U.S. cut off Anthropic’s Mythos access Friday, forcing Europe to accelerate “sovereign AI.”

On Friday, when the U.S. government pulled the plug on global access to Anthropic’s most powerful AI models, it confirmed a fear Europe has been nursing for a while: Washington can act like there is a “kill switch” for foreign access to U.S. AI systems.

The immediate consequence was panic and anxiety across Europe about dependence on American AI technology. And that is exactly what makes the episode different from a normal policy hiccup. This is not just about which model is better or whose lab is faster. It is about whether a continent’s critical digital capabilities can be unplugged by a foreign government, overnight, when geopolitics turns.

Europe is not starting from strength. The EU relies on non-EU countries for more than 80% of its technology and 70% of its cloud computing, according to the European Commission and the European Parliament. At the infrastructure level, the U.S. and China together control roughly 90% of global AI computing infrastructure, according to AI research firm Epoch AI. That leaves Europe with limited capacity to train or run AI models independently, even if politicians have been talking about strategic autonomy for years.

So when Anthropic’s shutdown hit, it re-ignited calls for what officials describe as “sovereign AI.” The idea is simple in concept and brutal in execution: countries should control the AI models, the computing infrastructure, and the data that underpin AI, instead of depending on systems that can be restricted or withdrawn by another government.

Europe has been taking small steps. The European Parliament replaced Google with Qwant, a French privacy-focused search engine, citing concerns over personal data collection. Germany, France, and the Netherlands have also moved Microsoft off public-sector infrastructure. But those moves do not fix the central problem the Anthropic event put a spotlight on: frontier AI requires compute at a scale Europe struggles to match, and it is not only a software story. It is a data center, semiconductor, and power story.

The scramble shows up in the details people are using to describe the stakes. In the U.K., Member of Parliament Al Carns said British hospitals, companies, and researchers had been suddenly stripped of access to Fable 5. Tom Tugendhat, a former U.K. security minister, framed the incident as evidence that national security is now more about “code than cannons,” and criticized the U.K.’s regulatory approach for prioritizing safety over building competitive AI capacity. Kanishka Narayan, the U.K.’s minister for AI, said the move has significant implications for defense, arguing that the most capable models used in drones, counter-drone defense systems, and cybersecurity now sit on the fundamental fault line for who wins and who loses in warfare.

Across the continent, politicians are pushing for faster European development. Many reactions focused on Mistral, described as the EU’s most credible AI contender, which is reportedly in talks to raise $3.5 billion in funding at a $23.2 billion valuation. Former French Prime Minister Édouard Philippe said the incident shows AI is now critical infrastructure, as essential as electricity or the internet, and that infrastructure controlled by others is infrastructure others can unplug. Bruno Retailleau, a French 2027 presidential candidate, called it a “wake-up call,” arguing that dependency makes a nation unplug-able overnight. In Finland, Member of the European Parliament Aura Salla said Europe cannot keep building its tech stack on access that can be switched off overnight by a foreign government. In Germany, MEP Sergey Lagodinsky pushed for the EU to team up with middle powers such as Canada, Australia, Singapore, and the U.K. to pool compute capacity through a consortium approach.

Critics, though, are warning that Europe’s “race” mindset may be expensive and not necessarily efficient. Sandra Wachter, a professor of technology and regulation at the University of Oxford, told Fortune that everyone was aware Europe depends heavily on U.S. technology, but the episode made it feel like “the side that actually has to face the kill switch.” She also suggested Europe could prioritize smaller, more efficient AI models that can run on smaller data centers and consume less electricity, instead of trying to match the scale of American frontier systems. If the capability gap is only a few percentage points, Wachter questioned whether closing it justifies the costs of an all-out race. She argued that the “racing” narrative can help the very companies developing those technologies, and said alliances with like-minded countries could be more sustainable than trying to outbuild the U.S. alone.

Policy is moving, but so far it is moving with questions attached. Just weeks before the shutdown, the European Commission unveiled what it billed as its answer to dependency risks: the European Technological Sovereignty Package, released on June 3. It targets cloud computing, AI, semiconductors, and open-source. Its two main legislative proposals are Chips Act 2.0 and the Cloud and AI Development Act. The Cloud and AI Development Act aims to triple data center capacity in the EU over the next five to seven years. “We cannot afford to depend on others for the technologies that keep our hospitals running, our energy grids stable, and our services secure,” European Commission president Ursula von der Leyen said.

The package targets an estimated €422 billion ($490 billion) in total investment across semiconductors, data centers, cloud, AI, and open-source software over the next decade. But the Commission has not confirmed how that will be financed, saying it will consult with member states and the European Investment Bank. Critics have questioned whether Europe can deliver. The Commission has deferred to member states on procurement rules enforcement, and under the proposed regulation, only around 10% of cloud contracts would carry a strong European sovereignty standard, with the remaining 90% open to all suppliers. Cristina Caffarra, head of the EuroStack Industry Initiative, told The Parliament the package was “very feeble,” arguing the Commission watered down previously stronger mechanisms for buying European under pressure from Washington.

Finally, the episode is not just Europe’s problem. Sovereign AI is quickly becoming a global power question. At February’s AI Impact Summit in New Delhi, conversations about sovereign AI and a potential coalition of “middle powers” were front and center. Yoshua Bengio, a Canadian computer scientist and AI pioneer, previously told Fortune that sovereignty is about global power, and about “equitable world order” where no one country can use technology to dominate the others, warning against ending up with “two hegemons” controlling part of the world. While the U.S. may have had legitimate national security grounds, the manner of intervention risks undermining American influence globally, according to Jonathan Iwry, a fellow at the Wharton Accountable AI Lab.

For executives and board members, the strategic takeaway is blunt. If frontier AI access can be revoked, supply chains are political. Compute is policy. And infrastructure sovereignty is no longer an abstract concept. It is the difference between being able to keep deploying and being forced into emergency workarounds while competitors sprint.