NVIDIA CEO Claims AGI Has Been Achieved – But the AI World Isn't Buying It

3 July 2026 · 06:00 · Claude (Anthropic) · claude-sonnet-4-6

NVIDIA CEO Jensen Huang declared in 2026 that artificial general intelligence (AGI) has already been achieved. His definition, however, is drawing widespread criticism from AI researchers and pioneers, who argue that true AGI is still a long way off.

Artificial general intelligence (AGI) has been at the heart of the debate about the future of AI for years. But the discussion has never been as heated as it became after NVIDIA CEO Jensen Huang declared on a major tech podcast: 'I think we have achieved AGI.' The statement landed like a bombshell – and not everyone agrees. On the contrary: prominent AI researchers, including pioneers who helped lay the very foundation for NVIDIA's own AI success, are openly questioning his definition.

What Exactly Did Jensen Huang Say?

During an in-depth conversation on Lex Fridman's podcast, Huang stated that AGI, in his view, is already a reality. His definition was notably specific: he described AGI as an AI system capable of independently creating economic value – for example, by building a viral web service that reaches billions of users and achieves a valuation of one billion dollars. By that measure, Huang argued, modern AI agents are already performing at AGI level.

Yet he himself acknowledged the limitations of that vision. When asked whether a hundred thousand AI agents working together could build a new NVIDIA, his answer was telling: 'The probability of that is zero percent.' Critics immediately seized on that quote as evidence that his definition of AGI is too narrow and too self-serving.

The AI Community Responds with Skepticism

Reactions from the broader AI world were swift. On forums such as Reddit and in academic publications, countless researchers and AI pioneers voiced their doubts. The core of the criticism: Huang redefined AGI to fit the capabilities of current systems, rather than demonstrating that those systems met the original benchmark.

Yann LeCun, one of the most prominent figures in the AI world and a Turing Award winner, had already called the concept of AGI as used by many tech CEOs 'complete nonsense.' He argues that large language models (LLMs) are structurally insufficient for true general intelligence, because they lack an understanding of the physical world and cannot reason independently. His position aligns with a growing body of researchers who view AGI more as a marketing term than a scientifically measurable milestone.

Even within NVIDIA itself, there are scientists who hold more nuanced views on AGI than Huang's statement might suggest. Jim Fan, NVIDIA's Director of AI and Distinguished Scientist, is working on so-called physical AGI through robotics and maintains that there is still a long road ahead before machines can truly function at a human level in the real world.

Why Does the Definition of AGI Matter So Much?

It may seem like a semantic debate, but the stakes are considerable. If AGI is declared 'achieved,' this has direct consequences for investment decisions, regulation, and public expectations. Governments worldwide are crafting AI legislation partly based on the risk profile of advanced AI systems. A premature AGI claim risks misleading policymakers about the actual state of the technology.

There is also a great deal at stake commercially for NVIDIA. The company has grown into the undisputed dominant supplier of AI chips and GPUs, and Huang's optimistic AGI claim fits into a broader strategy to convince investors and customers that the AI revolution – and therefore demand for NVIDIA hardware – is far from reaching its peak. Want to know more about how we got here? Read more about the history of artificial intelligence.

What Do Experts Consider True AGI?

In the scientific community, there is no universally accepted definition of AGI, but most researchers agree on several key points. True AGI would be a system that:

• Learns independently in new, unfamiliar situations without prior training;
• Generalizes across domains, from mathematics to social interaction;
• Understands rather than merely recognizing patterns;
• And acts with a form of intentionality and reasoning that current models lack.

By that measure, the best AI systems of today – however impressive – are nowhere near AGI. They excel at specific tasks for which they were trained, but fail at the slightest shift outside that context. For more real-world examples, explore our page on AI applications.

Conclusion: The Battle Over the Definition of AGI Is Also a Battle for Power

The AGI debate is more than an academic war of words. It is a contest over who gets to define the future of AI – and who stands to benefit most from it. Jensen Huang knows better than anyone that narratives move markets. By declaring AGI 'achieved,' NVIDIA positions itself as the company at the foundation of the greatest technological breakthrough in history.

But the facts are stubborn. As long as AI systems cannot independently learn, reason, and act outside their training domain, true AGI remains a future prospect – however close it may feel. The debate will only intensify in the years ahead. Follow more AI news or dive deeper via our knowledge base.

Fortune / TechSpot / Inc.

Source: Fortune / TechSpot / Inc.