30 September 2026
AI firms want a safety pass on antitrust law. Many lawmakers say no
AI companies have asked Washington for a narrow antitrust exemption to cooperate on safety, but senators from both parties and a former top enforcer say it is not needed.
The biggest AI companies say they want to slow down and coordinate on safety. To do that, some argue, they need permission from Washington to talk to each other. Several lawmakers who would write those rules are pushing back.
What happened
Earlier this month Anthropic CEO Dario Amodei wrote an essay suggesting the US government issue a narrow antitrust waiver so AI companies could hold certain safety conversations together. Antitrust laws stop competitors from teaming up in ways that hurt competition, such as fixing prices. Other AI executives backed his idea. On Tuesday, Amodei, Google's Sundar Pichai, Elon Musk, Meta's Mark Zuckerberg, OpenAI President Greg Brockman and Nvidia's Jensen Huang met President Trump at the White House.
Skeptics in both parties told CNBC they have doubts:
- Sen. Elizabeth Warren (D) said antitrust law already allows companies to cooperate on safety.
- Jonathan Kanter, who ran the Justice Department's antitrust division under President Biden, said firms can share threat information and set safety standards today. What the law might block, he said, is agreeing to stop competing, and he warned safety could become cover for other coordination.
- Sen. Richard Blumenthal (D) said no waiver is needed if companies work under some form of government oversight.
- Sen. Ruben Gallego (D) said he would need to see the details and would not hand the industry a blank check.
- Sen. John Hoeven (R) pointed to unhappiness with past carveouts given to internet companies as a lesson.
- Sen. Kevin Cramer (R) said he is a little suspicious of anything that keeps newcomers out of an industry.
The case for a carveout
The idea has backers too. A bill from Reps. George Whitesides (D) and Bob Latta (R) and Sens. Adam Schiff (D) and Jim Banks (R) would let AI companies share information on security threats and coordinate to delay or limit the release of high-risk models. Rep. Sam Liccardo (D), who pushed for it, said it should be drawn very narrowly, and argued the industry is so competitive that firms have little reason to collaborate beyond safety.
Some lawmakers worry about regulatory capture, where rules meant to protect the public end up serving the industry instead. Some analysts have floated less flattering motives for a slowdown, though CNBC noted there is no concrete evidence of such coordination.
Why it matters
Rules for AI could shape how fast the industry's largest players, including Nvidia, Alphabet and Meta, can release new products. Congress has little time left this year and is deeply divided, so even modest resistance can stall a bill. For now the debate splits along an unusual line: not left versus right, but supporters of a narrow safety exemption versus antitrust hawks in both parties.
Based on reporting by CNBC: AI's coming roadblock in regulation: Antitrust hawks. Facts, not advice.
More news
- Anthropic's IPO, in numbers: 825 million shares at $70 to $78, and a $1.4 trillion price tag
- Trump says the Nvidia scandal didn't endanger US security, and stands by Lutnick
- Nvidia's first earnings without Huang: what Kress will be asked on Tuesday
- Buffett gives away $5.5 billion of Berkshire stock in his Thanksgiving letter
- Banks are using AI to trim back offices. Proof that it adds revenue is harder to find
- IMF chief says the world economy leans too hard on AI as circular deals top $1 trillion