Political risk / Jun 28, 2026 / 5 min
Austria Asked the EU to Host Anthropic Inside the Bloc
On June 28, Austrian State Secretary Alexander Pröll asked the EU to explore hosting Anthropic inside the bloc — not petitioning Washington for Mythos access, but trying to relocate the kill switch under European law.
Austria just told Brussels the quiet part out loud: Europe's AI sovereignty problem isn't compute or capital — it's that Washington holds the kill switch on Anthropic's frontier models, and the only fix is relocating the company under EU jurisdiction.
What's new: On June 28, Austrian State Secretary for Digitalization Alexander Pröll wrote EU Tech Commissioner Henna Virkkunen urging member states to "jointly explore the strategic establishment and participation of Anthropic within the European Union." Bloomberg, Reuters, and Austria's ORF confirmed the letter the same day. Anthropic declined to comment.
Why it matters: This is not another Brussels delegation begging Commerce to restore Mythos. Vienna is lobbying to move the lab — because a U.S. export order can revoke foreign access to Fable 5 and Mythos 5 regardless of what any European customer signed.
The trigger:
- June 12: Commerce ordered Anthropic to cut off all foreign-national access to Fable 5 and Mythos 5.
- June 26: Commerce partially restored Mythos to roughly 100 vetted organizations. Fable stays dark.
- June 28: Pröll's letter lands as OpenAI also restricts GPT-5.6 Sol to government-vetted partners.
What Pröll said:
- "Let us jointly explore the strategic establishment and participation of Anthropic within the European Union — with legal certainty, market access, capital and a set of values that suits this company." — letter to Virkkunen, reported by Bloomberg and Firstpost
- Via ORF: Anthropic "puts safety over speed" — "a deeply European attitude. This company would not be constrained in Europe; it would be unleashed."
- He admitted the idea would draw skepticism and offered no roadmap for how to achieve it.
The structural read:
- Permission, not parameters: For two years the race was who builds the best model. June proved the scarcer input is who permits it to ship — and to whom.
- Jurisdiction shopping: Hosting Anthropic in the EU would put distribution decisions under Brussels law, not Commerce export controls. That's the only lever export orders can't reach.
- Labs as sovereign assets: Pröll's pitch — legal certainty, market access, capital — is the language governments use to court national champions, not SaaS vendors.
Europe's split screen:
- The Commission unveiled a tech-sovereignty package this month targeting U.S. cloud and chip dependence.
- Brussels signed Washington's Pax Silica chip alliance days earlier.
- A CUDO Compute survey cited by Firstpost found 43% of U.K. organizations still prioritize cost over sovereignty when deploying AI.
The honest limit: Austria cannot unilaterally relocate a San Francisco PBC. Anthropic's weights, talent, and capital sit in the United States. Pröll knows it. The letter is a flare, not a plan.
But the flare landed: UpCloud CEO Arno Schäfer told Fortune — cited by Firstpost — that the Anthropic episode turned digital sovereignty from a political slogan into a business-continuity risk. Pröll's letter makes that risk explicit: when the frontier goes dark, Europe's options are petition Washington or rewrite jurisdiction.
Convina's view: Austria probably cannot move Anthropic to Vienna. That is not the point. Pröll just named the new scarcity in AI — permission — and admitted Europe does not hold it. Every government outside Washington now faces the same choice: accept guest-list access, build open-weight rivals, or court frontier labs with the one offer export law cannot block — a new legal home. The AI sovereignty debate finally has a concrete test case, and Brussels has to answer whether it wants to compete for labs or keep negotiating with Lutnick.