Anthropic shutdown fuels overseas push for homegrown AI
Officials in Europe and Canada pointed to Anthropic’s model outage as evidence that access to advanced AI can be shaped by U.S. policy.
By James Whitfield · Staff Writer
3 min read
Anthropic took its Fable 5 and Mythos 5 AI models offline over the weekend after a White House demand to block access for foreign nationals, according to the company. The outage quickly became a warning sign for officials outside the United States who say critical AI tools should not depend on one country’s political decisions.
Anthropic said the U.S. request covered all foreign nationals, including some of its own employees. The Verge reported that the models were the company’s newest and most powerful systems and had already been subject to limits on use in high-risk areas.
The move sharpened an existing debate over “sovereign AI,” a term governments use for domestic control over AI infrastructure, models and access. Officials in Britain, France and Canada each pointed to the shutdown as a reason to build more capacity outside U.S. providers.
European officials cite strategic risk
In the UK, AI and online safety minister Kanishka Narayan did not name Anthropic, President Donald Trump or the United States directly. In a video posted on X, Narayan argued that Britain should treat AI dependence as a sovereignty and national security issue.
Narayan said Britain treats other threats to sovereignty with “deadly seriousness” and should apply the same approach to AI. He said the country must decide how AI affects its economy, security and sovereignty before others make that decision.
In France, former Prime Minister Gabriel Attal was more direct. Attal, who is running as the presidential candidate for Emmanuel Macron’s Renaissance party, called the shutdown the beginning of “the AI war” and said France’s reliance on foreign technology leaves it exposed.
Attal compared the loss of access to Anthropic’s models with a blockade of the Strait of Hormuz, framing AI as a strategic chokepoint. Le Monde reported that French politicians from across the political spectrum voiced similar concern.
The European Union has already put more emphasis on reducing dependence on outside suppliers in chips, cloud computing and AI, according to EU policy materials. The Anthropic episode gave that campaign a more immediate example for lawmakers who argue Europe needs its own technology base.
Canada and China enter the debate
Canadian Prime Minister Mark Carney also tied the outage to the risk of relying too heavily on one partner. The Associated Press reported that Carney said the Mythos and Fable situation shows why countries should build out and diversify their access to important technology.
China has pursued domestic AI champions for years, according to The Verge, and is one of the few markets with models that can compete with U.S. frontier labs. The Verge also reported that Anthropic has accused Chinese rivals of using its models to train their own systems at an “industrial” scale.
The Verge reported that part of the White House decision involving Mythos stemmed from a belief that a China-linked group had accessed the model. Anthropic has not been reported as saying that all foreign access posed the same risk.
Most governments and companies lack the resources of leading U.S. and Chinese AI labs. Still, The Verge noted that efforts such as France’s Mistral and Canada’s Cohere show that capable AI companies can emerge elsewhere, while countries including Singapore and the United Arab Emirates have focused on infrastructure or local-language systems.
Anthropic may restore access to Mythos and Fable, according to The Verge. The political effect may last longer, as officials outside the United States now have a concrete example of how quickly access to U.S.-controlled AI can change.
This story draws on original reporting from The Verge.