Jensen Huang Says Nvidia Has Achieved AGI, Downplays Milestone

Artificial general intelligence, or AGI, is commonly understood as AI that matches or surpasses human capabilities across a broad range of domains, but the technology industry has no agreed definition or test. OpenAI’s charter describes AGI as highly autonomous systems that outperform humans at most economically valuable work. A separate, reported commercial definition between OpenAI and Microsoft ties the threshold to at least $100 billion in profits, illustrating how technical and financial benchmarks can produce different claims about when AGI has been reached.

On Nvidia’s Aug. 26, 2026, earnings call, CEO Jensen Huang said that, for many tasks, “we could say that we’ve already achieved AGI,” while offering no benchmark. He immediately called such milestones “senseless,” arguing that the useful test is whether AI performs productive work and generates profitable tokens. Huang said systems are moving beyond simple prompt responses toward autonomous agents that can learn new skills and improve recursively. His remarks echoed a March appearance on the Lex Fridman podcast, where he also said AGI had arrived but rejected the idea that 100,000 agents could build another Nvidia.