It was supposed to be a big moment: May 21 was the scheduled signing of a new executive order on AI regulation. But Donald Trump hit the brakes — quite literally at the last minute.
What Happened
The planned order would have required frontier model developers to submit their models to the federal government for security review at least 90 days before release. The Office of the National Cyber Director would have coordinated the assessment.
Trump told reporters in the Oval Office: he didn’t like certain aspects and didn’t want to do anything that could jeopardize America’s lead in AI. AI is causing tremendous good, and the order could have been a blocker.
Who Was Against It — and Who Was For It
The opposition was high-profile: Elon Musk, Mark Zuckerberg, and David Sacks (Trump’s former AI advisor) had all spoken out against the order. Their argument: too much regulation slows innovation.
Interestingly, OpenAI supported the order. Sam Altman has been positioning himself as pro-regulation for months — which also happens to make strategic sense, since compliance requirements tend to hit smaller competitors harder than a company with OpenAI’s resources.
What This Means
Short term: relief for AI companies. No mandatory pre-release review, no 90-day waiting period. Existing voluntary frameworks remain in place.
Long term: the conversation isn’t over. Trump didn’t scrap the order — he postponed it. And the fact that an administration that generally avoids regulation got this far shows that even Washington is waking up to the security risks of frontier AI.
My Take
This delayed executive order captures the AI regulation dilemma in real time. Nobody wants to block the next big AI breakthrough. But even hardliners in the Trump administration seem to agree you can’t just let everything run unchecked. That the tech billionaires killed the order is a pattern we’ll see more of: whoever has the most compute also has the loudest voice in the regulation debate.
Sources: