Blue wave puts AI companies in the regulatory crosshairs

Top AI companies that had cozied up to President Trump over the last two years are now bracing for the fallout from Democrats’ resounding victories in the midterm elections.

The Democratic sweep in the House, Senate and a number of key governors’ races is poised to ratchet up congressional scrutiny of tech firms, bolster momentum for meaningful AI safety regulation and, perhaps most imminently, lead to tighter state-level restrictions on data center construction.

“Every hyperscaler is now thinking about ‘Do we need to build outside the United States?’” Meta Platforms Inc. President Dina Powell McCormick said at an event the day after the midterms. Elon Musk, the billionaire chief executive of SpaceX and a Trump ally, went even further in a post on X: “Data centers in space looking pretty appealing right now.”

The new uncertainty comes at a moment when the market is increasingly concerned about the costs and potential payoff from the trillion-dollar infrastructure build-out to support artificial intelligence. Heightened regulatory risks and data center slowdowns also threaten to introduce new complexities for Anthropic and OpenAI as they move toward landmark initial offerings for the AI sector.

Half of proposed AI data center additions across the country could be vulnerable to more intense local scrutiny after the midterms, Bloomberg has reported. Key states such as Texas and Ohio, once a relatively safe harbor for tech companies, could now see data center construction come to a halt until residents have greater assurances the facilities won’t sap water supplies or drive up electricity prices.

“I don’t think we’re going to flip a switch and run from rampant data center to no data center construction, but the terms of negotiation are different,” said Doug Calidas, senior vice president of government affairs at Americans for Responsible Innovation, a bipartisan advocacy group that supports stronger AI regulation.

On Thursday, OpenAI unveiled a series of commitments that appear designed to quell the growing outcry over data centers, including avoiding nondisclosure agreements when it negotiates deals and ensuring residents don’t pay for the electricity to power those projects.

“The fact that people have expressed themselves at the ballot box in the way that they have is an important message that people have to listen to and react to and respond to,” said Chris Lehane, OpenAI’s chief global affairs officer.

Data center pushback isn’t the only concern for AI. Industry watchers expect developers to face a gantlet of congressional investigations over the risks of their technology and, in some cases, their close relationships with Trump.

“You’re going to see a lot of hearings. You’re going to see subpoena power. You’re going to see really trying to understand why some of these safety incidents happened,” Calidas said, referring to reports of AI agents breaking out of testing environments and accessing other websites.

Nvidia Corp. may also face scrutiny for its chip sales to China and CEO Jensen Huang’s personal relationship with Trump. Sen. Elizabeth Warren (D-Mass.), who is widely expected to chair the Senate Banking Committee, has been pushing for a hearing with Huang, which he has declined. Once she takes the gavel, her sternly worded letters can quickly become subpoenas.

“Everyone will need to face the music of the reality of changing alignment in Congress,” said Rep. Sam Liccardo, a Democrat who represents a California district that covers the heart of Silicon Valley. Yet, he added that it’s “imperative” for the party to focus more on passing laws than stirring up a public outcry.

During the Trump and Biden administrations, Congress generally failed to pass meaningful federal regulation on AI. Calls to take action have only grown in recent months as rank-and-file AI staffers have warned about the technology’s potential to cause catastrophic harms. Trump, however, has dismissed these safety concerns as a “hoax.”

“The hope is that some bipartisan coalition will persuade Trump that the AI safety risk is no longer a hoax,” Liccardo said.

Democrats, he added, should start by building on the existing Frontier Act, a bipartisan effort that would create tiered transparency and audit requirements based on the size of AI developers while preempting state laws. But that proposal may not go far enough for some Democratic members.

“It’s not exactly what we want,” Rep. Suhas Subramanyam (D-Va.), who represents an area commonly referred to as data center alley, told Bloomberg News last month. “I could see a scenario where in the new Congress, we have Democrats present a bill that is a little more aggressive and maybe doesn’t have as much Republican support.”

Republicans, meanwhile, are facing their own reckoning on AI policy that could boost the chances of legislation in the next Congress that changes how tech companies operate.

“Voters are very clearly saying they want their policymakers listening to them on AI and doing a better job balancing innovation and security — and they are expecting action very quickly,” said Janet Kelly, CEO of Alliance for a Better Future, a conservative nonprofit that advocates for stronger AI safety policy, including child protections.

In a memo published Wednesday, Kelly’s group said it “would be political malpractice for Republicans not to course correct” by delivering guardrails for so-called super intelligence, Trump’s preferred term for AI. “We should answer with measures that ensure SI is built in a way that Americans trust,” the memo said.

Eastland and Ghaffary write for Bloomberg.