Move Anthropic to Europe. Sounds crazy? Unrealistic? Consider this: someone is currently planning to colonise Mars and build data centres in orbit. The market rewarded that vision with a two-trillion-dollar valuation. Against that backdrop, inviting a values-aligned AI company to open a legal headquarters in Europe sounds almost boring. It might also be the most pragmatic idea in the room.

A wake-up call Europe cannot afford to ignore
On 13 June 2026, the US Department of Commerce gave Anthropic ninety minutes' notice to suspend access to its two most capable AI models for every user outside the United States. No consultation. No right of appeal. Simply switched off.

European businesses that had built their workflows around these tools found out via a post on a website. No warning. No transition period. No alternative. This is not a hypothetical risk. It already happened. And nothing structural has changed to prevent it from happening again.

The problem is not political. It is structural.
European consumers and businesses rely on AI infrastructure they do not control. The contracts are governed by US law. The models run on US cloud infrastructure. The decisions are made in Washington.

The US CLOUD Act obliges American technology companies to grant federal authorities access to stored data on demand, regardless of whether that data sits in Frankfurt, Paris or Dublin. FISA Section 702 gives US intelligence agencies a parallel route to the same data, without an individual court order for each target. European data residency is not European data sovereignty. Most organisations have not yet grasped the difference, but they will.

The EU AI Act matters. The GDPR matters. But neither can protect a European company when a foreign government decides, for its own reasons, to pull the plug. This is structural dependency. And it has a straightforward solution.

What you are signing, and why it matters
This petition has two concrete goals, and one longer-term vision.

1) The first goal is to create awareness and open a serious public dialogue. Policymakers, European institutions, business leaders, the media and Anthropic itself need to engage with this question before it is too late. That conversation has not yet happened at the level it deserves. Signatures, coverage and public support are what make it happen.

2) The second goal is to invite Anthropic to establish its legal headquarters in the EU, while keeping its existing presence in the United States. A dual-headquarter model, of the kind operated by ASML, SAP and Novo Nordisk, would give Anthropic a stable European legal foundation without requiring it to abandon its current structure. This would mean European law applies. European institutions have standing. The decisions that affect European consumers and businesses are made on European soil.

This would be an important step. But it would not yet be the complete answer. Real digital sovereignty also requires European infrastructure: data centres, compute capacity, energy, talent. That takes time and money, and it requires meaningful support from the EU and national governments. That third step is the consequential one. But it is not utopian. If a two-trillion-dollar company can credibly plan data centres in orbit, Europe can credibly build sovereign AI infrastructure on the ground. The first step is what makes the rest possible.

Why Anthropic, specifically
No other frontier AI company has the record Anthropic has.

Anthropic was founded as a Public Benefit Corporation, legally required to prioritise long-term societal benefit alongside commercial returns. An independent Long-Term Benefit Trust can override the company's own board if it drifts from its safety mission. In February 2026, Anthropic refused a US Department of Defense contract worth $200 million rather than accept terms permitting its models to be used for mass surveillance and autonomous weapons. Federal Judge Rita Lin subsequently ruled the government's retaliation unlawful and described the administration's conduct as "Orwellian."

OpenAI accepted the Pentagon's terms. Google removed its own AI weapons pledge. xAI received the classified access Anthropic declined. There is no other frontier AI lab of which this picture can be drawn.

Anthropic's EMEA revenue grew more than ninefold in the past year. Its large European enterprise customer base grew tenfold. It already operates offices in London, Dublin, Zurich, Paris, Munich and Milan. The company is deeply European in its commercial reality. It simply does not yet hold a European legal address.

A homecoming
There is a detail rarely mentioned in the coverage. Dario Amodei, CEO of Anthropic, and his sister Daniela, President and co-founder, are the children of Riccardo Amodei, a craftsman from Massa Marittima, a small town in Tuscany. Their parents met in Italy and only moved to the United States in the 1980s. That is one generation. An invitation to Europe would not only be a strategic proposition. It would also be a homecoming.

What we are asking for
We are not issuing demands. We are asking for something more important: serious, open engagement with this question, with an honest weighing of the arguments for and against, and a clear-eyed comparison with the alternative, which is the status quo of structural dependency.

Specifically, we would love to see media, experts, organisations and European institutions engage with this publicly and bring it the attention it deserves. We would love to see supporters with networks and influence help bring Anthropic and European decision-makers into the same conversation. And we would love to see that conversation happen before October 2026, when Anthropic's anticipated NASDAQ IPO will make repositioning significantly harder.

This is not charity toward Anthropic. At a valuation approaching one trillion dollars, Anthropic does not need saving. This is Europe recognising a narrow strategic window and choosing to use it.

Sign this petition. Share it. And let's build the AI infrastructure Europe actually deserves.

---
Daniel Kugler is a brand strategist and a committed European. He writes about brand building, artificial intelligence and digital sovereignty, always placing people at the centre, whether founders, employees, consumers or users. For him, the societal dimension of technological change matters as much as the technological one.

This initiative is independent, unfunded and politically unaffiliated.

He would rather readers disagreed with him in the comments than simply followed along. John Stuart Mill put it better than most: "He who knows only his own side of the case knows little of that." That's the entire premise of this piece, and of the broader conversation around it. If you think the argument is weak, the numbers don't hold up, or something important has been missed, say so. That's not a courtesy. It's the only way any of this gets better.

Follow if you'd like to keep reading. Better still, get in touch.

LinkedIn: www.linkedin.com/in/kuglerdaniel