Hugging Face CEO: US Could Fall Behind China if AI Remains Closed-Source

The debate around the future of artificial intelligence in the United States has taken a sharp turn after strong remarks from Hugging Face CEO, Clement Delangue. According to him, the U.S.

risks being ‘cooked’ in the AI race if it continues to cling to a closed-source model. While the country has undeniably led the global charge in AI innovation, the conversation has shifted to whether its reliance on proprietary systems could become its Achilles’ heel.

At the center of this debate is the contrast between the U.S. and China. While American firms like OpenAI and Anthropic have built powerful models locked behind APIs and subscriptions, China has poured resources into building an open-source AI ecosystem. This approach has allowed developers, researchers, and even smaller companies to experiment with and build upon shared technologies. As Delangue argues, open-source AI is not just about access to code – it’s about fostering collaboration, transparency, and a shared advancement of knowledge that can accelerate innovation across industries.

The recent unveiling of Meituan’s 560-billion-parameter Mixture of Experts (MoE) model highlights China’s commitment to this open approach. The model is designed to streamline workflows such as customer service automation, but perhaps more importantly, it symbolizes a growing movement within China to democratize AI. By releasing these models openly, Chinese firms are creating ecosystems where progress multiplies collectively, rather than being siloed in corporate vaults.

In contrast, U.S. companies are incentivized to keep their work closed. With billions invested in training and infrastructure, their business models depend on exclusivity. The likes of OpenAI’s GPT series, Google’s Gemini, and Anthropic’s Claude are monetized through paid access, ensuring companies can recoup costs and maintain a competitive edge. However, Delangue warns that this mindset carries a long-term risk: by prioritizing revenue over collaboration, America might fall behind nations that see AI as a shared scientific resource rather than a walled garden.

Interestingly, even President Trump’s earlier AI action plan recognized the importance of fostering openness, suggesting a U.S. policy desire for broader availability. Yet, industry giants remain hesitant. While experiments like OpenAI’s gpt-oss project nod toward open collaboration, the momentum pales in comparison to China’s rapid expansion in the field.

The implications are profound. If open-source AI becomes the global standard, the U.S. could find itself on the defensive, forced to adapt to models shaped by foreign ecosystems. The tension between profitability and progress has never been more stark in the technology sector. Whether the U.S. recalibrates toward greater openness – or doubles down on exclusivity – may define the balance of global AI power in the coming decade.