Nvidia chief executive Jensen Huang told CBS News there is a "0% chance" that AI ends the world by 2030, calling extinction talk unscientific and "irresponsible" — a rebuke aimed less at philosophers than at rival labs asking the industry to ease off the throttle.
Coverage of the mid-September interview, including reports by Bloomberg, the BBC, and The Guardian, placed Huang against a week of renewed slowdown talk. A former Anthropic researcher, Jacob Coxon, had said some developers believe AI could kill everyone by the end of the decade. Anthropic's Dario Amodei, OpenAI's Sam Altman, and Elon Musk have all argued, in different registers, for more caution or coordination. Huang's answer was speed: companies should "go as fast as we can, irrespective of anybody else," while not shipping unsafe products.
He framed safety as an engineering and liability problem, not a case for new coordination waivers or a coordinated pause. That is convenient. Nvidia is the pick-and-shovel company of the buildout; a slowdown in frontier training is a slowdown in GPU demand. Calling doomsday narratives unfounded is also a bid to keep regulators from treating compute as a controlled substance.
The split is now commercial. Labs that sell models can afford to talk about pauses because their scarce asset is weights and talent. The firm that sells the racks cannot. Huang's 0% is not a scientific forecast. It is a statement that the safety camp's timeline is a market risk, and that Nvidia will not underwrite it.
Image source: i.ibb.co