Jensen Huang | Training Data Summary

1. Key Themes

The Computing Paradigm Shift: From Retrieval to Generation

The future of computing is fundamentally generative rather than retrieval-based. This represents a complete architectural shift in how we think about computation.

Substantiation: Jensen Huang explained: "Search is storage based computing. It's retrieval based computing. It's retrieving information for you to consume yourself. Perplexity or AI is generating. It goes and studies. It goes and reads all the content and it generates a for you... The future of computation is 100% generative. And behind it, you need an AI factory, which is the reason why I'm 100% certain we're at the beginning of this journey."

He further elaborated: "100% of what you and I just went through is generated. Every question you asked me, I didn't run back to my office and retrieve something and bring it to you... Today's computer is we just, we're just interacting. And so we are generating everything in real time based on the context."

AI as Digital Labor: The $100 Trillion Opportunity

AI will fundamentally augment and supplement the global labor market, creating a new category of "digital employees" that represents the largest addressable market in technology history.

Substantiation: Jensen stated: "This entire layer of AI model makers is also building AI factories. And these AI's are going to power the next generation of new opportunities... they're going to go after for the very first time in history, an industry that never was addressable. And that's the labor industry... These two industries represent about $100 trillion of the world's economy."

He provided concrete examples: "A robot taxi is essentially a digital chauffeur... future workforces in enterprise will be a combination of humans and digital humans. And some of them will be open AI based and some of it would be a Harvey based or open evidence or cursor."

On Nvidia's own adoption: "100% of our software engineers, 100% of our chip designers. Every single engineer today is augmented by cursor... We use cursor, originally, inside our company. And so, we now have AI's for all of our engineers."

The Physics of AI Economics: Performance Per Watt as Revenue Driver

The fundamental economic equation for AI factories is throughput per unit of energy, which directly governs customer revenues. This is the key metric that will determine winners and losers.

Substantiation: Jensen emphasized: "If your data center is 1 gigawatt, you're not going to get more than that, you're 1 gigawatt, and so if our per per watt, our energy performance per unit of energy used, is 3 times, your company can generate 3 times more revenues in that factory... Your throughput, token, it's called tokens, token generation rate per unit energy of your factory is your revenues."

He added: "In the future of AI factories, your throughput per unit energy governs the revenues of your customers. It's not just about selecting a better chip. It's about deciding what your revenues are going to be. And in fact, if you go back and look at all the CSPs, the ones that chose right, solid revenue growth."

2. Contrarian Perspectives

China Export Restrictions Hurt America More Than Help

Jensen argues that current China export policies are strategically damaging to U.S. interests, not protective of them—a view that runs counter to current government policy.

Substantiation: Jensen stated bluntly: "AI is a new technology and we have to think about before we, you know, we have to be thoughtful about about ultimately how to regulate it... However, it's important to be mindful that what is, what harms China could oftentimes also harm America and even worse."

He backed this with market facts: "At the moment, we are 100% out of China. And so China is 0% of, we went from 95% market share to 0%. So I can't imagine any policy maker thinking that that's a good idea."

His reasoning from first principles: "China has about 50% of the world's AI researchers, incredible schools, incredible focus in AI, lots of passion around AI. And I think it's a mistake to not have those researchers build AI on American technology. On first principles, I think that's a mistake... How do you balance winning, staying ahead, on the other hand, ensuring that the world builds on American tech stack."

The $1M Sequoia Investment Was Seen as Impossible (0% Odds)

The conventional wisdom in 1993 was that simultaneously inventing a new technology AND a new market was essentially impossible—yet that's exactly what Nvidia did.

Substantiation: Jensen recalled: "Sequoia capital is a big issue at the time with Nvidia's funding principles that we had to go invent the technology and the market simultaneously. The odds of that happening is approximately 0%."

He shared the legendary pitch story: "I still remember when I pitched the story and I said, in Don Valentine, at the time, you said, what's your app? Where's the killer app? I said, oh, yeah, there's this company called Electronic Arts. I didn't realize Don had just invested in Electronic Arts... He goes, you know, Jensen, I want you to know that we invested in Electronic Arts and their CTOs 14 years old and it's driven to work and you're telling me that's your killer app."

Most People Misunderstand What Makes Computing Platforms Successful

The conventional view is that better technology wins, but Jensen argues it's about the systematic invention of ecosystems, go-to-market strategies, and developer adoption pathways.

Substantiation: Jensen explained: "The KUTA's invention is part invention of the technology which is observation of how we can generalize our GPUs. But a lot of it is about the invention of new products, how to take it to market, invention of new strategies, how to get the market to adopt it and invention, inventing essentially ecosystems that ultimately creates the flywheel that makes a computing platform happen."

He provided perspective: "If you go back and you take a step back and you ask yourself, aside from ARM and SITEM, X86, what is another computing platform that exists in the world that almost everybody uses? It doesn't exist and so inventing a new computing platform rarely happens in our case at all it's almost 30 years."

The Current AI Infrastructure Build is "Extremely Small"

While most observers see hundreds of billions in AI infrastructure spending as massive, Jensen argues we're still at the very beginning.

Substantiation: Jensen stated: "We're a few hundred billion dollars, a few extremely small. We're only a few hundred billion dollars of infrastructure built. For what likely will be trillions of dollars of infrastructure built each year."

His reasoning: "AI needs to think. And so you can't pre compile it, put it into a binary, download it and use it. It's got a process all the time... It needs a machine. It needs computers to do that. And that's the reason why AI factories exist."

3. Companies Identified

Cursor

AI-powered coding tool that Nvidia uses company-wide for software development.

Quote: Jensen stated: "100% of our software engineers, 100% of our chip designers. Every single engineer today is augmented by cursor. We use cursor, originally, inside our company. And so, we now have AI's for all of our engineers. Productivity gains, the work that we do, and so much better."

When asked about undervalued technology: "We observed that general purpose technologies... tend not to be very good, extremely good at very hard problems... the existence of that says the rest of it is engineering."

OpenAI

AI research company and first customer of Nvidia's DGX-1 AI factory.

Quote: Jensen shared the origin story: "I announced this thing and I remember going, uh-huh. And literally on that GTC I was invited Elon to talk about the two of us were working on self-driving cars. And so he came on stage and he says, Jensen, what's the computer? And I said, DGX

Nvidia's Jensen Huang on AI & the Next Frontier of Growth