A 10-year moratorium on state AI regulation in the US, backed by President Donald Trump and many of his tech industry donors, has been lifted. The US Senate this week rejected what some critics dubbed the “AI amnesty” last summer. But now tech lobbyists and the Trump administration are trying to smuggle the same moratorium into a national defense spending bill, despite bipartisan resistance. Trump has even threatened to sue states — and choke off their federal funding — if they dare to pass their own AI laws in response to citizens’ concerns.

The AI moratorium is opposed by an unusually broad coalition which ranges from progressive Democrats to populist and conservative Republicans. It brings together figures as varied as Steve Bannon, senators Josh Hawley and Rand Paul, and governors Gavin Newsom, Ron DeSantis and Sarah Huckabee Sanders — an alliance that finds itself aligned against the tech lobby and libertarian ideologues who oppose any government regulation on principle.

Trump, whose 2024 presidential campaign was heavily bankrolled by libertarian tech billionaires including Elon Musk, has sided with advocates such as venture capitalists Marc Andreessen and Ben Horowitz. These moguls describe themselves as “single-issue donors”, willing to shower campaign finance money on any candidate of any party who shares their radical libertarian vision of an “optimistic” technology-enabled future.

As a result, Trump is aggressively promoting a policy which splits his own party while motivating Democrats. That shared anxiety leaves little partisan cover for the President’s stance. When it comes to public opinion, Republicans and Democrats share concerns about the possible dangers of AI. According to a poll by Pew last summer, Republicans (50%) and Democrats (51%) felt “more concerned than excited” about the growing use of AI. This shows that, if anything, voters want more guardrails, not fewer, which makes the moratorium a harder sell with each passing month.

Unfortunately, the Trump administration is not working on a comprehensive system of federal AI regulation that would address the legitimate concerns of the public and many policymakers. If there were a comprehensive system of federal AI regulation addressing the concerns of state lawmakers, the “pre-emption” of state laws by a single national standard might make sense so that companies would not have to deal with a patchwork of inconsistent state regulations.

The AI amnesty is justified neither by national security nor by legitimate concerns about innovation. Its survival has nothing to do with protecting Americans and everything to do with protecting industry interests. Seen in that light, the AI moratorium — and Trump’s attacks on American states and the EU — serves to shield US tech giants from accountable regulation at home and abroad. As Hawley has observed, the persistence of this bad idea “shows what money can do”.

Growing opposition in both parties makes it unlikely that a federal law pre-empting regulations of AI by American states can pass Congress. If Trump tries to impose the policy indirectly, by suing and defunding particular state governments, his actions are likely to be opposed by many Republican state officials as well as Democrats and stalled in federal courts. The President’s courting of Silicon Valley money may yet prove to have a high price.