According to the Wall Street Journal, on June 15, Anthropic sent its top technical team to Washington over the weekend to meet with Trump administration officials and seek to lift export controls on its most powerful AI model. Last Friday, Commerce Secretary Howard Lutnick ordered Anthropic to prohibit foreign citizens from accessing the Fable 5 and Mythos 5 models, citing “national security threats.”
Timeline of the June 12 Order Confirming Controls
According to a confirmation timeline disclosed to the Wall Street Journal by Anthropic sources:
Friday at 1:00 p.m. (ET): Anthropic received a government call instructing it to withdraw the release of the Mythos and Fable models on the grounds of a “national security threat,” with no further details
Background confirmation: The government and Anthropic had previously conducted pre-release testing of Fable, and the company had obtained explicit deployment approval
What the source said: “We immediately tried to understand the specific nature of the threat in order to eliminate it,” but the government insisted on its original request
Friday at 5:30 p.m. (ET): Anthropic received a letter from the Department of Commerce imposing controls on where the models could be used and by whom
Company action: The company immediately initiated a process to roll back deployment of the models and shut off access for all users (including U.S. users and company employees)
Background of the Confirmation Dispute Between Anthropic and the Trump Administration Since February 2026
According to the report, this conflict did not happen suddenly:
February 2026: Federal agencies were banned from using any Anthropic products; Anthropic was listed as a “supply chain risk” company. The rationale was not a security vulnerability or data breach, but its own policies— the company has long strictly limited its technology for military purposes, including surveillance and autonomous weapon applications
February 2026: President Trump publicly referred to Anthropic as an “anti-extremist left-wing, woke company”
June 2026: The Department of Commerce ordered the closure of access a few days after the release of Fable 5 and Mythos 5
Legal Framework Confirmation for Export Controls Applying to AI Models
According to the report, imposing export controls on AI models is a relatively unknown area of regulation: traditionally, these mechanisms have been used for physical goods (weapons, semiconductor equipment, nuclear materials, etc.). In this round of controls, the Department of Commerce treated Fable 5 and Mythos 5 as dual-use hardware, rather than consumer software, and required access control based on the users’ nationality.
Anthropic’s compliance options: If it chose to keep the service running and immediately verify each user’s nationality, it would face compliance risk; the company ultimately chose to shut off access for all users to avoid the risk of immediate penalties.
FAQ
What are the specific grounds the U.S. government cited as a “national security threat”?
According to Anthropic sources’ disclosure to the Wall Street Journal, the government raised “national security threats” as the rationale during the phone call, but did not provide specific details. Anthropic said it immediately sought to understand the specific nature of the threat, but the government stuck with its withdrawal request and provided no further explanation.
Why did Anthropic need to shut off access for all users, rather than just foreign users?
According to the report, the Commerce Department’s control rules for Fable 5 and Mythos 5 prohibit foreign citizens (including foreign citizens residing in the U.S.) from accessing them. Anthropic believed that, compared with immediately verifying each user’s nationality and bearing the risk of penalties, fully shutting off access was a more direct compliance approach. As a result, including paid U.S. users and company employees, access was temporarily lost for all.
If Anthropic had already obtained deployment approval for Fable, why was it still required to withdraw it?
According to Anthropic sources, the government and Anthropic had previously jointly conducted pre-release testing of Fable, and the company had obtained explicit deployment approval. The June 12 government withdrawal request went beyond the scope of the previous approval. Anthropic said it was sudden and that it tried to understand the specific nature of the threat.
