A recent US decision to cut AI firm Anthropic out of federal supply chains has prompted a bold proposal from Germany’s Social Democratic Party (SPD): invite the company to base operations in Europe. SPD digital policy spokesperson Matthias Mieves called the idea a “once-in-a-lifetime chance” to bolster European digital sovereignty and keep influential AI technology on the continent.
Anthropic, valued at roughly $380 billion (€327 billion), is the creator of the Claude family of large language models, which have been applied across many areas including drone piloting. The company began working with US military and government agencies in late 2024. US media reports have suggested Claude was used in operations in Venezuela and in strikes on Iran, though those accounts remain sparse on detail.
In late February, US Defense Secretary Pete Hegseth demanded Anthropic remove contractual clauses that prohibit the use of Claude for mass surveillance and for fully autonomous weapons. When co-founder and CEO Dario Amodei refused, Hegseth designated Anthropic a “supply chain risk,” a label not previously applied to a US firm. President Donald Trump then ordered all federal agencies to stop using Anthropic’s services. Anthropic has challenged the designation in court, arguing the blacklisting could cost the company billions and damage its reputation.
Anthropic says it declined the US government’s requests for two principal reasons: it believes frontier AI models are not yet reliable enough to be entrusted with fully autonomous weapons systems, and it views mass domestic surveillance as a violation of fundamental rights.
Mieves wrote to Chancellor Friedrich Merz and European Commission President Ursula von der Leyen proposing that Europe offer to host Anthropic to continue developing “human-centered, trustworthy AI.” He argued Anthropic’s commitment to individual protections aligns with European legal norms and suggested three concrete measures:
– Offer a major European city—Mieves named Berlin, Paris, or Munich—as a potential new base for Anthropic.
– Create an alliance of European investors, including public actors such as pension funds and instruments like euro-denominated bonds, to strengthen Europe’s digital-financial backbone.
– Secure guarantees from the European Commission and key member states that they would partner with Anthropic going forward.
The proposal has drawn skepticism from AI industry figures. Daniel Abbou, head of Germany’s AI industry association BVKI, described the plan as “very, very nice” but unrealistic. He pointed out that Anthropic’s major backers include Amazon and Google and that the company is tightly integrated with US cloud infrastructure and capital markets. He warned that an outright move to Europe could jeopardize commercial contracts and liquidity.
Abbou also doubted that losing US government contracts would immediately topple Anthropic’s business, citing continued strong use of products such as Claude and Opus 4.6. As an alternative, he recommended forming a European AI lab in collaboration with Anthropic to attract and retain talent in Europe—a “reverse brain-drain.” Abbou highlighted structural problems that push European AI startups to the US: a weaker venture capital ecosystem and limited scaling opportunities across Europe, which encourage roughly 600 German AI firms to seek US partners or buyers.
Mieves conceded that his proposal might seem far-fetched but said that was intentional. He framed the initiative as a way to push policymakers to “think much bigger” and to consider radical steps if Europe wants to avoid ceding leadership in AI to the US and China.
The backdrop to the debate is the EU’s rollout of the Artificial Intelligence Act as part of a broader “Apply AI” strategy, which aims to promote values-based regulation. The law classifies AI applications into unacceptable, high, limited, and minimal risk categories. Some industry voices have urged delaying implementation, arguing strict rules could stifle early innovation and hinder the EU’s ambition to become an “AI continent.”
Calls for “digital sovereignty” are common among European politicians and business leaders: the idea that data, expertise, and critical infrastructure should remain under European control while protecting intellectual property. Arthur Mensch, CEO of French model-maker Mistral—currently the only European firm capable of building large language models—has warned about dangerous dependencies on US and Chinese providers. He argues Europe must be able to build models and operate the data centers that run them.
Mieves said his proposal was intended to force a more assertive conversation about concrete measures to achieve such sovereignty, rather than the vague rhetoric he sees today.
Anthropic was contacted repeatedly for comment but had not responded by the time of publication.