Loading market data...

Nvidia, Meta, Microsoft Among 25 Firms Defending Open-Weight AI Models in Letter to Washington

Nvidia, Meta, Microsoft Among 25 Firms Defending Open-Weight AI Models in Letter to Washington

Twenty-five companies, including Nvidia, Meta, and Microsoft, have signed a letter to the U.S. government defending open-weight AI models. The letter was sent days after a Chinese AI system helped Hugging Face survive a security breach triggered by OpenAI's own infrastructure.

Who Signed the Letter

The coalition spans major tech players and smaller developers. Signatories include chipmaker Nvidia, social media giant Meta, and software heavyweight Microsoft. All three have invested heavily in open-weight models—AI systems where the trained parameters are publicly released, allowing others to run, modify, or build on them.

Other signatories weren't named in the available facts, but the group's size—25 companies—signals broad industry support for keeping such models open. The letter is addressed directly to Washington, indicating a coordinated push to shape upcoming policy.

Why Open-Weight Models Are Under Scrutiny

Open-weight models have become a flashpoint in the U.S. debate over AI safety. Critics argue that releasing model weights makes it too easy for bad actors to misuse powerful AI—running it on their own hardware without any guardrails. Proponents, including the signatory companies, counter that openness drives innovation, transparency, and security research.

The letter arrives amid a flurry of regulatory proposals in Congress and the White House. Some lawmakers want to require companies to vet users before releasing model weights, while others have called for outright restrictions on the most capable systems.

Timing and the Hugging Face Incident

The letter's timing is notable. Just days earlier, a Chinese AI model helped Hugging Face, a major platform for sharing AI models, survive a hack that was triggered by OpenAI's own systems. The incident highlighted both the vulnerabilities in closed AI ecosystems and the value of open models in mitigating damage.

While details of the hack remain sparse, the Chinese AI's role in aiding Hugging Face underscores a paradox: the very openness that some want to restrict can also serve as a defense. The signatory companies are likely pointing to this event as evidence that open-weight models can bolster security, not just threaten it.

The letter doesn't mention the Hugging Face incident directly, but the timing suggests it's part of the backdrop. The companies are betting that a clear message—keep open-weight models legal and accessible—will resonate in Washington.

What comes next is uncertain. The U.S. government has not yet responded to the letter. The companies are waiting to see whether lawmakers will embrace their arguments or move toward tighter controls.