Nvidia, Microsoft and 23 Other Companies Ask the US Not to Restrict Open Weight AI Models
A group of 25 technology companies, led by Nvidia and Microsoft, published a letter on July 24, 2026, asking the United States government to avoid premature restrictions on open weight AI models. The letter lands amid a heated debate over Chinese open models, after American officials accused Moonshot of distilling Kimi K3 from Anthropic's Claude Fable 5, a company that, along with OpenAI, did not sign the document.
25 companies ask the US not to restrict open weight models
A group of 25 technology companies, visibly led by Nvidia and Microsoft, published an open letter on July 24, 2026, titled Open Weights and American AI Leadership, asking US policymakers to avoid broad or premature restrictions on open weight artificial intelligence models, according to reporting from CNBC, Qz and Artificial Intelligence News.
What the letter defines as open weight and what it asks for
The document defines open weight models as AI systems whose parameters, the weights, can be downloaded, inspected, modified and run on local infrastructure. The letter asks Congress and the US government to reject early restrictions on this type of model and specifically opposes broad limits on distillation techniques, the process of training or fine tuning a new model using an existing model's outputs.
Who signed and who stayed out
Signatories include Meta, Palantir, IBM, Andreessen Horowitz, Hugging Face, Mozilla, the Linux Foundation, Dell and AMD, according to a roundup from Value Add Pulse. The most notable absence, however, was OpenAI and Anthropic: neither of the two largest US closed model companies signed the document, reinforcing the reading that the letter mainly represents the interests of infrastructure and hardware vendors and the open source community, not necessarily the frontier labs that sell closed API access.
The trigger: the Kimi K3 distillation accusation
The letter lands during a week of specific tension: White House official Michael Kratsios publicly accused Chinese company Moonshot AI of having built Kimi K3, its open model this outlet covers separately, by distilling outputs from Anthropic's Claude Fable 5 without authorization. Moonshot denies the accusation. The episode gave fresh momentum to the US debate over how to treat Chinese origin open models within a broader restriction policy.
Distillation sanctions enter the US Treasury's radar
US Treasury Secretary Scott Bessent went as far as publicly floating sanctions specifically aimed at model distillation practices, according to the same CNBC coverage. It is exactly that kind of measure, potentially broad enough to also hit American companies that use distillation legitimately to train smaller, cheaper models, that the 25 company letter is trying to head off before it becomes law or formal regulation.
200 startups had already made the same request
The big companies' letter was not the first of its kind: a day earlier, nearly 200 American startups had already sent a similar appeal directly to the White House, according to the same CNBC report. The repetition of the request, first by smaller startups and then by giants like Nvidia and Microsoft, signals a broad, coordinated front from the US technology industry in defense of open models, at a moment when the government is weighing tougher measures.
Why it matters to Brazilian agencies and SMBs
US restrictions on open weight models, whether limits on distillation or barriers on Chinese origin models, directly affect which affordable AI options will be available to Brazilian agencies and SMBs that today benefit from competition between open and closed models to pay less for customer service and marketing automation. If this regulatory pressure narrows the supply of frontier open models, the indirect effect in Brazil tends to be higher prices and fewer alternatives when choosing an AI vendor, even without any Brazilian law changing at all.