Jensen Huang rejects 2030 AI extinction risk and urges rapid development

Rejection of catastrophic risk forecasts

Nvidia chief executive Jensen Huang dismissed warnings regarding existential threats from artificial intelligence, asserting that the technology carries no risk of ending human civilization within this decade. In a televised interview excerpt with Jo Ling Kent for CBS News, Huang assigned a zero percent probability to artificial intelligence causing the destruction of the world by 2030. The executive argued that the technology industry should proceed at maximum velocity rather than holding back research and deployment. His statements directly countered arguments from researchers and rival executives who have advocated for temporary pauses to evaluate systemic hazards. Huang rejected those cautionary warnings during the preview, arguing that critics promoting catastrophic scenarios are guided by alternative motivations.

“There is 0 percent chance that the world ends in 2030.”

Shift from earlier summit remarks

Huang's CBS interview reflects a change in framing from statements he delivered in August. During an onstage telephone conversation with Donald Trump at the All-In Summit, Huang declined to adopt the term hoax to describe artificial intelligence safety concerns. At that time, Huang described apocalyptic forecasts as not grounded in science, addressing the debate through evidence rather than assigning zero probability. He also stated in August that technology companies should pause or pace their development if they felt their internal operations were becoming uncontrollable. His latest stance urges the sector to advance as quickly as possible irrespective of anyone else. The full CBS interview is scheduled for broadcast on Sunday, with initial details reported by Bloomberg.

Divided perspectives across the tech sector

Disagreements regarding artificial intelligence safety and deployment speed continue to divide corporate leaders and public officials. Former Anthropic employee Jacob Coxon previously stated that developers fear artificial intelligence could kill everyone before the end of the decade. Major research laboratories have responded by publishing safety frameworks, capability thresholds, and tripwires designed to operate under probabilistic uncertainty. In contrast, Huang shared a stage with Salesforce chief executive Marc Benioff on Tuesday at an industry conference to emphasize the positive potential of artificial intelligence. Donald Trump has also opposed slowing development, describing safety warnings as a hoax, announcing initiatives for an AI Force and an AI tsar, and arguing that regulation would risk ceding the technological lead to China.

“It makes no sense. They surely have other reasons for it, perhaps political or perhaps to get attention.”

Release standards versus external pacing

Alongside his calls for acceleration, Huang maintained that Nvidia and peer companies will continue to enforce internal readiness standards before distributing new systems. The Nvidia chief stated that commercial entities would not deliver products to customers if those systems remained incomplete or unsafe.

“We must go as fast as possible, but would never sell products that are not ready or safe.”

Safety advocates point out that relying on corporate discretion leaves readiness determinations entirely to the releasing company. The pacing frameworks proposed by safety researchers were structured specifically to introduce external evaluators and objective benchmarks rather than relying on corporate self-policing. The tension between accelerating development to outpace competitors and establishing external safety controls remains central to technology governance discussions.