Nvidia’s Huang calls new AI laws ‘completely unnecessary’ as Meta agrees

Nvidia CEO Jensen Huang spent Tuesday arguing that AI needs no new rules at all. He told Salesforce CEO Marc Benioff onstage at Dreamforce that safety is a matter for engineers, not legislators. Hours later he said the same thing to Jim Cramer on CNBC’s “Mad Money,” calling proposed AI laws “just completely unnecessary.”

Safety is paramount in a lot of ways. It’s job one. However, safety is an engineering problem. We’re developing computing systems after all … If you build a product or a service and you’re not confident in its functionality, capability or safety, then don’t release it.

Jensen Huang, CEO, via Nvidia’s own recap of the session

That’s a direct answer to Anthropic CEO Dario Amodei, who published an essay called “We Must Pace the Frontier” on Saturday. Amodei wrote that the industry “must slow the pace at which we improve the capabilities of AI models,” and he set out three steps to do it.

Step one puts third-party evaluators inside each lab on employee-like access, and Anthropic says it’s committing to that unilaterally now. Step two asks the US government for a narrow waiver so rivals can discuss safety without tripping the Sherman Antitrust Act. Step three reaches for global coordination, which Amodei concedes is far harder.

The waiver is what Huang objects to most. “We have plenty of laws. We have plenty of regulations that govern the reliability and the functionality of products,” he told Cramer.

He also brushed off the extinction warnings that have driven the past week of coverage. “We’re not going to die in 2030,” he said.

Safety is an engineering problem, not a legal one.

Jensen Huang, speaking at Dreamforce

Meta CEO Mark Zuckerberg landed on Huang’s side the same day. He posted that “trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models,” and that labs skipping alignment work fall behind. As his evidence, he said Meta delayed shipping its Muse AI technologies over safety and security, with nobody forcing it. “We just did it as part of our day-to-day work because it was clearly the right thing for people and for us,” he wrote.

Executive	Position this week	On new rules
Jensen Huang, Nvidia	“Safety is an engineering problem, not a legal one”	New AI laws are “completely unnecessary”
Dario Amodei, Anthropic	Labs “must slow the pace” of capability gains	Wants embedded evaluators plus a narrow antitrust waiver
Mark Zuckerberg, Meta	Alignment is a competitive advantage, not a brake	Liability and market pressure are enough
Sam Altman, OpenAI	Wants industry coordination on safety timing	Commits to “independent evaluators with employee-like access”

Both men have money riding on the answer. CNBC noted that Nvidia counts Anthropic and OpenAI among its important customers, and said a slower frontier may pose a risk to Nvidia’s future growth. The same essay pulled a response from Broadcom, which kept its AI chip target, and it landed in the week Altman ruled out an OpenAI listing this year.

The market-polices-itself argument has a visible test case. As TechCrunch pointed out, Meta has just paid $18 billion to settle a lawsuit over social media harms to children. Liability did bite there, though through a settlement rather than a decision not to ship.

There’s also a long record of these promises going nowhere. The Verge traced AI executives asking for regulation back to Elon Musk in July 2017, through Sundar Pichai in 2020 and Sam Altman’s Senate testimony in 2023. Most of what followed was voluntary accords rather than binding law. California’s SB 53, a transparency statute Anthropic endorsed, is one of the few that actually passed.

So the fight has narrowed to that antitrust waiver, because it’s the one piece Amodei can’t deliver alone. Anthropic didn’t respond to CNBC’s request for comment on Huang’s remarks. Watch whether anyone in Washington offers the waiver, and whether Nvidia’s customers keep buying at the current pace while they wait.