At the G7 luncheon, it’s a split of “ice and fire”: Altman, with his face brimming with joy, seems to be beaming directly at Trump, while Amodei—whose model is stuck—wears a grim, heavy expression.

According to Beating Monitoring, at the AI working luncheon during the G7 summit in Évian-les-Bains, France, the publicly released attendance list revealed the United States’ absolute dominance in AI technology. The tech sector had a total of 12 seats, with the US taking 5 seats by virtue of OpenAI’s Altman, Anthropic’s Amodei, DeepMind’s Hassabis, Scale AI’s Alexandr Wang, and Salesforce’s Benioff. By contrast, the other participating countries—including the UK, France, Germany, Japan, India, and Canada—were relegated to supporting roles: each received only a single, extremely scarce seat reserved for their leading domestic AI representative. The extreme imbalance in technical discourse power was laid bare at the dining table.

The three “big shots’” seating arrangement also suggested subtle political treatment. OpenAI CEO Altman sat to the right of Trump, while Google DeepMind’s head Hassabis sat immediately to the left of Trump; the placement of the two executives was highly indicative of core political influence. In contrast, Amodei, the founder of Anthropic—who had just been subjected to a model export ban imposed by the US government—was snubbed and placed far across from Trump, seated next to French President Macron, who was loudly protesting the US cut-off of supplies.

In footage leaked from the scene, Altman appeared in high spirits, while Amodei looked grim. Public opinion generally holds that the seating division at this G7 luncheon is not only about social etiquette, but also marks the fact that leading AI giants have officially entered the power room where geopolitical decisions are made—while the order of seats serves as a direct picture of Washington’s current reward-and-punishment mechanism regarding each company’s political compliance.