President Donald Trump canceled the signing of a major AI executive order on May 21, just hours before the ceremony. The abrupt reversal, confirmed by multiple administration officials to WIRED, has triggered a chaotic internal battle over whether to resurrect the regulation. The fight pits Chief of Staff Susie Wiles against former AI czar David Sacks, with the future of US AI oversight hanging in the balance.
The canceled order was not a minor policy tweak. At its center was a voluntary framework requiring AI labs like OpenAI, Anthropic, and Google to give the White House early access to powerful new models for cybersecurity testing. WIRED reported that the provision became the most section of the draft, sparking fierce opposition from industry leaders and Trump's own allies.
The timeline is tight. The draft suggested labs could submit models up to 90 days before public release. Several AI executives told WIRED their companies are not prepared to meet that window.
The uncertainty has left Silicon Valley guessing. "Some AI executives have privately told WIRED they are uncertain what a revised executive order might require, or whether one will end up being signed at all," the outlet reported. Trump himself explained the cancellation bluntly. He told reporters the order could stifle competition and erode the US advantage over China.
That logic resonated with David Sacks. The influential former AI czar reportedly implored Trump to kill the signing. Politico first reported the Sacks-Wiles dynamic.
Sacks made his position public. In a post on X last week, he wrote, "President Trump understands that unnecessary regulation is the biggest threat to innovation in America. Winning the AI race means not only beating China but also clearing bureaucratic hurdles thrown up by state legislatures and woke politicians in DC."
Susie Wiles is not backing down. She has assembled a counter-faction that includes Treasury Secretary Scott Bessent and National Cyber Director Sean Cairncross. Their goal is to revive the order, perhaps with its less controversial provisions intact.
The aides described the effort as a complete restart. "We're back to the drawing board, so everything is still to play for," a senior administration official told WIRED. Bessent's role is expanding. He has emerged as a key architect of the administration's AI policy, meeting with Anthropic CEO Dario Amodei and other executives in recent weeks.
The Treasury Secretary is also expected to lead negotiations on cross-border AI regulation with China. That portfolio gives him significant leverage in the internal debate. Here is what they are not telling you.
The push for regulation reflects a shift inside the White House. The administration that once eschewed AI oversight now views it as a national security imperative. The reason is specific: the capabilities of Anthropic's Mythos and OpenAI's GPT-5.5 models.
These systems excel at finding vulnerabilities in legacy software, a threat that has recast AI safety as a defense priority. The Pentagon's posture illustrates the tension. Undersecretary Emil Michael, a former top Uber executive, has shown little interest in crafting the executive order itself.
His focus is narrower and more urgent: ensuring the Defense Department gets early access to frontier models. A person familiar with the matter confirmed Michael's priority to WIRED. Commerce Secretary Howard Lutnick has similarly distanced himself from the regulatory scrum.
He already has early access to new AI models through a preexisting program called the Center for AI Standards and Innovation. Lutnick expanded the center's mandate at the start of Trump's second term, turning it into the government's main clearinghouse for testing frontier AI without formal preapproval requirements. The math does not add up.
The White House is simultaneously trying to regulate AI and avoid regulating it. The result is paralysis. "Resolving the infighting only matters if it gets Trump to yes," one administration official told WIRED, speaking anonymously to describe sensitive deliberations. The president remains the biggest hurdle.
White House spokesperson Liz Huston offered a statement that papered over the cracks. "The President's team is united in executing his bold agenda and maintaining this critical balance," Huston said. The reality, as described by aides across multiple agencies, is a factional war with no clear resolution. The stakes extend far beyond Washington.
China is not waiting for the US to sort out its internal debates. Beijing has accelerated its own AI development and regulatory framework, viewing the technology as a core pillar of strategic competition. Every month of US indecision is a month China uses to close the gap.
It reflected a growing consensus among national security officials that voluntary cooperation from AI labs is insufficient. The 90-day pre-release window was designed to give the government time to test for cyber vulnerabilities before models hit the open market. That window is now gone, and with it, a key tool for managing risk.
Industry reaction is mixed. Some AI leaders are hopeful the order can return in a stripped-down form. Others are preparing for a world with no federal AI regulation at all.
The uncertainty complicates long-term planning for companies that must decide whether to invest in compliance infrastructure or wait for clarity that may never come. Follow the leverage, not the rhetoric. The real power in this fight lies with Bessent and the Treasury Department.
By controlling the financial levers and leading China negotiations, Bessent can shape AI policy even without a formal executive order. His meetings with Amodei signal that the administration's AI strategy is being forged in Treasury's corridors, not the West Wing. The infighting also reveals a deeper ideological split within Trump's coalition.
One camp views any regulation as a betrayal of the anti-bureaucracy mandate that fueled Trump's return. The other sees unregulated AI as an existential risk that no president can ignore. Trump has not chosen a side.
He has simply canceled the meeting. Why It Matters: The collapse of the AI executive order leaves the US without a clear framework for governing technologies that can find and exploit software vulnerabilities at machine speed. For the average American, this means critical infrastructure—power grids, hospitals, financial systems—remains exposed to AI-augmented cyberattacks with no federal testing requirement in place.
The regulatory vacuum also hands China a competitive advantage, as Beijing deploys state-directed AI policy while Washington fights itself. - The canceled executive order centered on a voluntary 90-day pre-release review for frontier AI models, a provision that split the White House into warring factions. - Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent are pushing to revive the order; former AI czar David Sacks successfully lobbied Trump to kill it. - President Trump remains the decisive obstacle; one official said resolving the infighting only matters if it "gets Trump to yes." - China is accelerating its own AI regulatory and development efforts while the US remains paralyzed by internal conflict. What comes next is a race against time. The Wiles-Bessent faction must craft a revised order that can survive Sacks's opposition and win Trump's signature.
That means stripping the most provisions while preserving enough substance to matter. If they fail, US AI policy will default to the patchwork of state laws and voluntary industry commitments that Sacks prefers. The next milestone to watch is any public statement from Bessent following his meetings with AI CEOs.
A joint announcement with Anthropic or OpenAI would signal that a deal is close. Silence would mean the infighting continues, and China keeps gaining ground.
Key Takeaways
— - The canceled executive order centered on a voluntary 90-day pre-release review for frontier AI models, a provision that split the White House into warring factions.
— - Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent are pushing to revive the order; former AI czar David Sacks successfully lobbied Trump to kill it.
— - President Trump remains the decisive obstacle; one official said resolving the infighting only matters if it "gets Trump to yes."
— - China is accelerating its own AI regulatory and development efforts while the US remains paralyzed by internal conflict.
Source: WIRED
