According to TechCrunch, President Donald Trump said on Monday he plans to sign an executive order this week that would limit states from enacting their own AI regulations. He posted on social media that there must be “only One Rulebook” to maintain the U.S. lead in AI, arguing companies can’t get “50 Approvals” for everything. This follows a failed Senate attempt to insert a similar, deeply unpopular preemption proposal into a defense budget bill, which was rejected 99-1. The draft order would create an “AI Litigation Task Force” to challenge state laws in court and direct agencies to evaluate “onerous” state rules. It would also give White House ‘AI czar’ David Sacks direct influence over policy, superseding the usual office. States like California, with its SB 53 safety bill, and Tennessee, with its ELVIS Act against AI deepfakes, have already moved ahead with their own regulations.
The Silicon Valley push
Here’s the thing: this isn’t a new fight. Silicon Valley figures like OpenAI’s Greg Brockman and David Sacks have been arguing for this exact “one rulebook” approach for a while. Their position is simple: a patchwork of 50 different state laws would be a compliance nightmare, stifle innovation, and hand the advantage to China. And look, they have a point about complexity. But it’s also a classic play from the tech lobbying handbook. For years, the industry’s strategy has been to block meaningful federal regulation while simultaneously arguing that state laws are too fragmented. It keeps the regulatory landscape wide open. This executive order, if you read the leaked draft, is basically their wish list. It centralizes power, puts their guy (Sacks) in a key role, and sets up a legal task force to fight states in court. It’s a preemptive strike against any local rules that might cut into profits or slow down deployment.
Bipartisan backlash is real
But the reaction is where this gets fascinating. This isn’t a partisan issue—it’s a federalism issue. And it’s uniting some strange bedfellows. When the draft leaked, prominent Republicans immediately pushed back. Representative Marjorie Taylor Greene posted that “States must retain the right to regulate.” Florida Governor Ron DeSantis has been vocal, warning it would strip Florida’s ability to protect families and calling data centers resource drains. He also said letting tech companies “run wild” is an overreach. Senator Marco Rubio advised Trump to “leave AI to the states.” This mirrors the 99-1 Senate vote against a 10-year AI moratorium. That’s a staggering level of agreement in today’s Congress. A coalition of over 35 state attorneys general warned of “disastrous consequences,” and over 200 state lawmakers signed a letter in opposition. The states are digging in.
Why states aren’t backing down
So why are states so eager to regulate AI themselves? Because they’re on the front lines, and they don’t see the federal government acting. There’s a vacuum. When there’s no general federal consumer protection framework for AI, states feel they have to step in. The concerns are real and immediate: teen suicide linked to AI chatbots, the rise of what psychologists call “AI psychosis,” deepfakes destroying careers, and the massive resource consumption of data centers. Tennessee didn’t pass the ELVIS Act because it hates innovation; it did so because AI was actively harming its music industry workforce. California’s looking at safety and transparency. These are targeted responses to tangible problems. The argument that this will “destroy AI in its infancy,” as Trump’s post claims, seems like hyperbole. It seems more like it will destroy the idea of a completely unregulated wild west—which is exactly what some in tech want. The fight isn’t really about innovation versus stagnation. It’s about who gets to set the rules of the road, and whether those rules can vary by locality. This executive order, if signed, won’t be the end of it. It’ll just move the battle to the courts.
