Recently, a Chinese foundation model company has refreshed its market value record.
GLM's intraday market value exceeded HK$1 trillion, with a year-to-date increase of over 1900%. By the close, the company's total market capitalization reached HK$1.1 trillion, roughly equivalent to half of Alibaba's, nearly twice that of Meituan, and three times that of JD.com.
Having been listed for less than half a year, GLM has surged from a company valued at the HK$50 billion level into the trillion-HK-dollar valuation range in the Hong Kong stock market.
Above the threshold of the Hong Kong stock market's "Trillion Club" are industrial giants like Tencent, Alibaba, China Mobile, and ICBC, all backed by proven user entry points, cash flows, or infrastructure attributes.
GLM's entry into this coordinate system signifies that the public market's pricing of Chinese large model companies is continuously inflating.
However, the roller-coaster-like growth rate has sparked a debate over whether this is a myth or a bubble.
GLM's first financial report after listing disclosed its full-year 2025 revenue data: revenue of 724 million yuan, a net loss of 4.718 billion yuan, and R&D expenses of 3.18 billion yuan.
A company with revenue of 700 million yuan the previous year has now reached a trillion-level market cap. Bulls view it as a valuation breakthrough for China's AI industry; while bears believe the current stock price carries strong bubble attributes, having run far ahead of fundamentals.
To some extent, this debate has transcended GLM itself, rising to the valuation of China's large model industry as a whole.
Latest news indicates that DeepSeek's valuation exceeded $50 billion after a new round of financing, while the capitalization expectations for Stepfun and Moonshot AI continue to heat up, both targeting the Hong Kong stock market.
The valuation curve for Chinese large model companies is shifting from the primary market to the secondary market. With GLM becoming the first trillion-dollar market cap large model enterprise, this curve is constantly breaking new upward boundaries.
1
The direct catalyst for GLM's latest round of market value surge is GLM-5.2.
Since last year, the focus of mainstream model upgrades within the industry has centered on Coding, Agent, long-range tasks, and complex engineering capabilities.
In these scenarios, GLM's GLM series can rank in the first tier, with the global strongest reference being Anthropic.
This year, Anthropic has completed a $65 billion financing round, with a post-money valuation of $965 billion. The core reason for its追捧 is primarily that Claude Code has generated revenue in the developer scene.
Official disclosed data shows that Claude Code's annualized revenue has exceeded $2.5 billion, doubling since the beginning of the year.
This sample has provided a new valuation anchor for global AI companies.
C-end chat products prove user scale, while code and enterprise Agent prove payment capability. The former creates buzz, the latter approaches revenue; and the commercialization capability behind revenue is precisely the key for capital markets to evaluate large model enterprises.
GLM's revaluation in this round is precisely because the market is observing it within the coordinate of "China's Anthropic."
Recently, the indirect interaction between Tesla CEO Elon Musk and GLM founder Tang Jie amplified this imagination.
Musk predicted that China might see a model approaching Anthropic's Fable level in the first quarter of 2027.
Tang Jie responded that it wouldn't take that long. Although this reply itself is brief and lacks information, it highlights that domestic models are already within the narrative of catching up to global cutting-edge models, and it's a matter of who can keep pace with leading competitors.
This year, besides GLM's stock price surge, there have been frequent capital moves in the domestic model circle.
GLM passed a resolution for an A-share issuance earlier this month, planning to list on the STAR Market to raise no more than 15 billion yuan, with 12 billion yuan directed towards general foundation model R&D and 2 billion yuan towards the MaaS one-stop service platform.
Almost simultaneously, MiniMax signed a tutoring agreement with China Securities, officially launching its A-share IPO process and evaluating issuing RMB shares to list on the STAR Market. Two Hong Kong-listed large model companies simultaneously advancing A-share paths indicate that China's AI valuation curve is extending from Hong Kong stock elasticity to A-share industry financing.
Meanwhile, DeepSeek is疯狂吸纳 financing in the primary market.
Recently, DeepSeek completed over $7.4 billion in financing, with a valuation exceeding $50 billion. This AI giant from Hangzhou, relying on excellent model efficiency, global developer buzz, and open-source ecosystem, has pushed the primary market price of Chinese large models to a new height.
In the secondary market, following the successive listings of GLM and MiniMax, the Hong Kong stock market is also turning to embrace the AI industry.
This year, 65 companies have listed in Hong Kong, raising a combined total of over HK$176.5 billion. Hard tech and high-end manufacturing enterprises contributed nearly HK$129.3 billion, accounting for over 70%. Semiconductors, electrical equipment, and AI companies have become the main lines of issuance. Capital is treating the Hong Kong stock market as a repricing market for China's hard tech assets.
In other words, GLM's trillion-dollar market cap stems not only from the outstanding performance of GLM-5.2 but also from the simultaneous occurrence of several conditions: the rise in global AI asset valuations, the acceleration of capitalization for Chinese large model companies, the lack of pure large model targets in Hong Kong stocks, and the renewed aggregation of hard tech capital.
Therefore, what the market is truly buying is that the valuation boundary for China's AI can continue to widen. However, the wider this boundary expands, the debate over industry bubbles will persist.
2
Seeing GLM's trillion-dollar market cap, most investors intuitively judge that there must be a bubble component within.
A company with annual revenue of just over 700 million yuan, with an intraday market cap exceeding HK$1 trillion, would make any valuation model nervous.
Moreover, GLM's stock price had previously experienced significant volatility. In late May, its intraday market cap once exceeded HK$880 billion before quickly retreating; during today's surge above the trillion mark, intraday fluctuations remained intense, with over HK$100 billion in market value evaporating within just over ten minutes from the peak.
However, this is precisely what makes the AI行情 most特殊. The market acknowledges the high price on one hand while continuing to chase leading targets on the other.
Bubble warnings and valuation upgrades occur simultaneously, constituting the底色 of this AI capital cycle.
Meanwhile, differentiation has emerged in the Hong Kong新股 market.
In the first quarter of this year, the Hong Kong IPO破发 rate rose to 44.74%, significantly higher than last year. Traditional industries and companies with insufficient subscriptions are under pressure, while hard tech, AI, and new energy enterprises are still pursued by capital. Capital is not buying new shares indiscriminately; it's concentrating its冒险 in only a few sectors.
Clearly, the AI industry, represented by GLM, belongs to the assets being集中冒险 upon. Breaking through the trillion mark relies on the market's scarce expectations for Chinese AI infrastructure companies.
However, capital will not投入无止境. In early July, GLM and MiniMax will face their first post-listing限售股解禁.
GLM's解禁时间 is July 8th, with a scale of approximately 25.6816 million shares, accounting for about 5.76% of the total share capital. Based on recent stock prices, this corresponds to a解禁市值 of about HK$26.9 billion. Before解禁, GLM's流通股 were approximately 17.3508 million shares. In other words, even though the proportion of total shares is not high, the新增可交易筹码 will significantly alter the previous low流通盘结构.
MiniMax faces greater pressure. It will解禁 approximately 107 million shares on July 9th, accounting for about 34.25% of its total share capital. Among these, financial investors hold over one-third.
The stock prices of both companies have experienced substantial increases since their listings, so the market naturally worries提前 about减持 pressure.
In fact, it's not just domestic large models; the global market is also experiencing similar valuation conflicts.
Two weeks before the隔空对话 between Musk and Tang Jie, SpaceX's market cap迅速突破 $2 trillion after listing, with its stock price also climbing to the $210 range.
This aerospace and satellite internet company heavily consolidated resources before listing, strengthening its industry position within the narratives of AI infrastructure, xAI synergy, Starlink network, and future computing power.
For this company, Professor Damodaran of NYU Stern School of Business gave a valuation lower than the market pricing, warning that SpaceX's high valuation already incorporates大量未来假设.
In the past few trading sessions, SpaceX's stock price experienced a significant回调, falling to the $170 range.
In early June, Michael Burry, founder of Scion Asset Management and the原型投资人 portrayed in the movie "The Big Short," also directed质疑 at Anthropic.
He believes the研发成本 behind Claude's large model is too high, and a near-trillion-dollar valuation is difficult to支撑 solely by its current business model.
Such significant认知分歧 within the industry actually shows that seemingly opposing views are not contradictory; during a period of高速产业增长, bubbles have naturally become part of AI valuation.
The key factor determining whether this part of the "bubble" constitutes long-term溢价 hinges on the commercialization capability of AI enterprises.
3
After breaking the trillion-dollar market cap, the real issue before GLM is the speed of revenue growth.
In 2025, GLM's revenue was 724 million yuan, a year-on-year increase of 131.9%. Among this, localized deployment revenue was 534 million yuan, up 102.3% year-on-year; cloud deployment revenue was 190 million yuan, up 292.6% year-on-year; and enterprise-level智能体业务 revenue was 166 million yuan, up 248.8% year-on-year.
Specifically, GLM's MaaS API platform ARR is approximately 1.7 billion yuan, having increased 60-fold over the past 12 months, with platform毛利率提升 to 18.9%. This means the revenue proportion from MaaS is continuously increasing.
In Q1 2026, faced with旺盛需求 for model calls, GLM directly chose to—raise prices.
Since February this year, GLM has implemented structural adjustments to the pricing of its GLM Coding Plan套餐, with overall increases starting at 30%. The理由 given is the rapid提升 in user scale and call volume, necessitating保障 stability and service quality under high负载.
However, how much revenue growth this part of the营收 ultimately corresponds to will be revealed in GLM's next financial report.
Meanwhile, as an independent model company, GLM faces巨大的成本压力.
In 2025, GLM's net loss was 4.718 billion yuan, adjusted net loss was 3.182 billion yuan, and R&D开支 were 3.18 billion yuan. R&D投入相当于收入的4倍多.
Model companies can burn money, but they must let the capital market see the expectation of long-term profitability.
Fortunately, the foundation of调用量 across the entire Chinese market is continuously抬升.
IDC data shows that in 2024, the call volume of China's enterprise-level MaaS market was 114 trillion Tokens. In 2025, it跃升至 1,944 trillion Tokens, and is projected to reach 40,000 trillion Tokens in 2026. Tokens are transitioning from a technical metric to a commercial settlement unit.
ByteDance, with the largest体量 among domestic models, sees its Doubao large model's daily Token call volume突破 120 trillion, doubling in three months and growing 1000-fold compared to its release in May 2024. The number of enterprises with cumulative call volumes exceeding one trillion Tokens increased from 100 at the end of last year to 140.
Simultaneously, C-end scenarios are also expanding.
QuestMobile data shows that in March 2026, the MAU scale of China's AI-native Apps reached 440 million. Doubao, Qianwen, and DeepSeek ranked in the top three. AI applications have transformed from niche tools into大众产品.
However, user scale does not代表利润水平. The industry's泡沫担忧本质上 stems from a sense of无力感 towards AI's无底线烧钱 and the lack of visible盈利预期. On this issue, Anthropic remains the最强参照.
In Q2 this year, Anthropic is预计 to achieve revenue of $10.9 billion and approximately $559 million in operating profit. If this expectation materializes, it will become a全球转折样本 for AI company commercialization. The market will therefore believe that large model companies do not necessarily remain in the "burn money for scale" stage indefinitely.
This is also why Chinese AI companies are being重估.
If Anthropic can接近盈利 through code, Agent, and enterprise APIs, the domestic market will naturally search for corresponding Chinese标的. GLM, DeepSeek, MiniMax, Moonshot AI,以及 several leading AI大厂 will all be placed within this commercialization curve for comparison.
With GLM率先突破万亿, this has become the共同命题 for China's AI industry moving forward.
The valuation boundary has been opened,争议 over bubbles will not cease, and market volatility and uncertainty still exist.
To consolidate this valuation boundary, large model manufacturers must first convert model capabilities stably into revenue and further push that revenue towards毛利 and cash flow. Simultaneously, more domestic manufacturers need to跑出有效样本 on the path of catching up to Anthropic.
GLM's站上万亿市值 today marks the first peak in the重估 of China's AI industry. Whether this peak can be守住, whether it can transform into an industry curve, will ultimately be answered by the commercialization results of domestic large model manufacturers.
This article is from the WeChat public account "字母榜", author: Li Zhaofeng
