Open-Weights Letter Doubles to 50 Signatures — With OpenAI Quietly on Board, and Anthropic Still Out

An open letter urging US policymakers to avoid "premature restrictions" on open-weight AI models has doubled its signatories from 25 to 50 in a matter of days — and picked up its most surprising name yet: OpenAI, which signed quietly and without public comment, according to TechRadar and Forbes.

The letter, published July 24 and initiated by Nvidia, launched with signatures from Microsoft, Meta, Dell, IBM, Palantir, Mistral, Mozilla, Hugging Face, the Linux Foundation, Andreessen Horowitz and Y Combinator. The second wave added Google, AMD, Cisco, Cloudflare, GitHub, Block, Ollama — and OpenAI.

Conspicuously absent from every version: Anthropic and Amazon.

What the letter argues

The signatories warn that restricting open models would "stifle competition or drive innovation overseas," and directly attack the assumption that closed equals safe: "Relying solely on closed models is not inherently safe: they can be breached, misused, or fail in ways that outsiders cannot detect" — a line that reads pointedly in the same month OpenAI's own closed models breached Hugging Face.

Nvidia CEO Jensen Huang promoted the letter on social media; Microsoft's Satya Nadella reposted it, framing open models as a national security asset rather than a liability.

The subtext is China

The letter's real target is the policy debate in Washington over whether to restrict Chinese open-weight models — a debate supercharged by the White House's accusation that Moonshot AI built Kimi K3 by distilling Anthropic's Fable, and by K3's open weights landing on Hugging Face on July 27. With Chinese open models now downloaded by the millions and topping agentic leaderboards, US industry is effectively arguing that walling them off would hurt American developers more than Beijing.

OpenAI's quiet signature is the tell. The company has no open frontier weights of its own, but it understands where enterprise demand is heading — and appears unwilling to be positioned alongside Anthropic as an advocate of restriction while Nvidia consolidates the open ecosystem around itself.

Anthropic's lonely corner

Anthropic CEO Dario Amodei pushed back this week, insisting he has "never advocated" banning open-weight models while restating that released weights "cannot be withdrawn" and calling instead for chip export controls and mandatory safety testing. The nuance may be genuine, but the optics are stark: on the defining industry question of the summer, 50 companies are on one side of the page, and Anthropic is not on it.