Anthropic AI ban underscores EU push for tech sovereignty

EU Tech Sovereignty Package Gains Urgency as US Blocks Anthropic’s AI Models From Foreign Users

The European Commission’s Tech Sovereignty Package, published on June 3, has taken on new relevance after the Trump administration ordered Anthropic to suspend access to its most advanced AI models for all foreign nationals on Friday, underscoring the bloc’s vulnerability to American technology decisions.

The package comprises four initiatives: the Chips Act 2.0, the Cloud and AI Development Act (CADA), an EU Open Source Strategy, and a Strategic Roadmap for Digitalisation and AI in Energy. Together, they represent what the Commission called a “major shift” in its approach to technology, aimed at reducing structural dependencies on non-EU providers.europa

Sovereignty Framework for Public Cloud

At the heart of CADA is a four-tier sovereignty classification framework for cloud and AI infrastructure. Under the proposal, member states would be required to conduct sovereignty risk assessments of their cloud providers and procure services accordingly, with the most sensitive public-sector systems required to be hosted on services controlled by EU entities.jonesday

Level 4, the most demanding tier, requires full EU ownership and control, EU-cleared personnel, no transfer of AI inference data outside the bloc, and third-party audits validated by national authorities. The act also aims to triple EU data center capacity within five to seven years.iapp

The Chips Act 2.0, meanwhile, targets €120 billion in investment by 2035 and aims to produce at least 20 percent of the world’s cutting-edge semiconductors by 2030. Both legislative proposals now proceed to the European Parliament and Council for negotiation, a process that could take 18 to 24 months.reuters

Anthropic’s Forced Suspension Validates EU Concerns

On Friday, Anthropic announced it was disabling its Fable 5 and Mythos 5 models for all users after receiving a U.S. Commerce Department export control directive barring foreign nationals from accessing them, according to Reuters. The company said it was not given specific details regarding the national security concern.reuters

Amazon Web Services confirmed that Anthropic had requested the revocation of access “in all regions”. The suspension illustrated precisely the dependency risk the Commission’s package seeks to address: European governments and companies relying on American AI providers can lose access overnight due to decisions made in Washington.reuters

Building a European Alternative

The package allocates €2 billion over seven years for an open-source strategy designed to scale European alternatives in cloud computing, AI, and cybersecurity. A “Free Software first” principle would apply to public cloud and AI procurement, targeting 30 million active users of open-source collaboration tools by 2030.jonesday

The timing of Anthropic’s forced withdrawal — just ten days after the Commission’s proposals — has strengthened the hand of EU officials who argued the bloc needed enforceable sovereignty mechanisms rather than voluntary commitments. As one Commission official told Reuters, the proposals aim to ensure Europe “can develop, deploy and secure” technologies to satisfy its own needs.reuters