Z.AI Loses HK$1 Trillion in Market Value Over Five Months as Tang Jie's Fortune Shrinks by HK$70 Billion

Deep News
Yesterday

News of restricted access to computing power continues to weigh on Hong Kong-listed AI large-model company Z.AI (HKG: 2513), sending its shares even lower. On November 17, Z.AI shares opened low and drifted down, closing 5.26% lower at HK$335 per share, dragging the company's total market value down to HK$163.5 billion. During intraday trading, the stock touched HK$328 per share, marking a new low since March of this year. Behind the slump in Z.AI shares, the threshold for domestic AI companies to secure computing power has once again risen.

On November 16, the U.S. Department of Commerce's Bureau of Industry and Security issued an interim final rule (IFR) aimed at preventing data centers in third-party countries from circumventing existing access restrictions through remote computing arrangements. The market worries that this rule could affect domestic AI companies' ability to rent overseas computing power, which would in turn impact their large-model training and business operations. This decline also extends Z.AI's earlier downward trend.

As a well-known domestic AI large-model company, Z.AI listed in Hong Kong in January of this year. After listing, Z.AI shares climbed all the way, peaking at HK$2,980 per share in June, a gain of more than 25 times the issue price, with total market value reaching HK$1.33 trillion. Since then, Z.AI shares have begun to pull back, and compared with the June high, the latest share price has fallen nearly 90%, with total market value down nearly HK$1.1 trillion.

As Z.AI shares decline, the fortunes of the company's two founders, Liu Debing and Tang Jie, are also shrinking. Currently, Liu Debing holds about 1.54% of Z.AI through Beijing Lianpai and direct holdings, while Tang Jie directly holds 6.02% of Z.AI. At the June share price peak, the value of Liu Debing's and Tang Jie's stakes stood at HK$20 billion and HK$80 billion, respectively. Now, Liu Debing's stake has fallen to HK$2.3 billion, and Tang Jie's stake has dropped to HK$9 billion.

In recent months, shares of another Hong Kong-listed large-model company, MiniMax (0100.HK), have also been declining. On November 17, MiniMax shares fell 7.13% to HK$116 per share, with a total market value of about HK$40.8 billion. After hitting a high of HK$1,330 per share in March of this year, the company's shares have continued to slide, with the latest price down more than 90% from the peak. In October of this year, MiniMax shares broke below their issue price, and the stock is now down 30% from the issue price.

The AI sector is undergoing a broad revaluation

Behind the months-long plunge in Z.AI and MiniMax, the AI sector is facing a revaluation. In the first half of this year, the AI sector was at the center of attention, and as the only two AI large-model companies listed in Hong Kong, Z.AI and MiniMax were heavily sought after by the market, driving their share prices sharply higher. But this enthusiasm has significantly diverged from the sector's fundamentals. At MiniMax's March share price peak, the company's price-to-sales ratio (TTM) exceeded 600 times, and at Z.AI's June peak, its price-to-sales ratio (TTM) even exceeded 1,000 times.

Since then, enthusiasm for the sector has cooled markedly, and share prices of companies across the AI industry chain have begun to pull back. The commercialization progress of Z.AI and MiniMax has started to become the market's focus. The two companies did not disappoint, delivering financial reports with strong growth during the interim reporting season. In the first half of this year, MiniMax's business structure underwent a major shift, with B-end business replacing C-end products as its main source of revenue. In the same period, the company's total revenue rose 283.1% year-on-year to US$117 million (about RMB 778 million), while its adjusted net loss widened 52.7% year-on-year to US$293 million (about RMB 1.96 billion). Z.AI's performance also achieved substantial growth, with first-half revenue of RMB 954 million, up 399.74% year-on-year; its adjusted net loss was RMB 1.964 billion, widening 12.11% year-on-year.

However, even as business restructuring drove strong performance growth, the gross margins of both companies declined, which has become one of the factors behind market concerns about their competitiveness. More critically, as market competition intensifies, whether the leading domestic large-model companies can continue to maintain rapid growth is also being questioned.

A recent research report from HSBC showed that, in the face of factors such as the price war launched by international giants like Anthropic and OpenAI and high R&D expenses, Z.AI's next-generation large-model development cycle has lengthened, and short-term share price volatility will continue. As a result, HSBC lowered its medium-term profit expectations for Z.AI, adjusting its annualized recurring revenue (ARR) and total revenue forecasts for fiscal 2027 to 2028 by 7% to 27%, while sharply cutting its adjusted operating margin forecast by 15 to 19 percentage points. It also pushed back the company's expected timeline for achieving adjusted break-even by two years, from 2027 to 2029. HSBC also sharply lowered its target price for Z.AI, slashing it 64% from HK$1,375 to HK$500.

In addition to the above factors, the capabilities of flagship large models are also a key factor affecting the market's view of AI companies. Since the second half of this year, the pace of flagship large-model launches by domestic AI companies has slowed. In April, June, and August of this year, Z.AI successively launched three generations of flagship large models, GLM-5.1/5.2/5.3, while in March and June of this year, MiniMax also launched two generations of flagship large models, M2.7 and M3. But to date, the latest generation of flagship large models from both companies is still under development. On the Artificial Analysis leaderboard, Z.AI's flagship model GLM-5.3 (MAX) has fallen to 20th place, while MiniMax's flagship model has dropped off the list entirely.

In a July research report, JPMorgan said that open weights expand the scope of model distribution, but monetization capability still depends heavily on model quality, iteration speed, and endpoint reliability. For models with leading capabilities, open weights can expand usage scale through external GPU computing power such as cloud service providers, inference platforms, and private deployments. But for models with insufficient differentiation, broader access instead accelerates price comparison and traffic substitution.

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