G7 lunch puts Amodei, Altman in same room

G7 CEO – Dario Amodei of Anthropic and Sam Altman of OpenAI are both set to attend a CEO lunch at the G7 meeting in Evian-les-Bains today—an awkward moment for European leaders after Washington suspended access to Anthropic’s latest AI models for non-US users. Despite

Evian-les-Bains is preparing for a carefully managed moment: a CEO lunch at the G7 meeting where Dario Amodei, Sam Altman, Demis Hassabis and Arthur Mensch will sit together while Europe absorbs the shock of a US decision that has just restricted access to Anthropic’s newest frontier AI models.

The timing is hard to ignore. The European scramble is happening Wednesday, with European leaders meeting the US president and Anthropic CEO Amodei just days after Washington suspended access to Anthropic’s latest AI models.

But the room isn’t expected to turn into a public showdown with Donald Trump. Diplomats and officials attending the gathering insist they can still work with the US to minimize security risks posed by frontier AI—turning a fresh rupture into a test of cooperation rather than a flare-up that deepens mistrust between the continents.

“We are ready to engage and tackle these security risks together with our like-minded partners,” European Commission spokesperson Thomas Regnier said in Brussels on Tuesday ahead of the G7 get-together.

A European diplomat offered a sharper image of what success would look like: “On the topic of frontier models, we should be able to create unity before the end of the G7. The question is to recreate confidence, we need to recreate a circle of trust.”

The lunch itself—listed on the official agenda as a two-and-a-half-hour discussion—will focus on how AI can drive economic growth and how to keep societies resilient, especially for young people.

Still, the dispute between the Trump administration and Anthropic is the “elephant in the room,” according to one industry representative who declined to be identified because they could not comment on preparations for the meeting.

Anthropic confirmed that Amodei would attend the G7, but declined to comment further on the discussions.

Even as the CEO seats were being lined up, Anthropic moved to keep channels open with Europe. The company said it would meet Commission and EU cyber authorities in San Francisco on Thursday. An Anthropic spokesperson said the visit is part of the company’s ongoing engagement with the EU. allied democracies and international institutions on frontier AI’s implications for cybersecurity and opportunities for international collaboration.

Under the EU’s AI law. providers of frontier AI models already face strict obligations to test and evaluate their models for a series of risks. Brussels. however. has been navigating the controversy cautiously because there has been no formal communication or notification from the US government regarding export controls—only a statement that Anthropic issued on Friday.

The US restrictions followed the order aimed at barring non-US citizens from using Anthropic’s newest models. In response, Anthropic said it had cut off global access to Fable 5 and Mythos 5, its newest, most tightly controlled models with elite cyber capabilities.

EU tech leadership also signaled concern about how any response would be framed. Henna Virkkunen. the bloc’s tech chief. told European Parliament lawmakers on Tuesday that contingency measures taken in this light should not be discriminatory against partners. referring to the US order against Anthropic.

At the same time, a read-out of a bilateral meeting between Commission President Ursula von der Leyen and Trump at the G7 on Tuesday did not mention the issue.

Arianna Podestà, the European Commission’s deputy chief spokesperson, said discussions at G7 with tech companies were expected, and she would not pre-judge what would be raised in that context.

The lunch also carries another layer of political intent for Europe’s biggest member states. Italian Prime Minister Giorgia Meloni is expected to use the meal to defend a position Rome has pushed since it chaired the G7 back in 2024. The focus is “responsibility” for AI and the definition of truth in artificial intelligence, an Italian diplomat said.

That matters now because a debate is already underway over watermarking AI-generated content. and Meloni discussed the issue on Monday with Japanese Prime Minister Sanae Takaichi in Rome. The diplomat said it would be more and more difficult to make the difference between what is true and what is fake. and that this is definitely a topic Meloni will raise.

Not everyone at the table is trying to make the case for confrontation. One of the few European tech executives set to participate, Domyn’s Uljan Sharka, told POLITICO that Europe needs to find common ground with the US.

“This narrative of us versus them is completely wrong, and I hope that during the G7 this is going to be addressed,” Sharka said. He added that the transatlantic partnership, as it exists for defense with NATO, should also work on AI.

But Sharka still picked his moment to criticize the pressure behind the US actions. “I don’t blame the US administration for doing that,” he said. “They were pushed and forced to take action,” referring to Anthropic’s branding of its models as highly capable of finding software vulnerabilities.

Across the Atlantic, the UK has signaled it is watching the situation closely. The UK government said it had been in touch with the US government and Anthropic to understand the situation. Following the export control order. the UK’s AI Minister Kanishka Narayan said the main lesson was that as the future of national security and technological sovereignty is debated. access to AI capabilities is crucial.

The sequence is straightforward. even if the diplomatic choreography isn’t: Washington’s restriction arrived before the G7 CEO lunch; Brussels wants confidence rebuilt fast; and companies are appearing in the same room while preparing separate meetings to discuss frontier AI’s cybersecurity implications.

For EU leaders, that creates a narrow window. They are trying to keep the focus on governance and risk management without turning the dispute into a public rupture—because today’s shared table is also tomorrow’s test of whether “circle of trust” can be more than a line on a diplomatic agenda.

G7 Evian-les-Bains Anthropic Dario Amodei OpenAI Sam Altman export controls frontier AI EU AI law Henna Virkkunen Ursula von der Leyen Giorgia Meloni Kanishka Narayan