You Can’t Regulate Your Way to AI Sovereignty
August 9, 2026 – 5:00 pm
For years, critics of Europe’s AI strategy have argued that while America builds artificial intelligence, Europe writes laws for it. The EU AI Act, a landmark attempt to govern AI, exemplified this perceived divide and was seen as proof that Europe prioritized regulation over innovation.
Recent events, however, suggest a shift in this debate. In June, U.S. restrictions temporarily blocked access to Anthropic’s advanced Fable 5 model, disrupting businesses and governments across Europe before the ban was lifted. Around the same time, OpenAI faced pressure to limit access to its most advanced capabilities, while Anthropic itself reportedly faced political scrutiny over restrictions on using its models for autonomous weapons and mass surveillance.
These episodes indicate a broader change: frontier AI is no longer viewed merely as commercial software but as strategic infrastructure. This shift is why Austria urged the European Union to attract Anthropic to strengthen Europe’s presence, turning a previous criticism on its head.
The EU AI Act, once criticized for driving innovation away, might now become a valuable asset in attracting AI companies due to legal predictability and regulatory stability. However, a deeper question arises: how safe can this "harbor" truly be?
Even if Europe successfully attracts frontier AI companies, their dependence on U.S.-based ecosystems—advanced semiconductors, hyperscale cloud providers, computing capacity, and energy systems—remains significant. Restricting access to these essential dependencies could unravel Europe’s AI strategy, as governments don’t need to control the entire value chain to exert influence.
In essence, relocating a company doesn’t necessarily move its dependencies, exposing a weakness in Europe’s approach to AI sovereignty.