Nvidia CEO Claims AGI Achieved — But Scientists Strongly Disagree

Nvidia CEO Jensen Huang just dropped a bombshell that has the entire tech world arguing: he claims AGI achieved status is already here. Yep, you read that right. During a recent episode of the Lex Fridman Podcast, Huang stated that artificial general intelligence is here now, not in some distant future. According to Fortune, this claim has sparked intense debate among researchers and industry experts.

But before you start prepping for the robot apocalypse, there's a major catch. The definition of AGI Huang used is, well, pretty specific. According to the Fortune report, podcaster Lex Fridman asked Huang whether AI could build a technology business worth $1 billion within the next 5 to 20 years. Huang's response? "I think it's now. I think we've achieved AGI." He did add a hedge though, noting that the company didn't necessarily have to remain that valuable forever.

For more on AI breakthroughs, check out our coverage of ChatGPT apps from major brands and how AI is transforming the music industry.

Why Everyone's Arguing About What AGI Actually Means

Here's where it gets messy. Very few AI researchers actually agree with Fridman's billion-dollar-business definition of AGI. Most definitions of artificial general intelligence refer to AI that matches human cognitive abilities across a vast range of tasks, not just making money. The term has become incredibly slippery, which makes bold claims like Huang's both headline-grabbing and controversial.

Just days before Huang's podcast appearance, researchers at Google DeepMind published a serious research paper proposing a more scientific way to measure AGI. Reported by Google, the paper titled "Measuring Progress Toward AGI: A Cognitive Framework" outlines how true AGI achieved status requires matching human performance across 10 key cognitive faculties including perception, reasoning, memory, learning, attention, and social cognition.

The DeepMind researchers argue that today's AI models have what they call a "jagged" cognitive profile. These systems might crush humans at math or trivia, but they still struggle with learning from experience, maintaining long-term memories, or understanding social situations. An AI would need to hit at least median human performance across all 10 areas to truly earn the AGI label and be considered true artificial general intelligence with AGI achieved status.

What the Research Actually Shows

Other attempts to scientifically measure AGI tell a different story than Huang's bold claim. Last year, a team led by AI safety researcher Dan Hendrycks and deep learning pioneer Yoshua Bengio developed their own AGI framework. When they tested OpenAI's GPT-5, the most capable system available, it scored just 57%, falling well short of matching educated human performance across all cognitive dimensions.

Another ambitious benchmark called ARC-AGI, created by machine learning researcher François Chollet, tests whether AI can learn new skills efficiently rather than just recalling existing knowledge. The latest version, ARC-AGI-3, requires AI agents to explore new environments, acquire goals on the fly, and learn continuously over multiple steps, abilities that remain challenging for current systems.

So why would Huang claim AGI has been achieved when the research suggests otherwise? Well, Nvidia has transformed from a gaming graphics company into a $4 trillion AI infrastructure giant under his leadership. He's not just a CEO, he's the person who piloted Nvidia through near-bankruptcy to become one of the most valuable companies on Earth. In the same podcast where he declared AGI achieved, Huang also admitted that 100,000 AI agents could never replicate what he built at Nvidia, essentially acknowledging that current AI lacks the creative and strategic capabilities that define human intelligence.

The reality is that AGI achieved status remains a moving target. OpenAI CEO Sam Altman has called it a "very sloppy term" while simultaneously claiming his company knows how to build it. Microsoft's secret contract with OpenAI supposedly defines AGI as technology generating $100 billion in profits, a threshold OpenAI is nowhere near hitting despite $13 billion in annual revenue.

For Gen Z watching the AI revolution unfold, the takeaway is clear: the biggest names in tech are making bold claims about achieving human-level AI, but the actual research tells a more nuanced story. These systems are incredibly capable in specific domains while remaining surprisingly limited in others. Whether true AGI achieved status has been reached depends entirely on how you define it, and right now, nobody can agree on that definition.