Nvidia CEO urges open-weight AI models as U.S. debates restrictions

Nvidia chief Jensen Huang used his first post on the social platform X to argue that open-weight artificial-intelligence models should remain available as a cornerstone of U.S. technological leadership. In a letter signed by roughly 25 companies, including Microsoft, Meta, Hugging Face, Andreessen Horowitz, Perplexity and IBM, Huang said open models “strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.” He added that the world needs both “frontier closed models and frontier open models.”

The signatories contend that open models broaden economic access for startups and universities, prevent a handful of firms from monopolising the technology and improve security by allowing a larger community of researchers to audit code and fix vulnerabilities. The letter also defends “distillation,” the practice of training a model using another model’s outputs, as a legitimate research method rather than theft.

Huang drew a parallel with the 1980s open-source software movement, noting that open code became the foundation of the modern internet despite early scepticism. He argued that open AI models should be allowed to flourish rather than face “premature restrictions.”

The post appears amid a debate in Washington over whether to limit Chinese open-weight models such as Moonshot AI’s Kimi K3, released on July 16. White House adviser Michael Kratsios has accused Moonshot of using distillation to copy a U.S. model, intensifying the controversy.

A day before Huang’s message, Chinese President Xi Jinping delivered a keynote at the World AI Conference in Shanghai, calling for “a symphony of global cooperation” and urging opposition to “overstretching the national security concept in the field of AI.” Analysts interpreted the remarks as a challenge to U.S. export controls. The same week, 29 nations signed an agreement in Shanghai to create the World Artificial Intelligence Cooperation Organization, a body positioned as an alternative to U.S.-led AI governance.

Huang’s stance places him alongside Chinese officials on the theme of openness, contrasting with the more cautious positions of OpenAI and Anthropic, which are absent from the open-weights letter and have warned about the risks of powerful Chinese models.

“Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty,” Huang wrote.

“The world needs both frontier closed models and frontier open models.”

The article notes that Fortune journalists employed generative AI as a research tool, with an editor verifying the information before publication.