Nvidia CEO says AI can now make real money—and that changes everything

Dubai: Jensen Huang did not soften his words when asked about artificial general intelligence, a term the tech industry has spent years debating and, more recently, avoiding. Speaking on the Lex Fridman podcast on Sunday, the Nvidia chief said plainly: “I think we’ve achieved AGI.”

What he meant, though, was not that AI now thinks like humans. His point was simpler and more immediate: AI is reaching a stage where it can create real economic value, not just generate text, images, or code. That shift—from answering questions to making money—is what makes the claim land differently now.

The distinction matters because most people experience AI today as a tool that helps with tasks. Huang is pointing to something else emerging: systems that can build products, launch services, and potentially generate revenue with far less human input than before.

To understand the shift, it helps to look at how AGI is usually defined. In simple terms, it refers to AI systems that can perform a wide range of tasks at a human level, including reasoning, adapting, and solving unfamiliar problems.

Huang’s definition moves away from that idea and replaces it with something more concrete. In the podcast, Lex Fridman framed AGI as an AI system capable of launching and running a company worth more than $1 billion, essentially asking whether AI can now create real-world economic outcomes.

Huang agreed, but with a key clarification. “You said a billion, and you didn’t say forever,” he said, suggesting that even short-lived success would meet that threshold.

That framing shifts the conversation. Instead of asking whether AI thinks like a person, it asks whether AI can produce something valuable enough that people will pay for it.

Huang pointed to emerging systems like OpenClaw, an open-source platform where networks of AI agents can work together to build applications and digital products. These agents can handle tasks such as writing code, generating content, and managing workflows, which previously required teams of people.

The idea is not that AI is independently running companies today, but that the building blocks are already in place. A group of AI agents, guided by humans, could create a product that scales quickly enough to generate meaningful revenue or even reach a billion-dollar valuation, at least briefly.

That is a different kind of milestone. It is less about intelligence in the abstract and more about whether AI can participate directly in the economy.

Huang did not claim that AI has solved every problem associated with general intelligence. He acknowledged that even large-scale AI systems would struggle to build and sustain a complex company like Nvidia, which requires long-term planning, real-world awareness, and judgment shaped by experience.

Those limitations remain central to the skepticism from researchers. Current AI systems can pass exams, write production-level code, and process large amounts of data, but they still produce errors, struggle with unfamiliar situations, and lack the kind of contextual understanding humans rely on.

The gap, then, is clear. AI may be getting better at generating value, but it is not yet capable of fully replacing human decision-making across all domains.

Huang’s claim has drawn immediate skepticism from researchers and industry experts, many of whom argue the disagreement is not about progress, but about definitions. In academic and technical circles, artificial general intelligence is still widely understood as a system that can match human-level performance across a broad range of tasks, including reasoning, adapting to unfamiliar situations, and applying common sense.

By that standard, current AI systems fall short. They can write code, pass professional exams, and generate content at scale, but they still produce factual errors, struggle with novel scenarios, and lack the kind of contextual understanding humans build through experience.

Critics say Huang’s framing shifts the goalposts by focusing on economic output instead of cognitive ability. A system that helps create a valuable product or briefly reaches a high valuation, they argue, is not the same as one that can independently reason, plan, and operate across domains over long periods.

Even Huang acknowledged some of those limitations in the same conversation, noting that AI systems would struggle to build and sustain a complex organization like Nvidia. That distinction—between generating short-term value and demonstrating broad, human-like intelligence—sits at the center of the pushback.

The shift from capability to economic output has wider consequences than the debate over definitions might suggest. The term AGI appears in contracts and strategic agreements at companies such as OpenAI and Microsoft, where certain conditions are tied to whether AGI has been achieved.

If AGI is interpreted as the ability to generate significant economic value, those thresholds could be reached sooner than expected. That would affect partnerships, competition, and access to advanced AI systems.

It also changes how businesses think about AI adoption. A tool that can generate revenue is treated differently from one that simply improves productivity.

For Nvidia, the shift reinforces its central role in the AI economy. The company’s chips power the systems used to train and run advanced AI models, which means any increase in AI-driven economic activity translates into demand for its hardware.

The numbers reflect that expectation. Nvidia shares were trading around $176 on Monday, with a slight dip of about 0.3% in early Tuesday trading, even as long-term growth projections remain strong.

At its GTC conference earlier in March, Huang projected at least $1 trillion in chip revenue from its Blackwell and Vera Rubin platforms through 2027, adding roughly $500 billion in additional order visibility since October 2025.

He has also outlined ambitions to reach $3 trillion in revenue, compared with fiscal 2026 revenue of $215.9 billion, highlighting how closely Nvidia’s future is tied to the expansion of AI-driven economic activity.

Huang’s statement does not settle the question of whether AGI has truly arrived. What it does is shift the focus of the conversation toward something more tangible: whether AI can create value that people are willing to pay for.

For most people, that distinction is easier to grasp. AI is no longer just a tool that helps you write emails or generate images. It is moving toward systems that can contribute directly to how money is made.

That shift, more than the label itself, is what could shape how companies build, how work evolves, and how the next phase of the AI economy unfolds.