Nvidia, Microsoft, Meta warn against overregulating open-weight models

Nvidia, Microsoft, Meta, palantir and more than 20 other companies published a letter on Friday urging policymakers to avoid “premature restrictions” on vulnerability-heavy AI models that would “stifle competition or encourage innovation abroad.”

Vulnerability-heavy AI models are available for users to download, modify and run their own infrastructure, and have been the subject of fierce debate in the tech industry in recent weeks.

China’s deficit-heavy models gain ground against leading offerings from American companies OpenAI And anthropicwhich primarily develops proprietary, closed models. Officials and executives discuss whether to restrict access to Chinese models in the US

Moonshot AI, a Chinese startup, raised concerns earlier this month after launching a model called the Kimi K3 that beats cutting-edge American offerings on some industry benchmarks. The Trump administration will investigate whether Chinese companies are stealing American intellectual property, U.S. Treasury Secretary Scott Bessent told CNBC on Tuesday, adding that the government “has the ability to sanction them for that theft.”

But in Friday’s letter, the group of US technology companies warned against hasty actions. They wrote that vulnerability-heavy models foster competition and ensure that the benefits of technology are “shared widely rather than concentrated in a few hands.”

“Relying solely on closed models is inherently unsafe: they can be breached, misused, or fail in ways that outsiders cannot detect,” the letter said. “And concentrating advanced AI capabilities behind a small number of closed models increases risk.”

Nvidia CEO Jensen Huang and Microsoft CEO Satya Nadella shared the letter on their personal social media accounts.

OpenAI and Anthropic did not sign the letter. Both companies, each valued at about $1 trillion, are preparing for potentially large IPOs that could happen as soon as this year. Anthropic secretly filed its prospectus with the Securities and Exchange Commission in June, and OpenAI followed suit days later.

OpenAI president Greg Brockman said Thursday that the company believes in broad access and has not been involved in any discussions with the Trump administration about potentially banning China’s open weight models in the United States.

“I think fundamentally AI and the use of AI is a very important thing in terms of democratization,” Brockman told reporters during a briefing in New York City. he said. “And for me, on a deep level, I think having more models, more usage is a good thing.”

Moonshot AI is developing the Kimi K3 model, White House adviser Michael Kratsios said Wednesday. distillation Anthropic’s technology. Distillation is a term for an AI training method in which a smaller, less capable model is created using output from an existing, more powerful model.

Kratsios wrote in a post on

In the letter published Friday, US tech companies said concerns about illegal distillation should be addressed through “targeted legal and commercial frameworks” rather than “broad restrictions on techniques that play a key role in AI innovation.”

“Our AI leadership will be judged not by a single leading AI model, but by whether the United States has created a strong, open ecosystem that spans every sector,” the letter said. “This is crucial to creating opportunities for innovation and prosperity across the country.”

WRISTWATCH: Artificial intelligence is forcing the cyber industry to reconsider the way it operates, says TrustedSec’s David Kennedy