US Government Curbs Anthropic AI Access, Sparks Global Push for Sovereign AI
The U.S. government's decision to cut global access to Anthropic's most powerful AI models has prompted international concern, particularly in Europe, regarding technological dependence. This action, described by some as a "kill switch," has reignited calls for nations to control their own AI infrastructure, models, and data. European officials and politicians are now emphasizing the strategic vulnerability of relying on foreign AI systems, accelerating efforts to develop independent capabilities.
On Friday, the U.S. government restricted global access to Anthropic's most powerful AI models, confirming existing fears in Europe regarding technological dependence. This incident has been characterized by some as a "kill switch," demonstrating the ability to disconnect foreign entities from U.S. AI systems.
Politicians worldwide have expressed alarm about their countries' reliance on U.S. AI technology. In Europe, this event has amplified calls for "sovereign AI," advocating for nations to control their own AI models, computing infrastructure, and data, rather than depending on systems subject to foreign control.
Europe has a significant reliance on non-EU countries for its technological infrastructure, with over 80% of its technology and 70% of its cloud computing coming from outside the EU, according to the European Commission and the European Parliament. Furthermore, an AI research firm, Epoch AI, reports that the U.S. and China collectively manage approximately 90% of the world's AI computing infrastructure, leaving Europe with limited independent capacity.
Some European governments had already taken steps to reduce this reliance. For instance, the European Parliament replaced Google with Qwant, a French search engine focused on privacy. Germany, France, and the Netherlands have also transitioned Microsoft off public-sector infrastructure.
The Anthropic shutdown has intensified the focus on the AI race. UK Member of Parliament Al Carns stated that British hospitals, companies, and researchers abruptly lost access to Fable 5. Tom Tugendhat, a former UK security minister, emphasized that national security now increasingly hinges on "code than cannons." Kanishka Narayan, the UK's Minister for AI, highlighted the defense implications, noting that capable AI models are crucial for modern warfare.
Across Europe, politicians are urging for accelerated development of indigenous AI models. France's Mistral, considered a leading European AI contender, is reportedly in discussions to secure $3.5 billion in funding at a $23.2 billion valuation. Former French Prime Minister Édouard Philippe asserted that AI is now critical infrastructure, essential as electricity or the internet, and that control by others implies vulnerability. French 2027 presidential candidate Bruno Retailleau described the event as a "wake-up call," stressing that technological dependence leaves a nation susceptible to being disconnected overnight. Finland MEP Aura Salla also reinforced the need for European independence in AI.
According to Fortune, these developments underscore a growing geopolitical concern over technological sovereignty.
