July 17, 2026 · Executive Power & Institutional Control
· Official action
Developing
White House Quietly Takes Control of Frontier AI Model Access
In July 2026, the Trump administration began asserting new control over which companies and partners receive early access to the most advanced ("frontier") AI models, according to CNBC, which cited people familiar with the arrangements. Anthropic's and OpenAI's lists of partners with early access to models such as Claude Mythos 5 and GPT-5.6 reportedly now require government sign-off; the administration blocked release of two Anthropic models on stated national-security grounds last month before restoring access after weeks of negotiation.
If confirmed, the arrangement would give the executive branch informal leverage over which firms and countries can use cutting-edge AI models, without a public rulemaking or regulatory process establishing that authority.
A White House official told CNBC that participation is voluntary and that release timing remains the companies' decision, disputing the characterization of government control. No official policy document describing the arrangement has been published.
If confirmed, the arrangement would give the executive branch informal leverage over which firms and countries can use cutting-edge AI models, without a public rulemaking or regulatory process establishing that authority.
A White House official told CNBC that participation is voluntary and that release timing remains the companies' decision, disputing the characterization of government control. No official policy document describing the arrangement has been published.