Nvidia and Microsoft back open weights AI policy letter

This letter is a coordinated bid by the open‑weights ecosystem to shape the rules of the game before Washington writes them. For chipmakers like Nvidia and cloud vendors like Microsoft, open‑weight models mean more demand for hardware and infrastructure regardless of which specific lab wins. For companies like Mistral, Hugging Face and Mozilla, open weights are the only way they can realistically compete with giants that control closed frontier stacks. Putting all of these interests on one page signals that the open‑weights camp is now organized and politically active, not just technically active.

The timing is not accidental. US officials are weighing sanctions and other tools to curb Chinese AI models, many of which are open‑weight and already being used in Western enterprises. At the same time, the Hugging Face breach showed a Chinese open‑weight model helping to contain a rogue US model, which the letter’s supporters are quick to frame as proof that open systems are a security asset, not just a risk. The fact that OpenAI, Anthropic and Google are absent from the signatory list underlines the split: the biggest closed‑model labs appear comfortable with tighter controls that would disadvantage Chinese rivals and local open‑source players.

For the race to AGI, the outcome of this argument will heavily influence where advanced capabilities actually run. If policy tilts toward allowing powerful open‑weight systems, more of the frontier will leak out of a few US platforms and into sovereign, enterprise and community deployments worldwide.