Published on
Last Friday, one of the AI race leaders, Anthropic, pressed the “kill switch” on its most advanced models, Fable 5 and Mythos 5. Following a surprise directive from the Trump administration citing national security, the company was forced to ban all non-US citizens from its frontier technology.
Across the continent, research labs, tech firms, and even some hospitals have spent the last year piloting these US-made systems for core operations. Now, they have learned that their digital infrastructure can be unplugged overnight.
European leaders across the spectrum—from France’s far-right presidential candidate Jordan Bardella to the European Commission’s own spokesperson, Thomas Regnier—are in rare agreement, arguing that nations that do not develop and control their own AI models will remain entirely dependent on the choices of foreign powers.
What’s even more interesting, earlier this month the European Commission announced its ambitious plan for European Strategic Autonomy in technology.
Anchored by the new Cloud and AI Development Act, the plan aims to triple Europe’s data centre capacity, build massive “AI Gigafactories,” and push for the secure open-source alternatives.
However, Europe is currently trapped in a massive “compute gap.” While the EU scrambles to build infrastructure, the sheer scale of investment from US giants dwarfs European efforts. Relying on “moonshot” projects alone might be simply too slow.
Another option could be coalition of the “middle powers”—by teaming up with nations like the UK, Canada, and Japan, Europe could pool its fragmented computing power, creating a sort of AI insurance policy.
However, regardless of what the EU does, it better do it quickly. Otherwise, these sudden kill switches will become a regular feature of the continent’s digital future.
Watch the Euronews video in the player above for the full story.