Anthropic suspended access to its latest AI models for foreign nationals on Friday evening after receiving a U.S. government directive [3].
This move signals an intensifying effort by the U.S. government to treat high-capacity artificial intelligence as a strategic asset subject to strict export controls. By restricting access to the most powerful tools, the administration aims to prevent foreign adversaries from leveraging advanced AI for cyberattacks or intelligence gathering.
The U.S. Department of Commerce issued the export-control order on June 12, 2026 [3]. The directive specifically targeted two of the company's top-performing models, Fable 5 and Mythos 5 [3]. These tools were disabled to comply with national security concerns regarding cybersecurity, and hacking [1, 2].
The suspension occurred just three days after the public launch of Claude Fable 5 [1]. The rapid reversal highlights the volatile intersection of commercial AI deployment and federal security oversight — a tension that has grown as models gain capabilities in coding and system exploitation.
While the company acted on the directive, reports on the timing of the administration's orders vary. Some reports said the suspension followed a formal national-security order [1], while others suggested a delay in the signing of a broader AI executive order [1].
Anthropic has not provided a timeline for when access might be restored or if the restrictions will expand to other model versions. The company is now navigating a regulatory environment where the speed of innovation is being checked by the requirements of state security [2].
“Anthropic suspended access to its latest AI models for foreign nationals on Friday evening.”
The restriction of Fable 5 and Mythos 5 suggests that the U.S. government now views frontier AI models as dual-use technologies, similar to advanced semiconductors or military hardware. By imposing these controls, the Trump administration is establishing a precedent where the federal government can intervene in the commercial availability of AI to mitigate perceived cybersecurity risks, potentially leading to a fragmented global AI landscape where access is determined by geopolitical alignment.
