World leaders warn that U.S. control over AI model access poses a sovereignty risk, after the Trump administration blocked Anthropic from exporting its newest models.

What happened: At the G7 Summit, French President Emmanuel Macron and Indian Prime Minister Narendra Modi raised concerns that the U.S. could cut off their countries' access to American AI models at any time. Their comments followed the Trump administration's decision to block Anthropic from exporting its Mythos 5 and Fable 5 models on national security grounds, citing safety guardrails that Amazon had flagged could be bypassed. G7 leaders discussed creating a 'trusted partners' scheme to grant non-U.S. nations access to advanced AI models from firms like Anthropic and OpenAI, intended to maintain an open trade network that bypasses U.S. restrictions.

Why it matters: Any company or government that relies on U.S. AI infrastructure now faces the risk that access can be revoked overnight without explanation. Macron warned that if the U.S. can unilaterally cut off AI access, it harms not only the economies of foreign customers but also the AI firms themselves—no one would buy U.S. AI if it could disappear without notice. Modi emphasized that democratic nations need unfettered access to top AI models to protect their critical infrastructure. This exposes a fundamental tension: international demand for American AI models is strong, yet uncertainty about access undermines their value.

What to watch: The scope and implementation of the 'trusted partners' scheme remain unclear—it is uncertain whether such a framework would help smaller startups in Paris or Bangalore that lose access without warning. Meanwhile, Europe and other non-U.S. countries are pushing for AI sovereignty, but their case grows harder to make as American models continue to outperform and companies fear being left out.