KNOWLEDGE ATLAS's Trillion-Dollar Valuation Sparks Debate: Is It a True AI Unicorn or a Market Anomaly?

Deep News
Jun 22

The surge in KNOWLEDGE ATLAS (02513.HK) shares has ignited intense market discussion. The stock closed at HK$2,410 today, marking a 15% increase from the previous session. At this closing price, the company's market capitalization has soared to approximately HK$1.0745 trillion.

MINIMAX also saw a significant rise, closing at HK$616.5, up 23.89% from the previous day, giving it a market value of around HK$193.4 billion.

As two of the few listed large language model (LLM) concept stocks in the Hong Kong market, KNOWLEDGE ATLAS and MINIMAX have become standout performers against a generally weak market backdrop. It is noteworthy that KNOWLEDGE ATLAS initially faced a decline on its debut, indicating early skepticism from investors.

While KNOWLEDGE ATLAS and MINIMAX are soaring, the share prices of established tech giants like Tencent, Alibaba, Xiaomi, Meituan, and Li Auto are hitting new lows. KNOWLEDGE ATLAS's market cap now equals a quarter of Tencent's, half of Alibaba's, nearly double that of Xiaomi, twice that of Meituan, and three times that of either JD.com or Baidu.

The extreme rally has led some online commentators to jest about the company's name, suggesting it should be called "Unbelievable" instead.

Financial Performance: High Revenue Growth Amidst Substantial Losses

Founded in 2019 and originating from Tsinghua University's technology transfer, KNOWLEDGE ATLAS operates a business model centered on MaaS (Model-as-a-Service), primarily through API calls for its large models, similar to the monetization strategies of OpenAI and Anthropic. As one of China's earliest LLM developers, it pioneered the GLM (General Language Model) pre-training framework and has been dubbed the "Chinese OpenAI."

However, its commercialization progress lags behind that of its international counterparts. The company's annual report shows 2025 revenue of 724 million yuan, a 131.9% increase from 312 million yuan the previous year. Gross profit was 297 million yuan, up from 176 million yuan.

Cloud deployment revenue in 2025 reached 190 million yuan, a surge of 292.6% from 48.5 million yuan, attributed to increased model usage driven by continuous performance improvements. Localized deployment revenue was 534 million yuan, a 102.3% increase from 264 million yuan.

Despite the top-line growth, KNOWLEDGE ATLAS reported a net loss of 4.718 billion yuan for 2025, widening from 2.958 billion yuan. On an adjusted basis, the net loss was 3.182 billion yuan, compared to 2.466 billion yuan the previous year.

Rising Share Price Conceals Significant Underlying Risks

KNOWLEDGE ATLAS listed on the Hong Kong Stock Exchange on January 8, 2026, issuing 37.4195 million shares at HK$116.2 each, raising a total of HK$4.3 billion. After deducting listing expenses of HK$175 million, the net proceeds were HK$4.173 billion, making it the world's first publicly traded pure-play LLM company.

Its IPO attracted a star-studded lineup of cornerstone investors, including international long-term funds, prominent industrial capital, and investment institutions. Eleven cornerstone investors, such as JSC International Investment Fund SPC, JinYi Capital, Perseverance Asset Management, Shanghai Gaoyi, WT Asset Management, Taikang Life Insurance, GF Fund Management, and 3W Fund Management, subscribed to approximately HK$2.98 billion worth of shares, accounting for nearly 70% of the offering.

A critical point of concern is that the total shares issued in the IPO represented only 5.76% of the company's total share capital. After accounting for cornerstone holdings, the actual free float available for public trading is a mere 1.81%. This extremely limited free float means the stock price can be significantly influenced by relatively small amounts of capital.

A key factor supporting the high share price is that the company has not yet entered the post-IPO lock-up expiration period. Once the lock-up period ends in July 2026, a substantial number of shares will become eligible for trading, potentially leading to significant price volatility. Prior to this, the stock has already experienced sharp fluctuations, having fallen 30% from its peak, fueling speculation among some investors that it may be a manipulated "pump-and-dump" stock.

On the business front, the recent U.S. government restriction on foreign access to Anthropic's latest models, leading to the takedown of Fable 5 and Mythos 5, was followed by KNOWLEDGE ATLAS's launch of its new flagship model, GLM-5.2. This move garnered support from major international banks.

JPMorgan Chase maintained its "Overweight" rating on the stock and raised its December 2026 target price from HK$1,400 to HK$1,800. The bank noted that the GLM-5.2 launch and its pricing strategy indicate the Chinese LLM industry is not entirely engaged in a price war. While mature capabilities see price reductions, premium pricing for cutting-edge model upgrades is possible, potentially improving the commercial outlook for the model layer.

JPMorgan highlighted that compared to GLM-5.1, GLM-5.2 eliminated lower pricing tiers, applying a higher uniform rate, resulting in an average API price increase of 13%. Its pricing is estimated to be 1.2 times that of Kimi's K2.7-Code, 2.5 times that of MINIMAX's M3, 4.9 times that of DeepSeek's V4Pro, and 14.4 times that of V4 Flash. The performance gains for GLM-5.2 primarily stem from reinforcement learning and post-training optimization rather than a significant increase in model scale, suggesting potential for improved profit margins without a corresponding rise in cost base.

Nevertheless, the sustained share price climb harbors substantial risks. Post-lock-up, a significant increase in liquidity could lead to heavy selling pressure. A 1% divestment by major shareholders would represent over HK$10 billion in potential selling, a figure that dwarfs the stock's recent average daily trading volume of HK$11.3 billion (approximately 4.6 million shares).

The soaring valuation of KNOWLEDGE ATLAS has also provided opportunities for competitors like Moonshot AI (Kimi). Kimi has been actively raising capital, securing over $3.9 billion in the past six months, bringing its total funding to over 37.6 billion yuan, making it the most funded LLM startup in China. It is reportedly seeking new funding at a pre-money valuation of $30 billion, indicating that risks in the sector are also accumulating.

Lessons from History: The Volatile Fate of Tech IPOs

Many Hong Kong market participants recall the trajectory of Kuaishou. Its shares surged over 100% on its debut in 2021, reaching a peak of HK$417.8 and a market cap of HK$1.74 trillion in February 2021. Today, its shares trade around HK$45.52, with a market cap of approximately HK$197 billion.

A more recent example is Xiaomi. Its share price rallied as its electric vehicle deliveries began, surpassing HK$60 in September 2025 and achieving a market cap of around HK$1.6 trillion. However, within just six months, the hype subsided. The stock has since been labeled a "laggard," closing recently at HK$23.72 with a market cap of HK$613 billion, representing a loss of nearly HK$1 trillion in value. Stories of retail investors who bought at the peak are now common on online forums and social media.

Currently, KNOWLEDGE ATLAS has completed its IPO guidance and is preparing for a listing on Shanghai's STAR Market. If successful, it would become the first pure-play LLM stock on the A-share market, creating an "A+H" dual-listing structure.

A critical question for the market is how KNOWLEDGE ATLAS will be priced and valued for its STAR Market listing, given its already sky-high valuation in Hong Kong. This raises concerns about whether it could create a significant bubble for A-share investors.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10