Analysis of Market Cap Gap Between Zhipu AI and MiniMax Post-IPO: A Study on Capital Market Preferences for B2B vs B2C Business Models

Based on the collected data, I will provide you with a detailed analysis report.

---

In early 2026, Zhipu AI and MiniMax listed on the HKEX within 48 hours of each other, creating a historic moment of “Domestic AI Dual Champions”. However, the two companies showed a clear divergence in their capital market performances:

Metric	Zhipu AI (02513.HK)	MiniMax (00100.HK)
Listing Date	January 8, 2026	January 9, 2026
Offer Price	HK$116.2 per share	HK$165 per share
Funds Raised	HK$4.348 billion	HK$5.540 billion
First-Day Gain	+13.17%	+109%
Post-IPO Stock Performance	Three consecutive days of gains	Doubled on the first day
Latest Market Cap (as of January 12)	HK$91.7 billion	HK$123.1 billion
Oversubscription Multiple	1159x	1837x

: As of the close on January 12, the market cap gap between the two companies reached , with MiniMax’s market cap being approximately that of Zhipu AI [1][2].

From the perspective of IPO pricing, both companies had pre-IPO valuations around HK$50 billion, being evenly matched. Zhipu AI was issued at HK$116.2 per share, while MiniMax was issued at the upper limit of HK$165 per share. However, after listing, the capital market “voted with real money”, and MiniMax received stronger market recognition [3].

---

: Zhipu AI follows a typical B2B (enterprise) and G2G (government) route, with its core revenue source being localized large model deployment services.

:

• Localized Deployment
  : As of 2024, localized deployment contributed
  84.5%
  of revenue, with the proportion reaching 84.8% in the first half of 2025 [4]
• Cloud MaaS Business
  : Accounted for only 15.2% of revenue, with gross margin plummeting from 76.1% in 2022 to -0.4% in the first half of 2025, falling into gross profit loss
• Customer Structure
  : Core customers are concentrated in governments, large state-owned enterprises, and leading internet companies. In 2024, over 50% of revenue came from internet companies, with the top three industries being technology, public services, and telecommunications

:

Metric	2022	2023	2024	First Half of 2025
Revenue (CN¥100 million)	0.57	1.25	3.12	1.91
Revenue Growth Rate	-	119%	150%	325%
Gross Margin	54.6%	64.6%	56.3%	50.0%

:

• High Gross Margin, Steady Growth
  : B2B customized services have pricing power advantages
• Project-Based Model
  : Relies on labor input and delivery, with long project cycles
• High Customer Concentration
  : Has a clear risk of over-reliance on specific client segments

: MiniMax has developed a consumer-facing (C-end) driven global monetization path, with its products covering over 200 countries and regions.

:

• AI Native Product Revenue
  : Accounted for
  71.1%
  of revenue in the first three quarters of 2025 [5]
• Overseas Revenue Proportion
  :
  73%
  of revenue came from overseas markets (Singapore, the U.S., etc.) in the first three quarters of 2025
• Flagship Products
  : Talkie/Xingye (AI Companion) and Conch AI (Video Generation). Talkie contributed nearly 64% of revenue in 2024, while Conch AI’s revenue share rose to 33% in the first three quarters of 2025

:

Metric	2023	2024	First Three Quarters of 2025
Revenue (US$10,000)	346	3052	5343.7
Revenue Growth Rate	-	782%	174.7%
Gross Margin	-24.7%	-	23.3%

:

• Cumulative served users exceed
  212 million
• Average monthly active users reach
  27.64 million
• The single product Talkie/Xingye has accumulated
  147 million users
• Paid user count surged
  15 times
  from the end of 2023 to September 2025

:

• Low Gross Margin, High Growth
  : C-end business has a gross margin of only 4.7%, but has a steep growth curve
• High Marketing Investment
  : Marketing expenses reached US$87 million in 2024, 2.8 times the revenue in the same period
• Low Paid Conversion Rate
  : Talkie/Xingye has an overall paid conversion rate of only 1.4%, with an average annual per-user payment of US$36-530

---

From the perspective of capital market performance, investors have given drastically different valuation premiums to the two business models:

:

1. User Scale and Growth Potential
   : Network effects and growth imagination brought by a 200 million user base
2. Global Layout
   : 73% overseas revenue proportion, with “going global” narrative and exchange rate diversification risk mitigation
3. High Productization Level
   : Direct-to-consumer AI applications with clear monetization paths
4. Technological Barriers
   : Full-modal capabilities (text, voice, video), one of only four companies globally with such capabilities

:

1. Limitations of Project-Based Model
   : Relies on labor input for delivery, making it difficult to achieve marginal cost reduction
2. Lack of Explosive Growth
   : Concentrated customers, long delivery cycles, and lack of exponential growth potential
3. Customer Stickiness Risk
   : Cloud MaaS business has negative gross margin and faces price competition

Company	Annualized Revenue	Valuation	Market Cap-to-Revenue Ratio
OpenAI	US$20 billion	~US$83 billion	~40x
Anthropic	US$9 billion	~US$35 billion	~39x
MiniMax	~US$70 million (annualized)	~HK$120 billion	~150x
Zhipu AI	~CN¥400 million (annualized)	~HK$90 billion	~200x

Referring to overseas counterparts, OpenAI and Anthropic are valued at around 40 times their annualized revenue. However, the valuations of MiniMax and Zhipu AI have far exceeded this level, which reflects a “enthusiasm premium” for Chinese AI companies in the capital market [6].

“MiniMax focuses on a C-end global product matrix, with its applications such as Talkie and Conch AI accumulating over 200 million users. Its user scale and growth potential have received high valuation premiums from the capital market. Zhipu, on the other hand, focuses on the B-end government and enterprise market, with localized deployment services at its core. It has concentrated customers and long delivery cycles, and despite stable revenue, it lacks explosive growth momentum.” [7]

“Zhipu centers on government and enterprise customers, with a business model oriented towards long-term contracts and service delivery, which is more in line with the investment logic of steady growth and long-term value, and its market pricing is relatively rational. MiniMax has strengthened its high-growth narrative through its more visible business model, high proportion of overseas revenue, C-end consumer-grade products, and global layout. Its strong performance in the initial post-IPO period reflects capital’s risk preference for high-elasticity AI companies.” [8]

---

Metric	Zhipu AI	MiniMax
Cumulative Loss (2022 - Sep 2025)	Over CN¥6.2 billion	Approx. CN¥9.2 billion
Monthly Cash Burn Rate	CN¥221 million	Approx. US$27.9 million
Cash Reserves	Approx. CN¥2.457 billion (Dec 2024)	Approx. US$1.046 billion (Sep 2025)
Cash Runway	Approx. 11 months	Approx. 37.5 months
R&D Investment Intensity	2024 R&D investment of CN¥2.195 billion, 7 times the annual revenue	Cumulative investment of US$500 million, 77% used for computing power

MiniMax’s cash position is relatively ample. Based on the current monthly cash burn rate, its existing reserves can support operations for approximately 37.5 months even without counting IPO proceeds. Zhipu faces greater cash pressure, but its cash runway has been significantly extended due to continuous equity financing in 2025 [9].

MiniMax has demonstrated strong cost control and R&D efficiency advantages:

• MiniMax
  : Cumulative R&D investment from its establishment to September 2025 is only
  US$500 million
  , accounting for less than
  1%
  of OpenAI’s US$40-55 billion in related expenditures
• Zhipu
  : 2024 R&D investment reached CN¥2.195 billion,
  7 times
  its annual revenue

This “ultra-cost-effective” R&D model has reshaped the industry’s inherent perception that “AI R&D requires huge capital investment” [10].

---

• The C-end market has a potential user scale of hundreds of millions, while the B-end market has a limited number of customers
• MiniMax’s 200 million user base provides huge monetization imagination space

• C-end products can generate massive user behavior data to drive model iteration
• A positive cycle of “user usage → data training → model enhancement → more users”

• B-end project-based services have high marginal costs, while C-end products have relatively low marginal costs once a user base is established
• MiniMax’s 15-fold surge in paid users reflects its economies of scale potential

• 73% of MiniMax’s revenue comes from overseas, with dual advantages of exchange rate diversification and global market expansion
• Aligns with international capital’s valuation preference for “global companies”

• C-end companies usually have clearer IPO paths and valuation benchmarks (refer to U.S. C-end AI companies)
• B-end companies have geographically concentrated businesses and relatively low recognition from international capital

Although the C-end model has received higher valuation premiums in the capital market, the B-end model is not without value:

1. High Revenue Stability
   : Enterprise customers have long contract terms, with highly predictable cash flow
2. Relatively Stable Gross Margin
   : Customized services have pricing power
3. Policy Support Advantages
   : Zhipu has received strategic investments from state-owned capital in Beijing, Shanghai, Hangzhou, Chengdu, Zhuhai and other places, giving it a “national team” label
4. Data Quality Advantages
   : B-end customer data is usually of higher quality

---

1. Huge R&D Investment
   : Both companies are in a loss-making state, with high R&D expenditures continuously consuming cash
2. High Computing Power Costs
   : Huge cost pressure from upstream hardware such as GPUs
3. Business Model Verification Period
   : The AI large model industry is still in the stage of commercial exploration
4. Technological Iteration Risk
   : May face disruptive impacts from new technologies and models at any time

1. Insufficient User Payment Willingness
   : Paid conversion rate is only 1.4%, with low ARPU (Average Revenue Per User)
2. Fierce C-End Competition
   : The AI companion and video generation tracks are crowded with competitors
3. Regulatory Policy Uncertainty
   : Data security and cross-border flow regulations may restrict global layout

1. High Customer Concentration
   : Over-reliance on large customers, lack of explosive growth
2. Negative Gross Margin for Cloud Business
   : MaaS business faces fierce price competition
3. Limitations of Project-Based Model
   : Large-scale expansion is limited by labor costs

---

The listing of Zhipu and MiniMax marks:

1. Collapse of the “Six Little Dragons of Large Models” Structure
   : The market is shifting from “fierce competition among many players” to “duopoly”
2. Valuation Anchoring Effect
   : Provides a valuation reference frame for subsequent AI company listings
3. Capital-Driven Industry Integration
   : Promotes survival of the fittest and integration in the industry chain

According to a research report from Huachuang Securities, after the two companies listed, the transparency of their market caps has put hard-core indicators such as revenue growth rate and gross margin on the table, forcing startups to shift from parameter comparison to actual profitability, and accelerating business model convergence [11].

Dimension	Key Focus Areas for Zhipu AI	Key Focus Areas for MiniMax
Revenue Growth	Ability to break through B-end growth ceiling	Improvement of C-end paid conversion rate
Gross Margin	Ability to turn cloud business profitable	Improvement of C-end business gross margin
Cash Flow	Financing pace and cash burn rate	User growth and cost control
Technological Barriers	R&D progress of basic models	Breakthroughs in multi-modal technology
Policy Risk	Stability of cooperation with state-owned capital	Cross-border data compliance

Given that the AI large model industry is still in a period of rapid change, it is recommended that investors:

1. Continuously track user data
   : C-end indicators such as monthly active users, paid conversion rate, and retention rate
2. Pay attention to gross margin changes
   : Profit improvement of B-end cloud services and C-end products
3. Monitor R&D efficiency
   : Computing power cost per unit and model performance improvement speed
4. Assess policy impacts
   : Data security, cross-border flow, and AI regulatory policy trends

---

The market cap divergence between Zhipu AI and MiniMax reflects the deep-seated preferences of the current capital market for AI business models:

. MiniMax’s HK$100 billion market cap and its doubling on the first day of listing reflect investors’ risk preference for “high-elasticity AI companies”.

However, this does not mean that the enterprise-facing (B-end) model has lost its value. Zhipu AI’s steady growth, state-owned capital background, and policy support advantages have also built unique competitive barriers for it. At a time when the AI large model industry is transitioning from a technology competition to a commercial verification stage,

.

In the future, with the continuous maturity of AI technology and the gradual improvement of user payment willingness, the two business models may move towards integration – B-end enterprises may extend to the C-end, while C-end enterprises may strengthen their B-end service capabilities. Ultimately, companies that can achieve a balance between technological capabilities, commercialization efficiency, and global layout will win this AI capital race.

---

[1] Yicai - “Zhipu and MiniMax Target HK$100 Billion Market Cap: China’s Large Model Industry Enters a Watershed” (https://www.yicai.com/epaper/m/202601/12/content_49166.html)
[2] 21st Century Business Herald - “Zhipu and MiniMax Shine in Their Listings: Pressure on Other “Little Tigers”?” (https://www.21jingji.com/article/20260114/herald/e4dd26bc84b9bedb8be24d7c28e819ed.html)
[3] OFweek - “Two Bell-Ringing Ceremonies in 48 Hours: China’s AI Large Model Industry Faces a Valuation Reassessment” (https://m.ofweek.com/ai/2026-01/ART-201712-8120-30678981.html)
[4] Yicai - “Zhipu and MiniMax Target HK$100 Billion Market Cap” (https://www.yicai.com/epaper/m/202601/12/content_49166.html)
[5] Eastmoney - “MiniMax Officially Lists on HKEX, Opens 42% Higher with Market Cap Reaching HK$71.9 Billion” (https://finance.eastmoney.com/a/202601093613983298.html)
[6] Guancha.cn - “MiniMax’s Stock Doubles on First Day of Listing, Becoming the World’s First AI Listed Company with HK$100 Billion Market Cap” (https://www.guancha.cn/economy/2026_01_09_803325.shtml)
[7] Yicai - Analysis by Guo Tao (https://www.yicai.com/epaper/m/202601/12/content_49166.html)
[8] China Business Journal - “Zhipu vs. MiniMax: HKEX Showdown May Trigger 2026 AI Listing Wave” (http://www.cb.com.cn/index/show/zj/cv/cv135341401260)
[9] OFweek - “Sci-Tech Innovation: MiniMax and Zhipu Submit Listing Applications Successively, B2B vs B2C Business Showdown” (https://m.ofweek.com/ai/2025-12/ART-201700-8420-30677928.html)
[10] Eastmoney - “MiniMax, Which Successfully Listed with a Market Cap Over HK$80 Billion, Reshapes the Global AGI Landscape with “Extreme Efficiency”” (https://caifuhao.eastmoney.com/news/20260109192925003664600)
[11] Securities Times - “AI Companies Intensively List on Capital Markets, Commercialization Process Pressed the “Accelerator Button”” (https://www.stcn.com/article/detail/3582200.html)