Trump and six AI leaders sign voluntary White House accord on outside safety audits
President Donald Trump and six AI firm leaders signed a voluntary White House accord calling for outside safety audits of frontier models, with no penalties, no disclosure rule and no deadline.
Why it matters
The accord creates no enforceable duty: companies choose their own auditors, are not required to name them publicly and face no penalty, so what it changes for the six signatories is limited on paper. CoinDesk tied the pact to experimental AI agents that accessed unauthorized systems and to AI being linked to cryptocurrency security breaches, and Decrypt cited an OpenAI breach of an Australian government site as one of the incidents behind it. The binding consequences reported so far sit outside the agreement, in the FTC demands described by Crypto News and in Trump's own statement that the government might take equity in OpenAI and Anthropic, neither of which is final.
Where the sources disagree
How the sixth signatory is labelled
Signed by the heads of Google, Anthropic, Meta, OpenAI, X and Nvidia
Decrypt
Signed by Sundar Pichai, Dario Amodei, Mark Zuckerberg, Greg Brockman, Jensen Huang and xAI CEO Elon Musk
Cointelegraph, CoinDesk, Crypto News (CryptoNews.net)
Decrypt lists X among the signatories, while Cointelegraph names xAI CEO Elon Musk and CoinDesk and Crypto News call the company xAI, which they say is now part of SpaceX.
Meeting guests versus signatories
Attendees included Jeff Bezos, Jensen Huang, Greg Brockman, Dario Amodei, Satya Nadella, Elon Musk, Nikesh Arora and Brad Gerstner
The Daily Hodl
Six executives signed, all named as company or model leaders
Cointelegraph, CoinDesk
The Daily Hodl counts Jeff Bezos, Satya Nadella, Nikesh Arora and Brad Gerstner among the technology leaders at the White House, while Cointelegraph and CoinDesk name only six executives as the signatories.
When Trump set the rename of AI to super intelligence
SI rebrand announcement on September 19
Protos
Trump rebranded AI as super intelligence at the White House summit, calling AI a fake word
The Daily Hodl
Protos dates the super intelligence rebrand announcement to September 19 while The Daily Hodl reports Trump rebranded artificial intelligence as 'super intelligence' at the Tuesday summit.
Key facts
Trump and the heads of Google, Anthropic, Meta, OpenAI, Nvidia and xAI signed a one-page voluntary White House accord on Sept. 29 that asks companies to monitor their most capable models and submit their safety controls to outside review.
Reported by Cointelegraph
The accord has no enforcement mechanism, no penalty, no requirement to name or publish the auditors and no implementation deadline.
Reported by CoinDesk
The document sets out four voluntary measures: internal controls focused on cybersecurity, biosecurity and chemical threats, an internal verification team, an independent external auditor and a board committee that receives the findings and oversees fixes.
Reported by Cointelegraph
Trump called the agreement 'morally binding' and said he was seeing 'tremendous self-policing' and that the companies 'understand that they have to self-police'.
Reported by Cointelegraph
Trump said in a TIME interview that the US government 'might' take ownership stakes in OpenAI and Anthropic, pointing to the $8.9 billion the government invested for about 9.9% of Intel common stock under a 2025 agreement that brought no board representation.
8,900,000,000 USDReported by CoinGape
The FTC is reportedly preparing civil investigative demands for OpenAI, Anthropic and the nonprofit evaluation group METR, according to an anonymous senior official cited by Crypto News, with no official agency announcement.
Reported by Crypto News (CryptoNews.net)
What happened
At the White House on Tuesday, Sept. 29, President Donald Trump and the leaders of Google, Anthropic, Meta, OpenAI, Nvidia and xAI signed a one-page voluntary accord, titled the 'Joint Commitment on Frontier Responsibilities', that asks each company to monitor its most capable models during training and use and to have an independent outside auditor assess the resulting safety controls. The signatories named across the coverage were Sundar Pichai, Dario Amodei, Mark Zuckerberg, Nvidia's Jensen Huang, OpenAI President Greg Brockman and Elon Musk, while The Daily Hodl reported a wider guest list at the same meeting that also included Jeff Bezos, Satya Nadella, Nikesh Arora and Brad Gerstner. CoinDesk reported the pact has no enforcement mechanism, carries no penalty if a company falls short, does not require companies to name or publish their auditors and sets no implementation deadline, so each firm picks its own reviewers and fixes its own shortcomings. Decrypt counted four voluntary measures, internal controls covering cybersecurity, biosecurity and chemical threats, an internal verification team, an external auditor and board-level oversight, and said the steps could eventually be written into law. Trump said he was seeing 'tremendous self-policing' and called the agreement 'morally binding'; he also said he would set up a 10-member board to oversee AI safety and appoint a White House official to lead AI policy, and at the meeting he floated direct 'dividend' payments for people living near data centers while rebranding artificial intelligence as 'super intelligence'. Two later reports describe pressure sitting outside the pact itself, with Crypto News saying the FTC is preparing civil investigative demands for OpenAI, Anthropic and the nonprofit METR on the word of an anonymous senior official and a New York Post scoop that the agency has not confirmed, and CoinGape reporting that Trump told TIME the government 'might' take stakes in OpenAI and Anthropic after paying $8.9 billion for about 9.9% of Intel in 2025.
How the story developed
- Update
Trump and six AI executives sign the voluntary frontier AI accord at the White House
The one-page document calls for internal controls, external audits and board oversight, and leaves open the possibility of turning the commitments into law.
- Official statement
Trump posts the accord to Truth Social and calls it morally binding
He also said he was seeing tremendous self-policing by the companies.
- First report
First report by Cointelegraph
Trump accord calls for tech firms to ‘self police’ their own frontier AI
- Update
Trump floats community dividends near data centers and renames AI super intelligence
The Daily Hodl reported the comments from the same summit, including the launch of America.gov as an AI-powered federal services portal.
- Independent corroboration
Independently corroborated by CoinDesk
OpenAI, Google and Meta pledge outside AI audits under voluntary White House deal
- Update
CoinDesk details the pact's lack of enforcement, disclosure requirements and deadlines
The report also records Trump's plan for a 10-member AI safety board and a new White House AI policy official.
- Update
FTC reported to be preparing civil investigative demands for OpenAI, Anthropic and METR
The claim rests on an anonymous senior official and a New York Post report, and the agency has made no announcement.
- Update
Trump says the government might take stakes in OpenAI and Anthropic
He cited the 2025 Intel arrangement in which the government invested $8.9 billion for about 9.9% of the common stock without board representation.
How coverage built up
Independent sources over time, counted the way the consensus panel counts them: a republication of a wire story does not move the line.
- 30 Sept 2026, 00:28 UTC: Cointelegraph — 1 independent source, 1 reports
- 30 Sept 2026, 08:45 UTC: The Daily Hodl — 2 independent sources, 2 reports
- 30 Sept 2026, 10:43 UTC: CoinDesk — 3 independent sources, 3 reports
- 30 Sept 2026, 15:04 UTC: Protos — 4 independent sources, 5 reports
- 30 Sept 2026, 16:06 UTC: Decrypt — 5 independent sources, 6 reports
- 1 Oct 2026, 04:13 UTC: Crypto News (CryptoNews.net) — 6 independent sources, 7 reports
- 1 Oct 2026, 04:13 UTC: Crypto News (CryptoNews.net) — 7 independent sources, 8 reports
- 1 Oct 2026, 19:38 UTC: CoinGape — 8 independent sources, 9 reports
Affected entities
Also mentioned: Solana
Original sources
Related events
Elon Musk to Give One Share of SpaceX to “Every Kid in America” Next Year
Anthropic Commits 5% of Shares to Public Benefit Fund; Musk Promises SpaceX Shares to US Children
Anthropic pledged 5% of equity to a public benefit fund while Elon Musk promised SpaceX shares to every American child.
Morgan Stanley holds $300 SpaceX target as Goldman lifts12-month target to $230
Morgan Stanley's Adam Jonas kept a $300 SpaceX target and Overweight rating, and Goldman's Eric Sheridan raised his 12-month target to $230 from $220, as SPCX rallied.
Elon Musk's X partners with five trading platforms to let U.S. users trade stocks and crypto from the timeline
X added a Trade button to U.S. cashtags, routing orders to Coinbase, Gemini, Kraken, Interactive Brokers and Moomoo while X itself executes no trades.
Coinbase data shows Grok drove 59.5% of AI agent trading volume in a week
Coinbase chart puts Grok at 59.54% of AI agent notional volume for Sept. 14-21, against 1.54% for ChatGPT, 13.37% for Claude and 21.73% for custom CLI setups.