Austria urges EU to host Anthropic as U.S. AI curbs widen gap

Austria has urged Brussels to explore hosting Anthropic inside the European Union after U.S. restrictions on frontier AI access deepened fears that Europe could be locked out of the most advanced models. In a letter to EU Technology Commissioner Henna Virkkunen, State Secretary for Digitalization Alexander Proell said the bloc should not be left watching from the sidelines as the next generation of AI is built, financed and governed.

Proell framed the issue as more than a corporate relocation pitch. “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,” he wrote. The letter, released by the Austrian government, argued that Europe should use the moment to decide whether it wants to shape the AI race or merely react to it.

The Austrian push comes after Washington tightened controls on who can use some of the most powerful U.S.-made models. Anthropic said it would abruptly disable access to its most advanced systems after a U.S. government order restricted access for foreign nationals, and later coverage identified the affected models as Fable 5 and Mythos 5. The move has become a flashpoint in Europe, where officials worry that access rules could harden into a lasting competitive barrier for companies, researchers and governments outside the United States.

The timing also intersects with a broader European industrial strategy. On June 3, the European Commission unveiled a European technological sovereignty package that includes Chips Act 2.0, the Cloud and AI Development Act, an EU Open Source Strategy and a strategic roadmap for digitalisation and AI in the energy sector. Brussels said the package is meant to strengthen the European Union’s capacity in semiconductors, artificial intelligence, cloud and open source, streamline datacentre deployment and create a single EU-wide framework to assess cloud and AI sovereignty.

That agenda gives Austria’s intervention added weight. Proell’s letter effectively presses the case that Europe cannot regulate its way into leverage alone, and that it may need to welcome major AI firms on European terms if it wants strategic influence over infrastructure, investment and model access. The dispute widened further when OpenAI delayed a full public launch of GPT-5.6 at the U.S. government’s request, limiting early access to a small group of vetted partners whose details were shared with authorities. A U.S. legal technology company has also sued over the directive that forced Anthropic to cut off access to two of its most advanced models worldwide, underlining the legal and commercial blowback from Washington’s tighter grip on frontier AI.