How Silicon Valley Lost Its War Against State Ai Laws

The tech industry’s effort to stop states from regulating artificial intelligence is on the verge of collapse, as legislators from across the country lose their fear of Silicon Valley.

Interviews with more than a dozen state lawmakers, tech representatives and AI regulation advocates revealed a defiant mood sweeping through statehouses. Legislators are increasingly ignoring warnings from the tech lobby and its Washington allies about an unworkable “patchwork” of conflicting laws — and are instead preparing bills to address the safety of advanced models, new AI auditing standards, chatbot protections for kids and the impact of data centers.

The lawmakers are driven by surging public opposition to the AI industry and buoyed by the emergence of well-financed groups favoring tough regulations. Both the White House and Capitol Hill appear unable or unwilling to halt their momentum. And tech lobbyists who had planned to strong-arm or freeze out the states are now rushing to work with them before it’s too late.

“They saw the light and realized that the train was starting to move, so they better get on — or they’ll get run over,” said Mary Edly-Allen, a Democratic state senator from Illinois. She was the chief sponsor of a new law requiring top AI developers to submit to independent audits of their safety plans.

In Massachusetts, lawmakers are advancing an AI safety bill even tougher than those approved by Illinois, California and New York. Rhode Island passed a law in June requiring new guardrails for AI chatbots, part of a nationwide surge in state proposals addressing the technology’s psychological harms. California legislators are pushing through an aggressive package of bills, including rules on AI audits, chatbots and data privacy. Even deep-red states like Utah and South Carolina are spinning into action, as key legislators renew pushes for AI and kids’ safety proposals.

It’s a shift from late last year, when President Donald Trump’s threat to sue states over their AI laws sparked apprehension among legislators and governors of both parties. Leading the Future, a super PAC network flush with more than $100 million from OpenAI President Greg Brockman and venture capitalists Marc Andreessen and Ben Horowitz, sent an additional chill across statehouses as it prepared to back candidates who opposed state AI rules.

“There was this foreboding feeling at the beginning of the year that we weren't going to be able to get anywhere, because there was this mountain of money waiting to pounce,” said Adam Billen, co-executive director of the pro-regulation group Encode AI.

But when Billen attended a gathering of hundreds of state legislators in Chicago in late July, he found them eager to discuss their plans for AI laws — and dismissing potential threats from the tech sector.

“Legislators just didn't really seem to care,” Billen said. “If anything, they were kind of pissed off and emboldened.”

Lobbyists opposed to onerous AI regulations say they’ve also noticed the change, a shift they attribute to growing public pressure.

“As the polling got more negative or more people were supportive of some sort of regulation, I think that legislators, especially in an election year, they're drawn to that kind of like bugs drawn to light,” said Jeremy Kudon, executive director of the American Innovators Network, an AI lobbying group representing “little tech” and backed by venture capital giants Andreessen Horowitz and Y Combinator.

While stressing that AI rules “really should be a federal thing,” Kudon acknowledged that the argument has become untenable. His lobbyists have largely abandoned efforts to convince legislators to hold off, and are instead urging them to streamline AI laws across the states.

Kudon also suggested his lobbyists increasingly feel outgunned by an emerging army of regulation advocates — including some with ties to the top AI developer Anthropic, which favors strict regulations.

“We are the Davids now, against the Goliaths,” he said.

‘The landscape has changed dramatically’

Rebecca Bauer-Kahan, a Democratic assemblymember from California who is working on legislation to create AI auditing standards and chatbot protections for kids, said she heard rumors at the start of the year that the tech lobby would run someone against her.

“There were definitely threats,” she told POLITICO.

But a tech-backed challenger never materialized. And she instead found that lobbyists who previously ignored or fought against her were suddenly eager to work with her on AI legislation.

“The tech company involvement has been very different this year,” said Bauer-Kahan, who chalked up the shift in part to the public’s rapid swing in favor of stricter AI rules.

“They’re seeing the same polling I’m seeing,” she said.

Recent surveys commissioned by POLITICO suggest overwhelming public support for new AI regulations and surging opposition to the data centers powering the technology.

Doug Fiefia, a Republican state representative from Utah, is also feeling emboldened. Earlier this year, he abandoned his attempt at a sweeping AI and kids’ safety bill after facing pressure from industry and the White House. But in a state Senate primary in June, he handily defeated an incumbent backed by Leading the Future — and in an interview with POLITICO, Fiefia called his support for AI rules a deciding factor.

“People are worried about data privacy,” he said. “People are worried about AI and the new tools that are coming out. They're worried about their kids and how they will interact with AI. They're worried about jobs and how it will impact the workforce.”

The Utah Republican now plans to revive his AI legislation, and does not expect the White House or the tech lobby to stand in his way. “The landscape has changed dramatically, and I think we’re poised to put something across the finish line,” he said.

In Maryland, Democratic state Sen. Katie Fry Hester has worked for years to pass AI legislation, including bills that would rein in the private-sector use of high-risk AI systems and protect kids from harms caused by chatbot companions. Time and again, she said, the tech lobby thwarted those efforts.

But Hester said the tide is turning. Despite the failure of a kids’ chatbot bill earlier this year, she believes her state Senate committee will pass that legislation next year.

“I think in the past two years, peoples’ perception of the tech companies has gotten a whole hell of a lot worse,” said Hester. “And I felt like we started to see that shift with the chatbot bill.”

Other legislators cited polls and conversations with voters that illustrate a sharp turn in public sentiment against AI technologies, the data center lobby and Silicon Valley writ large.

“Things are changing so rapidly,” said Jacqui Irwin, a member of California’s Assembly and a Democratic nominee for the U.S. House.

“When you look at public perception of AI ... it went from probably 40 percent thought it was more bad than good, to now you're starting to get to the point that probably 80 or 90 percent are saying it's more bad than good,” she said. “We feel a lot of that angst from our constituents.”

The public pressure is undermining prior threats from the tech lobby to challenge lawmakers who advance objectionable AI laws.

“If you're fundamentally afoul of the sense of society or the electorate as a whole, it doesn't matter how much money you have,” said Steve Padilla, a Democratic state senator from California. “The industry understands where voters — by huge, bipartisan margins — are on these issues. And they're not going to win that argument.”

In fact, some legislators said industry efforts to punish pro-regulation lawmakers may be backfiring.

Alex Bores, a Democratic assemblymember from New York who narrowly lost a House primary race after Leading the Future spent millions against him, has emerged as an inspiration for legislators itching for a fight with the AI lobby. Similarly, Utah’s Fiefia said he believes his opponent in June’s state Senate primary was harmed by his pro-AI stance and endorsements from industry groups like Leading the Future.

In a statement, Leading the Future spokesperson Jesse Hunt said Americans “overwhelmingly support policies that promote innovation, allow the U.S. to remain the global leader in AI development, and protect families and communities.”

“We’ll continue to support lawmakers who back that mission,” Hunt added.

Washington (or governors) to the rescue?

Tech companies had hoped to get Washington’s help in thwarting state AI laws. And last year, they almost got their way.

A 10-year moratorium that would have banned states from regulating AI made it into last summer’s One Big Beautiful Bill Act, but the Senate removed the provision at the 11th hour amid opposition from state lawmakers, governors and attorneys general. That fall, congressional Republicans worked with venture capitalist and White House AI czar David Sacks on an abortive effort to shoehorn a state AI ban into the year-end defense bill.

But it was Trump’s December executive order on state AI laws that sparked the most anxiety in statehouses. Several lawmakers said the White House threat to sue and cut off funding for states that pursued AI rules temporarily chilled their colleagues’ desire to legislate — even in blue states like Vermont, where Democratic state Rep. Monique Priestley said leaders debated whether to slow down.

And yet as the months passed without Trump’s Justice Department suing anyone, states increasingly came to view the order as a paper tiger. Legislators like Utah’s Fiefia were also watching as Sacks stepped down as AI czar and the Trump administration took its first chaotic steps to crack down on AI’s safety risks.

“I’m not sure if we understand who's running the AI strategy federally,” Fiefia said.

Other legislators said the congressional push to preempt state AI laws seems to have petered out, with several citing the major backlash in June to a bipartisan House proposal that would have blocked some state rules.

“That failed. They tried that,” said Bauer-Kahan. “This is such a powerful bipartisan sentiment that I don’t see it being politically feasible.”

The vacuum created by Washington leaves governors as a “backstop” that can block onerous state AI laws, the American Innovators Network’s Kudon said.

Kudon touted his group’s success in persuading then-Republican Virginia Gov. Glenn Youngkin to veto a bill on high-risk AI systems. He pointed to last fall’s rejection of Bauer-Kahan’s chatbot bill by Democratic California Gov. Gavin Newsom, as well as Democratic Colorado Gov. Jared Polis’ push this year to replace his state’s landmark AI law. And he praised New York Gov. Kathy Hochul, another Democrat, for amending her state’s new AI safety law to make it more palatable to the tech lobby.

But governors won’t be able to protect Silicon Valley much longer, several state legislators said. Even some who had previously backed many of the AI industry’s priorities, such as Pennsylvania Democrat Josh Shapiro and Texas Republican Greg Abbott, are moving to curb new data center projects in the face of rising public outrage.

“If the temperature across the country has changed to the degree that the industry is now earnestly engaged at the table, what does that say about what the legislator and the governor are seeing?” said Padilla, the California lawmaker. “The chances that they’re going to be able to go to the governor and successfully say, ‘Rip this apart for us’ — I can’t imagine any governor thinking that’s politically wise for them.”

‘We can now pay to show up like the industry does’

Lobbyists from Kudon’s group were on the ground at July’s meeting in Chicago, working to convince legislators to harmonize their plans with existing state laws and warning that overly aggressive rules would harm AI startups.

But in a twist from the previous year’s gathering, they were met at nearly every turn by advocates promoting strict regulations.

Kudon said the surge in pro-regulation groups is driven in large part by money from Anthropic, which he warned is working to capture regulators and shield itself from competition. He accused the burgeoning campaign of spreading anxiety about the technology to state lawmakers.

“A lot of them have that fear,” said Kudon. “You didn't meet many who are like, ‘Hey, I think this is a phenomenal innovation that could fuel the economy and is an even greater development than the internet.’”

In a statement, Anthropic’s head of U.S. state and local government relations said the company has backed major new safety rules only for companies that use significant amounts of computing power and rake in more than $500 million a year.

“They don't affect startups, so it is unfortunate to see them invoked to fight third-party audits, independent evaluations, and other commonsense AI safety regulations that the majority of Americans support,” said the Anthropic executive, Cesar Fernandez.

Regulation advocates agree their influence is increasingly being felt in the states, in part due to a rapidly growing constellation of groups supporting their cause.

Riki Parikh, policy director at the Alliance for Secure AI, said regulation advocates are “fortunate to be well-financed, and so we can now pay to show up like the industry does.” But he rejected the notion that they now possess more firepower than the tech lobby.

Encode AI’s Billen worried that the industry could still persuade legislators to coalesce around relatively weak AI laws passed by other states — a plan epitomized by OpenAI’s “reverse federalism” strategy, which seeks to harmonize state AI safety rules around a single standard.

But he also said the momentum for state AI laws likely spells the end of the AI lobby’s plan to block new regulations entirely.

“Public sentiment is changing, the risks are changing, and elected officials are reacting to that,” Billen said. “In many ways, that's kind of exactly what you would hope to see in a functioning democracy.”

Tyler Katzenberger contributed to this report.