Analysis of Zhipu AI's First-Day Trading Performance: Comparing with AI Chip Companies to Reveal Deep-Seated Issues in the Large Model Industry

Based on the data I collected, it is necessary to first clarify a fact:

[1]. However, compared to AI chip companies listed in the same period, its gain is indeed relatively modest, and this comparison reflects the deep-seated issues faced by the large model industry.

Indicator	Data
Offer Price	116.2 HKD/share
Opening Price	120 HKD (+3.27%)
Closing Price	131.5 HKD (+13.17%)
Intraday High	129.8 HKD (+11.7%)
Total Market Capitalization	HKD 57.89 billion
Over-subscription	Public offering: 1159.46x, International placement: 15.28x

Notably, Zhipu AI’s first trading day saw a trend of “high opening with fluctuations, briefly breaking below the offer price before rebounding”, and intraday it briefly fell below the offer price of 116.2 HKD[2]. This volatile trend reflects the market’s game-like mentality regarding the pace of large model commercialization.

Company	First-Day Gain	Business Type
Muxi Co., Ltd.-U	+693%	AI Chip
Moore Threads-U	+425%	AI Chip
Biren Technology	+76%	AI Chip
Zhipu AI	+13.17%	Large Model

This comparison reveals several core issues faced by the large model industry:

Large model companies generally face the dilemma of “high investment, slow returns”. According to the prospectus, from 2022 to 2024, Zhipu AI’s R&D expenditure soared from RMB 84.4 million to RMB 2.195 billion[3], with cumulative R&D investment exceeding RMB 4.4 billion. However, its 2024 revenue was only RMB 312.4 million, resulting in a huge gap between revenue scale and R&D investment.

“Even OpenAI, which developed ChatGPT, once faced pressure from shareholders asking ‘when will you be profitable’. Even with its paid services now available, OpenAI is still in a phase where revenue is booming but losses are expanding simultaneously.”[3]

An analysis report from Huaan Securities points out that Zhipu AI’s listing on the Hong Kong Stock Exchange will guide the narrative logic of AI large model vendors from “telling technology stories” to “realizing commercial value”[4]. The market’s evaluation criteria for the AI industry have shifted from depicting “disruptive AI visions and long-term potential” to expecting “feasibility” as a key dimension[5].

An Yun, Deputy Director of the Artificial Intelligence Research Institute at SAI Intelligent Industry Research Institute, stated:

“The capital market’s recognition of the AI field has shifted from early concept verification to more rational commercial acceptance. Investors no longer value companies solely based on technical indicators such as model parameter scale, but instead focus more on AI enterprises’ actual ability to convert technology into stable revenue and cash flow.”[5]

Zhipu AI’s listing reflects profound differentiation in the AI large model industry[3]:

Enterprise	Strategic Choice
Zhipu AI	Adheres to general large models, listed on Hong Kong Stock Exchange
MiniMax	Multi-modal large model, concurrently listed on Hong Kong Stock Exchange
Moonshot AI	Abundant cash (over RMB 10 billion), no listing in the short term
01.AI, Baichuan Intelligence	Abandons pre-trained general models, shifts to vertical sectors
ByteDance, Alibaba	Continues to increase investment in large model business

Yang Zhilin, Founder of Moonshot AI, publicly stated:

“Compared to the secondary market, we can still raise larger amounts of capital from the primary market.”[3]

Although Zhipu AI has made significant technological breakthroughs—GLM-4.7 ranked first globally among open-source models in the Code Arena blind test, and set an industry record with 73.8% in the SWE-bbench-Verified test[2]—the conversion of technological advantages into commercial value still requires time to be validated.

Zhipu AI’s business model is mainly based on MaaS (Model as a Service), generating revenue through API calls and local privatized deployment. According to the prospectus, Zhipu AI’s local deployment revenue accounts for 85% of total revenue, while cloud-based revenue has increased from 0 to 15%[1]. The full-year MaaS year-on-year growth rate in 2025 exceeded 900%, showing strong growth momentum[6].

Despite short-term profit pressures, the market remains optimistic about the medium- and long-term prospects of the AI large model industry:

1. Continued Expansion of Market Scale
   : Data from the China Academy of Information and Communications Technology shows that in 2024, the scale of China’s core artificial intelligence industry exceeded RMB 900 billion, with a growth rate of 24%, and is expected to exceed RMB 1.2 trillion in 2025[5].

2. Gradual Maturity of Business Models
   : Zhipu AI’s annualized MaaS ARR (Annual Recurring Revenue) increased from RMB 20 million to over RMB 500 million, achieving a 25-fold growth in 10 months[6], indicating that the MaaS model has large-scale potential.

3. Global Competitive Landscape
   : The United Nations Conference on Trade and Development (UNCTAD) predicts that the global artificial intelligence market size will soar from USD 189 billion in 2023 to USD 4.8 trillion in 2033[6]. Zhipu AI has established a developer ecosystem in 37 countries, with European and American developers accounting for 42%[4].

The modest gain of Zhipu AI on its first trading day relative to AI chip companies actually reflects the market’s cautious wait-and-see attitude towards the commercialization process of the large model industry. The core issues reflected behind this include:

• Profit Model Validation Period
  : Large model companies are still in a critical stage of converting technological advantages into sustainable profitability
• Valuation Logic Restructuring
  : The capital market has shifted from chasing technological concepts to demanding tangible commercial returns
• Intensifying Industry Differentiation
  : Different companies choose differentiated development paths based on their own resource endowments
• Long Input-Output Cycle
  : The time gap between high R&D investment and revenue growth remains a common challenge for the industry

As the “world’s first large model stock”, Zhipu AI’s listing provides the industry with its first public valuation anchor, as well as a reference benchmark for the capitalization process of subsequent large model companies. Listing is just a new starting point; how to convert technological advantages into sustainable profitability will be a core proposition that Zhipu AI and even the entire large model industry need to continuously address[2].

---

[1] Securities Times Network - “World’s First Large Model Stock” Zhipu AI Exceeds HKD 57 Billion in Market Capitalization on First Trading Day (https://finance.eastmoney.com/a/202601083613015124.html)

[2] OFweek AI Network - Observing the Certainties and Uncertainties of the “World’s First Large Model Stock” from Zhipu AI’s First-Day Trading Performance (https://m.ofweek.com/ai/2026-01/ART-201700-12001-30678584.html)

[3] Sina Finance - The World’s First Large Model Stock is Born! Zhipu AI Surges 13.17% on First Trading Day (https://finance.sina.com.cn/jjxw/2026-01-08/doc-inhfqyhi5216174.shtml)

[4] Yicai Global - “World’s First Large Model Stock” Goes to Zhipu AI (https://www.yicai.com/news/102981627.html)

[5] Xinhua News - AI Enterprises Flock to the Capital Market, Pressing the “Accelerator Button” for Commercialization (https://www.news.cn/fortune/20260109/f063f5dd0e1f4c798dc26b7247b4200a/c.html)

[6] The Paper - “China’s Version of OpenAI” Zhipu AI Lists, the Gold Content of China’s AI Large Models Continues to Rise (https://www.thepaper.cn/newsDetail_forward_32343111)