How Big Tech Is Pitting Washington Against California

A group of powerful tech companies are working to outmaneuver Sacramento and other state capitals on artificial intelligence laws — using Washington as their weapon.

As California and other states move to regulate AI, companies like OpenAI, Meta, Google and IBM are all urging Washington to pass national AI rules that would rein in state laws they don’t like. So is Andreessen Horowitz, a Silicon Valley-based venture capitalist firm closely tied to President Donald Trump.

Tech lobbyists are pushing the White House to oppose state AI laws, sitting down with key lawmakers to emphasize the threat and — perhaps most crucially — huddling with fellow lobbyists to see if they can unite behind a national strategy.

Their effort puts the companies in the unusual position of supporting federal regulation on the fast-moving technology, which they have so far managed to stave off almost entirely.

It is also pumping oxygen into a long-running power struggle between Washington and the tech industry’s home base of California, where progressive lawmakers have emerged as the nation’s de facto tech regulators.

The tech lobby’s AI push in Washington has infused Sacramento lawmakers with a new sense of urgency, as they watch Big Tech firms try to cash in on the influence they’ve built with Trump.

California Democratic state Sen. Scott Wiener, who used to say he’d prefer Washington to take the lead on AI, now fears that the close connection between the tech industry and Washington Republicans means they “could try to pass very weak regulations and try to preempt stronger state laws.”

“Anything that [this Congress] might produce is unlikely to protect the public, and is very likely to be designed to give the large tech companies free rein to do whatever they want,” he told POLITICO.

At issue is who gets to set rules around the growth of a powerful new technology. California lawmakers are eyeing a number of new regulations on AI, including more oversight on advanced models and rules against discrimination or copyright violations. Worried those rules could hinder innovation and hurt their bottom lines, tech companies are starting to mobilize behind the idea of a milder national law — as long as it would supersede state rules.

Wiener is himself a big reason for the AI industry’s anxiety. The California legislator spearheaded a wide-ranging AI safety bill last year that nearly became law. After a furious lobbying campaign prompted a last-minute veto of the legislation by California Gov. Gavin Newsom, Wiener went back to the table and is now drafting a more modest set of rules that he hopes can make it past the governor’s desk, including protections for whistleblowers at AI companies.

Sacramento legislators are also mulling rules meant to prevent AI-enabled discrimination, enforce tough new copyright transparency requirements and protect children from being manipulated by the technology.

Other states — including Colorado, New York, Illinois and Connecticut — have launched their own campaigns to regulate AI, prompting lobbyists to worry about a tough-to-follow “patchwork” of competing state requirements. Tech lobbyists worry California could help kick things off by passing a flood of AI laws that other states emulate.

Many in Sacramento explicitly want to play that role. California state Sen. Josh Becker, a Democrat representing Silicon Valley, said his state has “an opportunity and a responsibility to lead the country” on AI. He’s pushing multiple bills this year, including one that would require companies to limit kids’ exposure to “addictive” AI chatbots and another imposing privacy rules on high-risk deployers.

“People look to California — and a lot of times when California does it, it does de facto become a national model,” Becker said.

Groups working to place strict regulations on AI have struggled to gain momentum in the new, GOP-dominated Washington. But they’re having better luck in California, and hope to use its influence to spread their ideas across the country.

Andrew Freedman, co-founder of new AI safety nonprofit Fathom, said legislation from California state Sen. Jerry McNerney that would establish voluntary review panels to test AI programs could be taken up by other states, creating an AI policy “that can be national, even if it's not federal.”

To block California and prevent an AI patchwork, tech lobbyists are working to unite key companies and trade groups behind a single Washington strategy.

But they face a challenge: Nobody quite knows what to ask for, and there’s disagreement within the industry over how strong federal AI rules should be.

“There are the folks that would prefer just no regulation at the state level,” said one representative from a leading AI company, granted anonymity to describe sensitive talks between top tech firms. “And then there are folks that are more comfortable with some regulation.”

Another tech industry representative said they’re “still having a conversation with our members about what the best approach looks like.” An industry lobbyist said they expect the White House to ultimately “[tell] Congress that this is where we need legislative action.”

Dean Ball, now a senior policy adviser on AI and emerging technology at the White House, warned in March that the tech industry’s preemption plan could backfire. He said lobbyists should “be careful” because Congress could override state AI laws with rules that companies like even less.

“Every issue that you want to preempt on, you're opening the can of worms,” said Ball, formerly a policy researcher who spoke to POLITICO before accepting the White House job.

Ball said he expects tech billionaire Elon Musk and venture capitalist Marc Andreessen to play crucial roles in any conversation about state preemption, calling their input “the big political X factor.” As top Trump administration players with strong — and possibly countervailing — views on AI regulation, Ball said Musk and Andreessen “are perceived as having effective veto power.”

The White House is expected to unveil an “AI Action Plan” by July, which tech lobbyists believe congressional Republicans will take as marching orders.

Sen. Ted Cruz (R-Texas), chair of the Senate Commerce Committee, said conversations about preempting state AI laws have already started on Capitol Hill. “Having a patchwork system of rules for AI, I think, does pose very real challenges,” the senator told POLITICO in March. In a statement released ahead of a Thursday hearing on AI, Cruz pledged to soon unveil a “light-touch” bill that would create a “regulatory sandbox for AI.”

The whole debate takes place in the long shadow of a previous California-vs.-Washington argument about tech — one that Sacramento won last time around.

Starting in 2018, strict data privacy laws passed by California prompted a wave of other states to approve their own privacy rules. Tech lobbyists lamented the flood of state privacy laws, and urged Congress to preempt them all.

But bipartisan privacy legislation was repeatedly killed on Capitol Hill — and more than once, then-House Speaker Nancy Pelosi and other Democrats from California contributed to its demise. Pelosi and others in the California delegation refused to preempt their state’s privacy law, which remains the de facto nationwide standard. The California Privacy Protection Agency arguably serves as the most significant privacy regulator in the U.S.

The tech industry believes this time will be different — in part because Washington perceives AI as a key part of America’s burgeoning conflict with China. As the preemption conversation heats up, its lobbyists plan to play the national security card early and often.

“The geopolitical and national security backdrop for AI is just different from the privacy conversation,” said the representative from a leading AI company. “Whereas the privacy conversation was almost completely grounded in domestic politics, this one can — and really should be — grounded in our ability to maintain our technical leadership in the world.”

Despite what Ball described as “internal conflict” within the AI industry over how Washington should preempt the states, he said he believes that “OpenAI, Anthropic, DeepMind and even Meta could get in a room together and come up with something.”

He said the “lowest-hanging fruit” for the tech industry to rally behind would be a federal “liability standard” that shields AI developers from most legal liability in exchange for their adherence to technical standards. Some state laws, including McNerney’s, envision similar protections as part of a broader regulatory framework for AI.

“I would guess that that is where the conversation is going to be in six months,” Ball said.