Huang Says Nvidia Has 'Achieved AGI' — and That the Milestone Is Senseless

On Nvidia's earnings call, Jensen Huang casually announced that the company had "achieved AGI" — the goal its biggest rivals have spent years and enormous sums chasing — then immediately dismissed the milestone as "senseless."

He is not wrong. For a supposed finish line, artificial general intelligence has no agreed definition and no test that would settle whether anyone has crossed it, which makes claiming it roughly as meaningful as declining to.

What he actually said

Asked about OpenAI's pursuit of AGI, Huang said that for Nvidia, "for many tasks, we could say that we've already achieved AGI." He offered no benchmark, and added: "I think of all of those milestones and all those, you know, they're kind of senseless at this point."

He pointed instead to AI moving past simple prompt-response into autonomous agents that learn new skills and improve themselves "recursively." What matters, he said, is that AI is "doing productive and useful work" and "generating profitable tokens" — with more compute producing more tokens.

The substitution being made

That last phrase is the tell. Huang is not resolving the AGI debate; he is replacing its metric with one Nvidia is positioned to win. Tokens require chips, and profitable tokens require more of them.

The rhetorical move is efficient. He gets to claim the milestone, deny that claiming it means anything, and redirect the conversation to the number on his own income statement — all in the space of a single answer to an analyst.