Core Viewpoint - Dongwu Securities expresses optimism about Zhipu AI's strengths in local model technology, open-source ecosystem, and local deployment capabilities, anticipating stable growth in local business and cloud services becoming the main driver, benefiting from the long-term trend of transitioning from local deployment to cloud services in China's large model industry [1] Company Overview - Zhipu AI, established in 2019, is a leading independent general large model developer in China, originating from Tsinghua University's Knowledge Engineering Laboratory. The company has developed its own GLM (General Language Model) pre-training framework, which differs from mainstream GPT architectures by employing an autoregressive fill-in-the-blank design, excelling in long text understanding, logical reasoning, and low hallucination rates [2] Market Position and Performance - According to Frost & Sullivan data, Zhipu AI ranks first among independent general large model developers in China and second overall, with a market share of 6.6%. By mid-2025, the company had served over 8,000 institutional clients, with 9 out of the top 10 internet companies in China using GLM models. The global download count for open-source models exceeds 45 million, with over 2.7 million registered developers on the MaaS platform, and daily token consumption rapidly increasing, reaching 4.2 trillion by November 2025. Paid API revenue surpasses the total of all domestic models [3] Business Model - The business model centers on a MaaS (Model as a Service) platform, driven by both localized and cloud deployments. Localized deployment targets government and enterprise clients, offering privatized operation and customization services, with a high customer price and stable gross margin, accounting for 84.8% of revenue in the first half of 2025, with a gross margin of 59%. Cloud deployment, through API calls and subscription services, has a low entry barrier and strong scalability, accounting for 15.2% of revenue, with rapid growth in revenue share. The company aims to increase the proportion of API revenue in the long term [4] Financial Performance - Historical financial performance shows high revenue growth, with revenues of 57 million yuan, 125 million yuan, and 312 million yuan for 2022-2024, reflecting a compound annual growth rate of over 130%. In the first half of 2025, revenue reached 191 million yuan, a year-on-year increase of 325%, surpassing the total revenue for 2023 [4] IPO Details - The IPO price is set at 116.20 HKD per share, with a global offering of 37.42 million H shares, raising approximately 4.3 billion HKD, leading to a post-fundraising market capitalization of about 51.1 billion HKD. The funds will primarily enhance general large model research (about 70%), optimize the MaaS platform infrastructure (about 10%), expand ecosystem cooperation and strategic investments (about 10%), and supplement working capital. Key investors include prominent institutions such as Shanghai Gao Yi, GF Fund, and Taikang Life, with subscription amounts accounting for about 70% of the offering size [5] Competitive Advantages - The company's core competitive advantages lie in its comprehensive self-research technology system, leading model performance, open-source ecosystem, and deep adaptation to domestic computing power. With 74% of its workforce in R&D, the core team has a strong academic background in natural language processing from Tsinghua KEG Laboratory. The rapid iteration of the GLM series, particularly GLM4.7, shows outstanding performance in programming scenarios, while AutoGLM enables AI to autonomously operate smartphones and computer GUIs, marking a new paradigm for agents. Multi-modal capabilities cover text-to-image, text-to-video, and visual understanding, with CogView-4 and CogVideoX ranking highly in open-source evaluations, translating these technological advantages into widespread applications across various industries [6] Revenue Forecast - Revenue projections for 2025-2027 are estimated at 790 million yuan (up 151% year-on-year), 1.55 billion yuan (up 97% year-on-year), and 3.22 billion yuan (up 108% year-on-year), with a gradual shift in revenue structure from localized to cloud-dominated. The overall gross margin is expected to reach 50% in 2025, stabilizing around 51% in 2026-2027, while cloud gross margins are anticipated to improve from low levels to 40%. The path to profitability is becoming clearer [7]

Core Viewpoint - Zhiyuan AI has officially listed on the Hong Kong Stock Exchange, becoming the first "domestic large model stock," reflecting strong investor interest in AI companies despite short-term profitability pressures [1][3]. Group 1: Company Overview - Zhiyuan AI, established in 2019 from Tsinghua University's Computer Science Department, is the largest independent general model developer in China, focusing on achieving Artificial General Intelligence (AGI) [3]. - The company has developed a comprehensive model matrix that includes language, vision, code, and agents, showcasing its capability from foundational algorithms to full-stack independent research and development [3]. Group 2: Business Model - Zhiyuan AI's business model features a "dual-drive" approach: high-margin localized deployments provide stable cash flow, while cloud-based API services offer future growth potential [4]. - Localized deployments, aimed at data-sensitive government and enterprise clients, are projected to have a gross margin of 59% by mid-2025, contributing approximately 85% of the company's revenue [4]. Group 3: Cloud Business Potential - The cloud business is seen as a key variable for long-term value, with a shift from localized to cloud services expected as model iterations accelerate [5]. - Cloud revenue share has rapidly increased from a low base in 2022 to 15.2% by mid-2025, with daily token consumption reaching 4.2 trillion by November 2025 [5]. Group 4: Technological Differentiation - Zhiyuan AI utilizes its self-developed GLM (General Language Model) architecture, which offers unique advantages in long text understanding, logical reasoning, and low hallucination rates [6]. - The flagship model GLM-4.7, set to release in December 2025, introduces innovative mechanisms that enhance performance in programming and complex task planning [6]. Group 5: Ecosystem Development - The company has built an "ecosystem flywheel" through a strategy of open-source and commercialization, attracting global developers and converting traffic into commercial orders [8]. - As of mid-2025, over 2.7 million developers are registered on the MaaS platform, with open-source model downloads exceeding 45 million [8]. Group 6: Financial Performance - Zhiyuan AI reported a revenue of 1.91 billion RMB in the first half of 2025, reflecting a year-on-year growth of 35.03%, but also incurred a net loss of 2.351 billion RMB due to increased costs in computing services and R&D [9][11]. - The cost structure indicates that computing service fees have become the second-largest expense after labor costs, accounting for 19% of total costs in the first half of 2025 [11]. Group 7: Valuation Insights - The market tends to use the Price-to-Sales (PS) ratio for valuation, with an expected PS of approximately 30 times for 2026 based on the IPO pricing [11]. - Despite a higher valuation compared to some peers, the scarcity of pure large model stocks and the explosive potential of cloud business justify this valuation, with expectations for rapid revenue growth leading to compression of valuation multiples [11].

Investment Rating - The report does not provide a specific investment rating for the company [1]. Core Insights - The company, Zhipu AI, is a leading independent general large model developer in China, established in 2019, and has developed a unique General Language Model (GLM) framework that excels in long text understanding and logical reasoning [6][12]. - Zhipu AI's flagship products, GLM-4.5 and GLM-4.7, have achieved top rankings in international benchmark tests and have gained significant recognition in the global developer community [6][14]. - The company has a strong market position, ranking first among independent general large model developers in China with a market share of 6.6% as of 2024 [6][15]. - Zhipu AI plans to go public on the Hong Kong Stock Exchange on January 8, 2026, with an IPO price of HKD 116.20 per share, aiming to raise approximately HKD 4.3 billion [6][15]. Summary by Sections 1. Company Overview - Zhipu AI is built on technology from Tsinghua University and aims to compete with OpenAI, focusing on a self-regressive fill-in-the-blank GLM framework [12][16]. - The company has released several significant models, including the GLM-130B, which marked its entry into the mainstream large language model market [12][16]. 2. Business Model and Operations - The business model is centered around Model as a Service (MaaS), offering both localized and cloud deployment options [21][24]. - Localized deployment accounts for a significant portion of revenue, with high margins, while cloud deployment is rapidly growing and aims to capture a larger market share [24][25]. 3. Historical Financial Analysis - Revenue has shown rapid growth, with projections indicating revenues of CNY 785 million in 2025 and CNY 1.55 billion in 2026, reflecting a compound annual growth rate of over 130% from 2022 to 2024 [1][30]. - The company has been operating at a loss due to substantial R&D investments, with cumulative R&D expenses exceeding CNY 4.4 billion from 2022 to 2024 [6][30]. 4. Core Competitiveness - Zhipu AI's competitive edge lies in its fully self-developed technology system, leading model performance, and a robust open-source ecosystem [38][39]. - The GLM series models have demonstrated significant advantages in various applications, including multi-modal understanding and generation [39][40]. 5. Profitability Forecast and Investment Suggestions - The company is expected to achieve revenues of CNY 7.9 billion in 2025, CNY 15.5 billion in 2026, and CNY 32.2 billion in 2027, with a gradual shift towards cloud-driven revenue [6][7]. - The overall gross margin is projected to reach 50% by 2025, with improvements in cloud margins as the business scales [6][7].