When Washington pulled Anthropic’s most capable models from foreign hands, a small EU member did something no one expected: it asked to move the company in.
How We Got Here
June 13, 2026
A U.S. Commerce Department export-control order forces Anthropic to disable Claude Fable 5 and Mythos 5 — its frontier, “Mythos-class” models — for all foreign users on national-security grounds.
Days later
Commerce allows a limited restoration of access — but the precedent is set: a single government can switch the frontier off for everyone outside its borders.
June 28, 2026
Austria’s State Secretary for Digitalization, Alexander Pröll, writes EU Tech Commissioner Henna Virkkunen urging the bloc to “explore the strategic establishment and participation of Anthropic within the European Union.”
Earlier this month, the U.S. Commerce Department ordered Anthropic to cut off foreign access to Fable 5 and Mythos 5 — the two most powerful systems the company has ever shipped — citing national security. Anthropic complied, disabling the models for everyone outside the United States. Washington has since permitted a limited restoration, but the message landed: access to the frontier is now a lever a government can pull.
Austria’s response was not to lobby Washington. It was to lobby Brussels. In a letter to EU Tech Commissioner Henna Virkkunen, released by the Austrian government, State Secretary Alexander Pröll argued that Europe cannot afford to be cut off from major innovations — and floated hosting Anthropic itself inside the bloc.
Alexander Pröll · Austrian State Secretary for Digitalization
“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.”
Pröll did not spell out how it could be done, and conceded there would be scepticism about whether it is possible at all. That candor is the point: this is less a workable plan than a flare sent up by a government that just watched its access to frontier AI flicker off without warning.
The Structural Read
For two years the debate over AI power has been about who can build the best model. This episode quietly changes the question to who can permit it to ship — and to whom.
We call this the Permission Layer: above the compute, the models, and the applications sits a governance layer where states decide which capabilities cross which borders. Export controls turn that layer from theory into infrastructure. Once a model is classified as a national-security asset, the company that built it no longer fully controls its distribution — the government does.
The key insight: Austria isn’t trying to buy a model. It’s trying to relocate the permission. Hosting Anthropic in the EU would put the on/off switch under European jurisdiction — which is the only thing an export order can’t reach.
That reframes the whole move. “Host Anthropic in Europe” sounds like industrial policy. It’s really jurisdiction-shopping for the kill switch — an attempt to put the company under a legal regime where Brussels, not Washington, holds the lever.
Three Implications
1 · Frontier access is now a sovereign dependency
Every government and enterprise outside the U.S. just learned that its access to the best models can be revoked by a third country. That is a supply-chain risk in the boardroom sense, not a policy footnote.
2 · Labs get courted like sovereign assets
Pröll’s offer — “legal certainty, market access, capital” — is the language used to attract a national champion, not a SaaS vendor. Expect more jurisdictions to dangle similar terms at frontier labs.
3 · “Sovereign AI” gets a concrete test case
The slogan has floated around for two years. Anthropic-in-the-EU is the first time a real company, a real model, and a real export shock collide — making the abstraction suddenly legible to every regulator watching.
Business Engineer Framework
The Permission Layer
Above compute, models, and apps sits a governance layer where states decide which AI ships and to whom. Map where every player — and every dependency — sits in the stack before you bet a business on it.
Austria probably can’t move Anthropic to Vienna, and Pröll all but admits it. But the letter matters anyway, because it names the new reality out loud: in 2026, the scarcest thing in AI isn’t compute or talent — it’s permission. The country that controls the kill switch controls the frontier, and the rest of the world is only now realizing it doesn’t hold the switch.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
