Nvidia, Microsoft, Meta and dozens of tech firms urge Washington not to restrict open-weight AI
Fifty technology companies, venture-capital firms, and nonprofits signed an open letter dated July 24, 2026, calling on Washington to avoid restricting open-weight AI models. Nvidia CEO Jensen Huang shared the letter on X, his first post on the platform, and within a day the list of signatories doubled as OpenAI, Google, AMD, Cisco, and others added their names. Anthropic is the only major U.S. frontier lab that did not sign.
What is the letter and who signed it?
The letter, titled Open Weights and American AI Leadership, was published as a PDF on Nvidia’s site. It argued that open-weight models, meaning AI models that anyone can download, inspect, modify, and run on their own infrastructure, are central to U.S. competitiveness in artificial intelligence.
Huang’s first post on X carried 25 initial signatories. By the next day, the count had grown to roughly 50, with OpenAI, Google, AMD, and Cisco among the additions. Anthropic, which has publicly warned about the risks of open-weight releases, was not on the list.
What does the letter argue?
The signatories made four core claims.
- Broad access to the AI economy. Open weights let startups, businesses, universities, and public institutions build on advanced models without training one from scratch or paying frontier-model prices for every task, allowing them to match the right model to the right job at the right cost.
- Stronger competition. Open weights strengthen competition across model developers, cloud providers, chips, applications, and services, which the letter says keeps the gains of AI broadly shared and drives down costs.
- Customer control. Open weights give organizations greater control over their own data, let them adapt models to their needs, and help avoid vendor lock-in.
- Safer AI. Openness may be one of the most important paths to AI safety and security, the letter argues, because concentrating advanced capability behind a few closed models creates single points of failure, and a world where attackers use advanced AI requires defenders with comparable capabilities.
Does the letter acknowledge the risks of open weights?
Yes. The letter states that released weights are beyond the original developer’s control and that modified versions are hard to trace. It argues the answer is not prohibition. Instead, it says open models broaden defensive capability and let vulnerabilities be found and remediated across many teams, and that relying solely on closed models is not inherently safe.
What policy actions are signatories asking for?
The letter asks Washington to expand access to compute for startups and researchers, invest in shared training assets such as datasets, tools, and evaluation frameworks, and keep the frontier plural by avoiding premature restrictions that could stifle competition or drive innovation overseas. It also asks policymakers not to conflate legitimate model-development techniques with misappropriation, specifically calling out distillation, the practice of using one model’s outputs to train another, and arguing it should not be confused with misappropriation.
Why is this letter coming out now?
The letter follows reporting that Chinese labs extracted intelligence from U.S. models, including Anthropic’s Fable capstone project. Washington’s policy response to those allegations remains unclear, and Anthropic’s absence makes it the lone U.S. frontier lab not represented among the signatories.
FAQ
What is an open-weight AI model?
An open-weight AI model is one whose trained parameters can be downloaded by anyone, allowing the user to inspect, modify, and run the model on their own infrastructure rather than relying on a closed API.
Which companies signed the Open Weights and American AI Leadership letter?
The letter, dated July 24, 2026, was shared by Nvidia CEO Jensen Huang on X. Initial signatories included Nvidia, Microsoft, and Meta, and within a day the list expanded to roughly 50 names, adding OpenAI, Google, AMD, and Cisco. Anthropic did not sign.
Why did Anthropic not sign the open-weight letter?
Anthropic has publicly warned about the risks of releasing open-weight models. The letter was framed as a defense of open-weight AI, and Anthropic was the only major U.S. frontier lab whose name did not appear on the list.
Source: Open Weights and American AI Leadership (Nvidia PDF).
Related coverage
- Poolside releases Laguna S 2.1, an open-weight coding model it says beats rivals 10x its size
- Introducing Antares: Open-Weight AI Models for Vulnerability Localization
Related coverage
- Poolside releases Laguna S 2.1, an open-weight coding model it says beats rivals 10x its size
- Introducing Antares: Open-Weight AI Models for Vulnerability Localization
This article summarizes reporting from images.nvidia.com.