Anthropic CEO Dario Amodei will dine with President Donald Trump at the White House on the evening of September 27, 2026 (US Eastern Time). It will be the first one-on-one meeting between the two men — Axios broke the news of the dinner first, and TechCrunch subsequently confirmed it with a source familiar with Amodei's plans. The same weekend, Amodei was lampooned on the season premiere of Saturday Night Live, which portrayed him as a study in contradictions: a man warning that AI could destroy humanity while continuing to build it.
Why tonight's dinner is worth watching
The dinner's weight lies not in the meal but in who sits on either side of the table. In recent months, Amodei and Trump have ended up on opposite sides of the AI safety debate: Amodei published a plan arguing that AI development should slow down, or at least proceed with more caution; Trump has insisted, without evidence, that the backlash against AI is a hoax manufactured by Democrats, and he wants to rebrand the technology as "super intelligence."
Their disagreement is not academic — it touches federal money and contracts directly. A private meeting means the argument is moving from public sparring to face-to-face negotiation.
From "supply-chain risk" label to the dinner table: what happened this year
The relationship soured earlier this year. The Pentagon designated Anthropic a supply-chain risk after the company tried to put guardrails around how its technology is used, and Anthropic is now challenging that designation in court. Even so, the Trump administration is not a monolith — other officials have taken a friendlier stance toward the company.
Notably, the White House has been taking an increasingly hard line on AI model exports and sharing — it had already moved to require OpenAI and Anthropic's new models to clear US review before being shared with the UK. Viewed in that context, the dinner looks more like direct policy reconnaissance: the president wants to know how far the lab most associated with "slow down" is willing to go along with the administration's pace.
What will really be on the table
Neither side has disclosed the dinner's agenda, but the likely topics are not hard to list: the aftermath of the Pentagon's supply-chain-risk designation, the prospects for federal agencies buying Anthropic's models, and just how far the government intends to regulate AI safety commitments.
The dinner will not settle anything in one sitting, but it sends a signal: after trading public barbs, both sides think it is time to talk in person. For Amodei, who has consistently argued for the cautious route, it is a chance to put his safety case directly in front of the president; for Trump, it is a chance to find out whether this lab will be a problem for his policy agenda. The outcome will become clear enough once the two sides make their post-dinner statements.
