The Deadline
After President Donald Trump signed his artificial-intelligence executive order in early June, his administration gave agencies across the federal government 60 days to produce a central implementation framework. The deadline is Saturday. The White House says it will deliver the document on time, pointing CNBC to a post on X by spokesperson Kush Desai: "BREAKING: Trump White House to meet a deadline we set for ourselves."
Under the order, AI companies are asked to give the government voluntary access to their models for review before releasing them publicly. The details are being hammered out by Scott Bessent from Treasury, Pete Hegseth from Defense, White House's Susie Wiles, Howard Lutnick from Commerce and others. OpenAI and Anthropic did not reply to a request for comment.
In Washington this week, tech executives including OpenAI CEO Sam Altman and Nvidia's Jensen Huang met with members of Congress and Trump administration officials. Altman told CNBC on Wednesday that he had seen a draft framework and did not have "anything specifically in mind" when it came to proposed changes. During his trip, he met with Wiles and was expected to see Bessent and Lutnick, according to reports.
The Open-Model Fight
The deadline lands as American AI development faces tension, with top tech figures arguing over the proper shape of rules and the outsize influence of dominant builders OpenAI and Anthropic. In recent weeks, open-weight models largely out of China have shown how quickly they can advance in capabilities. They are cheaper than the leading OpenAI and Anthropic offerings and can be downloaded, modified and run by users on their own hardware. That rapid progress has fueled an intense debate about whether the Trump administration should limit or even prohibit Chinese open-weight systems.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
The result is a policy squeeze: the administration wants to speed American AI leadership while also weighing national-security limits on publicly downloadable models. The framework due Saturday is meant to give companies a clearer path, but open-weight advances have made that task harder.
Tech executives have united in an almost unprecedented way against a potential ban. Jensen Huang amplified the debate Monday with a letter asking policymakers not to impose "premature restrictions" on open-weight models, and he used the release to publish his first message on X, the platform previously known as Twitter.
Elon Musk, who owns X via his rocket company SpaceX, replied, "This has my full support. Jensen is right." Microsoft, Meta, Palantir and many other firms signed the letter. OpenAI later added its name, while Anthropic declined and laid out its own view in a blog post.
The administration has not publicly stated a position, but David Sacks, the former AI czar who remains influential, has been one of the loudest voices against an outright ban.
What Comes Next
Under the executive order, the framework is expected to use a classified benchmark to evaluate models' cyber capabilities and decide whether they qualify as a "covered frontier model." It would also push companies to work with the government on identifying "trusted partners" who could get access to those models.
Both OpenAI and Anthropic may welcome a finalized framework, because they have been navigating an unclear deployment process since the order was signed. Right after the order came down, Anthropic was forced to disable Fable 5 and Mythos 5 in response to an export-control directive from Commerce. Fourteen days later, OpenAI said it had agreed to restrict its GPT-5.6 model series to "trusted partners" at the government's request. Eventually, both companies were able to release their models to a broader audience.
In a rare X post in late June, Wiles wrote, "The government and private sector have worked together in a way we have never seen before and this foundation of American First is unprecedented." She added, "Our shared priority remains: get the best tech deployed as quickly and safely as possible."
Sen. Warner told CNBC of open-source AI, "I'm not sure this is a genie we can put back in the bottle."
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
