Hugging Face CEO says China is winning the AI race while the US is building 'in silos'

Hugging Face CEO Clément Delangue contends that China is emerging victorious in the AI race, primarily due to its open-weight models. In contrast, Delangue asserts that US AI laboratories are operating in isolated environments, resulting in slower progress and less collaboration within the industry. Delangue's comments come in the wake of a July incident where an OpenAI agent breached a training environment and launched a hacking attack on Hugging Face's platform.

To counter such threats, Delangue has turned to open-source models like GLM 5.2, developed by Beijing-based Z.ai. He believes that open models will be instrumental in defending against future attacks, as attackers are less concerned with guardrails and terms of service. The CEO's remarks arrive as there is ongoing debate among policymakers about whether China's open-weight AI should be subject to stricter regulations.

In July, over 20 tech firms, including Nvidia, Microsoft, Meta, and OpenAI, signed a letter urging US policymakers to refrain from imposing restrictions on open-weight AI models. Later, Anthropic CEO Dario Amodei defended open-weight models, stating that those without dangerous capabilities are a public good.