Nvidia’s Jensen Huang challenged Anthropic chief executive Dario Amodei over his public warnings about artificial intelligence at a closed-door White House meeting on 29 September 2026, The Wall Street Journal reported.
The newspaper published the account on 30 September, citing several sources. It shows a split among AI leaders on the same day the Trump administration brought them together to back voluntary rules rather than federal regulation.
The exchange took place in the Roosevelt Room, in a meeting with White House staff straight after a lunch hosted by President Donald Trump and House Speaker Mike Johnson. Huang asked Amodei why he issues such stark public warnings about the risks of AI models, the Journal reported.
Amodei held his ground. He told the other executives that the industry had to be honest with the public about what its models can do and should not play down the dangers. He pointed in particular to cyber threats and disruption to jobs.
A divided guest list
About 20 executives attended the lunch. They included:
- Elon Musk of SpaceX and Tesla.
- Sundar Pichai of Alphabet.
- Satya Nadella of Microsoft.
- Mark Zuckerberg of Meta.
- Greg Brockman, president of OpenAI.
South Korea’s Seoul Economic Daily reported that Trump seated Huang on his right and Musk on his left, with Amodei further away.
Zuckerberg also took on Amodei’s safety concerns during the lunch, according to the Journal. He argued that the industry could deal with them by keeping to its own principles, and opposed new regulation.
Speaking to reporters afterwards, Amodei said that if industry and government work together, “we can win safely”. Trump called the remark excellent.
A failed industry body
According to the Journal, the companies had earlier discussed setting up a self-regulatory organisation for the industry:
- Backers: Anthropic, OpenAI and Google.
- Opponents: Huang, Zuckerberg and Musk, who argued it would concentrate too much power in the hands of a few leading companies.
The plan was dropped.
The administration’s approach rests on voluntary commitments and testing of AI models before release.
Anthropic’s bid to mend ties
The meeting came two days after Trump and Amodei held their first one-on-one meeting, a private dinner at the White House on 27 September. The dinner followed Amodei’s absence from the state dinner for Chinese President Xi Jinping, which a source told CNBC was due to a scheduling conflict.
Relations between Anthropic and the administration have been strained:
- Export controls: In June 2026, Anthropic suspended access to two of its most advanced models, Claude Fable 5 and Claude Mythos 5, to comply with US Department of Commerce export controls. The department lifted the controls on 30 June, and the company restored access on 1 July.
- Defence policy: Media reports have described separate disputes over defence-related policy.
Investing.com, summarising the Journal’s report, said Anthropic is preparing for a stock market listing that could value it at about $2 trillion.
