SAN FRANCISCO, Jan 16, 2026, 06:41 PST
- Huang said no researchers can build a so-called “god AI” with today’s technology.
- He also pushed back on “doomer” talk and warned against rules that could “suffocate startups.”
- Nvidia shares were up about 0.9% in early U.S. trading.
Nvidia chief executive Jensen Huang said the tech world is nowhere close to creating a so-called “god AI” — shorthand for a single, all-knowing system that can fully “understand” everything from human language to biology and physics. “I don’t see any researchers having any reasonable ability to create god AI… That god AI just doesn’t exist,” Huang said. 1
The remarks land as companies rush to roll out AI features and build out the computing to run them, while regulators and workers ask the same question in different ways: what happens when the models get stronger, cheaper and more common.
Huang’s voice carries weight because Nvidia’s chips have become a backbone for much of the AI buildout. His view also runs against a louder strand of the conversation that skips from today’s chatbots to omnipotent machines, or to a labor market free-fall.
In a separate interview, Huang called the “doomer narrative” — end-of-the-world talk about AI — “extremely hurtful,” and said it has “done a lot of damage.” He also questioned the push for tighter AI rules from parts of the industry, asking: “Why are they talking to governments about these things to create regulations to suffocate startups?” 2
Huang also tried to reframe the jobs debate with a distinction between “tasks” and “purpose,” arguing AI will chew through repetitive work without wiping out most roles. “I spend most of my day typing,” he said, adding that automating that work lets him do more — “I become more busy because I’m able to do more work.” 3
The same logic applies, he said, in jobs from software engineering to restaurants: even if AI takes an order or drafts code, the core job is still solving problems and being accountable for outcomes.
Huang has used a more sweeping line to describe what he thinks happens next: “Software is eating the world, but AI is going to eat software.” In plain terms, that means the software layer itself changes — less hand-written instruction, more systems that generate, adapt and automate on the fly. 4
Nvidia shares were up about 0.9% at $188.78 in early trading on Friday.
The company is not alone in trying to steer the narrative. Rivals such as AMD are pushing their own data-center chips, while big tech groups continue to design more silicon in-house to cut costs and reduce reliance on outside suppliers.
But there is an obvious uncertainty: the gap between flashy demos and hard productivity gains, plus the risk that governments clamp down after high-profile failures, misuse, or a wave of AI-linked layoffs. Huang’s argument depends on companies using AI to expand output — not to simply shrink headcount.
For now, he is trying to drag the discussion back to earth: focus on narrow systems that work, don’t sell sci-fi, and don’t pretend “god AI” is next quarter’s product cycle.
