Nvidia CEO Jensen Huang says human-level artificial intelligence has arrived, pointing to OpenAI's GPT-6 Astra as the proof. The declaration comes with no universally accepted test for AGI, and Huang's company sells the chips that power the AI boom.
A declaration with a stake
Huang's financial stake in Nvidia gives him a direct interest in keeping the AI spending spree alive. Four days before his latest statement, he told a G20 meeting that AGI would be achieved in the next couple of years and that "we're practically there today." He added that the term "either means a lot or it doesn't mean anything."
That ambiguity hasn't stopped him from making the call. GPT-6 Astra, he said, is the milestone.
The contract that once made AGI a trigger
For six years, saying "AGI has arrived" would have triggered termination of the Microsoft-OpenAI partnership under the 2019 contract. The original deal tied AGI to systems capable of generating about $100 billion in profit. An October 2025 revision required an independent expert panel to verify any declaration from OpenAI.
Then, in April, the companies rewrote the agreement. OpenAI's payments to Microsoft now continue through 2030 regardless of technological progress. The trigger is gone.
Believers and skeptics
Huang isn't alone in his timing. Sequoia partners published "2026: This is AGI" in January. Physicist Mark Gubrud, who coined the term in 1997, also says AGI has arrived.
But the bar isn't universal. Anthropic CEO Dario Amodei set a higher threshold in February, saying "we don't have a country of geniuses in a data center." Cognitive scientist Gary Marcus remains unconvinced, calling any declaration made before AGI is actually achieved a premature one.
With the contract rewritten, Huang's declaration carries no contractual weight. The question of what AGI really means — and whether GPT-6 Astra qualifies — remains open.
