Anthropic took its Fable 5 and Mythos 5 models offline on June 13, 2026 after the White House ordered the company to block all foreign nationals from accessing them — including Anthropic's own employees outside the US. The shutdown was imposed over a weekend with little warning or public explanation, and it converted a long-running European talking point about technological dependence into an immediate political emergency in London, Paris, and Ottawa.
Fable 5 and Mythos 5 were already operating under safeguards limiting their use in what Anthropic calls high-risk areas. The administration's order pulled them entirely from non-US users, and part of the reasoning, according to reporting around the shutdown, is a White House belief that a China-linked group had accessed Mythos. Anthropic has separately accused Chinese rivals of using its models to train competing systems at industrial scale.
The reaction abroad was faster and sharper than the shutdown itself. Within hours, sitting ministers and presidential candidates in three NATO countries were citing Fable and Mythos by name as evidence that access to American frontier AI is now a strategic chokepoint, not a commercial product.
Key facts
- 01Anthropic pulled Fable 5 and Mythos 5 offline on June 13, 2026 after the White House demanded a block on all foreign nationals, including Anthropic's own staff abroad.
- 02UK AI minister Kanishka Narayan called AI 'the central political question of our time' and pushed for British capacity without naming the US directly.
- 03Former French PM Gabriel Attal called the move the start of 'the AI war' and compared it to Iran's blockade of the Strait of Hormuz.
- 04Canadian PM Mark Carney cited Mythos and Fable as proof of the danger of overreliance on a single AI partner.
- 05France's Mistral, Canada's Cohere, and state-backed efforts in Singapore and the UAE are now positioned as the sovereign alternatives.
In the UK, AI and online safety minister Kanishka Narayan used the moment to argue that Britain has to build its own frontier capacity. He did not name Anthropic, Donald Trump, or the US directly, and instead framed AI dependence as a national security gap on par with any other sovereign threat.
“We treat every other threat to our sovereignty with deadly seriousness, but we haven't learned to treat this one in the same way.”— Kanishka Narayan, UK AI and online safety minister
Narayan said Britain must decide how AI shapes its economy, security, and sovereignty "before someone else decides the answer for us." The speech was paired with imagery of British police and military, a deliberate signal that the UK government now treats foreign-controlled AI access on the same footing as defense procurement.
France was more explicit. Former Prime Minister Gabriel Attal, running as the presidential candidate for Emmanuel Macron's Renaissance party, called the shutdown the start of "the AI war" and compared it to Iran's blockade of the Strait of Hormuz. Le Monde reported similar alarm across the French political spectrum, and members of the European Parliament cited Fable and Mythos as proof that tech sovereignty has to move from policy paper to procurement decisions.
Canadian Prime Minister Mark Carney drew the same conclusion in plainer language. He framed the episode not as a wrong committed by Anthropic or by Washington, but as a structural lesson about concentration risk in the AI supply chain.
The sovereign AI argument is not new in Europe — the EU has spent years pushing to reduce reliance on US chips, cloud, and models. What changed on June 13 is that the hypothetical became a live case study. Trade disputes and NATO funding fights with the Trump administration had already softened the ground; the Anthropic shutdown gave every European industrial-policy hawk a single, datable event to point at.
The practical alternatives remain uneven. France's Mistral and Canada's Cohere are the most credible non-US, non-Chinese frontier labs, though neither matches OpenAI or Anthropic on capability today. Singapore and the UAE have focused on narrower bets — local-language models, AI infrastructure, sovereign compute — rather than chasing the frontier directly. Open-source releases from China and elsewhere offer another route, one that is harder for any single government to switch off.
China is the obvious exception to American AI dominance, and Beijing has spent years funding domestic champions for exactly the scenario that just played out against Anthropic's foreign customers. The Verge notes that Chinese models still lag US frontier systems in some areas, but the gap matters less when the alternative is a model that can be revoked by executive order.
Anthropic may restore Fable 5 and Mythos 5 access soon — the company has said it had little choice in complying with the order, which implies it is pushing to reverse it. Even a quick reinstatement does not undo the demonstration effect. Every CIO, defense ministry, and sovereign wealth fund outside the US now has a concrete reference point for what export-grade American AI looks like when Washington decides to pull the plug.
The business consequence for US frontier labs is the part Washington appears not to have priced in. OpenAI, Anthropic, and Google sell into a global market where enterprise and government buyers are now being told, on the record, by their own heads of state, to diversify away from US-controlled models. That is a multi-year tailwind for Mistral, Cohere, UAE- and Singapore-backed efforts, and any credible open-weights program — paid for, ironically, by an administration whose stated goal is American AI dominance.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
