AI Policy

Nvidia, Meta and Microsoft Tell Washington: Don't Restrict Open-Weight AI

Newaiera Desk · 2026-07-28 · 4 min read

25 companies — including Nvidia, Meta, Microsoft, IBM, Hugging Face and Mistral — signed a joint letter urging U.S. policymakers to avoid 'premature restri

A split is forming in how the AI industry wants Washington to treat open models — and this week, one side put its name on paper.

On July 24, 25 companies and organizations — including Nvidia, Meta, Microsoft, IBM, Dell Technologies, CrowdStrike, Palantir, ServiceNow, Hugging Face, Perplexity, Mistral, Andreessen Horowitz, Y Combinator, the Linux Foundation and Mozilla — signed a joint open letter urging U.S. policymakers to protect open-weight AI models from new restrictions.

What's actually being asked for

Open-weight models are systems whose trained parameters are published for anyone to download, inspect, modify and run on their own hardware — as opposed to closed models accessible only through an API. The letter makes three concrete asks: expand compute access for startups and researchers, fund shared training datasets and evaluation frameworks, and avoid what it calls "premature restrictions" on open models, in the context of ongoing debate in Washington over whether to restrict or ban certain foreign open-weight models, particularly from China.

The letter's central line frames the stakes directly: 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.

Who signed, and who didn't

The signatory list is a genuinely unusual coalition — chipmakers, cloud platforms, cybersecurity firms, open-source foundations and venture capital all in the same letter. Nvidia CEO Jensen Huang and Microsoft CEO Satya Nadella both personally shared the letter on their social accounts, putting individual executive weight behind it rather than leaving it as a press-office statement.

What's more notable is who isn't on the list: OpenAI, Anthropic, and Google did not sign. All three are primarily closed-model labs — their flagship products are proprietary systems accessed via API, not open-weight releases. Their absence isn't necessarily opposition to open models existing at all, but it does mean the companies with the most to lose competitively from a stronger open-weight ecosystem chose not to publicly endorse protecting it.

Why the split makes commercial sense

This isn't really an ideological argument, whatever the letter's framing — it maps closely onto business model. Nvidia sells chips regardless of which lab's model runs on them, so a thriving open-weight ecosystem with many players buying compute is unambiguously good for Nvidia. Meta has bet its own AI strategy on open-weight releases (Llama) as a distribution and ecosystem play. Hugging Face's entire business is hosting and distributing open models. Mistral competes with the closed labs specifically on being more open.

OpenAI, Anthropic and Google, by contrast, compete on the strength of proprietary frontier models that customers pay to access — a stronger open-weight ecosystem, especially one with government support, is competitive pressure on that model, not a boost to it. Their non-participation is best read as neutrality-by-omission on a policy fight where they have less to gain than the signatories do.

The China question underneath

The letter's timing tracks a live Washington debate over whether to restrict access to certain foreign open-weight models. The signatories' ask — expand compute access, fund shared evaluation infrastructure, avoid premature restriction — reads as an argument that the U.S. competes with open Chinese models by out-building the open ecosystem domestically, not by banning the category of open-weight AI itself. Whether that argument holds depends largely on how national-security-focused the eventual policy response ends up being, which is exactly the fight this letter is trying to get ahead of.