Trump’s executive order on artificial intelligence promises a “single rulebook.” Startups may get into legal hot water instead.

President Donald Trump on Thursday evening signed an executive order directing federal agencies to target state AI laws, arguing that startups need relief from a “patchwork” of rules. But legal experts and startups say the order could prolong the uncertainty, sparking court battles that leave startups navigating changing state requirements while waiting to see if Congress can agree on a single national framework.

The order, titled “Ensuring a National Policy Framework for Artificial Intelligence,” directs the Justice Department to convene a task force within 30 days to challenge certain state laws on the grounds that artificial intelligence is interstate commerce and should be regulated at the federal level. It gives the Commerce Department 90 days to compile a list of states’ “onerous” AI laws, an assessment that could affect states’ eligibility for federal funds, including broadband grants.

It also asks the Federal Trade Commission and the Federal Communications Commission to explore federal standards that could preempt state rules and directs the administration to work with Congress on a uniform AI law.

The order comes amid a broader push to rein in state-by-state AI rules after efforts in Congress to temporarily halt ISIS regulation faltered. Lawmakers in both parties have argued that without a federal standard, preventing states from acting could leave consumers and businesses largely unchecked.

“This executive order led by David Sachs is a gift to the Silicon Valley minority who are using their influence in Washington to shield themselves and their companies from accountability,” Michael Kleinman, head of US policy at the Future of Life Institute, which focuses on reducing extreme risks from transformative technologies, said in a statement.

Sachs, the Trump administration’s AI and cryptocurrency policy official, has been a leading voice behind the administration’s preemptive AI push.

Even supporters of the national framework acknowledge that the system does not create a national framework. Because state laws are still enforceable unless courts block them or states temporarily suspend their enforcement, startups may face an extended transition period.

Sean Fitzpatrick, CEO of LexisNexis North America, UK and Ireland, tells TechCrunch that states will defend their consumer protection authority in court, with cases potentially escalating to the Supreme Court.

While supporters argue the order could reduce certainty by focusing the fight over AI regulation in Washington, critics say the legal battles will create immediate headwinds for startups navigating conflicting federal and state demands.

“Because startups prioritize innovation, they typically don’t have … robust organizational governance programs until they get to scale that requires one,” Hart Brown, lead author of Oklahoma Gov. Kevin Stitt’s task force on artificial intelligence and emerging technology, told TechCrunch. “These programs can be expensive and time-consuming to meet a very dynamic regulatory environment.”

Arul Nigam, co-founder of Circuit Breaker Labs, a startup that is creating red-team AI-powered chatbots related to mental health, echoed these concerns.

“There is uncertainty regarding implementation [AI companion and chatbot companies] Should it regulate itself? Nigam told TechCrunch that pointing out the state’s patchwork of AI laws hurts small startups in his field. “Are there open source standards they should adhere to? Should they keep building?”

He added that he hopes Congress can move more quickly now to pass a better federal framework.

Andrew Gamino Cheung, CTO and co-founder of AI governance firm Trustible, told TechCrunch that an EO would be counterproductive to AI innovation and pro-AI goals: “Big tech companies and big AI startups have the money to hire lawyers to help them figure out what to do, or they can simply hedge their bets. Uncertainty hurts startups the most, especially those that can’t get billions in funding almost at will,” he said.

He added that legal ambiguity makes it difficult to sell to risk-sensitive clients such as legal teams, financial firms and healthcare organizations, increasing sales cycles, system work and insurance costs. “Even the perception that AI is unregulated will reduce trust in AI,” which is already low and threatens adoption, Gamino Cheung said.

Businesses would welcome a single national standard, but “an executive order is not necessarily the appropriate vehicle to override laws duly enacted by states,” said Gary Cable, a partner at Davis + Gilbert. He warned that the current uncertainty leaves the door open to two extremes: imposing highly restrictive rules or taking no action at all, or creating a “Wild West” that favors Big Tech’s ability to absorb risks and wait things out.

Meanwhile, Morgan Reed, president of the Applications Association, urged Congress to quickly enact “a comprehensive, targeted, risk-based national framework for AI. We cannot have a patchwork of state AI laws, and a protracted court battle over the constitutionality of an executive order is never better.”