NVIDIA and Microsoft Lead 230-Strong Coalition Backing Open-Weight AI Models

More than 230 companies and organisations have thrown their weight behind a push to keep AI model weights open and accessible. NVIDIA posted on X that the coalition backing the “Open Weights and American AI Leadership” letter has now passed that milestone, framing it as a clear signal of where the tech industry stands.

The letter calls on US lawmakers and regulators to avoid what signatories describe as “premature restrictions” on open-weight AI models. Its core argument: openness expands access, sharpens competition, and keeps AI innovation from concentrating in the hands of a few closed providers.

From 25 to 230 in Under a Week

The letter first appeared on 24 July 2026, posted by NVIDIA chief executive Jensen Huang in what was notably his first post on X. At launch, 25 organisations had signed — among them NVIDIA, Microsoft, Meta, IBM, Dell Technologies, Palantir, Mistral AI, Mozilla, Hugging Face, the Linux Foundation, Andreessen Horowitz, Y Combinator and Perplexity.

Meanwhile, the number doubled to 50 within days. Then the floodgates opened.

By 30 July 2026, public statements from NVIDIA and other coalition members, including AI Native Foundation and Genspark, put the count at more than 230 signatories. OpenAI, Google, AMD, Cisco, Cloudflare, GitHub, Block and Ollama are among those reported to have joined after the initial launch. That figure comes from the companies’ own communications rather than an independently audited registry, so should be read with that in mind.

What “Open Weights” Actually Means

The term matters here. Open-weight models are AI systems whose numerical parameters — the weights — can be downloaded, inspected, modified and run on an organisation’s own hardware. That’s a different beast from a closed system accessed only through a proprietary API, where the provider controls everything.

NVIDIA describes this distinction plainly: open weights give users real control. They can run the model on their own kit, adapt it for their needs, and keep their data within their own infrastructure. For businesses wary of sending sensitive information to an overseas cloud provider, that’s not a small thing.

The letter ties this to broader policy goals. Signatories argue that open weights can enhance security through independent inspection and collaborative testing — the idea being that more eyes on a model’s internals means problems get found and fixed faster.

A Coalition Spanning the Industry

The breadth of the signatory list is striking. It reaches across chip makers, cloud providers, enterprise software firms, AI model builders, open-source foundations, venture capital funds and start-ups. Jensen Huang said: “We thank Microsoft for its partnership and are grateful for the more than 230 organisations that have rallied around open weights.”

But not everyone has signed. Anthropic and Amazon are conspicuous by their absence from any published version of the signatory list. Analysts have noted this gap, reading it as a sign that not all major AI players share the same view on how open access and safety concerns should be balanced. Anthropic in particular has built its public identity around safety-first AI development and tends to favour more controlled release strategies.

Security researchers have raised a parallel concern. Open weights can be a double-edged tool — the same transparency that lets defenders stress-test a model also lets bad actors probe it for weaknesses or adapt it for harmful uses. The letter acknowledges these risks but argues they’re better handled through safeguards and responsible practice than through blanket restrictions.

The Policy Battle Behind the Letter

This isn’t just a trade association sign-on exercise. The letter lands in the middle of live debates in Washington about regulating AI — including discussions about restricting access to Chinese AI models and whether open-source or open-weight AI poses national security risks.

NVIDIA’s position aligns with its commercial interests. A healthy open-weight ecosystem means more demand for the GPUs and computing infrastructure the company sells. Some analysts have pointed this out, though NVIDIA has also contributed to open AI security research through an industry alliance, sharing models, weights, data and agent research.

The coalition’s message to policymakers is direct: regulate carefully, don’t act prematurely, and consider what the US stands to lose if innovation is pushed elsewhere. Whether Washington listens is another matter.

What This Means for Kent Residents

For businesses, councils and public-sector bodies across Kent, a growing library of open-weight AI models could mean more affordable and flexible options for deploying AI without being locked into expensive proprietary platforms. Organisations handling sensitive data — local authorities, education providers, or those working alongside NHS Kent and Medway — may find open-weight models especially useful for keeping data within UK infrastructure rather than routing it through overseas closed systems. Any US regulatory decisions shaped by this letter could also affect how easily UK organisations access or licence these tools through American cloud providers, so it’s a policy debate worth watching even from this side of the Atlantic.