France to Boost AI Funding After US Blocks 2 Models; Startups Seek Independence

The G7 summit in Evian-les-Bains on June 17, 2026, became a flashpoint for the escalating struggle over artificial intelligence governance, pitting European calls for multilateral regulation against a sudden, nationalist move by the United States. French President Emmanuel Macron opened the high‑level meeting by urging wealthy democracies to jointly shape rules for advanced AI, warning that leaving safety solely to tech companies or individual nations risks fragmentation and economic instability. OpenAI CEO Sam Altman reinforced that message, proposing an 'international forum' to draw up AI guardrails, insisting that industry cannot alone manage the risks of frontier models. The backdrop to these calls was the Trump administration’s directive the previous week—enforced on Friday, June 12—that barred foreign nationals from using Anthropic’s newest models, Fable 5 and Mythos 5, forcing the company to abruptly take them offline. Anthropic publicly disagreed with the necessity of the step, stating the government’s security concern did not warrant such a severe restriction. Macron acknowledged the U.S. recognition that frontier AI can be dangerous—calling it a 'good thing'—but sharply criticized the move as a 'strictly nationalist' reaction that damages trust and could backfire economically. He warned that American AI firms might suffer a sudden drop in valuation if they are perceived as unreliable partners that can switch off access 'like a light switch.' The working lunch that followed included the three most powerful AI CEOs—Altman, Google DeepMind’s Demis Hassabis, and Anthropic’s Dario Amodei—discussing safe, rapid, and effective deployment. For Europe, the episode crystallized fears of dependency on U.S.-controlled AI infrastructure. Macron responded with an 'insurance policy,' pledging that France would significantly boost funding for its own AI industry to ensure sovereignty should international cooperation collapse. This parallel move signals a potential decoupling of AI ecosystems, with Europe accelerating investment in homegrown models and compute infrastructure. The directive, issued under national security authority, marks a new phase in AI export controls, moving beyond hardware restrictions to directly limit access to software models. For the broader industry, the incident raises urgent questions: if frontier models can be abruptly revoked based on nationality, global enterprises and SaaS providers relying on those APIs face operational and legal chaos. The tension between the U.S. administration’s unilateral security approach and the multilateral regulatory vision favored by Europe is now out in the open. The G7 meeting, intended to foster coordination, instead illuminated a deepening rift. Macron’s dual strategy—pushing for international rules while building domestic capability—reflects a pragmatic recognition that the window for collective action may be closing. The coming months will test whether Altman’s proposed international forum can gain traction, or whether the world will fragment into competing AI blocs with incompatible regulatory and access regimes, raising costs and slowing innovation globally. Ultimately, the G7 served as a stage for a larger debate about whether AI safety will be achieved through cooperation or through fortress-building, with profound consequences for investment, research collaboration, and the geopolitical balance of technology power.