Altman says he was wrong about AI’s adoption speed

OpenAI CEO Sam Altman has acknowledged that he misjudged how quickly society and the economy would absorb artificial intelligence, conceding that the disruption he expected to follow GPT-4's release in 2023 has been far slower to materialize.

"I thought when we got to GPT-4, which was back in 2023, that very quickly after that there was going to be much more disruption, software businesses up for grabs right away, than it turned out to be," Altman said in a recent interview that circulated widely on Sunday.techcrunch+1

The admission represents the most direct version yet of a message Altman has been gradually sharpening throughout 2026. In February, he told reporters that AI adoption had been "slower than he thought it would be," framing the challenge as "an application problem, not a training problem." By late July, he went further, telling TechCrunch that OpenAI may need to "pace the rate of AI development" so the world can catch up.wccftech+2

A Dig at Anthropic

In the same interview, Altman laid out what he sees as AI's two failure modes — the technology becoming uncontrollably powerful, or power concentrating in too few hands — and called both outcomes "very anti-human."officechai

"It's a very misanthropic thing to say we're going to just put all of our trust in this model and let it have all the power and decision making over the world," Altman said, adding that he believes "there are some people in the world who think that's the right outcome."officechai

Though he did not name Anthropic or its co-founder Dario Amodei directly, the framing was widely interpreted as a pointed reference to Anthropic's safety-first philosophy, which emphasizes centralized oversight of powerful systems.officechai

Industry Context

Altman's recalibration arrives amid growing public resistance to the physical infrastructure of AI. A Gallup poll found that roughly 70 percent of Americans oppose data center construction in their communities. In a July podcast appearance, Altman said he understands the sentiment and suggested data centers should be built "off in the desert, around no one."theatlantic+2

The concession also comes as NVIDIA ↘ 1.84% moved to hedge its own exposure to AI demand uncertainty, earmarking $7 billion — $6 billion in technology licensing and $1 billion in equity — for the AI startup Poolside, a deal reported by The Wall Street Journal News Corp ↗ 1.88% that would bolster NVIDIA's open-weight Nemotron AI models.wccftech

Altman's slower-than-expected timeline, paired with NVIDIA's defensive positioning, suggests that even the industry's most prominent boosters are preparing for a world where AI transforms the economy more gradually than the hype cycle promised.