China's AI race just printed its first public earnings report, and the numbers tell you everything about who's winning the global agent economy.

The Summary

  • Zhipu AI posted $105M revenue in 2024, up 132% year-over-year, with net losses widening on R&D spend
  • First earnings disclosure since January Hong Kong IPO, giving rare transparency into Chinese AI economics
  • Revenue growth at 132% while burning cash signals the same pattern as Western AI labs: land grab first, profitability never

The Signal

This is the first time we've seen actual financials from a Chinese AI foundation model company post-IPO, and it looks exactly like the Western playbook, just with different logos. Zhipu pulled in 724 million yuan ($105M) in 2024, more than doubling revenue while losses ballooned on R&D. That's the same growth-at-all-costs math OpenAI, Anthropic, and every other model lab is running.

What matters here isn't the absolute numbers. It's that Zhipu went public in Hong Kong in January and is now competing for the same global enterprise dollars as Western labs, but with access to Chinese compute clusters, different regulatory constraints, and a government that sees AI infrastructure as strategic. The 132% growth rate suggests they're winning deals, likely in China first but increasingly in markets where Western AI tools face either regulatory friction or customer skepticism about data sovereignty.

The widening losses on R&D tell you Zhipu is in the same arms race. Training larger models, building better inference infrastructure, hiring talent. The difference is they're doing it with Hong Kong capital markets access and Beijing's implicit backing. That's a different risk profile than venture-backed Western labs that need to show a path to profitability before the music stops.

This earnings report also sets a baseline. Now we can track whether Chinese AI labs can achieve commercial scale faster than their Western counterparts, or if they hit the same unit economics wall. If Zhipu's revenue growth continues at triple digits while Western labs plateau, that's your signal that the agent economy is fragmenting along geopolitical lines, with different winners in different markets.

The Implication

Watch Zhipu's next quarter. If revenue growth stays above 100% while Western labs slow, you're seeing a bifurcated AI market emerge. For builders, this means planning for a world where your AI stack might need regional variants. For investors, Chinese AI labs with public float just became trackable plays on the agent economy outside the U.S. regulatory perimeter.

Source: The Information