The G7 summit opens today in Évian-les-Bains — and for the first time, the three most important AI lab chiefs are in the room together. Dario Amodei (Anthropic), Sam Altman (OpenAI) and Demis Hassabis (Google DeepMind) all accepted host France’s invitation. It’s the first G7 meeting with all three major labs represented.
What’s on the agenda
France has built its presidency this year around AI openness, governance and safety. Sounds like clear principles — but the reality is messier. Two years ago, gatherings like this still produced relatively hard commitments on the risks of generative AI. The 2026 edition, by contrast, is likely to focus mostly on the economic upside of the technology.
Observers expect any language on AI governance to be heavily watered down. The reason: the US has made clear it opposes multilateral agreements that could threaten its industrial lead. That fits the picture — just last week the US government used an export-control order to force Anthropic to block Fable 5 and Mythos 5 for all foreign users.
A rare moment of unity
As fiercely as the three labs compete in the market, on one point they’re currently pulling in the same direction. Altman, Amodei and Hassabis co-signed a letter to the US Congress calling for stricter rules on synthetic DNA and AI-enabled biological threats. Three rivals teaming up in public is rare enough to make you look twice.
My take
The summit shows just how deeply AI has entered international diplomacy. When the heads of three private companies address the world’s leading economies, the message is clear: the question of who builds frontier models — and who gets to access them — is now foreign and security policy. The interesting part is whether anything concrete comes out of it, or just a soft closing communiqué — while every country quietly defends its own AI sovereignty. My bet’s on the latter. But I’d be happy to be surprised.
Sources: