Hugging Face CEO: Chinese Open-Source AI Models Already Leading, May Dominate Frontier AI by Year-End
Miles Bennett
Hugging Face CEO Clément Delangue says China already dominates open-source AI and could lead frontier models by year-end or next year. This means the US-China AI race is shifting from compute power to a contest of ecosystem models.
What exactly did he say?
Hugging Face CEO Clément Delangue told CNBC on Monday that China already dominates open-source AI models.
He went further: at the current pace, China could start leading frontier models by year-end or next year.
This means → This is not a throwaway remark. The head of the world's largest open-source AI platform is formally acknowledging China has pulled ahead on the open-source track.
Why is China pulling ahead?
Delangue attributes the edge to ecosystem philosophy: Chinese AI labs tend to open-weight their models — releasing parameters publicly so anyone can use and improve them — and actively share research.
US model developers, by contrast, are "building in silos, keeping things closed," a fundamentally different path.
In plain terms = China is building one wheel together; the US is building separate wheels behind closed doors. In Delangue's view, the collective approach is currently faster.
How much weight does this call carry?
Hugging Face is the world's largest open-source AI model hosting platform — the central infrastructure for gauging open-source ecosystem activity.
Delangue's statement is not an academic forecast. It is a front-line observation drawn from model-release trends on the platform — and it aligns with the recent surge of open-weight model releases from Chinese AI labs.
This reflects a deeper signal: the real battlefield in US-China AI competition may not be who has the stronger chips, but whose ecosystem attracts more developers and iterates faster.
What to watch next?
Delangue set a clear timeline: year-end to next year. If Chinese open-source models match or beat US closed-source models on frontier benchmarks, his call is validated.
This means → The pace of model releases and benchmark scores from AI labs in the second half will be the most direct signal of how the US-China AI race is shifting.
For investors, the question is not just any single company's model performance — it is whether the open-source or closed-source approach can sustain a durable lead.
Content is for reference only, not financial advice.