The CEOs of OpenAI, Anthropic, and Google DeepMind attended a working lunch with G7 leaders at the summit in Evian-les-Bains, France on Wednesday, joining approximately a dozen tech executives in direct talks with heads of state about AI governance, infrastructure sovereignty, and frontier model risks.
Sam Altman (OpenAI), Dario Amodei (Anthropic), and Demis Hassabis (Google DeepMind) were among the attendees, according to CNBC. Mistral CEO Arthur Mensch, Cohere CEO Aidan Gomez, Salesforce’s Marc Benioff, Meta’s Alex Wang, Synthesia’s Victor Riparbelli, and representatives from Black Forest Labs, Domyn, Sarvam (India), and Sakana (Japan) also joined the session.
Export Controls as Backdrop
The summit takes place while Anthropic remains locked in negotiations with the Trump administration over export controls imposed on its Fable 5 and Mythos 5 models on June 13, CNBC reported. The release of Mythos marked an “inflection point” in AI development, Cameron Kerry, a visiting fellow at the Brookings Institution, told CNBC, adding that it prompted the Trump administration to consider regulating the technology.
U.S. export controls on Anthropic’s models have “changed everything,” Emerson Brooking, senior fellow at the Atlantic Council, told CNBC. “Multiple G7 nations have previously alluded to the need for sovereign AI investment, but there was always an assumption that this would take place alongside access to the U.S. tech stack,” Brooking said. “Now the U.S. has indicated a willingness to cut off the G7 and even treaty allies from certain AI capabilities.”
Voluntary Commitments Before Binding Rules
Frontier AI risks, infrastructure, and sovereignty were all on the agenda, along with online child safety protections, according to a press briefing from the Élysée Palace cited by CNBC.
OpenAI told CNBC earlier in June that it expected a set of “voluntary commitments” to be reached by tech companies during the summit. “The frontier labs want to shape this debate before any binding rules exist,” Brooking told CNBC.
The expected commitments cover youth safety, frontier risk in cyber and biological domains, and are “likely to become the de facto global baseline,” according to Jessica Brandt, senior fellow for technology and national security at the Council on Foreign Relations, speaking to CNBC.
Power at the Table
“It just shows that in order to make credible commitments on AI, heads of state now need the cooperation, if not endorsement, of a handful of private sector executives actually building the technology,” Brandt told CNBC. “We’re seeing a shift in who gets a seat at the table and a signal of where power sits.”
The G7 includes the U.S., U.K., Canada, France, Germany, Italy, Japan, and the EU. The working lunch format places tech executives in the same room as heads of state, a step beyond the advisory panels and separate-track summits that characterized AI governance discussions through 2024 and 2025.
For the agent infrastructure market, the summit’s outcomes could determine whether frontier model access fragments along national lines or remains interoperable across allied markets. The Anthropic export control situation, where Washington restricted allied access to specific AI capabilities, has already reshaped how European and Asian governments evaluate dependence on U.S.-based AI providers.