Jensen Huang says AGI has already been achieved — and explains what he means
Nvidia CEO Jensen Huang said on Lex Fridman's podcast: "I think we've achieved AGI." The problem is that AGI is a fuzzy term, and every major player interprets it differently. Some drop it altogether, others rebrand it to suit themselves, and the substance stays the same. Huang's remark came out casually, but it touches a question the industry has failed to settle for years.
AI-processed from The Verge; edited by Hamidun News
Nvidia CEO Jensen Huang said a line on Lex Fridman's podcast that instantly spread across the industry: "I think we've achieved AGI." The statement sounded offhand, but behind it lies a question the industry has been unable to resolve for years. AGI, or artificial general intelligence, formally means a system capable of solving any intellectual task at or above the human level.
It is the word "any" that makes the term so slippery: it has no single measurable definition, which means anyone can consider it achieved by their own criteria. Huang is neither the first nor the last tech leader to manipulate the concept in public. In recent months, executives at the largest AI companies have increasingly distanced themselves from the term AGI: some call it overheated and marketing-driven, others introduce their own labels — ASI, "frontier AI", "human-level AI" — which essentially describe the same thing but sound more technical and less sensational.
The context of Huang's statement matters: Nvidia is the main supplier of chips for training and inference of modern language models. The company earns tens of billions of dollars precisely from the race for AGI launched by OpenAI, Google, Meta*, Anthropic and others. For him, the claim that "AGI has been achieved" is less a technical finding than a signal to the market: demand for compute will not fall, it will only grow.
At the same time, judging by the context of the interview, Huang himself put a pragmatic meaning into the words: modern models already pass tests that until recently were considered markers of human intelligence, from coding tasks to medical diagnosis. Whether that counts as AGI depends on whose dictionary you use. All of this exposes a fundamental problem: the industry has not agreed on what exactly it is building.
While the biggest players compete over who will announce a breakthrough first, the very concept of AGI is turning into a marketing tool rather than a scientific criterion. Huang's statement is a vivid symptom of this confusion, not a resolution of it.
*Meta is designated an extremist organization and banned in Russia.
- Nov 19, 2026Two pods share one A10G GPU on Amazon EKS: NVIDIA Run:ai gave each 8 GiB
- Nov 19, 2026Moonshot AI cancels $150 million GPU rental after US closes cloud loophole
- Nov 19, 2026Anthropic goes public and raises $81.3 billion, beating SpaceX's $75 billion record
- Nov 19, 2026Open DaisyRose-27B solves 12% of Terminal-Bench tasks, up from 5.6%
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.