Austria proposes EU hosts Anthropic after US export controls cut off Europe's access to top AI models

Austria's proposal: Host Anthropic on European soil

On June 28, 2026, Austrian State Secretary Alexander Proell formally urged EU Technology Commissioner Henna Virkkunen to explore "the strategic establishment and participation of Anthropic within the European Union" — effectively proposing that the EU host Anthropic on European infrastructure to insulate its models from unilateral US export controls . The proposal, first reported by Bloomberg, is a direct response to the June 12 directive . Anthropic had not responded to the proposal at the time of publication .

EU Commission pushes back and investigates

The European Commission moved quickly on multiple fronts. On June 14, the Commission issued a statement saying the restrictions "should not be discriminatory against partners" and opened an investigation into the practical consequences for European users . It acknowledged it has no legal authority to compel the US to restore access .

On June 15, EU Executive Vice-President for Tech Sovereignty Henna Virkkunen publicly rejected the US national security justification, stating: "Europe represents an economic opportunity, not a security risk. We are and will remain a trusted partner" . She raised the model-access cutoff directly with the White House during a Washington trip the week of June 22 .

The incident supercharged the EU's existing technological-sovereignty agenda, with Virkkunen framing the episode as proof that Europe must achieve "technological autonomy" .

US partial reversal: Mythos 5 restored for vetted US entities

On June 26, Commerce Secretary Howard Lutnick partially reversed the ban. He sent Anthropic a letter exempting "certain trusted partners" — over 100 named companies, government agencies, and critical-infrastructure providers — from export-license requirements for Mythos 5 only, the cybersecurity model . Fable 5, the consumer version, remained restricted for foreign nationals .

Anthropic welcomed the turnabout, but the relief was narrow: most European commercial users still lacked access to either model . President Trump told Axios on June 19 that he no longer viewed Anthropic as a national security threat, two days after meeting CEO Dario Amodei at the G7 summit, though the Commerce Department's order remained in force .

Parallel sovereignty moves by France, Germany, and Mistral

At the VivaTech conference in Paris on June 17, the French and German economy and digital ministers issued a joint call to action for European AI sovereignty, warning that the US demonstrated "rules can change overnight" . The crisis dominated both the G7 summit and VivaTech, with policymakers and executives accelerating risk-distribution strategies and calls for local infrastructure investment .

French AI rival Mistral AI leveraged the incident as validation of its long-standing warnings against European dependency on US-owned models . CEO Arthur Mensch argued the sudden cutoff confirmed the need for European-hosted, sovereign alternatives . France announced €655 million for AI on June 15, including a government chatbot powered by Mistral AI .

Who still lacks access

As of the partial reversal on June 26, most European commercial users remained locked out of both Fable 5 and Mythos 5 . The partial carve-out for Mythos 5 covered only US-based trusted partners, and Fable 5 had no confirmed release date for non-US users . The European Commission's AI Office, which gains enforcement powers in August 2026, could theoretically demand model access, but the Commission has made clear it has no current legal authority to compel the US to restore access .