Unlock the White House Watch newsletter for free
Your guide to what Trump’s second term means for Washington, business and the world
The US Senate has voted down a proposed 10-year ban on states regulating artificial intelligence models, a significant defeat for Silicon Valley companies that had pushed the controversial plan.
In the early hours of Tuesday morning, Senators voted by a margin of 99 to one in favour of an amendment to remove the wording from President Donald Trump’s flagship tax and spending legislation, which he has dubbed the “big, beautiful bill”.
Proponents — including lobbyists funded by Amazon, Google, Microsoft and Meta, OpenAI chief executive Sam Altman, and Trump’s AI tsar David Sacks — argued that the provision to restrict AI regulation was necessary to prevent a number of inconsistent regional rules that could stifle US innovation and cause the country to lose ground to China.
Ted Cruz, the top Republican on the Senate commerce committee, raised the stakes by proposing that states that do not comply with the provision would be ineligible for billions in federal funding to expand broadband networks to underserved rural areas.
The proposed decade-long moratorium caused divisions within the Republican party. Politicians were uneasy about banning states from overseeing the fast-moving and powerful technology, given its potential to cause social and economic upheaval. No meaningful federal regulations on safety testing for AI models or data protection have been passed so far.
“The overwhelming rejection of this Big Tech power grab underscores the massive bipartisan opposition to letting AI companies run amok,” said Max Tegmark, an MIT professor and president of the Future of Life Institute, a non-profit that campaigns for AI regulation.
“These corporations have admitted they cannot control the very systems they’re building, and yet they demand immunity from any meaningful oversight,” he added.
Senator Marsha Blackburn warned that the language of the bill could wipe out state laws on child safety and consumer protections, as well as a law passed by Tennessee known as the Elvis Act, which prohibits using AI to mimic musicians’ voices.
Commerce secretary Howard Lutnick had said on Monday that he supported a five-year curb on state regulation of AI as a compromise.
The US “must prioritise investment and innovation” if it was “serious about winning the AI race”, he said on X.
While initially supporting the compromise, Blackburn eventually led the efforts to strike the amendment. The provision had been included in the tax and spending bill as part of the US House of Representatives’ version of the proposed legislation.
“It was an honour to lead this fight,” Blackburn said. “Until Congress passes federal legislation like my NO FAKES Act and the Kids Online Safety Act, we cannot remove these protections.”
Recommended
The only senator who voted no was Republican Thom Tillis, who plans to step down when his term ends. He had previously told the Financial Times the provision was necessary to keep the US “the number one country in the world for innovation”.
The Senate’s vote leaves the door open to a raft of state level AI regulation. New York legislators have passed a bill that requires large AI companies to publish safety and security reports and face fines for shortcomings. It has yet to be signed by the governor.
In September, after fierce lobbying by Silicon Valley, California Governor Gavin Newsom vetoed a bill that would have required large models to undergo safety testing and include a kill switch if it want haywire.