The Pentagon's recent ban on Anthropic's Claude AI system marks a new phase in AI geopolitical fragmentation, placing European regulators in an uncomfortable position between competing governance models.
Arthur Mensch, founder of French AI firm Mistral, reframed the debate at Modi's chaotic AI summit. "The fight for AI supremacy is between open versus closed systems rather than where those systems are built," he said. His position challenges the EU's regulatory approach, which treats AI governance primarily as a safety and rights issue.
India's integration into US AI supply chains and America's move to restrict Chinese AI access creates a bipolar tech landscape. Europe's AI Act, focused on risk categories and transparency requirements, addresses neither the open-source question nor strategic autonomy concerns.
DeepSeek researcher Luke Sernau claims "an open-source free-for-all is threatening Big Tech's grip on AI." This aligns with Mensch's view but contradicts Europe's cautious regulatory stance. The EU has backed open AI development through LAION and other initiatives, yet its compliance requirements favor large, resource-rich closed systems.
The regulatory mismatch creates competitive pressure. US firms like Anthropic and OpenAI operate under minimal federal oversight while securing Pentagon contracts. Chinese companies face domestic control but receive state backing. European firms navigate the AI Act's compliance costs while competing globally.
NTT scientist Hidenori Tanaka highlighted the knowledge gap: "AI is becoming ubiquitous, but how these computational engines actually work remains, to a surprising degree, a mystery." This uncertainty complicates risk-based regulation when fundamental AI mechanisms remain unclear.
Indian billionaire Mukesh Ambani's $110 billion AI infrastructure investment underscores how capital follows less restrictive regulatory environments. Europe attracts 8% of global AI investment compared to 40% for the US and 35% for China.
The EU must decide whether AI governance prioritizes safety controls, open access, or strategic competitiveness. Current policy attempts all three but achieves none decisively. As geopolitical fragmentation accelerates, Europe's middle path risks leaving it dependent on foreign AI systems: both open and closed.