A coalition of 25 technology companies led by Nvidia, Microsoft, and Meta has signed an open letter urging U.S. policymakers to avoid "premature restrictions" on open-weight artificial intelligence models, warning that broad limits would "stifle competition or drive innovation overseas."

The joint letter, released on July 24, 2026 and first reported by CNBC and Fox Business, marks one of the largest coordinated industry pushes to defend open-weight AI just as the Trump administration weighs whether to crack down on the technology. The signatories — which also include Palantir and more than 20 other companies — argue that open models are critical to maintaining America's lead in the global AI race. For ongoing coverage of the policy battles reshaping the industry, follow our AI regulation reporting.

'A strong, open ecosystem'

The letter's central argument is that U.S. competitiveness depends on openness, not lockdown. "Our AI leadership will be judged not by one frontier AI model, but by whether the United States builds a strong, open ecosystem that diffuses into every sector," the signatories wrote, according to Fox Business.

The companies contend that open-weight models — whose parameters are publicly downloadable — accelerate innovation, strengthen cybersecurity, and help ensure the country remains competitive as nations race to develop increasingly powerful AI systems. Notably, the coalition framed its defense in economic and strategic terms rather than purely technical ones, casting restrictions as a threat to both startup competition and national advantage.

The Moonshot and distillation flashpoint

The industry appeal did not emerge in a vacuum. It came as the administration debates its approach to open-AI models following alarms about Chinese developer Moonshot AI and its open-weight Kimi K3 model, which has rapidly gained users worldwide.

Fox Business reported that White House Office of Science and Technology Policy Director Michael Kratsios accused Moonshot of developing Kimi K3 by distilling Anthropic's Fable 5 model — a technique that trains a smaller AI model using the outputs of a more advanced one. Distillation has become a flashpoint in the frontier-AI race, raising questions about intellectual property and whether open-weight releases from abroad are built on stolen capabilities.

Treasury Secretary Scott Bessent went further, suggesting that the United States could sanction companies found to have improperly used distillation to appropriate proprietary work.

Acknowledging the concerns

In a notable concession, the coalition did not dismiss the IP concerns outright. While defending open development, the letter acknowledged that companies found to have illegally extracted value from proprietary models should face consequences.

"Unlawful efforts to extract value from closed models raise legitimate concerns," the letter states, according to Fox Business. "Those concerns should be addressed through targeted legal and commercial frameworks rather than sweeping restrictions on techniques that play an important role in AI innovation."

That framing is deliberate: the signatories are trying to separate legitimate enforcement against IP theft from blanket limits on the distillation techniques and open-weight releases that underpin much of the startup ecosystem. Their argument is that punishing bad actors surgically is preferable to restricting the entire category.

The absent giants

The coalition's makeup is itself a story. The companies that did not sign the letter are as telling as those that did. CNBC noted that OpenAI and Anthropic — the two leading closed-model labs, both reportedly preparing for potentially massive initial public offerings — declined to add their names.

For OpenAI and Anthropic, open-weight models represent a competitive threat: freely downloadable alternatives can undercut pricing and erode the moat around proprietary frontier models. Their absence underscores how the open-versus-closed debate now cuts across traditional industry lines, pitting infrastructure and platform companies (Nvidia, Microsoft, Meta) against the labs selling closed access.

A debate with no easy answers

The letter arrives at a moment of genuine tension. Chinese open-weight models have surged in capability and adoption, narrowing gaps that U.S. firms once assumed were durable. At the same time, U.S. officials worry that freely released weights could be weaponized or built atop stolen IP. The signatories' pitch is that the answer is not to wall off the technology, but to invest in a domestic open ecosystem while enforcing against theft.

Whether Washington agrees will shape the next phase of the AI race. For now, the industry's biggest hardware, cloud, and social platforms have made their position clear: keep the weights open, target the bad actors, and let the ecosystem compete.

Stay Ahead of AI

The fight over open-weight AI is one of the defining policy battles of the decade. AI Buzz Wire covers every letter, hearing, and proposal as they land.

Explore the latest AI developments →