Trump Postpones AI Executive Order Signing Hours Before Ceremony

President Donald Trump postponed the signing of a highly anticipated AI oversight executive order on Thursday, just hours before a planned Oval Office ceremony that had been set to bring together the CEOs of nearly every major US AI company.

Speaking to reporters in the Oval Office, Trump said he was shelving the event "because I didn't like certain aspects of it." He added that he was worried the order could become "a blocker" on an industry where the US is currently leading.

"We're leading China, we're leading everybody, and I don't want to do anything that's going to get in the way of that lead," Trump said.

What the Order Would Have Done

The draft executive order, which had been circulating in Washington for several weeks, would have tasked the Office of the National Cyber Director and other federal agencies with developing a process to evaluate covered AI models for security risks before they are released to the public.

The framework was voluntary. Under the draft, AI developers would have been asked — not required — to share new models with the government 90 days before public release, giving federal evaluators a window to test for cybersecurity and national security concerns.

The order was billed inside the White House as a "light touch" response to safety concerns without imposing the kind of binding regulations that AI labs and industry groups have been lobbying against.

Who Was Supposed to Be There

According to multiple reports earlier in the week, the White House had been working to assemble a who's-who of US tech leadership for the signing ceremony.

The invite list included Elon Musk (xAI/Tesla), Alphabet CEO Sundar Pichai, Amazon CEO Andy Jassy, Meta CEO Mark Zuckerberg, Microsoft CEO Satya Nadella, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Apple CEO Tim Cook.

According to Axios, an unofficial second reason for the delay was logistical: not enough of those CEOs could make it to Washington on short notice. Several were reportedly already committed to Google I/O follow-up events and earnings preparation.

Why Trump Pulled It

Trump's public reasoning leaned hard on competitiveness with China. AI is "causing tremendous good," he said, and the executive order "could have been a blocker."

That framing tracks with private lobbying that AI executives have been making to the White House for months. Both Andreessen Horowitz and a16z-aligned policy groups have argued that any federal pre-release review framework, even a voluntary one, would slow US labs while Chinese labs operate without comparable friction.

A senior administration official, speaking on background to the Washington Post, said the postponement was "not a cancellation" and that the text would be revised and signed in the coming weeks.

Why this matters: The voluntary pre-release framework was the centerpiece of the Trump administration's AI security agenda. Postponing it sends a clear signal that the White House will prioritize industry concerns over safety review when the two come into tension.

Industry Reaction

AI labs have been publicly silent. None of the major frontier labs — OpenAI, Anthropic, Google, xAI, Meta — issued a statement within the first few hours of the postponement.

That silence is itself telling. Several of the same companies had been preparing internal talking points to support the order publicly, and the abrupt cancellation appears to have caught their policy teams off-guard.

On the other side, AI safety advocacy groups expressed frustration. The Center for AI Safety and several civil society organizations had supported the voluntary framework as a "floor, not a ceiling" for pre-release model review. Postponing the signing leaves the US federal government with no formal mechanism — voluntary or mandatory — for evaluating frontier models before release.

The Political Context

Trump's AI policy has zigzagged since he took office in January 2025. In his first weeks he tore down the Biden-era executive order on AI safety, including its reporting requirements for large training runs. He then directed agencies to draft a replacement framework, which has gone through multiple internal revisions.

Today's postponement is the second time a planned signing has been pulled back. An earlier version, expected in late March, was withdrawn after concerns about a clause requiring AI companies to share training data with federal evaluators.

The pattern reflects a White House caught between two factions of its own coalition: national security hawks who want more pre-release review of AI models, and AI accelerationists who view almost any federal involvement as a brake on US competitiveness.

What's Next

The administration says the order will be revised and re-introduced. There is no firm timeline, but officials suggested "weeks, not months." Expect the revised text to scale back any provisions that could be read as imposing new obligations on AI labs, and to lean harder on China-competitiveness framing.

For now, the practical effect is that US frontier AI labs continue to operate without any federal pre-release review process, voluntary or otherwise. Whatever framework eventually emerges will likely be substantially friendlier to industry than the draft Trump pulled today.

The Bottom Line

The most consequential US AI policy action of the year just got postponed at the last minute over concerns it would slow the industry down. For AI developers, that's a short-term win — but it also raises the stakes for whatever comes next. A White House that pulled this order once will be under pressure from both sides to land the next version cleanly.