Jensen Huang Says Nvidia Achieved AGI—But Calls Milestone 'Senseless'

Jensen Huang Declares AGI Achievement During Nvidia Earnings Call

During Nvidia's earnings call on August 26, 2026, CEO Jensen Huang made a startling claim: the company had "achieved AGI" for many tasks1

1

. Almost immediately, the Nvidia CEO dismissed artificial general intelligence as a "senseless" milestone1

1

. This isn't the first time Huang has made such claims. In March 2026, during an appearance on the Lex Fridman podcast, he plainly stated, "I think we've achieved AGI"2

2

. However, he qualified this by noting that while AI might create a viral billion-dollar app, "the odds of 100,000 of those agents building Nvidia is zero percent"4

4

. The declaration comes as Nvidia reported $96.2 billion in fiscal Q2 revenue, up a staggering 106% year-over-year2

2

, though the company faces extreme pricing conditions in memory that could remain a bottleneck through early 20282

2

.

The AI Industry Debate Over AGI's Definition Intensifies

The definition of AGI remains frustratingly nebulous across the AI industry. OpenAI, founded with the explicit goal of building artificial general intelligence, defines it as "highly autonomous systems that outperform humans at most economically valuable work"1

1

2

2

. Yet Sam Altman himself has acknowledged this is "not a super useful term"1

1

. Complicating matters further, OpenAI reportedly worked out a different, financially-driven definition of AGI with Microsoft—systems that can generate at least $100 billion in profits1

1

. In a recent Time interview, OpenAI chief research officer Mark Chen estimated the company is "80% of the way" to AGI, while Altman claimed OpenAI would achieve it by the end of the year1

1

4

4

. Other tech giants have retreated to their own terminology: Anthropic CEO Dario Amodei calls AGI "imprecise," even a "marketing term," preferring "powerful AI"1

1

. Meta discusses "personal superintelligence," Microsoft talks about "humanist superintelligence," Amazon pursues "useful general intelligence," and Google DeepMind's Demis Hassabis speaks of reaching the "foothills of the singularity"1

1

.

What Really Matters: Profitable Tokens Over AGI Milestones

Jensen Huang redirected focus away from the AGI milestone toward more tangible metrics. "I think the most important thing that matters for the industry is that one, AI is doing productive and useful work," Huang explained during the earnings call2

2

. "Two, AI is generating profitable tokens. And three, if we had more compute, we could generate more profitable tokens, which results in more profit for all of the services"2

2

. The Nvidia CEO pointed to AI's evolution beyond simple prompt-response interactions, highlighting how autonomous AI agents can now learn new skills and improve "recursively" by running tasks repeatedly1

1

2

2

. This shift in focus might respond to growing concerns about an AI bubble amid unprecedented investment in AI data centers, even as leading players like OpenAI and Anthropic struggle to prove profitability2

2

. Critics argue that large language models will never achieve AGI because they lack persistent memory, struggle with logic, and remain prone to hallucinating incorrect information2

2

.

Real-World Risks Beyond the AGI Hype

While tech leaders debate abstract AGI milestones, today's AI already poses concrete dangers that don't require human-level intelligence. AI can produce endless fodder for scams, clone voices and faces well enough to trick loved ones out of money through voice cloning, deploy phishing messages to hack secure servers, and flood social networks with synthetic accounts3

3

. None of this qualifies as AGI under any serious definition, but the distinction offers little comfort to victims3

3

. The fixation on AGI may actually complicate regulatory efforts, as lawmakers struggle to address immediate harms while companies obsess over hypothetical future scenarios3

3

. Huang's suggestion that profitability should replace AGI as a yardstick raises its own concerns. Scams are profitable, as is workforce displacement through replacing human employees with AI models that are 30% as effective at 10% the cost3

3

. Regulators need to focus on liability, identity verification, ownership, and other legal details rather than getting distracted by AGI's perpetually delayed arrival3

3

. So long as AGI remains poorly defined and carelessly used, the whole concept remains senseless—unless companies want a handy tool for AI hype1

1

. Expect the AI industry, including Huang, to keep the AGI talk coming regardless1

1

.