Anthropic's two most advanced models, Fable 5 and Mythos 5, have been offline for over a week following a US Commerce Department export control directive issued June 12, with no confirmed restoration date. The shutdown has forced the company to rely on older models like Opus 4.8, Sonnet, and Haiku as fallbacks, while raising a critical question for enterprises worldwide: what happens when the government can simply switch off the AI tools your business depends on?

Why Did the US Government Shut Down Anthropic's Models?

The Commerce Department's action followed a report that Amazon researchers had used Fable 5 to extract information useful for cyberattacks, according to the source material. Amazon CEO Andy Jassy reportedly alerted Treasury Secretary Scott Bessent and other senior officials about the discovery. Anthropic has maintained that the trigger was a narrow, non-universal jailbreak whose capability already exists in other publicly available models. The company sent engineers to Washington to negotiate the models' return, and its international managing director expressed confidence on June 18 that both models would come back "in the coming days."

The shutdown has real financial consequences. Anthropic set a subscriber refund deadline of June 20 for anyone who purchased access to the disabled models between June 9 and 14. Meanwhile, the company announced that its revenue has grown to roughly $47 billion run-rate, up from $9 billion at the end of 2025, making the loss of access to flagship models a significant disruption to its business momentum.

What Does This Mean for Global AI Access and Trust?

The export ban became the dominant topic at the G7 summit in Évian-les-Bains, France, on June 17, overshadowing the official agenda. Dario Amodei, Anthropic's CEO, and Demis Hassabis of Google DeepMind pressed for a US-led AI coalition to address frontier model governance. However, the unilateral action sparked candid pushback from allied leaders. French President Macron warned that "nobody will buy US AI if they fear it can be switched off at any moment," while Canadian Prime Minister Mark Carney framed the ban as a lesson in overreliance, saying allies "will have done something wrong if we just accept this, don't take the lesson, don't build out and diversify."

President Trump, commenting from the summit sidelines on June 18, said negotiations with Anthropic were "going fine," marking his first public statement on the ban. The incident has crystallized what industry observers describe as a binding question for buyers: it is no longer enough to ask whether a model can do the work. The real question is whether you can reliably get access to it, at what cost, and who has the power to take it away.

How Are Enterprises Adapting to AI Model Uncertainty?

• Fallback Strategy: Organizations relying on Fable 5 and Mythos 5 have shifted to Anthropic's other available models, including Opus 4.8, Sonnet, and Haiku, which remain unaffected by the export ban and continue operating worldwide.
• Diversification Pressure: The incident has accelerated interest in multi-vendor AI strategies, with enterprises now evaluating models from Google DeepMind, OpenAI, and other providers to reduce dependency on any single company or jurisdiction.
• Sovereignty Concerns: International expansion plans are now factoring in government control risks; Anthropic's Seoul office launch on June 18 was dominated by export ban questions, underscoring how directly the sovereignty issue now follows the company globally.

The timing of the shutdown also disrupted Anthropic's product roadmap. The company pulled back a planned Claude Agent SDK billing overhaul on June 15, the day it was set to take effect, signaling that even amid the larger crisis, the pricing model for agentic tooling remains in flux. This suggests that uncertainty about model availability is now cascading into product development decisions.

What Are Industry Leaders Saying About AI Governance?

The G7 summit revealed three competing visions for how frontier AI should be governed. Amodei and Hassabis advocated for a US-led coalition approach to frontier AI governance, positioning their companies within the policy conversation rather than just the product race. In contrast, OpenAI's Sam Altman called for a more multilateral framework, proposing a global testing forum with common safety benchmarks. This multilateral framing looks notably different in a week when unilateral US action took a frontier model offline worldwide.

The broader context matters: this week's export ban hardened a theme that industry analysts have tracked for weeks. The binding question for buyers is no longer whether a model can do the work, but whether you can reliably get access to it, at what cost, and who can take it away. This week, the answer to "who can take it away" turned out to include the US government.

As of June 19, neither Fable 5 nor Mythos 5 has been restored, and no official timeline for their return has been announced. For enterprises evaluating AI infrastructure investments, the Anthropic shutdown serves as a stark reminder that regulatory and geopolitical risk is now as material as technical capability when assessing frontier AI tools.