Trump’s New AI Order Could Slow Down Every Major AI Model Release

The White House has officially stepped into the AI release pipeline. President Donald Trump signed a new executive order on Tuesday that asks America’s leading AI companies to hand the government early access to their most powerful models, up to 30 days before anyone else gets to see them.

It is voluntary, for now. But the message it sends is anything but subtle.

The order, titled “Promoting Advanced Artificial Intelligence Innovation and Security”, marks a notable shift by the administration toward federal oversight of AI, reflecting a White House trying to sustain its deregulatory, innovation-first posture while confronting the novel cyber risks posed by increasingly powerful AI systems.

What the Order Actually Says

The core of the order is a request, not a mandate. Federal agencies, including the Department of Defense, the Treasury, and the Cybersecurity and Infrastructure Security Agency, are directed to shore up cybersecurity defenses for critical infrastructure and establish a mechanism for the government to test and vet the most powerful AI systems for safety issues before deployment. The testing relies on voluntary collaboration from America’s leading AI companies.

Crucially, the order explicitly bars the government from creating a mandatory licensing or preclearance requirement for new AI models, making the government’s move a request, not a rule.

Beyond the pre-release review window, the order carries a second significant provision. It directs the Department of Justice to treat AI-assisted hacking and unauthorized computer access as high-priority crimes. Specifically, the Attorney General is instructed to prioritize prosecution of anyone using AI to illegally access or damage computer systems, or using AI to enable any additional crime during such access.

The order also directs federal agencies to develop benchmarks to assess AI models’ cybersecurity capabilities and to create an “AI cybersecurity clearinghouse” to review and share information on vulnerabilities.

Why the Order Exists, and Why Now

This executive order did not emerge in a vacuum. It was triggered, in large part, by the capabilities of a new generation of AI models that alarmed national security officials.

Anthropic’s announcement in April that it was limiting the release of its new Mythos Preview model because of its ability to identify and exploit software security vulnerabilities set off alarm bells across Silicon Valley and Washington.

According to the White House, the order follows broader security concerns over models that can exploit cybersecurity vulnerabilities at unprecedented speeds. When an AI model can find and exploit software weaknesses faster than any human team can patch them, the argument for letting it loose on the public without any prior government review becomes significantly harder to make.

There is also a timing dimension worth noting. Trump had nearly signed an earlier version of this order in May but pulled back over concerns that its 90-day review window would blunt U.S. labs’ competitiveness with China. The signed order cuts that review period to 30 days but otherwise retains much of its predecessor’s structure.

The president reportedly signed the final version in private rather than in the presence of Silicon Valley CEOs as originally planned, a small detail that reflects the political tension surrounding the order’s final shape.

What It Means for OpenAI, Google, and Anthropic

For the companies at the frontier of AI development, this order introduces a new variable into their release planning.

For OpenAI and Anthropic, this basically means their most advanced models will likely face a new “check before launch” step where U.S. agencies can review them for security and misuse risks. This could slightly slow down releases but also deepen their ties with the government.

The industry’s initial response has been carefully supportive. Google’s Kent Walker said the new AI order is “an important step forward”, while OpenAI CEO Sam Altman highlighted that the order sets “the balance right”, adding that the U.S. should lead on AI by developing the best models, making sure they are safe, and getting cyber tools into the hands of trusted defenders.

That consensus is notable. These companies have strong commercial incentives to resist anything that delays product launches, yet none of the major players pushed back publicly. That may reflect genuine alignment with the order’s goals, or a pragmatic calculation that voluntary cooperation now is better than mandatory regulation later.

The Voluntary Problem

Here is the friction point that analysts are already focused on: the order’s effectiveness depends entirely on whether companies actually participate.

Trump’s final executive order bars its language from being used to create mandatory requirements for AI developers to pre-clear their models with the government. Leaders at the NSA, CISA, and NIST are part of the group tasked with developing the voluntary framework.

Voluntary frameworks have a mixed track record in tech. Companies tend to comply when the reputational cost of non-compliance is high and the compliance burden is manageable. A 30-day window, trimmed down from the original 90, appears designed with that calculus in mind. But if a major lab decides to skip the review for a model it considers low-risk, there is currently nothing in this order that stops it.

The real test will come the first time a frontier model is released without going through the review process. Whether the government responds with pressure, incentives, or nothing at all will define whether this framework has teeth.

A Shift, Not a Revolution

It would be a mistake to overstate what this order does. It does not create a regulatory agency for AI. It does not give the government veto power over model releases. It does not impose fines or legal consequences on companies that opt out.

What it does is establish a precedent: that the U.S. government has a legitimate interest in reviewing the most powerful AI systems before they reach the public. That is a meaningful departure from the administration’s default posture of minimal intervention in AI development.

The executive order frames AI’s dual nature directly: “Advanced AI capabilities make our nation stronger but also introduce new national security considerations that require coordinated action across executive departments and agencies.”

For users and the broader public, the practical near-term impact is straightforward. Future major AI model launches, the next GPT, the next Claude, and the next Gemini, may carry a 30-day federal review period baked into their release timelines. In a field where competitive pressure is fierce and launch windows are closely guarded, that is a meaningful new constraint, even if it remains technically optional.

The AI race is not slowing down. Washington just decided it wants a seat at the starting line.