Nvidia CEO Jensen Huang calls AI innovation and safety a false choice

Nvidia CEO Jensen Huang said at Dreamforce on September 15, 2026 that choosing between AI innovation and safety is a false choice, arguing engineering solutions, market incentives, and independent evaluators should serve as guardrails instead of new government regulation. Huang called predictions of human extinction risk from AI "made up," a direct counterpoint to Anthropic CEO Dario Amodei's recent essay urging a more measured approach to frontier AI development. President Donald Trump echoed the sentiment by phone at the All-In Summit, calling AI safety concerns a "hoax" and pivoting to data center expansion.

Nvidia CEO Jensen Huang calls AI innovation and safety a false choice Huang pushed back against calls for a slowdown at Dreamforce, arguing the industry can move fast and still pause when it matters Jensen Huang has a simple answer for anyone who thinks AI companies must choose between moving fast and staying safe: they are asking the wrong question. The Nvidia https://cryptobriefing.com/markets/nvidia/ CEO took the stage at Dreamforce on September 15, 2026, and made his position clear. The framing of innovation versus safety, he argued, is a false choice. Companies can accelerate development and still pump the brakes when genuine safety concerns surface. The timing was deliberate. Anthropic CEO Dario Amodei had recently published an essay calling for a more measured approach to frontier AI development, one that would build safety infrastructure before pushing capability further. Huang’s remarks landed as a direct counterpoint. What Huang actually said Huang’s core argument rests on a few pillars. First, he believes engineering solutions and market incentives are better guardrails than new government regulation. Second, he thinks independent evaluators and testing regimes, something closer to financial audits than federal mandates, can handle accountability without slowing the industry down. On the more extreme end of the safety debate, Huang was blunt. He called predictions of human extinction risk from AI “made up,” a phrase that will not endear him to researchers who spend their careers modeling those scenarios. That framing got more political air cover shortly after Dreamforce, when Huang appeared at the All-In Summit. President Donald Trump joined by phone and echoed the sentiment, describing AI safety concerns as a “hoax” and pivoting to the economic importance of data center expansion. AI, tech, and the markets they move—in one daily briefing. Daily. Free. Join 34,000+ readers across crypto, finance, and policy. Huang’s prescription, in short: keep the innovation engine running, use existing market mechanisms to filter bad outcomes, and rely on developers to police themselves with help from independent testers rather than regulators. The commercial subtext It would be incomplete to cover Huang’s remarks without noting where Nvidia sits in this ecosystem. The company is the dominant supplier of AI accelerators, the specialized chips that power the data centers training and running large AI models. When AI development accelerates, demand for Nvidia hardware goes up. When development slows, so does that demand. Where Huang’s argument gets thinner is on the question of who decides when to pause. Huang’s framework relies on companies voluntarily recognizing safety issues and stopping. The entire debate over AI safety regulation exists partly because critics, including some inside major AI labs, believe that market incentives push in the opposite direction. Amodei’s essay was, in part, an acknowledgment of that dynamic from inside one of the leading frontier labs. A divide that will shape the regulatory landscape The Huang-Amodei disagreement is playing out against a backdrop of active regulatory discussions in Washington and in Brussels. Huang’s preference for self-regulation and independent audits rather than government mandates aligns with a familiar tech industry posture. The companies arguing for regulatory caution, including Anthropic and voices within OpenAI, are not just worried about near-term harms. They are focused on longer-horizon risks from systems that may become substantially more capable than anything currently deployed. Dismissing those concerns as “made up” forecloses a serious conversation rather than advancing one. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .