‘The AI War Has Already Begun’: Europe Confronts the Reality of American Tech Dependency

EBM NEWSDESK ANALYSIS

The US government’s decision to block foreign access to Anthropic’s most powerful AI models has done what years of policy debate could not: forced European governments to acknowledge that their digital infrastructure can be switched off overnight by Washington.

A Friday Evening That Changed Everything

At 5:21pm on Friday 13 June, Anthropic received an order from the US government directing the company to prevent all foreign nationals from accessing its AI models Fable 5 and Mythos 5, citing national security concerns. The letter did not explain the government’s specific security concern in detail.

To comply, Anthropic had to abruptly disable access to the models for all customers — including its own non-US employees — at short notice. Banks, law firms, government agencies, research institutions and businesses across Europe, Canada and the Middle East lost access without warning to technology that had been embedded in live workflows.

The administration cited a security concern that a China-linked group had accessed Anthropic’s new AI model. US national security adviser David Sacks revealed in a post on X that the government had received a warning that Fable 5 could be jailbroken, and that when Anthropic’s CEO Dario Amodei was notified, the company did nothing to fix it.

Anthropic rejected that characterisation. The company said it had reviewed the report which it assessed had triggered the order, and its experts concluded it referred to a limited capability to use the AI to review specific programme code and correct errors — a capability also possessed by models from other providers including OpenAI’s GPT-5.5. The company said it had received only verbal notice of a narrow, non-universal jailbreak and disagreed that it should be grounds for a recall of a model deployed to hundreds of millions of people.

The dispute between Anthropic and the Trump administration predates this incident. The company refused to allow the US military to use its AI models for fully autonomous weapons systems, after which the Pentagon placed Anthropic on a supply chain blacklist deeming it too dangerous for government use. With these export controls, it has now also been deemed too dangerous for foreign use.

Europe’s Moment of Reckoning

The reaction across Europe was swift and, in some quarters, alarmed. French presidential candidate Gabriel Attal declared over the weekend that “the AI war has already begun.” A European Commission spokesperson said the move clearly underlines the continent’s urgent need for tech sovereignty. Alexandra Geese, a European parliamentarian from Germany, said the US export ban “shows how the US government views Europe: as an enemy, not as a friend and ally.”

Mathilde Velliet, a research fellow at the French Institute for Foreign Relations who focuses on US tech strategy, said the ban “materialises a risk that had been on everyone’s mind” outside the US. What had previously been a theoretical concern — the existence of an effective American kill switch over European AI infrastructure — had become a documented reality.

The practical consequences were immediate. In the UK, former Armed Forces Minister Al Carns noted that British hospitals, researchers and tech companies had their pilot projects halted overnight by a foreign government, warning: “This is not a story about AI. It is the story of every industry in which we were leaders.” tradingeconomics

The Dependency Problem in Numbers

The Anthropic episode is an acute expression of a structural problem that has been building for years. Europe, the Middle East and Canada have struggled to create their own AI companies of a similar scale to US-based frontier models, meaning local companies and governments often choose dominant American products that make them more vulnerable to the Trump administration’s decisions. mexc

That dependency is now a live geopolitical liability. The EU’s AI Act, whatever its merits as a regulatory framework, does not solve the fundamental problem that the most capable AI systems in deployment across European institutions are American — and therefore subject to American export control law. Friday’s events demonstrated in concrete terms that this exposure is not hypothetical.

The Beneficiaries

The disruption has created immediate commercial opportunity for alternatives. Cohere, a smaller Anthropic competitor, reported a “huge” amount of inbound interest since the export ban. The Canadian startup recently acquired Aleph Alpha, a struggling German AI firm, in a deal pitched as a diplomatic alliance with Germany. mexc

The clearest potential beneficiary in Europe is Mistral. France’s Prime Minister Sébastien Lecornu announced that the civil service would roll out a tool based on Mistral AI. Chapsvision, a French company, was selected by the French domestic intelligence agency DGSI to replace US software firm Palantir. Those decisions had been in discussion before the Anthropic ban. The ban has dramatically accelerated the political momentum behind them. mexc

Mistral, with a sales pitch that emphasises business continuity, is in talks with investors about nearly doubling its valuation to €20 billion. The timing is not coincidental. European enterprises that had been evaluating the trade-offs between capability and sovereignty are now recalibrating that calculation under conditions that make the sovereignty argument considerably more compelling.

What the Ban Reveals About US Strategy

The Anthropic episode sits within a broader pattern of American AI policy that European governments are now being forced to assess with fresh urgency. The Trump administration’s dispute with Anthropic, which insisted that its models should not be used without guardrails against fully autonomous weapons and mass domestic surveillance, ended with the Pentagon designating the company a supply chain risk. The message from the world’s largest military power is that normative constraints on military AI are obstacles to innovation rather than preconditions for lawful use.

Canadian Prime Minister Mark Carney said the Anthropic ban shows how over-reliance on certain AI models leaves countries vulnerable, and advocated for a coalition of middle powers — the UK, Canada and European nations joining with Korea and Japan to share computing resources and use their firms in the tech supply chain as bargaining chips.

That coalition logic is gaining traction. Dex Hunter-Torricke, president of the London-based Center for Tomorrow and a former Google DeepMind executive, said: “We are going to be at the whims and the interests of Washington and Beijing. Every leader has to admit we need something different if we want to preserve our autonomy in the future.

The Strategic Imperative

The Anthropic ban is, in one sense, a narrow commercial and security dispute between a US AI company and its government. In a larger sense, it is the most clarifying moment yet in the question of whether Europe can afford to build its economic and institutional future on American technology infrastructure.

The answer that European governments are now arriving at — with a speed that years of policy discussion never produced — is that it cannot. The question is whether the alternatives can be built quickly enough to matter.