Chinese open-weight models took 41% of Hugging Face downloads, and the top six on OpenRouter
Clément Delangue says China is clearly dominating on open models and would not be surprised to see it dominate at the frontier by the end of this year or next. His diagnosis of the cause is that US labs are building in silos.
The download share is the checkable part. Chinese open-weight models passed US models on Hugging Face in the spring at 41 percent, and the six most popular models on OpenRouter are all open releases from Chinese firms — Tencent, Xiaomi, DeepSeek, MiniMax and Z.ai among them.
Downloads measure adoption rather than capability, and the distinction matters. What they establish is which models developers actually reach for, which is a better predictor of ecosystem gravity than a leaderboard position. Tooling, fine-tunes and deployment know-how accumulate around whatever people run.
Delangue's causal claim is about structure rather than talent: open collaboration and sharing on one side, siloed building on the other. He is an interested party — an open-model platform benefits from open models winning — and the download figures are his own platform's. Both things can be true.
CNBC — Hugging Face CEO says China is winning the AI race and dominating on open models → · TechCrunch — The real AI race may no longer be at the frontier →