Nvidia, Microsoft, Meta warn against 'premature restrictions' of open-weight models - CNBC

What the report says

CNBC reported that Nvidia, Microsoft, Meta, Palantir and more than 20 other technology companies issued a letter on Friday asking policymakers not to impose early, broad limits on open-weight artificial intelligence models. The companies argued that such restrictions could weaken competition, push innovation abroad and concentrate AI capabilities among a small set of closed-model providers.

Open-weight models can be downloaded, adjusted and operated on users’ own systems, a feature that has made them central to a growing policy fight. CNBC said the debate has intensified as Chinese open-weight models challenge leading U.S. proprietary systems, including those from OpenAI and Anthropic. The article cited Moonshot AI’s Kimi K3 release as one development that raised concern, and noted U.S. Treasury Secretary Scott Bessent said the Trump administration would examine whether Chinese firms had taken American intellectual property.

The letter also argued that closed models are not automatically safer, because outsiders may be unable to identify failures or misuse. Nvidia CEO Jensen Huang and Microsoft CEO Satya Nadella shared the letter on social media, while Elon Musk expressed support, though SpaceX was not listed as a signer. OpenAI and Anthropic did not sign; CNBC reported both are preparing for potentially large IPOs. OpenAI President Greg Brockman separately said he supports broad AI access, and CEO Sam Altman wrote that he wants the U.S. to succeed with both open-weight and proprietary approaches.

CNBC also described concerns over AI “distillation,” a training method involving outputs from stronger models. White House adviser Michael Kratsios accused Moonshot of using Anthropic technology, while the company letter urged targeted legal and commercial responses rather than sweeping limits on techniques used in AI development.