Hugging Face CEO Clément Delangue said China is winning the artificial intelligence race with open-weight models and could catch up to U.S. makers this year [2].

This shift in momentum suggests a potential pivot in global AI dominance. If China continues to lead in open-source development, the U.S. may lose its strategic advantage in how AI is built and deployed globally.

Delangue detailed these concerns during a CNBC Television interview on Monday. He said that the strong concentration of top open-source AI models coming from China is a point of concern [1]. This trend could allow China to overtake the United States in overall AI capability if the U.S. does not keep pace with open-model innovation [4].

Open-weight models, often referred to as open-source, allow developers to access the underlying parameters of a model. This transparency fosters rapid iteration, and widespread adoption. Delangue said that open-source AI is vital for U.S. innovation and that the U.S. risks losing the race if it falls behind in this specific sector [3].

While the U.S. has historically led in proprietary, closed-source models, the rise of Chinese open-weight alternatives is closing the gap. Delangue said that China could match U.S. capabilities as early as 2026 [2].

Hugging Face, the platform where these models are often hosted and shared, currently holds a valuation of $4.5 billion [5]. The company serves as a central hub for the global AI community, providing a unique vantage point on which nations are contributing the most impactful code.

Delangue has previously expressed worry regarding the concentration of these models in China. He said the current trajectory indicates that the U.S. must accelerate its efforts to maintain a competitive edge in the open-source ecosystem [3].

“China is winning the artificial intelligence race with open‑weight models”

The transition of AI leadership toward open-weight models signifies a move away from the 'black box' approach favored by major U.S. corporations. If China dominates the open-source layer, it can set the global standards for AI architecture and accessibility, potentially bypassing U.S. sanctions or proprietary lockdowns on technology.