Amazon CEO Alerts Government, Restricts Anthropic Fable 5 Access

The US government issued an export control directive on June 12, 2026, requiring suspension of access to Anthropic's Fable 5 and Mythos 5 for any foreign national, whether inside or outside the United States, including foreign national Anthropic employees, according to Anthropic's official statement. The net effect was that Anthropic disabled both models for all customers globally to ensure compliance (Anthropic statement; TechCrunch). Amazon CEO Andy Jassy was on a pre-scheduled call with White House officials on June 11 when he raised security concerns from Amazon researchers; he subsequently spoke directly with Treasury Secretary Scott Bessent about a jailbreak technique enabling users to obtain information usable in cyberattacks, per TechCrunch and Fortune. The administration then contacted Anthropic CEO Dario Amodei, issuing a 90-minute deadline to fix the jailbreak or remove the models from the market; Amodei declined, and the Commerce Department issued strict export controls (Fortune).

Anthropic's official statement disputed the severity of the finding, describing it as a "narrow, non-universal jailbreak" - essentially asking the model to read a specific codebase and fix software flaws - and argued that applying such a recall standard industry-wide would effectively halt all new frontier model deployments. Anthropic added that other models were not affected by the directive (Anthropic statement).

The incident extends a pattern of escalating tensions between Anthropic and the Trump administration: the Pentagon ordered federal agencies in February 2026 to stop using Anthropic models, and in March the administration declared Anthropic a "supply chain risk" (Fortune). Amazon has invested in Anthropic since 2023 and provided additional infrastructure funding as recently as April 2026 (TechCrunch). Industry observers note this case establishes a precedent in which a major cloud investor can escalate model vulnerability findings directly to executive-branch officials, triggering regulatory action against the model developer (Gizmodo).

Watch for Anthropic's path to redeploying Fable 5 and Mythos 5, formal export-control rule-making around AI model access for foreign nationals, and whether other frontier lab investors adopt similar vulnerability-escalation channels with regulators.