U.S. President Donald Trump points while signing an executive order on AI next to Senate Commerce Chairman Ted Cruz (R-Texas) and U.S. Secretary of Commerce Howard Lutnick in the Oval Office at the White House in Washington, DC, on December 11, 2025.
Al Drago | Reuters
President Donald Trump signed an executive order Tuesday requiring artificial intelligence companies to provide models to the federal government to evaluate their capabilities before full release.
The order requires companies to voluntarily participate in a benchmarking process to assess a model’s “advanced cyber capabilities” and determine whether it should be considered a “covered frontier model.” It would then seek access to those models up to 30 days before companies plan to release them more broadly, and allow governments to help select “trusted partners” to receive early access.
“Nothing in this provision shall be construed to authorize the creation of mandatory government licensing, prior approval, or permitting requirements for the development, publication, release, or distribution of new AI models, including Frontier Models,” the order states.
Trump told reporters at the time that he signed the order privately, just weeks after postponing a signing ceremony with prominent technology CEOs because he “didn’t like certain aspects.”
Tuesday’s order does not go into specific details, but it comes at a pivotal moment for U.S. AI development.
On Monday, Claude developer Anthropic secretly filed for an IPO with the Securities and Exchange Commission, and rival OpenAI also announced that it is preparing to go public this year.
SpaceX, which owns Elon Musk’s AI lab SpaceXAI, is poised to beat both companies on the public market, with a debut expected as early as next week that could value it well over $1 trillion.
The tech industry, whose wealth has soared during the AI boom, has played a central role in the White House’s position on AI.
Venture capitalist David Sachs, a longtime Musk ally, served as the first crypto and AI czar until he ended that role earlier this year. However, the sax, along with the mask, meta CEO Mark Zuckerberg reportedly called the Trump administration last month to lobby against a prior AI executive order the president was prepared to sign.
Tuesday’s order also comes after Anthropic attracted government officials and Wall Street earlier this year with its Claude Mythos Preview, a model that excels at identifying weaknesses and security flaws in software. The company limited the rollout to a select group of companies as part of a cybersecurity initiative called “Project Glasswing,” and expanded its efforts on Tuesday.
The launch of Mythos led to several high-profile meetings between Anthropic and senior members of the Trump administration, including White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent.
President Trump’s AI order provides several deadlines for developing directives and other guidance, specifically asking the Defense Department to prioritize cyber defense of information systems.
The Department of Defense has been actively trying to distance itself from Anthropic’s frontier model, labeling the startup a supply chain risk shortly before releasing Mythos. The designation means Anthropic poses a threat to U.S. national security and prohibits defense contractors from using the company’s technology in their work with the company.
Antropic sued the Trump administration to revoke that designation, and the lawsuit is still ongoing.
WATCH: President Trump signs AI executive order requiring companies to give government early access to models
