Jensen Huang tells worried AI critics to just stop building it

By Splitfeed EditorialLast updated Sep 26, 2026, 19:17 UTC

Around September 23 to 25, 2026, Nvidia CEO Jensen Huang made a series of public statements on a podcast that put him in direct conflict with prominent AI safety advocates. He told developers worried about AI harm that they should simply stop building it, singled out Geoffrey Hinton by name and argued that Hinton's predictions have all been wrong and that he should stop making them, and separately stated that AI firms should not receive regulatory waivers. Hinton, widely described as the Godfather of AI for his foundational work on neural networks, has repeatedly warned about existential risks from the technology.

The remarks matter because they come from the chief executive of the company whose chips underpin the vast majority of today's AI infrastructure, giving Huang unusual standing to shape how the industry frames its own risks. His dismissal of catastrophe forecasts is not simply a philosophical disagreement; it lands in the middle of an active policy debate about whether AI developers should face mandatory safety disclosures, liability rules, or exemptions from regulation. His stance against regulatory waivers complicates any easy reading of his position, placing him at odds with parts of the industry while also dismissing the safety advocates who most loudly demand those same regulations.

Pro-AI voices read Huang's remarks largely as validation. Dare Obasanjo argued that concerns at companies like OpenAI amount to an engineering problem being dressed up as existential risk, and a New York Times framing circulating in that camp held that acceleration is itself the path to safety. One account put the probability of an AI-caused civilizational collapse by 2030 at zero. The Anti-AI camp is less interested in whether Huang is right about Hinton and more focused on who is speaking. Seamus Byrne described the interview as involving extensive hand-waving and a troubling case for what he called cognitive surrender. Dave Karpf acknowledged the argument could have been a legitimate counterweight to fearmongering from other AI executives, but concluded it reads as thinly veiled product promotion. Farrukh flagged Huang's apparent indifference to children forgetting basic mathematics as a telling tell. Middle Ground observers offer a more textured read: Ryan Moulton noted that Huang is the only major AI figure who did not arrive at the field through deep early thinking about it, while Jeremy Diamond described him as the sanest AI CEO precisely because he treats regulators as stakeholders rather than obstacles.

The clearest next test will be whether Huang's stated opposition to regulatory waivers translates into any concrete policy position or testimony before legislators, which would allow observers to judge whether his remarks represent a principled stance or a situational talking point.