Much of the US AI industry signs a letter defending open-weight models

On 24 July 2026, Nvidia published an open letter, Open Weights and American AI Leadership, urging the US government not to impose broad restrictions on open-weight AI models — those anyone can download, inspect, modify and run on their own infrastructure. Chief executive Jensen Huang promoted it in what was widely described as his first post on X, and the document circulated quickly among technology firms. It launched with roughly 25 corporate signatories, passed 50 within two days and reached around 200 within a week, spanning frontier labs, chipmakers, cloud providers, open-source foundations and application startups.

The letter opens by comparing open weights to the 1980s open-source software movement, and argues that American 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.” It makes four claims for openness: that it widens access to advanced models, sustains competition, protects customers from lock-in to a single vendor, and — against a common assumption — can strengthen security, because defenders need models as capable as those attackers use, and concentrating capability in a few closed systems creates “single points of failure.”

The timing was pointed. The letter followed reports that the Trump administration was weighing restrictions on Chinese open-weight models, after Moonshot AI’s Kimi K3 reached the frontier on several benchmarks and Anthropic accused Moonshot and other Chinese labs of extracting its models through distillation. The letter addresses that dispute directly, defending distillation as legitimate and warning against conflating it with theft:

Distillation … is a widely used technique for model improvement, evaluation, and validation. … unlawful efforts to extract value from closed models … should be addressed through targeted legal and commercial frameworks rather than sweeping restrictions.

— Open Weights and American AI Leadership

According to TechCrunch, the letter was organised by Hugging Face, Meta, Microsoft, Mistral, Nvidia and Replit; later signatories included OpenAI, Google, Amazon and SpaceX. The most conspicuous absentee was Anthropic — the company that had made the distillation allegations — which did not sign and published a separate statement disputing the letter’s cybersecurity argument; xAI also did not appear on the list.

Coverage noted that many signatories had commercial reasons to favour open models: Nvidia sells more chips as models proliferate, and tool and cloud firms benefit from cheaper alternatives to frontier APIs. A Fortune analysis questioned whether the appeals to competition and sovereignty were made in good faith. Either way, the letter marked an unusually broad show of industry alignment on a live policy question, and a public marker of the open-versus-closed divide — with Anthropic on the far side of it.