OpenAI, Anthropic CEOs break bread with G7 leaders as US AI export controls rattle allies

The CEOs of the two most powerful AI companies on Earth sat down for lunch with the leaders of the world’s seven largest economies on June 17. The menu wasn’t the interesting part.

Sam Altman of OpenAI and Dario Amodei of Anthropic attended a working lunch at the G7 summit in Evian-les-Bains, France, alongside executives from Google DeepMind and Mistral.

What happened, and why the timing matters

Around June 12-13, the US government issued directives restricting foreign access to Anthropic’s flagship models, Fable 5 and Mythos 5. Anthropic responded on June 13 by suspending global access entirely, a blunt move that effectively cut off allied nations from tools they had been using.

Four days later, the people who built those tools were sitting across the table from the people whose countries just lost access to them.

The lunch wasn’t just a photo op. G7 leaders used the session to explore what’s being called a “trusted partner” framework, essentially a tiered system that would allow vetted allies to regain selective access to cutting-edge US AI models.

Europe’s sovereignty push

The restrictions landed particularly hard in Europe, where officials have been vocal about the risks of depending on American AI infrastructure. The response from European leaders at the G7 was predictable but significant: renewed calls for sovereign AI development.

Mistral, the French AI startup whose CEO was also at the lunch, represents exactly this kind of European ambition. The company has positioned itself as the continent’s best hope for a homegrown competitor to OpenAI and Anthropic. Having its leadership in the same room as G7 heads of state wasn’t an accident.

Why crypto watchers should pay attention

Look, this summit wasn’t about crypto. Not a single blockchain mention made it into the discussions. But the underlying pattern, governments asserting control over which technologies cross borders and who gets access, has direct implications for the digital asset space.

For investors in AI-adjacent crypto projects, particularly those building decentralized compute networks or on-chain AI inference platforms, the “trusted partner” framework is worth watching closely. If access to frontier AI models becomes gated by government-to-government agreements, decentralized alternatives could see increased demand from countries left outside the velvet rope. Conversely, those same projects could face regulatory pressure if they’re seen as circumventing export controls.

The most important signal from Evian-les-Bains isn’t what was decided. It’s that the CEOs of private AI companies are now regular attendees at gatherings previously reserved for heads of state.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.