On September 29, 2026, Nvidia CEO Jensen Huang and other executives reportedly questioned Anthropic CEO Dario Amodei about his public warnings on AI capabilities and risks, after a White House lunch attended by about two dozen technology-company CEOs. The account became public on September 30, a day after the reported exchange.
Huang and other executives questioned Amodei after a White House lunch
The lunch took place in the White House East Room. A smaller discussion followed in the Roosevelt Room, where executives and White House aides and advisers discussed AI-safety principles. It was in that setting that Huang and other executives reportedly asked Amodei why he was speaking so forcefully about AI risks.
Amodei reportedly argued for candor about AI risks
Amodei reportedly defended speaking openly about what AI models can do and the risks they may pose. The exchange put a familiar disagreement about AI safety into a private conversation: how directly should leaders discuss the technology’s capabilities and potential dangers?
Executives had agreed to sign AI-safety principles
Before the smaller discussion, executives had agreed to sign a statement of principles intended to keep AI models safe. That commitment formed part of the gathering’s context as they discussed AI risks.
Amodei’s public argument for pacing AI development
The reported exchange followed Amodei’s public call to slow the pace of AI capability development so safety work could keep up. His proposal also called for independent evaluators and international cooperation. NeoTeo has covered Amodei and Huang’s public positions on AI safety; the White House discussion adds a reported private exchange to that debate.
