Chinese AI startup Zhipu valued at 1280 times annual revenue: bubble or new reality?
The market for public AI companies continues to surprise analysts. Chinese developer of the GLM-5.2 model, Zhipu (Z.ai), which listed on the Hong Kong Stock Exchange on January 8, 2026, is showing anomalous multiples. According to my calculations, based on data from partner Delphi Ventures, Zhipu's market capitalization has exceeded $118 billion, with annual revenue for 2025 of only about $107 million. Its net loss for the same period was 4.7 billion yuan.
The P/S (price/sales) multiple reaches a shocking 1280 times annual revenue. For comparison: OpenAI, with revenue of $25 billion and a private valuation of $852 billion, trades at approximately 34 times revenue, while Anthropic, with $47 billion in revenue and a valuation of $965 billion, trades at 21 times revenue. The gap is colossal: Chinese "labs" are valued tens of times higher than their American counterparts relative to actual financial performance.
Why is the market overheated?
For Zhipu to trade at even 50 times annual revenue, it would need to grow sales to $2.7 billion per year — 26 times its current level. For a multiple of 20, it would require $6.9 billion, or growth of 65 times. A similar picture is seen with MiniMax: a market capitalization of about $23 billion with revenue of $79 million gives a multiple of 290.
The only exception is Alibaba with its Qwen model. With a market capitalization of $245 billion and revenue of $151 billion for the 2026 fiscal year, the company trades at just 1.6 times revenue. But Alibaba is not a pure AI company; it is an e-commerce giant, so the comparison is conditional.
Where does the revenue go?
A key problem for Chinese developers is that a significant portion of revenue goes to third-party inference providers (services that run other companies' AI models on their servers and sell access via API), such as OpenRouter, Venice, and BaseTen. Users want to work with these models but are unwilling to send data directly to China due to privacy concerns, so they turn to intermediaries.
To turn the situation around, Chinese companies will have to prove they do not store user data and offer lower prices than competitors. Doing so is extremely difficult for cultural and social reasons.
Alternative scenario: Chinese developers could take equity stakes in American inference providers and sign agreements for early access to top models in exchange for a share of revenue. Then money would flow to model developers — through a percentage from providers, but amid overall market growth.
My professional opinion: The current valuations of Zhipu and MiniMax are a classic sign of an overheated market, fueled by speculative interest in Chinese AI. Until these companies solve the problems of monetization and data trust, their stocks remain an extremely risky asset, reminiscent of the dot-com bubble of the early 2000s.