Europe's anxiety over its dependence on American artificial intelligence has produced an unusual proposal: bring the Americans in. Austria has urged the European Union to explore courting Anthropic, the US AI company, and even helping establish it within the bloc, Bloomberg reported.
What Austria is proposing
In a letter to the European Commission's executive vice-president for tech sovereignty, Henna Virkkunen, Austrian State Secretary for digitalization Alexander Pröll suggested member states examine "the strategic establishment and participation of Anthropic within the European Union." The pitch, per Bloomberg, is to offer the company "legal certainty, market access, capital and a set of values that suits this company." Pröll did not spell out how it would work, and acknowledged there would be skepticism about whether it's even possible.
Why now: the export-control shock
The idea is a direct reaction to a jolt earlier this month. Washington ordered Anthropic to cut non-US users off from its two most capable models — Fable 5 and Mythos 5 — on national-security grounds. (An export control is a government rule restricting the sale or transfer of sensitive technology abroad.) Rather than carve out exceptions, Anthropic disabled the models for foreign users worldwide, abruptly cutting off institutions across roughly 15 countries — including close US allies — before access was later partially restored. Boursel has covered the episode and the diplomatic strain it caused.
For European policymakers, the lesson was uncomfortable: the bloc has written the world's most detailed AI rulebook, yet it depends on American companies for the frontier models, American clouds to run them, and largely American chips underneath. A single decision in Washington could unplug European researchers and businesses overnight.
What "courting Anthropic" would mean
Concretely, drawing Anthropic into the EU could involve a European data-center footprint, regulatory comfort, capital, or partnership arrangements that give the bloc more reliable, harder-to-sever access to top models. The appeal for Anthropic would be hedging against US policy risk while deepening its foothold in a large market; the trade-off for Europe is that its "sovereign" answer to AI dependence would still run on a US-founded company.
That tension explains why Europe is also backing home-grown options. France has thrown money and contracts behind Mistral AI, the bloc's most prominent domestic model-builder, as part of a broader AI sovereignty push. But few in Brussels believe Europe can match the leading US labs quickly, which is exactly why a stop-gap like luring Anthropic is on the table.
The bigger picture
The episode marks a shift in how governments treat AI. Cutting-edge models and the computing power to run them are increasingly handled like strategic resources — think energy or advanced weapons — rather than ordinary software. The same logic runs through the US move to restrict Anthropic, China's drive to build domestic models, and now Austria's scramble to anchor frontier AI on European soil. Austria's letter is only a proposal, vague on mechanics and far from EU policy. But the instinct behind it — that a country or bloc which can be "switched off" from the best AI has a strategic problem — is quickly becoming conventional wisdom in capitals worldwide.
