Mark Zuckerberg reportedly proposed and circulated the draft of the White House Accord on Super Intelligence, a voluntary AI agreement that Donald Trump and six technology executives signed on September 29, 2026. The White House posted the accord the next day.
How the draft reportedly took shape
Zuckerberg reportedly discussed AI regulation with House Speaker Mike Johnson, then talked through principles with Nvidia CEO Jensen Huang. He circulated a draft before the White House lunch where the accord was signed. Five people familiar with the meeting later corroborated the account.
What the accord asks companies to do
The accord describes four layers of oversight. Companies would monitor models during training and deployment, assign an internal team to check controls, use an independent external auditor or evaluator, and have an independent board committee receive reports. Monitoring concerns include cybersecurity, biosecurity and chemical threats.
The signatories also agreed to meet regularly to develop shared safety standards and best practices.
Who signed—and who attended
The six technology companies named as signatories are Meta, Nvidia, Anthropic, Google, OpenAI and xAI. They were represented at the signing by Mark Zuckerberg, Jensen Huang, Dario Amodei, Sundar Pichai, Greg Brockman and Elon Musk; Trump also signed the accord.
Microsoft CEO Satya Nadella and Amazon executive chairman Jeff Bezos attended the White House lunch, but Microsoft and Amazon did not sign.
What the accord says about enforcement
The voluntary accord sets no penalties or timetable, does not require companies to publish audit results, and has no clear enforcement mechanism. It does not legally compel signatories to change their behavior.
