Zhipu AI HK IPO Valuation Logic and Impact on Chinese AI Big Model Companies' Listing

Based on collected data and market information, I systematically analyze the valuation logic of Zhipu AI’s Hong Kong IPO and its supporting role in the listing of Chinese AI big model companies:

• Offer Price
  : 116.20 HKD
• Number of Shares Offered
  : 37.4195 million H-shares
• Expected Fundraising
  : ~4.35 billion HKD
• Overall Valuation
  : ~4 billion USD (31.2 billion HKD) [1]
• Expected Listing Date
  : January 8, 2026

• Revenue Scale
  : ~150 million USD in 2024 (significant growth compared to 2023) [2]
• Revenue Growth Rate
  : From 20 million USD in 2022 to 150 million USD in 2024, with a CAGR exceeding 150%
• Gross Margin
  : Increased from 52% in 2023 to 58% in 2024 [2]
• Loss Situation
  : With increased R&D investment, net loss in 2024 was ~120 million USD [2]
• Revenue Structure
  : 85% from local deployment, 15% from API services [3]

According to data analysis, Zhipu AI’s

, and the expected 2025 PS ratio is about 13x, which is within a among AI big model companies [0]:

Company	Valuation (100M USD)	2024/2025 Revenue (100M USD)	PS Ratio
OpenAI	1,000	13 (2025E)	77x
Anthropic	300	9 (2025E)	33x
Zhipu AI	40	1.5 (2024)	27x
MiniMax	40	0.8 (2024)	50x

:

• Zhipu AI’s PS ratio is
  significantly lower
  than OpenAI’s 77x and Anthropic’s 33x
• Compared to domestic peer MiniMax (50x PS), Zhipu’s valuation is more conservative
• Reflects investors’
  cautiously optimistic attitude
  towards Chinese AI big model companies

[1,5]:

• SenseTime
  : Current PS is ~11x, but has entered the profit phase
• 4Paradigm
  : PS ~9x, enterprise-level AI platform
• Zhipu AI
  : PS 27x, in high-growth phase

:

1. Technological Leadership
   : GLM-4.5 ranks 3rd globally and 1st domestically in comprehensive scores across 12 authoritative evaluations [3]
2. Open Source Strategy
   : GLM series is fully open-sourced under MIT License, which helps with ecosystem building [3]
3. Growth Potential
   : 2025 revenue is expected to reach 300 million USD, a year-on-year increase of 100% [0]
4. Tsinghua-affiliated Team
   : Core technical team comes from Tsinghua University, with strong R&D capabilities [3]

Although specific DCF model data is not available, we can infer from the market environment:

:

• 2024-2027 CAGR is expected to remain at 80-100%
• 2025 revenue target: 300 million USD
• 2027 revenue forecast: 800 million-1 billion USD

:

• Currently in high-investment phase, R&D accounts for over 80% of revenue
• Expected to break even in 2027-2028
• Gross margin continues to improve (58% in 2024), indicating economies of scale are emerging

:

• Considering the high-risk nature of the AI industry, WACC may be between 12-15%
• Terminal value multiple may use 3-4x revenue (conservative estimate)

• AI boom continues
  : In 2025, Hong Kong-listed AI concept stocks will see an IPO boom, with at least 6 AI companies planning to list [1]
• High investor enthusiasm
  : Moore Threads’ GPU IPO retail subscription reached 2751x, a three-year high [5]
• Policy support
  : China’s “15th Five-Year Plan” emphasizes technological self-reliance, with AI as a key area [5]

• GLM-4.5 Leading Performance
  : Ranks 3rd globally in 12 benchmark tests including MMLU Pro and AIME24 [3]
• High Parameter Efficiency
  : GLM-4.5’s parameter count is only 1/2 of DeepSeek-R1 and 1/3 of Kimi-K2, but with better performance [3]
• Fast Inference Speed
  : The high-speed version has a generation speed of over 100 tokens per second [3]

• Local Deployment
  (85% revenue): For B-end customers, providing private deployment
• API Services
  (15% revenue): Standardized services with obvious economies of scale
• Agent Capabilities
  : GLM-4.5 natively integrates reasoning, coding, and agent capabilities [3]

• Invested by internet giants like Alibaba and Tencent [1]
• Top VCs like Hillhouse Capital have dual-line bets [3]
• Supported by government-backed funds

• Domestic Competition
  : Giants like ByteDance (Doubao), Baidu (Ernie), Alibaba (Tongyi Qianwen) continue to invest
• International Competition
  : OpenAI, Anthropic, Google (Gemini) lead in technology
• Open Source Competition
  : Companies like DeepSeek use “open source + high cost-performance” strategies to disrupt the market [1]

• Project-based Business Characteristics
  : 85% of revenue comes from local deployment, facing project-based issues similar to SenseTime [3]
• Dependence on Sales Capabilities
  : Local deployment model highly relies on sales teams and customer relationships
• Low Standardization
  : Difficult to form network effects and economies of scale

• Continuous Losses
  : 2024 net loss of 120 million USD, with high cash flow pressure [2]
• High R&D Investment
  : R&D accounts for over 80% of revenue, making it difficult to profit in the short term
• Valuation Volatility Risk
  : If 2025 revenue falls short of expectations, it may face a valuation downgrade

• Zhipu AI is not currently on the “Entity List”, but there is uncertainty in the future
• Computing power supply chain relies on U.S. companies like NVIDIA

Zhipu AI’s listing with

provides an important valuation reference point for Chinese AI big model companies:

:

• Proves that the market is willing to pay high PS multiples for AI big model companies
• Validates the valuation logic of “technological strength + high growth”
• Provides pricing reference for subsequent IPOs of companies like MiniMax

:

• If Zhipu performs poorly after listing, it may affect the valuation of subsequent companies
• Investors may give different valuation premiums to different technical routes

:

1. Retail Subscription Multiple
   : Whether it reaches Moore Threads’ level of 2751x [5]
2. Quality of Cornerstone Investors
   : Whether international top institutions participate
3. Post-Listing Performance
   : First-day rise/fall and aftermarket stability

:

• If subscription is enthusiastic and post-listing performance is good → Strengthen market confidence and increase subsequent IPO valuations
• If subscription is cold or breaks below the offer price → Trigger a re-examination of AI valuations by the market

Zhipu AI’s

will undergo market testing:

:

• Local deployment customer renewal rate remains high
• API service revenue share increases
• Gross margin continues to improve

:

• Validate the sustainability of the local deployment model
• Explore the best commercialization path for Chinese AI companies
• Provide business model references for other companies

Valuation methods that Zhipu AI’s IPO may adopt:

:

1. PS-led
   : Suitable for high-growth, unprofitable companies
2. DCF Supplement
   : Based on long-term profit forecasts
3. Benchmarking
   : Reference international peers like OpenAI and Anthropic

:

• Token Economy
  : Valuation based on API call volume and token count
• User Value
  : Number of enterprise customers and ARPU value
• Technical Assets
  : Model performance, number of patents, quality of R&D team

1. Valuation Rationality
   :
   • 27x PS is within the reasonable range of the AI industry, lower than OpenAI but higher than traditional software companies
   • Considering technological leadership and high growth potential, the valuation
     has certain support

2. Supporting Logic
   :
   • Technological strength (GLM-4.5 ranks 3rd globally)
   • Growth potential (2025 revenue is expected to double)
   • Market environment (AI boom + policy support)
   • Shareholder background (Alibaba, Tencent, Hillhouse)

3. Risk Warnings
   :
   • Business model not fully validated
   • Increased competition may affect profitability
   • Geopolitical uncertainty
   • Valuation is at a historical high, with high correction risk

:

1. Technological Leadership
   : Achieve excellent results in authoritative benchmark tests
2. Open Source Strategy
   : Build ecosystem and expand influence
3. Hybrid Model
   : B-end local deployment + C-end API services
4. Top Shareholders
   : Gain support from internet giants and top VCs

:

1. Overpromising
   : Avoid setting unrealistic revenue targets
2. Single Dependency
   : Do not over-rely on a single customer or revenue source
3. Ignoring Cash Flow
   : Focus on cash flow health while pursuing growth

:

1. Zhipu AI’s post-listing performance (Jan-Feb 2026)
2. MiniMax’s IPO pricing and subscription situation
3. Whether Q1-Q2 2026 financial reports meet growth expectations
4. IPO progress of international giants like OpenAI and Anthropic

---

[0] Jinling API Data - Python Calculation Analysis
[1] Yahoo Finance - “Chinese AI Unicorns MiniMax and Zhipu Rush for IPO, Revenue Details Disclosed in Prospectus” (https://hk.finance.yahoo.com/news/中國ai獨角獸minimax和智譜衝刺ipo-招股書中披露營收詳情-064313689.html)
[2] Yahoo Finance - “Zhipu Achieves Dual Growth in Revenue and User Scale, Rushing for China’s First AI Startup IPO” (https://hk.finance.yahoo.com/news/智譜實現收入和用戶規模雙成長-衝刺中國首個ai初創公司ipo-081000321.html)
[3] Zhihu - “Zhipu and MiniMax Compete for ‘Global Big Model First Stock’, Hillhouse Capital Makes Dual-Line Bets” (https://www.zhihu.com/question/1985979705279673214)
[4] Yahoo Finance - “OpenAI Selected as Yahoo Finance 2025 Company of the Year” (https://hk.finance.yahoo.com/news/openai當選yahoo財經2025年度企業-051905048.html)
[5] Yahoo Finance - “Chinese GPU Newcomer Moore Threads, Retail Subscription Multiple Hits Three-Year High” (https://hk.finance.yahoo.com/news/中國gpu新貴摩爾線程-散戶認購倍數創三年來之最-233148786.html)
[6] Yahoo Finance - “6 Mainland Tech Companies to List in Hong Kong Within Weeks, AI Concept Stocks Rush for Boom, Raising Hundreds of Billions” (https://hk.finance.yahoo.com/news/內地6科企數周內港上市-ai概念股趕熱潮-集資料達數百億-185800289.html)