In January 2025, just three days after entering office for his second term, President Donald Trump signed an executive order that revoked many of the Biden administration’s efforts on AI. The order was part of Trump’s promise to take a light-touch approach to AI regulation, rolling back what he would later describe as his predecessor’s attempt to “paralyze” the industry through its emphasis on safety. Flanked by venture investor David Sacks, who had joined the administration as the AI and crypto czar, Trump scratched his signature on the order and then turned to his new adviser.

“That should take us to the forefront, right?” he asked Sacks at the Oval Office event. “Absolutely,” Sacks responded. “We got to win.”

Nearly halfway into Trump’s second term, though, his ambitious AI agenda has been upended. The advent of powerful frontier models like Anthropic’s Mythos has fueled fears among everyone from banking executives to intelligence agencies that they could lead to cybersecurity catastrophes. Public sentiment toward AI has shifted dramatically amid concerns about the technology’s impact on everything from jobs to electricity prices, prompting hundreds of states, cities and counties to pass data center moratoriums.