NextFin News - The Trump administration has descended into a period of internal policy paralysis following U.S. President Trump’s abrupt decision on May 21 to cancel a landmark executive order on artificial intelligence regulation. The move, which occurred just hours before a scheduled signing ceremony, has exposed a deep rift within the White House between national security pragmatists and a faction of Silicon Valley donors who view any federal oversight as a gift to foreign competitors.
At the center of the conflict is a proposed voluntary framework that would require leading AI labs, including OpenAI, Anthropic, and Google, to grant the White House early access to new models. This "red-teaming" period, potentially lasting up to 90 days, is designed to allow federal experts to evaluate cybersecurity risks before public release. While national security officials argue such measures are essential to prevent the misuse of advanced models like OpenAI’s GPT-5.5, critics within the administration’s inner circle have successfully framed the proposal as a "regulatory tax" that would slow American innovation.
The internal battle lines are clearly drawn. White House Chief of Staff Susie Wiles is currently leading a coalition to resurrect a revised version of the order, supported by Treasury Secretary Scott Bessent and National Cyber Director Sean Cairncross. According to aides familiar with the matter, Bessent has taken an increasingly prominent role in AI policy, recently meeting with Anthropic CEO Dario Amodei to find a middle ground that satisfies both security concerns and the industry’s need for speed. Bessent’s involvement suggests the administration is beginning to view AI not just as a tech issue, but as a core component of U.S. economic and financial stability.
Opposing this group is David Sacks, the influential former AI czar and venture capitalist. Sacks, who has long advocated for a "permissionless innovation" model, reportedly convinced U.S. President Trump that the executive order was too onerous. His intervention reflects a broader sentiment among a specific wing of the tech industry that fears domestic regulation will inadvertently hand a strategic advantage to China. This perspective has resonated with U.S. President Trump, who publicly stated that the order could "stifle competition" and erode the American lead in the global AI race.
The current deadlock has left Silicon Valley in a state of high uncertainty. Executives from major AI labs have privately expressed confusion over whether any federal guidelines will emerge in 2026. The stakes are particularly high as the capabilities of new models, such as Anthropic’s Mythos, demonstrate an unprecedented ability to identify vulnerabilities in legacy software systems. Without a clear federal framework, the burden of safety testing remains entirely with the private sector, a situation that some security experts warn is unsustainable as the technology approaches "frontier" levels of autonomy.
While the Wiles-led faction hopes to salvage less controversial provisions of the order—such as those focused on infrastructure and talent recruitment—the core dispute over pre-release disclosure remains unresolved. The outcome of this internal struggle will likely determine the trajectory of U.S. tech policy for the remainder of the term. For now, the administration’s AI strategy appears to be a tug-of-war between the instinct to protect national security and the political commitment to deregulation, leaving the world’s most consequential technology in a regulatory vacuum.
Explore more exclusive insights at nextfin.ai.
