be liable for compensation in accordance with the law. Where shareholders of the Company abuse the Company's position as an independent legal person and the limited liability of shareholders for the purposes of evading repayment of debts, thereby materially impairing the interests of the creditors of the Company, such shareholders shall be jointly and severally liable for the debts owed by the Company. The controlling shareholder or the de facto controller of the Company shall exercise their rights and perform their obligations in accordance with the laws, administrative regulations, the provisions of the CSRC and the stock exchanges to protect the interests of the company. Where the controlling shareholders and de facto controllers of the Company do not act as directors but actually carry out the Company's affairs, the provisions of the Articles of Association on fiduciary and diligence obligations of directors shall apply.
SHAREHOLDERS' MEETINGS The shareholders' meeting shall consist of all shareholders. The shareholders' meeting shall be the authority of power of the Company and shall exercise the following functions and powers in accordance with laws:
(i) to elect and change Directors who are not appointed as employee representatives, and decide on the remunerations of Directors;
(vi) to resolve on the merger, division, dissolution, liquidation or change in the form of the Company;
(viii) to resolve on the Company's appointment or dismissal of accounting firms that provide audit service to the Company;
(x) to consider the purchase or sale of significant assets by the company within one year that exceed 30% of the company's latest audited total assets;
(xi) Equity incentive plans and employee share schemes requiring approval by the shareholders' general meeting as prescribed under the Listing Rules;
(xii) to consider other matters which shall be resolved at the shareholders' meeting in accordance with laws, administrative regulations, departmental rules or the Articles of Association.
The shareholders' meeting may authorize the Board of Directors to make resolutions on issuance of company's bonds. The Company may issue shares or corporate bonds convertible into shares upon resolution of the shareholders' general meeting or resolution of the board of directors authorized by these Articles of Association or the shareholders' general meeting, the implementation of which shall comply with the provisions of laws, administrative regulations, the CSRC and the securities regulatory rules of the jurisdiction where the Company's shares are listed.
Save as otherwise provided by laws, administrative regulations, departmental rules or the securities regulatory rules of the jurisdiction where the Company's shares are listed, the aforementioned powers of the shareholders' general meeting shall not be delegated to the board of directors or any other institution or individual to exercise on its behalf.
Shareholders' meetings consist of annual shareholders' meetings and extraordinary shareholders' meetings. The annual shareholders' meeting shall be held once a year within six months after the end of the previous accounting year.
The Company shall convene an extraordinary shareholders' meeting within two months upon occurrence of the following events:
(i) when the number of Directors falls below the minimum requirement of the Company Law, or is less than two thirds of the number specified by the Articles of Association;
(ii) when the unrecovered losses of the Company amount to one third of the total amount of its paid-up share capital;
(iii) when shareholder(s) severally or jointly holding at least ten percent of the Company's voting shares request(s) to convene such meeting (including preferred shares with resumed voting rights);
(vi) other circumstances stipulated by laws, administrative regulations, departmental rules, the Listing Rules and the listing rules of the places where the shares of the Company are listed or the Articles of Association.
Shareholders' meeting shall be convened by the Board of Directors within the prescribed period. Independent directors shall have the right to propose to the Board of Directors to convene an extraordinary shareholders' meeting, provided that such a proposal is agreed to by a majority of all independent directors. For the aforesaid proposal, the Board of Directors shall, in accordance with laws, administrative regulations, the Listing Rules, the Articles of Association and other regulations where the shares of the Company are listed, give a written feedback on whether or not it agrees to hold an extraordinary shareholders' meeting within 10 days of receipt of the proposal. Where the Board of Directors agrees to hold an extraordinary shareholders' meeting, it will send out a notice thereon within 5 days after the relevant resolution of the Board of Directors is made. If the Board of Directors does not agree to hold an extraordinary shareholders' meeting, it shall state reasons.
The Audit Committee may propose to the Board of Directors to hold an extraordinary shareholders' meeting and shall put forward the proposal to the Board of Directors in written form. The Board of Directors shall, in accordance with laws, administrative regulations, the Listing Rules, other regulations where the shares of the Company are listed and the Articles of Association, give a written feedback on whether or not it agrees to hold an extraordinary shareholders' meeting within 10 days of receipt of the proposal.
Where the Board of Directors agrees to hold an extraordinary shareholders' meeting, it shall send out a notice thereon within 5 days after the relevant resolution of the Board of Directors is made; any change to the original proposal in the notice is subject to the consent of the Audit Committee. In case the Board of Directors refuses to convene an extraordinary shareholders' meeting, or does not give any response within 10 days upon receipt of the proposal, the Board of Directors shall be deemed to be unable or have failed to perform its duty to convene the shareholders' meeting, and the Audit Committee may convene and preside over the meeting by itself.
Shareholder(s) severally or jointly holding at least 10% of the voting shares of the Company, including preferred shares with resumed voting rights, shall be entitled to request the Board to convene an extraordinary shareholders' meeting, and shall put forward such request to the Board in written form. The Board shall, in accordance with laws, administrative regulations, the Listing Rules, other regulations where the shares of the Company are listed and the Articles of Association, inform in writing whether it agrees or disagrees to convene an extraordinary shareholders' meeting within 10 days upon receipt of the request.
If the Board of Directors agrees to convene an extraordinary shareholders' meeting, a notice for convening such meeting shall be issued within 5 days after the date of the resolution of the Board of Directors and any changes to the original proposal contained in the notice shall be subject to the approval of the relevant shareholders.
If the Board of Directors does not agree to convene such meeting, or fails to give a response within 10 days after receipt of the request, shareholders holding at least 10% of the voting shares of the Company (including preferred shares with resumed voting rights) separately or in aggregate shall have the right to propose to the Audit Committee to convene an extraordinary shareholders' meeting, and shall put forward such request to the Audit Committee in writing.
If the Audit Committee agrees to convene an extraordinary shareholders' meeting, a notice for convening such meeting shall be issued within 5 days after receipt of the request and any changes to the original proposal contained in the notice shall be subject to the approval of the relevant shareholders.
If the Audit Committee fails to issue a notice convening the shareholders' meeting by the prescribed period, the Audit Committee shall be deemed to refuse to convene and preside over such meeting, and shareholders holding at least 10% of the voting shares of the Company (including preferred shares with resumed voting rights) separately or in aggregate for no less than 90 consecutive days shall have the right to convene and preside over the meeting on their own.
The proposals put forward to the shareholders' meetings shall fall within the scope of functions and powers of the shareholders' meeting, have clear issues for discussion and specific matters to be resolved, and comply with the laws, administrative regulations, the Listing Rules, other regulations where the shares of the Company are listed and the Articles of Association.
When the Company convenes a shareholders' meeting, the Board of Directors, the Audit Committee and shareholders holding 1% or more of the shares of the Company (including preferred shares with resumed voting rights) separately or in aggregate shall be entitled to put forward proposals to the Company.
Shareholders individually or jointly holding 1% or more of the shares of the Company (including preferred shares with resumed voting rights) may submit ad hoc proposals which shall contain a clearly defined agenda and specific matters for resolution, to the convener of a shareholders' meeting in writing ten days prior to shareholders' meeting. The convener shall issue a supplementary notice of the shareholders' meeting to announce the information of such ad hoc proposals within 2 days after receipt thereof, and submit such proposals to the shareholders' meetings.
Except as provided in the preceding paragraph, the convener of a shareholders' meeting shall not amend the proposals set out in the notice of the shareholders' meeting or put up any new proposals after the issuance of the notice of the shareholders' meeting.
The Company shall convene an annual shareholders' meeting by notifying the shareholders in writing 20 days prior to the meeting and an extraordinary shareholders' meeting by notifying the shareholders in writing 15 days prior to the meeting.
When calculating the time limit of the notice, the date of the shareholders' meeting convened shall not be included but the issue date of such notice shall be included. Where otherwise provided by laws, regulations or the securities regulatory authorities of the jurisdiction where the Company's shares are listed, such provisions shall prevail.
After the notice of the shareholders' meeting is issued, the meeting shall not be postponed or cancelled and the proposals set out in the notice shall not be cancelled without proper reasons. In the case of any postponement or cancellation of the meeting, the convenor shall make a supplementary notice at least two working days prior to the original date of the convening and state the reasons therefor.
All ordinary shareholders (including preferred shareholders with resumed voting rights) and preferred shareholders with weighted voting rights registered on the record date or their proxies shall be entitled to attend the shareholders' meeting. They shall exercise their voting rights in accordance with the relevant laws and regulations, listing rules of the places where the shares of the Company are listed and the Articles of Association of the Company.
Shareholder may attend the shareholders' meeting in person, or appoint a proxy to attend and vote on his/her behalf.
A shareholders' meeting shall be presided over by the chairman of the Board of Directors. If the chairman is unable or fails to discharge his/her duties, half or more of the directors shall designate a director to preside over the meeting.
If a shareholders' meeting is convened by the Audit Committee, the chairman of the Audit Committee shall preside over the meeting. If the convenor of the Audit Committee is unable or fails to discharge his/her duties, half or more of the Audit Committee members shall designate a member of the Audit Committee to preside over the meeting.
If a shareholders' meeting is convened by the shareholders themselves, the convener or a representative nominated by the convener shall preside over the meeting.
When a shareholders' meeting is convened, if the presider of the meeting contravenes the rules of procedure, rendering the meeting impossible to proceed, with the consent from more than half of the attending shareholders with voting rights, one person may be nominated at the shareholders' meeting to serve as the presider and the meeting may proceed.
Individual shareholders attending a shareholders' meeting in person shall produce their identity cards or other valid proof or evidence of their identities, and in the case of attendance by proxies, the proxies shall produce valid proof of their identities and the proxy forms from shareholders.
For a corporate shareholder, its legal representative or a proxy appointed by such legal representative shall attend the shareholders' meeting. In the case of attendance by legal representatives, they shall produce their identity cards and valid proof of their capacities as legal representatives and, in the case of attendance by proxies of such legal representatives, such proxies shall produce their identity cards and the letters of authorization issued by such legal representatives according to the laws.
Resolutions of the shareholders' meetings shall be divided into ordinary resolutions and special resolutions.
Ordinary resolution at a shareholders' meeting shall be adopted by more than one half of the voting rights held by shareholders (including their proxies) attending the shareholders' meeting. Special resolution at a shareholders' meeting shall be adopted by at least two thirds of the voting rights held by shareholders (including their proxies) attending the shareholders' meeting.
(ii) the profit distribution plans and plans for making up losses drafted by the Board of Directors;
(iii) the dismissal and remuneration of the members of the Board of Directors and the method of payment of the remuneration;
(iv) to resolve on the Company's appointment or dismissal of accounting firms that provide audit service to the Company;
(v) to consider and approve the transactions, financial assistance and guarantees which shall be approved at the shareholders' meeting;
(vi) equity incentive plans and employee share schemes requiring approval by the shareholders' general meeting as prescribed under the Listing Rules;
(vii) 法律、行政法规、《上市规则》、公司股份上市地上市规则或《公司章程》规定须以特别决议通过的事项以外的其他事项。
(viii) 法律、行政法规、《上市规则》、公司股份上市地上市规则或《公司章程》规定的其他事项，以及在股东大会以普通决议通过后将对公司产生重大影响、需以特别决议批准的事项。
股东（包括代理人）按其所代表的有表决权股份数行使表决权，每股一票，但拥有加权投票权的优先股股东除外。
股东大会审议涉及中小投资者利益的重大事项时，中小投资者所投票数应单独计算。单独计算的结果应及时向公众披露。
公司持有的本公司股份不享有任何表决权，且不计入出席股东大会的股东所代表的有表决权股份总数。
股东大会审议关联交易相关事项时，关联股东不得参与表决，其所代表的有表决权股份数不计入有效表决权股份总数。股东大会决议应充分披露非关联股东的表决情况。
董事由股东大会选举或更换，股东大会亦可在其任期届满前将其免职。董事任期为三年，任期届满后如获重新选举可续任。独立董事的任期与其他董事相同，任期届满后如获重新选举可续任。
董事的任期自其就任之日起计算，至本届董事会任期届满时止。如届时未能及时完成换届选举，现任董事应在任期届满后依据法律、行政法规、部门规章及《公司章程》的规定继续履行董事职责。
总经理及其他高级管理人员可兼任董事，但兼任高级管理人员的董事与职工代表董事的合计人数不得超过公司董事总数的50%。
公司设董事会，向股东大会负责。董事会由九名董事组成，其中包括一名职工代表董事及三名独立董事。董事会设董事长一名，由全体董事过半数选举产生。
(xii) 管理本公司的信息披露事务； (xiii) 向股东大会提议聘请或更换担任公司审计工作的会计师事务所； (xiv) 听取公司总经理的工作报告并对其工作进行审查； (xv) 决定设立董事会专门委员会； (xvi) 行使法律、行政法规、部门规章、上市规则及公司股票上市地点的上市规则或本章程规定的其他职权。
公司董事会应设立审计、提名、薪酬、战略及ESG专门委员会，各委员会应依照本章程及董事会授权履行职责，并向董事会负责。专门委员会的详细实施细则由董事会制定。
公司设董事会秘书一名。董事会秘书为公司高级管理人员。董事会秘书应遵守法律、法规、部门规章及本章程的相关规定。
公司可聘任证券代表协助董事会秘书履行职责；董事会秘书无法履行职责时，由证券代表行使其权利并履行其职责。
董事会秘书负责股东大会及董事会会议的筹备工作、会议记录及文件的保管、股东信息管理、信息披露及其他日常事务。
公司设总经理一名，由董事会聘任或解聘。公司设若干副总经理，由董事会聘任或解聘。
总经理任期三年，任期届满后可续聘。
(viii) 公司章程或董事会授权的其他职权。
总经理应出席董事会会议。
公司依照法律、行政法规及国家有关部门的规章建立本公司的财务会计制度。若公司股票上市地点的上市规则或证券监管机构另有规定，则从其规定。
财务会计报告依照相关法律、行政法规、部门规章、上市规则及公司股票上市地点的其他证券监管规则编制。
公司除法定会计账簿外，不得另立会计账簿。公司资产不得以任何个人名义存储。
公司应从每年税后利润中提取10%作为公司法定公积金。当公司法定公积金累计金额达到注册资本的50%或以上时，无需再行提取。
若公司法定公积金不足以弥补上一年度亏损，则当年利润应首先用于弥补亏损，之后方可依照前款规定提取法定公积金。
公司从税后利润中提取法定公积金后，经股东大会决议，还可从税后利润中提取任意公积金。
弥补亏损并提取公积金后，剩余税后利润按股东持股比例分配给股东。
如果公司在经营管理中遭遇严重困难，持续存续将给股东带来无法通过其他途径解决的重大损失，持有公司全部表决权百分之十以上的股东可向人民法院申请解散公司。
就上述第（i）项、第（ii）项情形而言，若尚未向股东分配财产，公司可通过修改章程或经股东决议批准的方式继续存续。修改章程须经出席股东大会的股东所代表的三分之二以上表决权通过。
如公司在上述第（i）项、第（ii）项、第（iv）项及第（v）项所列情形下解散，公司应当进行清算。董事为清算义务人，应当在解散情形出现之日起十五日内成立清算组，开始清算。清算组由董事组成，但章程另有规定或股东大会决议另行选任他人者除外。
（vii） 代表公司参与民事诉讼活动。
清算组应当自成立之日起十日内通知债权人，并于六十日内在报纸上或国家企业信用信息公示系统上进行公告。债权人应当自接到通知书之日起三十日内，未接到通知书的自公告之日起四十五日内，向清算组申报其债权。
清算组清理公司财产、编制资产负债表和财产清单后，应当制定清算方案，并报股东大会或人民法院确认。
支付清算费用、职工工资、社会保险费用和法定补偿金，缴纳所欠税款，清偿公司债务后的剩余财产，可依照股东的出资比例进行分配。
清算期间，公司继续存续，但不得开展与清算无关的经营活动。公司财产在依照前款规定清偿前，不得分配给股东。
（iii） 股东大会决议修改章程。
本公司于2019年6月11日在中华人民共和国设立为有限责任公司，并于2025年3月26日依据中华人民共和国法律转制为股份有限公司。
截至最后实际可行日期，本公司注册资本为人民币40,281,069元，划分为40,281,069股非上市股份，每股面值人民币1.00元。于全球发售前及股份细分完成后，本公司注册股本将为人民币40,281,069元，由402,810,690股已发行非上市股份组成，每股面值人民币0.10元。
本公司已于香港湾仔皇后大道东248号大新金融中心40楼设立主要营业地点，并于2025年7月11日依据《公司条例》第16部在香港公司注册处登记为非香港公司。本公司公司秘书郑正杰先生已获委任为本公司授权代表，负责在香港代表本公司接受法律程序文件及通知书的送达。
由于本公司在中华人民共和国设立，本公司的公司架构及章程细则须受中华人民共和国相关法律法规的规管。本公司章程细则相关条文的摘要载于本招股章程附录五「章程细则摘要」。
除本招股章程「历史、发展及公司架构——本公司的设立及主要股权变动」一节所披露者外，本招股章程日期前两年内本公司股本概无其他变动。
2024年5月14日，上海智谱寰宇科技有限公司成立，注册资本为人民币1,000万元。
2025年6月26日，上海智谱寰宇科技有限公司注册资本由人民币1,000万元增加至人民币5亿元。
2024年12月16日，珠海智谱领航科技有限公司成立，注册资本为人民币500万元。
2025年4月24日，珠海智谱领航科技有限公司注册资本由人民币500万元增加至人民币1亿元。
2024年10月25日，天津智谱华章科技有限公司成立，注册资本为人民币5,000万元。
2024年12月20日，天津智谱华章科技有限公司注册资本由人民币5,000万元增加至人民币9.5亿元。
2024年5月22日，北京智谱领航科技有限公司成立，注册资本为人民币3,000万元。
2024年9月24日，北京智谱兴曜科技有限公司成立，注册资本为人民币3亿元。
2024年3月15日，北京智谱未来科技有限公司成立，注册资本为人民币1,000万元。
2024年12月18日，珠海智谱未来科技有限公司成立，注册资本为人民币500万元。
2025年4月23日，珠海智谱未来科技有限公司注册资本由人民币500万元增加至人民币1,500万元。
2024年12月27日，成都智谱华章科技有限公司成立，注册资本为人民币500万元。
2024年12月31日，成都智谱华章科技有限公司注册资本由人民币500万元增加至人民币3亿元。
2025年2月24日，浙江智谱新篇科技有限公司成立，注册资本为人民币4.5亿元。
2024年11月6日，北京智谱海英教育科技有限公司成立，注册资本为人民币500万元。
2025年3月13日，北京智谱清影科技文化传媒有限公司成立，注册资本为人民币500万元。
2024年10月29日，北京智谱慧兴科技有限公司成立，注册资本为人民币1,000万元。
2024年11月20日，ZYNIX LIMITED成立，股本为1英镑。
Corethinks Technology SDN. BHD.
2025年6月30日，Corethinks Technology SDN. BHD.成立，股本为1马来西亚林吉特。
2025年7月1日，SUPER CONVERGENCE SARL成立，股本为3,500欧元。
2025年8月6日，深圳智谱聆心智能科技有限公司成立，注册资本为人民币100万元。
2025年11月11日，黄石智谱华章科技有限公司成立，注册资本为人民币2,000万元。
有关本公司股份回购限制的详情，请参阅本招股章程附录五——《公司章程摘要》。
(p) 在全球发售完成的前提下，授予董事会一般授权，以便在下次股东周年大会结束之日或股东通过特别决议撤销或更改该授权之日（以较早者为准）之前的任何时间内，配发、发行股份，或出售及/或转让作为库存股持有的库存股；但须从库存股中发行或出售及/或转让的股份数目不得超过上市日期已发行股份数目的20%（不包括根据超额配股权发行的任何股份）；
(r) 本公司董事会及/或其授权人士已获授权处理与全球发售、H股发行及上市等相关事宜。
以下段落载有（其中包括）联交所规定须载于本招股说明书中有关回购本公司证券的若干资料。
董事会认为，回购H股将对本公司及其全体股东有利，符合本公司及其全体股东的最佳利益。此举可增强投资者对本公司的信心，并对维护本公司在资本市场的声誉产生积极作用。此类回购仅在董事会认为对本公司及其全体股东有利时方予进行。
回购H股后，本公司可注销任何已回购H股及/或将其作为库存股持有，但须受（其中包括）市场状况及回购时其资本管理需求所限，而上述情况可能因环境变化而改变。
(ii) 本公司任何股东大会上通过特别决议撤销或更改该授权时。
此外，我们须就实际向董事会授予回购授权（如适用）向相关政府机构完成登记及审批手续。在全面行使回购H股一般授权的情况下（基于上市日期已发行215,701,705股H股（假设超额配股权未获行使），且在下次股东周年大会举行之日或之前本公司不会配发及发行或回购任何H股），
根据公司章程、上市规则及任何其他适用法律法规，公司回购的H股将予以注销或作为库存股保留。
如因任何H股回购导致某股东在本公司投票权中所占比例增加，该增加将被视为根据《收购守则》所作的收购行为。因此，某股东或一组一致行动的股东可能取得或巩固对本公司的控制权，并须根据《收购守则》第26条规则作出强制要约。
除上述情况外，董事并不知悉根据回购H股一般授权进行任何回购将会在《收购守则》下引起的任何后果。
(iii) 采取任何其他措施，以确保本公司不会行使任何股东权利或收取任何权益，而该等权利或权益若该等股份以本公司名义登记为库存股，则根据适用法律须予暂停。
于最后实际可行日期，本公司并无持有任何库存股，且于上市时亦不会持有任何库存股。
若回购股份一般授权获全额行使，则与本公司最近期已发布经审计账目所披露的状况相比，可能对本公司的营运资金或资产负债状况产生重大不利影响。然而，董事并无建议将回购H股一般授权行使至对本公司营运资金或资产负债状况造成重大不利影响的程度。
董事已向联交所承诺，将按照上市规则及中国适用法律行使回购H股一般授权。关于回购本公司证券的说明声明及拟议股份回购均无任何异常特征。
B.
Listing) would result in a maximum of 21,750,170 H Shares being repurchased by the Company during the relevant period, being the maximum of 10% of the H Shares in issue (excluding any treasury shares) as of the Listing Date.
In repurchasing its H Shares, the Company intends to apply funds from the Company's internal resources (which may include surplus funds and retained profits) legally available for such purpose in accordance with the Articles of Association and the applicable laws, rules and regulations of the PRC. The Company is empowered by its Articles of Association to repurchase its H Shares. Any shares to be repurchased will be canceled or kept as treasury shares if allowed by the Articles of Association and applicable laws and regulations. The Company may not purchase securities on the Stock Exchange for a consideration other than cash or for settlement otherwise than in accordance with the trading rules of the Stock Exchange from time to time.
A listed company shall not repurchase its shares on the Stock Exchange at any time after inside information has come to its knowledge until the information is made publicly available. In particular, during the period of one month immediately preceding the earlier of: (i) the date of the board meeting (as such date is first notified to the Stock Exchange in accordance with the Listing Rules) for the approval of the company's results for any year, half-year, quarterly or any other interim period (whether or not required under the Listing Rules); and (ii) the deadline for the issuer to announce its results for any year or half-year under the Listing Rules, or quarterly or any other interim period (whether or not required under the Listing Rules), until the date of the results announcement, the company may not repurchase its shares on the Stock Exchange unless there are exceptional circumstances.
None of our Directors or, to the best of their knowledge having made all reasonable inquiries, any of their close associates have a present intention, in the event the general mandate to repurchase H Shares is approved, to sell any Shares to our Company.
No core connected person of our Company has notified our Company that they have a present intention to sell H Shares to our Company, or have undertaken to do so, if the general mandate to repurchase H Shares is approved.
A listed company shall not knowingly purchase its shares on the Stock Exchange from a core connected person (namely a director, supervisor, chief executive or substantial shareholder of the company or any of its subsidiaries, or a close associate of any of them), and a core connected person shall not knowingly sell their interest in shares of the company to it.
Subject to the Articles of Association, the Listing Rules and any other applicable laws and regulations, the H Shares repurchased by the Company will be canceled or kept as treasury shares.
If, as a result of any repurchase of H Shares, a Shareholder's proportionate interest in the voting rights of our Company increases, such increase will be treated as an acquisition for the purposes of the Takeovers Code. Accordingly, a Shareholder or a group of Shareholders acting in concert could obtain or consolidate control of our Company and become obliged to make a mandatory offer in accordance with Rule 26 of the Takeovers Code.
Save as aforesaid, our Directors are not aware of any consequences which would arise under the Takeovers Code as a consequence of any repurchases pursuant to the general mandate to repurchase H Shares.
For any treasury shares of the Company deposited with CCASS pending resale on the Stock Exchange, the Company shall, upon approval by the Board, implement the below interim measures which include (without limitation):
(i) procuring its broker not to give any instructions to HKSCC to vote at general meetings for the treasury shares deposited with CCASS;
(ii) in the case of dividends or distributions (if any and where applicable), withdrawing the treasury shares from CCASS and either re-register them in its own name as treasury shares or cancel them, in each case before the relevant record date for the dividend or distributions; or
(iii) taking any other measures to ensure that it will not exercise any Shareholders' rights or receive any entitlements which would otherwise be suspended under the applicable laws if those Shares were registered in its own name as treasury shares.
The Company did not hold any treasury shares as of the Latest Practicable Date and will not hold any treasury shares upon Listing.
If the general mandate to repurchase Shares were to be carried out in full at any time, there may be a material and adverse impact on our working capital or gearing position (as compared with the position disclosed in our most recent published audited accounts). However, our Directors do not propose to exercise the general mandate to repurchase H Shares to such an extent as would have a material and adverse effect on our working capital or gearing position.
Our Directors have undertaken to the Stock Exchange that they will exercise the general mandate to repurchase H Shares in accordance with the Listing Rules and the applicable laws in the PRC. Neither the Explanatory Statement on Repurchase of Our Own Securities nor the proposed share repurchase has any unusual feature.
B.
1.
We have entered into the following contracts (not being contracts entered into in the ordinary course of business) within the two years preceding the date of this prospectus that are or may be material: (a) a capital increase agreement dated February 7, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清 言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恒耀科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德 兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui – VI-6 –
Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥 企業(有限合夥)), Suzhou Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯 錦帆創業投資合夥企業(有限合夥)), and Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限 合夥)), pursuant to which (i) Suzhou Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯錦帆創業投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB116,173 at a consideration of RMB49,957,107; and (ii) Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB118,474 at a consideration of RMB50,946,934; (b) a capital increase agreement dated February 7, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能 科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓 子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有 限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業 管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Shanghai Feiya Technology Co., Ltd. (上海飛玡科技有限公司), pursuant to which Shanghai Feiya Technology Co., Ltd. (上海飛玡科技 有限公司) agreed to subscribe for our Company's newly increased registered capital of RMB465,090 at a consideration of RMB200,000,000; (c) a capital increase agreement dated February 7, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清 言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德 兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥 企業(有限合夥)), and AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限 合夥)), pursuant to which AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基 金(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB465,090 at a consideration of RMB200,000,000; (d) a capital increase agreement dated February 7, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能 科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓 子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing – VI-7 –
Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Aramco Ventures Investments Limited, pursuant to which Aramco Ventures Investments Limited agreed to subscribe for our Company's newly increased registered capital of RMB495,390 at a consideration of US$30,000,000 (equivalent to RMB213,030,000);
(e) a capital increase agreement dated April 18, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Hubei Yangtze CITIC Technology Mobile Communication Industry Investment Fund Partnership (Limited Partnership) (湖北長江中信科移動通信技術產業投資基金合夥企業(有限合夥)), pursuant to which Hubei Yangtze CITIC Technology Mobile Communication Industry Investment Fund Partnership (Limited Partnership) (湖北長江中信科移動通信技術產業投資基金合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB209,290 at a consideration of RMB90,000,000;
(f) a capital increase agreement dated August 8, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業(有限合夥)), pursuant to which Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB826,211 at a consideration of RMB500,000,000;
(g) a capital increase agreement dated August 8, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co.,
Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能 科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合 夥)), and Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發 展基金合夥企業(有限合夥)), pursuant to which Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發展基金合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB495,726 at a consideration of RMB300,000,000;
(h) a capital increase agreement dated August 8, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能 科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限 合夥)), and Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北 京星連肇基企業管理合夥企業(有限合夥)), pursuant to which Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB247,863 at a consideration of RMB150,000,000 (the "Xinglian Zhaoji Series B6-1 Capital Increase Agreement");
(i) a capital increase agreement dated August 8, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能 科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng
Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合 夥)), Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相 道股權投資合夥企業(有限合夥)), and Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有 限合夥)), pursuant to which (i) Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB89,700 at a consideration of RMB54,284,213; and (ii) Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB34,231 at a consideration of RMB20,715,787;
(j) a convertible note agreement dated September 3, 2024 entered into among Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創 業投資合夥企業(有限合夥)), our Company, Tang Jie (唐傑), and Liu Debing (劉德兵), pursuant to which Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)) agreed to provide a convertible loan to the Company in the principal amount of RMB130,000,000;
(k) a capital increase agreement dated November 27, 2024 entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Tianjin Haihe Fuxin Youda Venture Capital Fund Partnership (Limited Partnership) (天津海河富新優達創業投資基金合夥企業(有限合夥)), pursuant to which Tianjin Haihe Fuxin Youda Venture Capital Fund Partnership (Limited Partnership) (天津海河富新優達創業投資基金合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB1,569,800 at a consideration of RMB950,000,000;
(l) a supplemental agreement to the Xinglian Zhaoji Series B6-1 Capital Increase Agreement dated November 29, 2024, entered into among our Company, Nanjing Zhihu Information Technology Co., Ltd. (南京知乎信息科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恆耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)),
Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業(有限合夥)), pursuant to which Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業(有限合夥)) agreed to reduce the amount of subscribed newly increased registered capital to RMB160,285 at a reduced consideration of RMB97,000,000;
(m) a capital increase agreement dated December 31, 2024 entered into among our Company, Nanjing Knowledge Xinglan Technology Co., Ltd. (南京智譜星瀾科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恒耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Chengdu High-tech Orrino Youchan Equity Investment Fund Partnership (Limited Partnership) (成都高新策源優產股權投資基金合夥企業(有限合夥)), pursuant to which Chengdu High-tech Orrino Youchan Equity Investment Fund Partnership (Limited Partnership) (成都高新策源優產股權投資基金合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB495,726 at a consideration of RMB300,000,000;
(n) a convertible note agreement dated January 25, 2025 entered into among Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)), our Company, Tang Jie (唐傑), and Liu Debing (劉德兵), pursuant to which Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)) agreed to provide a convertible loan to the Company in the principal amount of RMB100,000,000;
(o) a convertible note agreement dated January 26, 2025 entered into among Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)), our Company, Tang Jie (唐傑), and Liu Debing (劉德兵), pursuant to which Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)) agreed to provide a convertible loan to the Company in the principal amount of RMB100,000,000;
(p) a capital increase agreement dated January 27, 2025 entered into among our Company, Nanjing Knowledge Xinglan Technology Co., Ltd. (南京智譜星瀾科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing
(q) a capital increase agreement dated February 7, 2025 entered into among our Company, Nanjing Knowledge Xinglan Technology Co., Ltd. (南京智譜星瀾科技有限公司), Shenzhen Knowledge Future Technology Co., Ltd. (深圳智譜未來科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhipu Overseas Innovation Technology Limited (智譜海外創新科技有限公司), Zhipu HengYao Technology Limited (智譜恒耀科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)), and Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)), pursuant to which (i) Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB240,784 at a consideration of RMB145,715,755; and (ii) Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB48,390 at a consideration of RMB29,284,213;
(r) a convertible note agreement dated February 7, 2025 entered into among Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)), our Company, Tang Jie (唐傑), and Liu Debing (劉德兵), pursuant to which Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)) agreed to provide a convertible loan to the Company in the principal amount of RMB100,000,000;
(s) a convertible note agreement dated March 17, 2025 entered into among Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)), our Company, Tang Jie (唐傑), and Liu Debing (劉德兵), pursuant to which Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)) agreed to provide a convertible loan to the Company in the principal amount of RMB400,000,000;
a capital increase agreement dated March 19, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限合夥)), pursuant to which AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB330,484 at a consideration of RMB200,000,000;
(u) a capital increase agreement dated May 13, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)), pursuant to which Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered
Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)), pursuant to which Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司) agreed to subscribe for our Company's newly increased registered capital of RMB165,242 at a consideration of RMB100,000,000, which was satisfied through the conversion of convertible note;
(v) a capital increase agreement dated May 13, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)), pursuant to which Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB214,815 at a consideration of RMB130,000,000, which was satisfied through the conversion of convertible note;
(w) a capital increase agreement dated May 13, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao
Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)), pursuant to which Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB165,242 at a consideration of RMB100,000,000, which was satisfied through the conversion of convertible note;
(x) a capital increase agreement dated May 13, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恆曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), and Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)), pursuant to which Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)) agreed to subscribe for our Company's newly increased registered capital of RMB826,211 at a consideration of RMB500,000,000, which was satisfied through the conversion of convertible note;
(y) a capital increase agreement dated May 23, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司),
Enterprise Management Co., Ltd. (杭州朵象企業管理有限公司), Shenzhen Zhonghang Hongyue Equity Investment Partnership (Limited Partnership) (深圳中航弘粵股權投資合夥企業(有限合夥)), Wuhan Luoshi Investment Partnership (Limited Partnership) (武漢洛石投資合夥企業(有限合夥)), Guizhou Transportation Industry Investment Fund (Limited Partnership) (貴州交通產業投資基金(有限合夥)), CICC Qizhi (Shanghai) Equity Investment Fund (Limited Partnership) (中金啓智(上海)股權投資基金(有限合夥)), CICC Qizhi No. 2 (Shanghai) Equity Investment Fund (Limited Partnership) (中金啓智二號(上海)股權投資基金(有限合夥)), Yangpu District State-owned Assets Supervision and Administration Commission of Shanghai (上海市楊浦區國有資產監督管理委員會), Chengdu Tianfu New Area Science and Technology Investment Group Co., Ltd. (成都天府新區科技投資集團有限公司), Shandong Cultural Tourism Equity Investment Fund (Limited Partnership) (山東文旅股權投資基金合夥企業(有限合夥)), Xi'an Zhongke Xinsheng Investment Management Co., Ltd. (西安中科新聲投資管理有限公司), Suzhou Yunchuang Tianshu Intelligent Technology Equity Investment Partnership (Limited Partnership) (蘇州雲創天樞智能科技股權投資合夥企業(有限合夥)), and Zhihui Linghang Venture Capital Partnership (Limited Partnership) (上海浦東智慧領航創業投資合夥企業(有限合夥)), pursuant to which Zhihui Linghang Venture Capital Partnership (Limited Partnership) (上海浦東智慧領航創業投資合夥企業(有限合夥)) agreed to subscribe for our Company's newly registered capital of RMB826,211 at a consideration of RMB500,000,000;
(z) a shareholders' agreement dated May 23, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), Zhipu HengYao Technology Pte. Ltd. (智譜恒曜科技有限公司), Tang Jie (唐傑), Li Juanzi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Kaiaigeer Technology Development Centre (Limited Partnership) (北京凱愛格爾科技發展中心(有限合夥)), Ningbo Huihui Enterprise Management Partnership (Limited Partnership) (寧波慧惠企業管理合夥企業(有限合夥)), Ningbo Zhideng Enterprise Management Partnership (Limited Partnership) (寧波智登企業管理合夥企業(有限合夥)), Tsinghua Control Technology Transfer Co., Ltd. (華控技術轉移有限公司), Beijing CAS Star Hard Technology Venture Capital Partnership (Limited Partnership) (北京中科創星硬科技創業投資合夥企業(有限合夥)), Beijing Innovation Zhiyuan Technology Co., Ltd. (北京創新智源科技有限公司), Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業(有限合夥)), Zaozhuang Tongzhi Equity Investment Partnership (Limited Partnership) (棗莊通智股權投資合夥企業(有限合夥)), Beijing Huakong Industrial Investment Fund (Limited Partnership) (北京華控產業投資基金(有限合夥)), Qingdao Huakong Growth Equity Investment Partnership (Limited Partnership) (青島華控成長股權投資合夥企業(有限合夥)), Beijing Rongpin Investment Management Co., Ltd. (北京榮品投資管理有限公司), Beijing The Jiangmen Venture Capital Center (Limited Partnership) (北京將門創業投資中心(有限合夥)), Shenzhen Dachen Chuanghong Private Equity Investment Partnership (Limited Partnership) (深圳市達晨創鴻私募股權投資企業(有限合夥)), Shenzhen Caizhi Chuangying Private Equity Investment Partnership (Limited Partnership) (深圳市財智創贏私募股權投資企業(有限合夥)), Luster LightTech Co., Ltd. (凌雲光技術股份有限公司), Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)), Suzhou Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯錦帆創業投資合夥企業(有限合夥)), Suzhou Qiming Rongqian Equity Investment Partnership (Limited Partnership) (蘇州啓明融乾股權投資合夥企業(有限合夥)), Kunshan Qiming Rongkai Equity Investment Partnership (Limited Partnership) (昆山市啓明融凱股權投資合夥企業(有限合夥)), Tianjin Sankuai Technology Co., Ltd. (天津三快科技有限公司), Shanghai Yunya Enterprise Management Consulting Co., Ltd. (上海雲玡企業管理諮詢有限公司), Hangzhou Duoxiang
Network Technology Co., Ltd. (杭州多項網絡科技有限公司), Trend Mega Limited (全德美嘉有限公司), Tianjin Heyuan Youze Yihao Venture Capital Partnership (Limited Partnership) (天津合遠優擇壹號創業投資合夥企業(有限合夥)), Hangzhou Guanghe II Venture Capital Partnership (Limited Partnership) (杭州光合貳期創業投資合夥企業(有限合夥)), Guangxi Tencent Venture Capital Co., Ltd. (廣西騰訊創業投資有限公司), TAL Education (Beijing) Co., Ltd. (欣欣相融教育科技(北京)有限公司), Beijing Xiaofeng Technology Co., Ltd. (北京小鋒科技有限公司), Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)), Xiamen HongShan Yaheng Equity Investment Partnership (Limited Partnership) (厦門紅杉雅恒股權投資合夥企業(有限合夥)), Ningbo Meishan Free Trade Port Zone Mingheng Enterprise Management Consulting Partnership (Limited Partnership) (寧波梅山保稅港區明恒企業管理諮詢合夥企業(有限合夥)), Beijing Shunying Equity Investment Partnership (Limited Partnership) (北京順贏股權投資合夥企業(有限合夥)), Shanghai Feiya Technology Co., Ltd. (上海飛玡科技有限公司), Wuxi Yunhui Digital Economy Investment Management Partnership (Limited Partnership) (無錫雲暉數字經濟投資管理合夥企業(有限合夥)), Shenzhen Zhaoshang Shuke Innovation Private Equity Fund Partnership (Limited Partnership) (深圳市招商數科創新私募股權投資基金合夥企業(有限合夥)), AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限合夥)), Beijing Lianrong Zhiyuan Equity Investment Partnership (Limited Partnership) (北京聯融致遠股權投資合夥企業(有限合夥)), P7 China Holdings PCC Limited (acting solely in respect of the P7CH Direct P7 I cell), Hubei Yangtze CITIC Technology Mobile Communication Industry Investment Fund Partnership (Limited Partnership) (湖北長江中信科移動通信技術產業投資基金合夥企業(有限合夥)), Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業(有限合夥)), Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發展基金合夥企業(有限合夥)), Tianjin Haihe Fuxin Youda Venture Capital Fund Partnership (Limited Partnership) (天津海河富新優達創業投資基金合夥企業(有限合夥)), Hangzhou Chengtou Industrial Development Investment Partnership (Limited Partnership) (杭州城投產業發展投資合夥企業(有限合夥)), Hangzhou Shangcheng Linghang Venture Capital Co., Ltd. (杭州上城領航創業投資有限公司), Chengdu High-tech Orrino Youchan Equity Investment Fund Partnership (Limited Partnership) (成都高新策源優產股權投資基金合夥企業(有限合夥)), Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)), Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)), Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)), Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)), and Zhihui Linghang Venture Capital Partnership (Limited Partnership) (上海浦東智慧領航創業投資合夥企業(有限合夥)), pursuant to which certain shareholder rights were agreed among the parties;
(aa) a shareholder special rights termination agreement dated June 27, 2025 entered into among our Company, Nanjing Knowledge Atlas Technology Co., Ltd. (南京智譜華章科技有限公司), Shenzhen Knowledge Atlas Technology Co., Ltd. (深圳智譜華章科技有限公司), Hangzhou Knowledge Atlas Technology Co., Ltd. (杭州智譜華章科技有限公司), Beijing Knowledge Qingyan Technology Co., Ltd. (北京智譜清言科技有限公司), Beijing Lingxin Intelligent Technology Co., Ltd. (北京聆心智能科技有限公司), Shanghai Knowledge Huanyu Technology Co., Ltd. (上海智譜寰宇科技有限公司), Beijing Knowledge Future Technology Co., Ltd. (北京智譜未來科技有限公司), Beijing Knowledge Haiying Education Technology Co., Ltd. (北京智譜海英教育科技有限公司), Beijing Knowledge Linghang Technology Co., Ltd. (北京智譜領航科技有限公司), Beijing Knowledge Xingyao Technology Co., Ltd. (北京智譜興曜科技有限公司), Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司), Tianjin Knowledge Atlas Technology Co., Ltd. (天津智譜華章科技有限公司), Zhejiang Knowledge Xinpian Technology Co., Ltd. (浙江智譜新篇科技有限公司), Chengdu Knowledge Atlas Technology Co., Ltd. (成都智譜華章科技有限公司), Zhuhai Knowledge Linghang Technology Co., Ltd. (珠海智譜領航科技有限公司), Zhuhai Knowledge Future Technology Co., Ltd. (珠海智譜未來科技有限公司), Jincheng
Yaoda Technology Limited, Hong Kong Xiangtai Ruifeng Technology Limited (香港祥泰瑞豐科技有限公司), JINGSHENG HENGXING TECHNOLOGY PTE.LTD (景盛恒興科技有限公司), Tang Jie (唐傑), Li Junazi (李涓子), Liu Debing (劉德兵), Xu Bin (許斌), Zhang Peng (張鵬), Wang Shaolan (王紹蘭), Beijing Lianpai Technology Development Center (Limited Partnership) (北京鏈湃科技發展中心(有限合夥)), Zhuhai Hengqin Huihui Enterprise Management Partnership (Limited Partnership) (珠海橫琴慧惠企業管理合夥企業(有限合夥)), Zhuhai Hengqin Zhideng Enterprise Management Partnership (Limited Partnership) (珠海橫琴智登企業管理合夥企業(有限合夥)), Tsinghua Control Technology Transfer Co., Ltd. (華控技術轉移有限公司), Beijing CAS Star Hard Technology Venture Capital Partnership (Limited Partnership) (北京中科創星硬科技創業投資合夥企業(有限合夥)), Beijing Innovation Zhiyuan Technology Co., Ltd. (北京創新智源科技有限公司), Beijing Xinglian Zhaoji Enterprise Management Partnership (Limited Partnership) (北京星連肇基企業管理合夥企業(有限合夥)), Zaozhuang Tongzhi Equity Investment Partnership (Limited Partnership) (棗莊通智股權投資合夥企業(有限合夥)), Beijing Huakong Industrial Investment Fund (Limited Partnership) (北京華控產業投資基金(有限合夥)), Qingdao Huakong Growth Equity Investment Partnership (Limited Partnership) (青島華控成長股權投資合夥企業(有限合夥)), Beijing Rongpin Investment Management Co., Ltd. (北京榮品投資管理有限公司), Beijing The Jiangmen Venture Capital Center (Limited Partnership) (北京將門創業投資中心(有限合夥)), Shenzhen Dachen Chuanghong Private Equity Investment Partnership (Limited Partnership) (深圳市達晨創鴻私募股權投資企業(有限合夥)), Shenzhen Caizhi Chuangying Private Equity Investment Partnership (Limited Partnership) (深圳市財智創贏私募股權投資企業(有限合夥)), Luster LightTech Co., Ltd. (凌雲光技術股份有限公司), Suzhou Junlian Xiangdao Equity Investment Partnership (Limited Partnership) (蘇州君聯相道股權投資合夥企業(有限合夥)), Suzhou Junlian Jinfan Venture Capital Partnership (Limited Partnership) (蘇州君聯錦帆創業投資合夥企業(有限合夥)), Suzhou Qiming Rongqian Equity Investment Partnership (Limited Partnership) (蘇州啓明融乾股權投資合夥企業(有限合夥)), Kunshan Qiming Rongkai Equity Investment Partnership (Limited Partnership) (昆山市啓明融凱股權投資合夥企業(有限合夥)), Tianjin Sankuai Technology Co., Ltd. (天津三快科技有限公司), Shanghai Yunya Enterprise Management Consulting Co., Ltd. (上海雲玡企業管理諮詢有限公司), Hangzhou Duoxiang Network Technology Co., Ltd. (杭州多項網絡科技有限公司), Trend Mega Limited (全德美嘉有限公司), Tianjin Heyuan Youze Yihao Venture Capital Partnership (Limited Partnership) (天津合遠優擇壹號創業投資合夥企業(有限合夥)), Hangzhou Guanghe II Venture Capital Partnership (Limited Partnership) (杭州光合貳期創業投資合夥企業(有限合夥)), Guangxi Tencent Venture Capital Co., Ltd. (廣西騰訊創業投資有限公司), TAL Education (Beijing) Co., Ltd. (欣欣相融教育科技(北京)有限公司), Beijing Xiaofeng Technology Co., Ltd. (北京小鋒科技有限公司), Social Security Zhongguancun Innovation Investment Fund (Beijing) Partnership (Limited Partnership) (社保基金中關村自主創新投資基金(北京)合夥企業(有限合夥)), Xiamen HongShan Yaheng Equity Investment Partnership (Limited Partnership) (厦門紅杉雅恒股權投資合夥企業(有限合夥)), Ningbo Meishan Free Trade Port Zone Mingheng Enterprise Management Consulting Partnership (Limited Partnership) (寧波梅山保稅港區明恒企業管理諮詢合夥企業(有限合夥)), Beijing Shunying Equity Investment Partnership (Limited Partnership) (北京順贏股權投資合夥企業(有限合夥)), Shanghai Feiya Technology Co., Ltd. (上海飛玡科技有限公司), Wuxi Yunhui Digital Economy Investment Management Partnership (Limited Partnership) (無錫雲暉數字經濟投資管理合夥企業(有限合夥)), Shenzhen Zhaoshang Shuke Innovation Private Equity Fund Partnership (Limited Partnership) (深圳市招商數科創新私募股權投資基金合夥企業(有限合夥)), AI Fund Partnership (Limited Partnership) (北京市人工智能產業投資基金(有限合夥)), Beijing Lianrong Zhiyuan Equity Investment Partnership (Limited Partnership) (北京聯融致遠股權投資合夥企業(有限合夥)), P7 China Holdings PCC Limited (acting solely in respect of the P7CH Direct P7 I cell), Hubei Yangtze CITIC Technology Mobile Communication Industry Investment Fund Partnership (Limited Partnership) (湖北長江中信科移動通信技術產業投資基金合夥企業(有限合夥)), Beijing Zhongguancun Science City Phase II Technology Growth Equity Investment Partnership (Limited Partnership) (北京中關村科學城二期科技成長股權投資合夥企業(有限合夥)), Beijing Daxing Industrial Fund Partnership (Limited Partnership) (北京市大興區產業發展基金合夥企業(有限合夥)), Tianjin Haihe Fuxin Youda Venture Capital Fund Partnership (Limited Partnership) (天津海河富新優達創業投資基金合夥企業(有限合夥)), Hangzhou Chengtou Industrial Development
Investment Partnership (Limited Partnership) (杭州城投產業發展投資合夥企業(有限合夥)), Hangzhou Shangcheng Linghang Venture Capital Co., Ltd. (杭州上城領航創業投資有限公司), Chengdu High-tech Orrino Youchan Equity Investment Fund Partnership (Limited Partnership) (成都高新策源優產股權投資基金合夥企業(有限合夥)), Tianjin Tianchuang Haihe Yongtai Puxin Venture Capital Partnership (Limited Partnership) (天津天創海河永鈦譜鑫創業投資合夥企業(有限合夥)), Zhuhai Huafa New Quality Productivity Investment Fund Partnership (Limited Partnership) (珠海市新質生產力投資基金合夥企業(有限合夥)), Hainan Xiarui Investment Partnership (Limited Partnership) (海南夏睿投資合夥企業(有限合夥)), Lenovo Small and Medium Enterprise Development Venture Capital Fund (Tianjin) Partnership (Limited Partnership) (聯想中小企業發展創業投資基金(天津)合夥企業(有限合夥)), and Zhihui Linghang Venture Capital Partnership (Limited Partnership) (上海浦東智慧領航創業投資合夥企業(有限合夥)), pursuant to which the parties agreed to terminate certain shareholder rights;
(bb) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, JSC International Investment Fund SPC (acting for and on behalf of Qizhi SP) and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$179.0 million;
(cc) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, JinYi Capital Multi-Strategy Fund SPC Ltd. acting for and on behalf of Structured Credit SP Fund and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$7.0 million;
(dd) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, Perseverance Asset Management International (Singapore) Pte. Ltd. and China International Capital Corporation Hong Kong Securities Limited, with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$29.0 million;
(ee) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, CICC Financial Trading Limited and China International Capital Corporation Hong Kong Securities Limited, pursuant to which CICC Financial Trading Limited has agreed to subscribe for H Shares at the Offer Price in the aggregate amount of Hong Kong dollar equivalent of US$9.0 million and hold such H Shares on a non-discretionary basis to hedge a series of cross border delta-one OTC swap transactions entered into by CICC Financial Trading Limited, China International Capital Corporation Limited and Shanghai Gaoyi Asset Management Partnership (Limited Partnership) (上海高毅資產管理合夥企業(有限合夥)) as investment manager for and on behalf of certain investment funds;
(ff) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, WT Asset Management Limited and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$44.9 million;
(gg) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, Taikang Life Insurance Co., Ltd (泰康人壽保險有限責任公司) and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$30.0 million;
(hh) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, GF Management Co., Ltd. (廣發基金管理有限公司) and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$38.45 million;
(ii) a cornerstone investment agreement dated December 29, 2025 entered into among our Company, GF International Investment Management Limited (廣發國際資產管理有限公司) and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$3.55 million;
(jj) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, 3W Fund Management Limited and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$18.0 million;
(kk) a cornerstone investment agreement dated December 29, 2025 entered into among our Company, RIME Capital Limited and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of HK$75.0 million;
(ll) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, Optimas Capital Limited and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$10.0 million;
(mm) a cornerstone investment agreement dated December 28, 2025 entered into among our Company, Luster LightTech International Limited (凌雲光技術國際有限公司) and China International Capital Corporation Hong Kong Securities Limited with respect to a subscription of H Shares at the Offer Price in the aggregate amount of the Hong Kong dollar equivalent of US$5.0 million; and
(nn) the Hong Kong Underwriting Agreement.
As of the Latest Practicable Date, we had registered the following trademarks which we consider to be or may be material to our business:
| No. | Trademark | Registered Class | Owner | | Registration No. | Expiry Date | |-----|-----------|-----------------|-------|---|-----------------|-------------| | 1. | | 38 | Our Company | PRC | 74735814 | 04/13/2034 | | 2. | | 45 | Our Company | PRC | 74721479 | 06/06/2034 | | 3. | | 36 | Our Company | PRC | 74721479 | 06/06/2034 |
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
(b) Patents As of the Latest Practicable Date, we had registered the following patents which we consider to be or may be material to our business:
| No. | Patent | Type | Registered Owner | Place of Registration | Patent Number | Date of Registration | |-----|--------|------|------------------|-----------------------|---------------|----------------------| | 1 | A method, device and medium for optimizing the instruction-following capability of LLMs (一種優化大語言模型指令遵循能力的方法、設備及介質) | Invention patent | Our Company | PRC | 2024115864607 | February 25, 2025 | | 2 | An intelligent extraction method and system for input text containing mathematical formulas (一種含數學公式的輸入文本的智能提取方法及系統) | Invention patent | Our Company | PRC | 2024103497310 | August 23, 2024 |
| No. | Patent | Type | Registered Owner | Place of Registration | Patent Number | Date of Registration | |-----|--------|------|------------------|-----------------------|---------------|----------------------| | 3 | A training method for web navigation AI agents based on LLMs (一種基於大語言模型的網頁導航智能體的訓練方法) | Invention patent | Our Company | PRC | 2024103093694 | October 1, 2024 | | 4 | A method, device, equipment and medium for dynamically adjusting the depth of LLMs (一種動態調整大語言模型深度的方法、裝置、設備和介質) | Invention patent | Our Company | PRC | 2024102713774 | October 1, 2024 | | 5 | A training method and device for MoE models based on decision trees (一種基於決策樹的混合專家模型的訓練方法和裝置) | Invention patent | Our Company | PRC | 2024102713030 | August 23, 2024 | | 6 | An application-oriented LLM interface system, method, device and medium (面向應用的大語言模型調用接口系統、方法、設備及介質) | Invention patent | Our Company | PRC | 2024100405834 | August 9, 2024 | | 7 | An automatic parallelised language model text generation method (一種自動並行化的語言模型文本生成方法) | Invention patent | Our Company | PRC | 2023118355360 | March 29, 2024 | | 8 | A method, device and storage medium for generating instruction fine-tuning data (一種指令微調數據的生成方法、設備和存儲介質) | Invention patent | Our Company | PRC | 2023117958411 | August 9, 2024 | | 9 | A general text quality evaluation method based on LLMs (一種基於大語言模型的通用文本質量評價方法) | Invention patent | Our Company | PRC | 2023116186705 | May 28, 2024 | | 10 | An alignment evaluation method for Chinese LLMs (一種針對中文大語言模型的對齊測評方法) | Invention patent | Our Company | PRC | 2023116210193 | May 28, 2024 | | 11 | A fair and efficient multi-dialogue system evaluation system and method (一種公平高效的多對話系統測評系統及方法) | Invention patent | Our Company | PRC | 2023115438272 | February 20, 2024 | | 12 | A generative information extraction method and device based on pre-trained models (基於預訓練模型的生成式信息抽取方法和裝置) | Invention patent | Our Company | PRC | 2021110162958 | January 3, 2025 | | 13 | An evaluation method, device and electronic equipment for LLMs (大語言模型的評估方法、裝置及電子設備) | Invention patent | Beijing Lingxin Intelligent Technology Co., Ltd. | PRC | 2023109676521 | July 12, 2024 |
| No. | Patent | Type | Registered Owner | Place of Registration | Patent Number | Date of Registration | |-----|--------|------|------------------|-----------------------|---------------|----------------------| | 14 | A personality test method, device and electronic equipment based on human-machine dialogue (基於人機對話的性格測試方法、裝置及電子設備) | Invention patent | Beijing Lingxin Intelligent Technology Co., Ltd. | PRC | 2023108615338 | March 22, 2024 | | 15 | A model optimisation training system, method and relevant device (模型的優化訓練系統、方法以及相關裝置) | Invention patent | Beijing Lingxin Intelligent Technology Co., Ltd. | PRC | 2023108092439 | October 20, 2023 | | 16 | A method, device, medium and computing equipment for initial dialogue content generation (初始對話內容生成方法、裝置、介質和計算設備) | Invention patent | Beijing Lingxin Intelligent Technology Co., Ltd. | PRC | 2023106006354 | August 22, 2023 | | 17 | A safety evaluation method and relevant device based on LLMs (基於大語言模型的安全測評方法以及相關裝置) | Invention patent | Beijing Lingxin Intelligent Technology Co., Ltd. | PRC | 2024119331870 | April 29, 2025 | | 18 | A method for extracting key information frames from surveillance videos (一種監控視頻中抽取關鍵信息幀的方法) | Invention patent | Nanjing Knowledge Atlas Technology Co., Ltd. | PRC | 201510062263X | March 27, 2018 |
(c) Copyrights As of the Latest Practicable Date, we had registered the following copyrights which we consider to be or may be material to our business:
| No. | Copyright | Registered Owner | Type | Place of Registration | Copyright Number | Date of Registration | |-----|-----------|------------------|------|-----------------------|------------------|----------------------| | 1 | AutoGLM software V1.1.01 (AutoGLM軟件V1.1.01) | Our Company | Software | PRC | 14591247 | 12/25/2024 | | 2 | GLM embedded large model system V1.0 (GLM嵌入式大模型系統V1.0) | Our Company | Software | PRC | 12267756 | 12/19/2023 | | 3 | Zhipu QingYan software V1.0 (智譜清言軟件V1.0) | Our Company | Software | PRC | 11783805 | 12/08/2023 | | 4 | ChatGLM-6B software V1.1 (ChatGLM-6B軟件V1.1) | Our Company | Software | PRC | 11479077 | 08/04/2023 | | 5 | Large model GLM3.5_130B platform V0.8 (大模型GLM3.5_130B平臺V0.8) | Our Company | Software | PRC | 11279154 | 06/20/2023 | | 6 | Large-scale pre-trained model application platform V1.0 (大規模預訓練模型應用平臺V1.0) | Our Company | Software | PRC | 9568431 | 05/20/2022 | | 7 | Large-scale pre-trained model system V1.0 (大規模預訓練模型系統V1.0) | Our Company | Software | PRC | 9568432 | 05/20/2022 |
(d) Domain names As of the Latest Practicable Date, our Group had registered the following domain names which we consider to be or may be material to our business:
| No. | Domain name | Registered Owner | Date of Registration | Date of Expiry | |-----|-------------|-----------------|---------------------|----------------| | 1. | aminer.cn | Our Company | 12/19/2012 | 12/19/2026 | | 2. | bigmodel.cn | Our Company | 11/10/2022 | 11/10/2031 | | 3. | codegeex.cn | Our Company | 10/22/2022 | 10/22/2034 | | 4. | chatglm.com | Our Company | 12/07/2022 | 12/07/2031 | | 5. | chatglm.cn | Our Company | 12/09/2022 | 12/09/2026 | | 6. | zhipuai.com | Our Company | 06/21/2021 | 06/21/2026 | | 7. | zhipuai.cn | Our Company | 06/21/2021 | 06/21/2029 |
(a) Directors, Supervisor and the chief executive of our Company Save as disclosed in the section headed "Substantial Shareholders" in this prospectus, immediately following the completion of the Global Offering and conversion of Unlisted Shares into H Shares, so far as our Directors are aware, none of our Directors, Supervisor or chief executive of our Company has any interests or short positions in the Shares, underlying Shares and debentures of our Company or its associated corporations (within the meaning of Part XV of the SFO) which will be required to be notified to our Company and the Stock Exchange pursuant to Divisions 7 and 8 of Part XV of the SFO (including interests or short positions which they were taken or deemed to have under such provisions of the SFO) or which will be required, under section 352 of the SFO, to be entered in the register referred to in that section, or which will be required, under the Model Code for Securities Transactions by Directors of Listed Issuers as set out in Appendix C3 to the Listing Rules (the "Model Code"), to be notified to our Company and the Stock Exchange once the H Shares are listed.
(b) Substantial Shareholders For the information on the persons who will, immediately following the completion of the Global Offering, have interests or short positions in our Shares or underlying Shares which would be required to be disclosed to our Company and the Stock Exchange under the provisions of Divisions 2 and 3 of Part XV of the SFO, or directly or indirectly be interested in 10% or more of the nominal value of any class of share capital carrying voting rights in all circumstances at general meetings of our Company, see the section headed "Substantial Shareholders" in this prospectus.
Save as disclosed in the section headed "Substantial Shareholders" in this prospectus, our Directors are not aware of any persons (other than our Directors and chief executive) who will, immediately following the completion of the Global Offering, have or be deemed or taken to have interests and/or short position in our Shares or underlying Shares which would be required to be disclosed under the provisions of Divisions 2 and 3 of Part XV of the SFO, or who is, directly or indirectly, interested in 10% or more of the nominal value of any types of the issued voting shares of any member of our Group.
As of the Latest Practicable Date, so far as our Directors are aware, the following persons (other than our Directors, Supervisor or chief executive of our Company) were interested in 10% or more of the nominal value of any class of share capital carrying rights to vote in all circumstances at general meetings of other members of our Group:
| Name of Member of our Group | Name of Shareholder | Percentage of Shareholding | |-----------------------------|--------------------|-----------------------------| | Beijing Knowledge Huixing Technology Co., Ltd. (北京智譜慧興科技有限公司) | Hainan Hezun Investment Co., Ltd. (海南何尊投資有限公司) | 30% |
Each of our Directors has entered into a service contract with our Company. The principal particulars of these service contracts comprise (a) a term of three years which is equivalent to the term of the Board; and (b) termination provisions in accordance with their respective terms. Our Directors may be re-appointed subject to Shareholders' approval. The service contracts can be renewed pursuant to our Articles of Association and applicable rules.
Save as disclosed above, we have not entered, and do not propose to enter, into any service contracts with any of our Directors or Supervisor in their respective capacities as Directors (other than contracts expiring or determinable by the employer within one year without any payment of compensation (other than statutory compensation)).
Save as disclosed in "Directors, Supervisor and Senior Management" and Note 8 to "Appendix I—Accountants' Report," for the three financial years ended December 31, 2024 and the six months ended June 30, 2025, none of our Directors received other remunerations or benefits in kind from us.
(a) None of our Directors, Supervisor or any of the parties listed in "—E. Other Information—4. Qualification and Consents of Experts" of this Appendix is:
(i) interested in our promotion, or in any assets which, within the two years immediately preceding the date of this Prospectus, have been acquired or disposed of by or leased to us, or are proposed to be acquired or disposed of by or leased to our Company; or
(ii) materially interested in any contract or arrangement subsisting at the date of this Prospectus which is significant in relation to our business;
(b) Save as disclosed in the section headed "Substantial Shareholders" in this prospectus or in connection with the Hong Kong Underwriting Agreement and the International Underwriting Agreement, none of our Directors or any of the parties listed in "—E. Other Information—4. Qualification and Consents of Experts" of this Appendix:
(ii) 拥有任何（无论是否具有法律强制执行效力的）权利，可认购本集团任何成员公司的证券，或提名他人认购本集团任何成员公司的证券； (c) 据本公司董事所知，本公司董事或监事或其紧密联系人，以及持有本公司已发行股本逾5%的股东，在过往记录期间各年度/期间内，均与本公司前五大客户或供应商无任何利益关系；及
(d) 除本招股说明书"主要股东"一节所披露的内容外，本公司任何董事或监事均非持有本公司股本权益之公司的董事或雇员，而该等权益一旦H股在香港联合交易所上市，须根据《证券及期货条例》第XV部第2及第3分部予以披露。
D.
董事会分别于2021年12月17日及2023年1月15日通过了2021年员工激励计划（"2021年计划"）及2023年员工激励计划（"2023年计划"）。2025年6月5日，董事会批准对2021年计划及2023年计划进行修订，并通过了2025年员工激励计划（"2025年计划"，连同经修订的2021年计划及2023年计划，统称"员工激励计划"）。根据员工激励计划，合资格参与者获授予员工持股平台或持有员工持股平台有限合伙权益的任何其他主体（"子平台"）的直接或间接合伙权益（"限制性奖励"）。
于最后实际可行日期，员工持股平台合共持有6,667,904股标的股份，约占本公司已发行股份的16.55%。全球发售完成后（假设超额配股权未获行使），员工持股平台将持有本公司全部已发行股份的约15.15%。详情请参阅"历史、发展及公司架构——员工持股平台"。
以下为员工激励计划主要条款的摘要。员工激励计划的条款不受《上市规则》第17章规定的约束，因为在上市后，员工激励计划不涉及本公司授出新期权或奖励或发行新股份。由于员工激励计划项下的所有股份均已发行予员工持股平台，因此员工激励计划不会导致上市后股东持股比例遭到摊薄。
员工激励计划的主要目的在于优化治理结构、提升本集团业绩、激励参与者、协调本集团与参与者之间的利益，并促进本集团的长期发展。
在适用法律法规及公司章程的规定下，员工激励计划的合资格参与者包括本集团的中高层管理人员、核心技术及业务人员，以及对本集团发展具有重要作用的其他雇员、顾问及人员（"受让人"）。
经董事会授权，公司首席执行官办公室（"首席执行官办公室"）为员工激励计划的管理机构，负责按照计划规定实施员工激励计划，包括确定计划的合资格参与者、限制性奖励应付代价及限制性奖励的授予数量。
自受让人登记为员工持股平台或子平台合伙人之日起至上市日期后12个月止（或适用法律法规可能规定的更长期间），各受让人不得转让限制性奖励，员工持股平台亦不得转让、质押或以其他方式处置相关标的股份。
锁定期届满后，受让人可要求相关员工持股平台在二级市场出售标的股份，并按照员工激励计划的条款将所得款项分配给受让人，从而实现限制性奖励附带的经济利益。
如受让人退出本集团，受让人可继续持有已归属的限制性奖励，而任何未归属的限制性奖励则由相关员工持股平台的普通合伙人或其指定主体以原始认购价格回购，或以普通合伙人批准的任何其他方式处置。
E.
1.
本公司董事获告知，本公司或本公司附属公司不太可能承担重大遗产税责任。
2.
据本公司董事所知，本集团任何成员公司均无针对本集团任何成员公司的重大未决或被威胁的诉讼或索赔。
3.
独家保荐人符合《上市规则》第3A.07条所载适用于保荐人的独立性准则。
独家保荐人已代表本公司向联交所申请，就由非上市股份转换而成的H股及根据全球发售发行的H股，申请其上市及买卖许可。独家保荐人担任本次上市保荐人将收取500,000美元的费用。
4.
Name | Qualification ---|--- China International Capital Corporation Hong Kong Securities Limited | Licensed to conduct Type 1 (dealing in securities), Type 2 (dealing in futures contracts), Type 4 (advising on securities), Type 5 (advising on futures contracts) and Type 6 (advising on corporate finance) regulated activities under the SFO KPMG | Certified Public Accountants; Public Interest Entity Auditor registered in accordance with the Accounting and Financial Reporting Council Ordinance
Name | Qualification ---|--- Tian Yuan Law Firm | Legal adviser to our Company as to the PRC laws Bayfront Law LLC | Legal adviser to our Company as to Singaporean data protection laws King & Wood Mallesons | Legal adviser as to International Sanctions laws Frost & Sullivan (Beijing) Inc., Shanghai Branch Co. | Independent industry consultant
Each of the experts has given and has not withdrawn its written consents to the issue of this prospectus with the inclusion of its reports, letters, opinions or summaries of opinions (as the case may be) and the references to its names and logos included herein in the form and context in which it is respectively included.
As of the Latest Practicable Date, none of the experts named above has any of our shareholding interests in any member of our Group or rights (whether legally enforceable or not) to subscribe for or to nominate persons to subscribe for our securities in any member of our Group.
Our Company has appointed Maxa Capital Limited as our Compliance Advisor in compliance with Rule 3A.19 of the Listing Rules.
Hong Kong stamp duty, currently charged at the ad valorem rate of 0.10% on the higher of the consideration for or the market value of the H Shares, will be payable by the purchaser on every purchase and by the seller on every sale of any Hong Kong securities, including H Shares (in other words, a total of 0.20% is currently payable on a typical sale and purchase transaction involving H Shares). In addition, a fixed stamp duty of HK$5.00 is currently payable on any instrument of transfer of H Shares. Where one of the parties is a resident outside Hong Kong and does not pay the ad valorem duty due by it, the duty not paid will be assessed on the instrument of transfer (if any) and will be payable by the transferee. If no stamp duty is paid on or before the due date, a penalty of up to ten times the duty payable may be imposed.
This prospectus shall have the effect, if any application is made pursuant hereto, of rendering all persons concerned bound by all the provisions (other than the penal provisions) of sections 44A and 44B of the Companies (Winding Up and Miscellaneous Provisions) Ordinance so far as applicable.
The English language and Chinese language versions of this prospectus are being published separately, in reliance upon the exemption provided by section 4 of the Companies Ordinance (Exemption of Companies and Prospectuses from Compliance with Provisions) Notice (Chapter 32L of the Laws of Hong Kong).
The promoters of our Company comprised all of the 47 then shareholders of our Company as of March 26, 2025 before our conversion into a joint stock company with limited liability. Within the two
years immediately preceding the date of this prospectus, no cash, securities or benefits have been paid, allotted or given, or are proposed to be paid, allotted or given to the promoters named above in connection with the Global Offering or the related transactions described in this prospectus.
Our Company did not incur any material preliminary expenses.
Our Directors confirm that, as of the date of this prospectus, there has been no material adverse change in our financial or trading position or prospects since June 30, 2025 (being the date to which the latest audited consolidated financial statements of our Group were prepared).
(i) save as disclosed in "History, Development and Corporate Structure" in this prospectus, no share or loan capital of our Company or our subsidiaries had been issued or agreed to be issued or proposed to be fully or partly paid either for cash or for a consideration other than cash;
(ii) no share or loan capital of our Company or our subsidiaries is under option or is agreed conditionally or unconditionally to be put under option;
(iii) save as disclosed in "Underwriting—Underwriting Arrangements and Expenses—Commissions and Expenses" in this prospectus, no commissions, discounts, brokerages or other special terms have been granted or agreed to be granted in connection with the issue or sale of any share or loan capital of our Company or our subsidiaries; and
(iv) save as disclosed in "Underwriting—Underwriting Arrangements and Expenses—Commissions and Expenses" in this prospectus, no commission has been paid or is payable for subscription, agreeing to subscribe, procuring subscription or agreeing to procure subscription of any share in our Company or our subsidiaries;
(b) there are no founder, management or deferred shares nor any debentures in our Company or our subsidiaries;
(c) there has not been any interruption in the business of our Group which may have or has had a significant effect on the financial position of our Group in the 12 months preceding the date of this prospectus;
(d) no company within our Group is presently listed on any stock exchange or traded on any trading system;
(e) all necessary arrangements have been made to enable our H Shares to be admitted into CCASS for clearing and settlement;
(g) there is no arrangement under which future dividends are waived or agreed to be waived, and there is no restriction affecting the remittance of profits or repatriation of capital by us into Hong Kong from outside Hong Kong; and
(h) none of the equity and debt securities of our Company, if any, is listed or dealt with in any other stock exchange nor is any listing or permission to deal being or proposed to be sought.
The documents attached to the copy of this prospectus delivered to the Registrar of Companies in Hong Kong for registration were:
(a) the written consents referred to in "Appendix VI—Statutory and General Information—E. Other Information—4. Qualification and Consents of Experts" in this prospectus; and
(b) a copy of each of the material contracts referred to in "Appendix VI—Statutory and General Information—B. Further Information about Our Business—1. Summary of Material Contracts" in this prospectus.
Copies of the following documents will be published on the websites of the Stock Exchange (www.hkexnews.hk) and our Company (www.zhipuai.cn) up to and including the date which is 14 days from the date of this prospectus:
(b) the Accountants' Report from KPMG, the text of which is set out in Appendix I in this prospectus;
(c) the report from KPMG in respect of the unaudited pro forma financial information, the text of which is set out in Appendix II in this prospectus;
(d) the audited consolidated financial statements of our Group for the three years ended December 31, 2022, 2023 and 2024 and the six months ended June 30, 2025;
(e) the material contracts referred to in "Appendix VI—Statutory and General Information—B. Further Information about Our Business—1. Summary of Material Contracts" in this prospectus;
(f) the service agreements entered into between our Company and each of our Directors referred to in "Appendix VI—Statutory and General Information—C. Further Information about Directors, Supervisor and Substantial Shareholders—2. Service Contracts" in this prospectus;
(g) the legal opinion issued by Tian Yuan Law Firm, our PRC Legal Advisors, in respect of certain general corporate matters and property interests in the PRC of the Group;
(k) the written consents referred to in "Appendix VI—Statutory and General Information—E. Other Information—4. Qualification and Consents of Experts" in this prospectus; and
(l) the PRC Company Law, the PRC Securities Law, the Trial Administrative Measures of Overseas Securities Offering and Listing by Domestic Companies, together with their unofficial English translation.
北京智譜華章科技股份有限公司 (A joint stock company established in the People's Republic of China with limited liability)
Sole Sponsor, Sponsor-Overall Coordinator, Overall Coordinator, Joint Global Coordinator, Joint Bookrunner and Joint Lead Manager