Chinese AI startup Zhipu is valued higher than OpenAI and Anthropic: a multiplier of 1,280 times annual revenue
The market is going crazy over Chinese open-source AI models. A prime example is the company Zhipu (Z.ai), developer of the GLM-5.2 model, which, after listing on the Hong Kong Stock Exchange on January 8, 2026, and the subsequent release of its flagship model, demonstrated phenomenal but highly controversial market capitalization growth.
As of June 22, Zhipu's market value exceeded $118 billion. Meanwhile, the company's annual revenue for 2025 was only about $107 million, with a net loss reaching 4.7 billion yuan. A simple calculation gives us a multiple of approximately 1280 times annual revenue. This is an absolute record, even for an overheated sector.
Multiples Detached from Reality
For comparison, American "labs" look much more modest. OpenAI, with an annual revenue of $25 billion and a private valuation of $852 billion, trades at 34 times revenue. Anthropic, with $47 billion in revenue and a valuation of $965 billion, trades at 21 times revenue. Even these figures are considered inflated, but Zhipu has surpassed them by orders of magnitude.
The situation with Zhipu is not an isolated case. Another Chinese developer, MiniMax, with a market capitalization of about $23 billion and revenue of $79 million, trades at a multiple of 290. Even Alibaba, with its Qwen model, which is not a pure AI company, with a market cap of $245 billion and revenue of $151 billion, trades at only 1.6 times revenue.
Why such a gap? Analysts see the root of the problem in the fact that a significant portion of Chinese developers' revenue goes to third-party inference providers—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. Therefore, they turn to intermediaries, who "eat up" the margin.
How to Turn the Situation Around?
To justify current valuations, Zhipu needs to increase revenue to $2.7 billion per year—26 times its current level. For a multiple of 20, it would require $6.9 billion, or a 65-fold increase. This looks extremely ambitious, if not unrealistic.
I see two scenarios. The first is that Chinese companies will have to prove they do not store user data and offer prices lower than competitors. The second, more likely scenario, is strategic alliances: taking equity stakes in American inference providers in exchange for early access to top models and a share of revenue. This would allow money to flow to developers, albeit through intermediaries, but with a growing market volume.
My conclusion: The current valuations of Zhipu and MiniMax are a classic sign of an overheated market and expectations that are not yet supported by fundamental indicators. Investors should exercise extreme caution until these companies prove their ability to monetize traffic, rather than handing it over to third-party platforms.