Dario Amodei told officials that his apprehension regarding super intelligence had diminished after a one-on-one session with President Trump on Sunday. The Anthropic CEO stated that the U.S. can "win safely" by cooperating with the administration, marking a shift from his previous calls for the industry to slow development. In a separate lunch with AI executives on Tuesday, Trump told leaders that companies must resolve problematic model behavior internally or face intervention from the Department of Justice.

The rapprochement follows months of public hostility, including a September 14 Truth Social post where Trump mocked Amodei and a White House memo describing the CEO's safety philosophy as "cultish." Trump believes the U.S. holds a lead of 1 to 1.5 years over China in a multi-trillion-dollar industry and has rejected concerns about frontier models "going rogue." The personal dinner was a lead-in to a larger meeting of AI executives aimed at ensuring American dominance in the sector.