#494 – Jensen Huang: NVIDIA – The $4 Trillion Company & the AI Revolution

Yeah, thanks for that question. The reason extreme co-design is necessary is because the problem no longer fits inside one computer. You want to go faster than the number of computers you add. So you have to break up the algorithm, refactor it, shard the pipeline, data, and model. When you distribute the problem, everything gets in the way due to Amdahl's law. Not only do you have to distribute computation, but also solve networking, switching, and distributed computing at scale. We have to bring every technology to bear, otherwise we scale up linearly or based on Moore's law, which has slowed.

That's why my staff is so large?

The first question is what is extreme co-design? You're optimizing across the entire stack from architectures to chips to system software to algorithms to applications. The second is why: you want to distribute the workload to exceed the benefit of increasing the number of computers. The third is how: it's the miracle of this company. When designing a computer, you have to have an operating system of computers. When designing a company, you should think about what you want it to produce. My direct staff is 60 people. I don't do one-on-ones because it's impossible. We present a problem and all of us attack it. The company is doing extreme co-design all the time.

I can reason through it systematically. We started as an accelerator company, but accelerators have a narrow application domain. The problem is market size dictates R&D capacity, which dictates influence. We had to find a way to become accelerated computing, struggling with the tension between specialization and generality. The first step was the programmable pixel shader, then FP32, which led to Cg and eventually CUDA. Putting CUDA on GeForce was a strategic decision that cost the company enormous profits, but it was necessary to build an installed base. That decision was existential: Nvidia's market cap went down to $1.5 billion, but we believed in it. GeForce took CUDA to everyone, and it became the foundation for deep learning.

It was the first strategic decision that was an existential threat. The importance of install base is everything. GeForce was successful, selling millions of GPUs a year. We decided to put CUDA on GeForce and put it into every PC, even if customers didn't use it, to cultivate an installed base. We went to universities, wrote books, taught classes. It increased the cost of the GPU so much that it consumed all the company's gross profit dollars. Our market cap dropped, but we carried CUDA on GeForce. Nvidia is the house that GeForce built.

First, I'm informed by curiosity. I reason about the future and become convinced. I believe it in my mind and manifest that future. But I also shape the belief system of everyone around me step by step. I don't make sudden announcements. I lay down bricks over time so that when I announce something, everyone is already bought in. For example, I talk about initiatives for two and a half years before announcing. Leadership sometimes looks like leading from behind. By the time I declare something, it's obvious.

We have more scaling laws now. Pre-training: we can use synthetic data. Post-training: inference is thinking, which is computationally intensive. Test time scaling: it's about reasoning, planning, search. Agentic scaling: agents spawn sub-agents, multiplying AI. The cycle feeds back. The ultimate blocker is compute, but we push extreme co-design to improve tokens per second per watt a million times over the last 10 years. Power is a concern, but we also improve energy efficiency. Supply chain is critical; I inform all CEOs of the dynamics so they can invest. For example, I convinced DRAM CEOs to invest in HBM memory, which became mainstream.

Total footprint of whatever data center you're going to build, let's say you would like to have 50 gigawatts of supercomputers running simultaneously, and it takes one week to manufacture that 50 gigawatts of supercomputers. Then each week in the supply chain, the supercomputers are going to need a gigawatt of power. So we're going to need the supply chain to increase the amount of power it has to build and test the supercomputers in the supply chain before I ship it.

Well, MVLink 72 literally builds supercomputers in the supply chain and ships them two or three tons at a time per rack. It used to be they came in parts and we used to assemble them inside the data center, but that's impossible now because MVLink72 is so dense. That's an example. I would have to go into the supply chain, meet my partners, and say, 'Guess what? Here's what we're going to do. This is the way we used to build our DGXs. We're going to build them this way. This is going to be so much better because we're going to need them for inference. The market for inference is coming, the inflection point is coming, it's going to be a big market.' So I first explain to them what's going on, why it's going to happen, and then I ask them to make several billion dollars of capital investments each. Because they trust me, I'm very respectful of them. I give them every opportunity to question me, I spend time to explain things, draw pictures, reason about it from first principles, and by the time I'm done, they know what to do.

Yeah.

No.

Because I told them what I needed, they understood what I need. They told me what they're going to do, and I believe in what they're going to do.

Well, 99% of the time we're nowhere near the worst case condition because the worst case is a few days in the winter, a few days in the summer, extreme weather. Most of the time we're nowhere near the worst case, probably running around 60% of peak. So 99% of the time our power grid has excess power sitting idle. But it has to be there just in case hospitals, infrastructure, airports need to be powered. So the question is whether we can go and help them understand and create contractual agreements, design computer architecture systems and data centers such that when they need maximum power for infrastructure in society, the data centers would get less.

But that's a very rare instance anyway. During that time, we either have backup generators for that little part, or we just have our computers shift the workload somewhere else, or run slower. We could degrade our performance, reduce power consumption, provide slightly longer latency response when somebody asks for an answer. I think that way of using computers, building data centers instead of expecting 100% uptime with very rigorous contracts, puts a lot of pressure on the grid. They'll have to increase from their maximum, but I just want to use their excess.

I think it's a throughway problem. It starts with the end customer. The end customer puts requirements on the data centers that they can never not be available. They expect perfection. To deliver that perfection, you need a combination of backup generators and grid power to deliver on perfection. So everybody has to have six nines.

Mhm.

As soon as you can specify it, you can engineer it beautifully, so long as it obeys the laws of physics. On first principles, I think we're good.

The third thing is we need the utilities to also recognize that this is an opportunity. Instead of saying it will take five years to increase grid capability, if you're willing to take power at this level of guarantee, I can make it available next month at this price. If utilities offered more segments of power delivery promises, then everybody will figure out what to do. There's way too much waste in the grid right now. We should go after it.

And instead of saying it will take me five years to increase my grid capability, if you're willing to take power at this level of guarantee, I can make it available next month at this price. So if utilities also offered more segments of power delivery promises, then I think everybody will figure out what to do. Yeah, but there's just way too much waste in the grid right now. We should go after it.

First of all, Elon is deep in so many different topics, yet he's also a really good systems thinker. He's able to think through multiple disciplines. He obviously pushes things, questions everything: Is it necessary? Does it have to be done this way? Does it have to take this long? He has the ability to question everything down to the minimal amount that's necessary. You can't take anything else out, yet the necessary capabilities of the product retain. He's as minimalist as you could imagine at system scale. I also love that he is present at the point of action. He'll just go there and if there's a problem, he'll show me the problem. When you do all of that in combination, you overcome a lot of 'this is just the way we do it.' Everybody has a lot of excuses. The last thing is when you act personally with so much urgency, it causes everybody else to act with urgency. Every supplier has many customers and projects. He makes it his business that he's the top priority of everyone else's projects, and he does that by demonstrating it.

Yeah.

Well, first of all, co-design is the ultimate systems engineering problem. We approach the work from that first principle. Another thing we do, a philosophy, a state of mind I started 30 years ago, is called the speed of light. Speed of light is shorthand for: what's the limit of what physics can do? Everything we do is compared against the speed of light: memory speed, math speed, power, cost, time, effort, number of people, manufacturing cycle time. When you think about latency versus throughput, cost versus throughput, cost versus capacity, you test against the speed of light for each constraint separately. Then when you consider them together, you have to make compromises because a system that achieves extremely low latency versus one that achieves very high throughput are architected fundamentally differently. You want to know the speed of light for a high-throughput system and for a low-latency system. Then with the total system, you can make trade-offs. I force everybody to think about first principles, the physical limits, before we do anything. We test everything against that. That's a good frame of mind. I don't love continuous improvement. You should engineer something from first principles at the speed of light, limited only by physics. After that, you improve it over time. But I don't like someone saying, 'It takes 74 days today, and we can do it in 72.' I'd rather strip it back to zero. First, explain why it's 74 days. Then think about what's possible today. If I build it from scratch, how long would it take? Often it might be six days. The rest of the gap could be well reasoned compromises and cost reductions. But at least you know what they are. Once you know six days is possible, the conversation from 74 to 6 is much more effective.

That's just one pod.

The phrase I use most often is: we need things to be as complex as necessary but as simple as possible. The question is whether all that complexity is necessary. We ought to test that and challenge it. Everything else above that is gratuitous. But this is some of the most incredible engineering in history. These systems are truly marvels of engineering. It is the most complex computer the world has ever made. The engineering teams—I don't mean to make it a competition, but if it were an Olympics of engineering teams, TSMC and ASML do incredible work, but NVIDIA is giving them a run for their money. Incredible teams, gold medalists in every sport, all assembled right here.

First, some facts: 50% of the world's AI researchers are Chinese plus or minus, mostly still in China. Their tech industry showed up at precisely the right time, during the mobile cloud era. Their way of contributing with software—this is a country with incredible science and math, well-educated kids. Their tech industry was created during the era of software, so they're very comfortable with modern software. China is not one giant economy; it has many provinces and cities with mayors all competing with each other. That's why there are so many EV companies, AI companies, every company you can imagine. They create some of them, and as a result, there's insane competition internally. What remains is an incredible company. They also have a social culture where it's family first, friends second, company third. The amount of conversation back and forth is essentially open source all the time. They contribute more to open source because they think, 'What are we protecting? My engineers' brothers are in that company, their friends are in that company, they're all schoolmates.' The schoolmate concept is like a brother for life. They share knowledge very quickly, so there's no sense keeping technology hidden. The open source community then amplifies and accelerates innovation. So you get rapid innovation from great talent, open source, and insane competition among companies. What emerges is incredible stuff. China is the fastest innovating country in the world today. Everything I've said is fundamental to how kids are raised: excellent education, parents wanting them to do well, a culture that values education. They showed up at the right time when technology is going exponential. Plus, culturally, it's pretty cool to be an engineer. It's a builder nation. Our country's leaders are mostly lawyers, while their leaders are mostly engineers who built the country out of poverty.

First, if we're going to be a great AI computing company, we have to understand how AI models are evolving. One thing I love about Nemotron 3 is it's not a pure transformer; it's transformer and SSM. We were early in developing conditional GANs and progressive GANs, which led to diffusion. Doing basic research in model architecture across different domains gives us visibility into what kind of computing systems will be needed for future models. That's part of our extreme co-design strategy. Second, we recognize that on one hand we want world-class models as proprietary products. On the other hand, we want AI to diffuse into every industry, country, researcher, and student. If everything is proprietary, it's hard to do research and innovate. So open source is fundamentally necessary for many industries to join the AI revolution. NVIDIA has the scale, skills, and motivation to build these models for as long as we live, so we ought to do that. We can open up and activate every industry, researcher, and country. A third reason is that AI is not just language. These AIs will use tools, models, and sub-agents trained on other modalities like biology, chemistry, physics, fluids, thermodynamics. Someone has to ensure that weather prediction, AI for biology, physical AI, etc., get pushed to the frontier. We don't build cars, but we want every car company to have access to great models. We don't discover drugs, but we want Lilly to have the world's best biology AI systems. So those three reasons—recognizing AI is broad, wanting to engage everyone, and co-design—drive our open source strategy.

Yeah.

The deepest misunderstanding about TSMC is that their technology is all they have—that they have a great transistor and if someone shows up with another transistor, game over. But it's more than the transistor and metalization. Their packaging, 3D packaging, silicon photonics—that technology makes the company special. But their ability to orchestrate the dynamic demands of hundreds of companies—moving up, shifting out, increasing, decreasing, pushing out, pulling in, changing wafer starts and stops—all this complexity while running a factory with high throughput, high yield, great cost, and excellent customer service. They take their promises seriously. When wafers are promised to show up, they show up so you can run your company. Their manufacturing system is miraculous. Second is their culture, simultaneously technology-focused and customer service-oriented. Many companies are one or the other, but TSMC is world-class at both. Third, they have created an intangible called trust. I trust them to put my company on top of them. That's a big deal. We have a close relationship built over many years. We've done tens of hundreds of billions of dollars of business without a contract. That's pretty great.

The story is true. I didn't dismiss it. I was deeply honored. I knew then, as I know now, TSMC is one of the most consequential companies in history. Morris is one of the highest-regarded executives and a personal friend. For him to ask, I was humbled and honored. But the work I'm doing here is really important. I had a vision of what NVIDIA could be and the impact we could have. It was my sole responsibility. So I declined. Not because it wasn't an incredible offer—it was unbelievable—but I simply couldn't take it.

Yeah.

Our single most important property is the install base of our computing platform—the install base of CUDA. Twenty years ago there was no installed base. If someone came up with another CUDA, it wouldn't make a difference because it's never been just about the technology. The technology was visionary, but the company was dedicated, stuck with it, expanded its reach. It wasn't three people that made CUDA successful; it was 43,000 people and several million developers who believed in us and trusted that we would continue to make CUDA better. They ported their mountain of software on top of it. That install base is the number one advantage. When you amplify it with our velocity of execution at this scale—no company in history built systems of this complexity, let alone once a year—that velocity combined with the installed base means a developer knows if they support CUDA, it will be 10 times better in six months. They reach a few hundred million computers, every cloud, every computer company, every industry, every country. If they create an open source package on CUDA first, they get both attributes. And they trust 100% that NVIDIA will keep CUDA maintained and optimized for as long as we live. If I were a developer today, I would target CUDA first. That is our first core advantage. Our second is the ecosystem. We vertically integrated this complex system but horizontally integrated into every company's computers. We're in Google Cloud, Amazon, Azure, ramping up AWS, new companies like CoreWeave and Nscale, supercomputers at Lilly, enterprise computers, edge radio base stations, cars, robots, satellites. One architecture covers every industry in the world.

The unit of computing used to be a GPU, then a computer, then a cluster. Now it's an entire AI factory. In the old days, when I announced a new product like 'Ampere,' I'd pick up the chip. That was my mental model. Today, picking up the chip is still adorable, but it's not my mental model. My mental model is a giant gigawatt thing with power generation, connected to the grid, with cooling systems and networking of incredible complexity. Thousands of people install it, hundreds of networking engineers, thousands behind it trying to power it up. Powering up one of those factories takes thousands of people. So when I go to bed, I'm thinking about collections of racks, pods, not individual chips. My next click is planetary-scale computing.

Yeah, we're already there. NVIDIA GPUs are the first GPUs in space. I didn't realize it, but we've been in space. It's the right place for a lot of imaging. Satellites have high-resolution imaging systems sweeping the Earth continuously. You want centimeter-scale imaging continuously for real-time telemetry. You don't want to beam petabytes of data back to Earth; you do AI at the edge, throw away everything you don't need. So AI ought to be done at the edge. Obviously, you have 24/7 solar if you put it at the poles, but there's no conduction or convection in space, just radiation. Space is big, so we'll put big radiators out there.

I'm much more practical. I look for the next bucket of opportunities first. Meanwhile, I'm cultivating space. I send engineers to work on the problem. We're learning a lot about radiation, degrading performance, continuous testing and qualification of defects, redundancy, graceful degradation. We can start engineering exploration up front. But my favorite answer is to eliminate waste. We have all that idle power on Earth. I want to use it as fast as possible.

I think NVIDIA's growth is extremely likely and in my mind inevitable. Let me explain why. We're the largest computer company in history. There are two foundational technical reasons. First, computing went from being a retrieval-based system (pre-record, pre-write, put it in a file, retrieve with a smart filter) to a generative-based system that processes and generates tokens in real time, contextually aware. We're going to need much more computation in this new world. We fundamentally changed computing. The only thing that would reverse that is if this way of computing isn't effective. But after 15 years of deep learning, I've never concluded it's a dead end. If anything, the last five years gave me more confidence than the previous ten. The second idea is that computers were storage systems, warehouses. We're now building factories. Warehouses don't make much money, but factories directly correlate with revenues. The computer's purpose changed from storage to generation of revenues. Now, the factory generates commodities that people want. The tokens are starting to segment like iPhones: free tokens, premium tokens. Intelligence is a scalable product. People will pay for high-intelligence tokens. The idea of paying $1,000 per million tokens is around the corner. So the commodity this factory makes is valuable and revenue-generating. The question is how many factories does the world need? How many tokens? What will society pay? What happens to the world's economy with such productivity improvement? I am absolutely certain the world's GDP will accelerate, and the percentage devoted to computation will be 100 times more than the past. When you back into what NVIDIA does and how much of that new economics we can address, I think we'll be a lot bigger. Is $3 trillion revenue possible? Of course yes, because there are no physical limits. Our supply chain burden is shared by 200 companies. We scale on the backs of this ecosystem. The only question is energy, and surely we'll have it. So that number is just a number. I remember when we first crossed a billion dollars. A CEO told me it's theoretically impossible for a semiconductor company to exceed a billion. Then someone said I'd never be more than $25 billion. Those aren't first-principle reasons. Think simply: what do we make and how large is the opportunity we can create? NVIDIA is not in the market share business—almost everything I just talked about doesn't exist yet. That's the hard part. If we were a $10 billion company trying to take share, it's easy to imagine growth. But it's hard to imagine how large we can be because there's nobody to take share from. That's the challenge: imagination of the future. But I have plenty of time. I'll keep reasoning and talking about it, and every GTC will make it more real. We'll get there.

I'm conscious about the fact that Nvidia's success is very important to the United States. We generate enormous amounts of tax revenues. We establish technology leadership for our nation. Technology leadership is important for national security, not just in one aspect but all aspects. When our country is more prosperous, we can do a better job with domestic policies and social benefits because we're generating so much reindustrialization in the United States. We're creating mountains of jobs, helping shift how we build things back to the US in so many different plants, chips, computers, and of course these AI factories. I'm completely aware that I have the benefit of mainstream investors, teachers, policemen who have somehow invested in Nvidia and are now millionaires. I am completely aware that Nvidia is central to a very large network of ecosystem partners. The way I deal with that is exactly what I just did: I reason about what it is we're doing, what it's causing, what impact it has on other people positively or through great burden, like the supply chain. Then I ask what I'm going to do about it. I break it down, decompose the problem, and the decomposition turns it into manageable things I can do. The only thing after that is: did you do it or get somebody else to do it? If you didn't, stop crying about it. I'm fairly tough on myself, but I also break things down so I don't panic. I go to sleep because I've made a list of things that need to be done and ensured everything that could put our company, partners, or industry in harm's way I've told someone who can do something about it. After that, Lex, what else can you do?

Oh yeah, sure, all the time. You just break down the problem into pieces, see what you can do about it. Part of it is forgetting. One of the most important attributes of AI learning is systematic forgetting. You need to know when to forget. I decompose the problem, reason about it, and share the load. I tell everybody, essentially sharing that burden as quickly as possible. Whatever worries me, tell somebody else. Decompose the problem into smaller parts and inspire people to do something about it. Part of it is just forgetting, being tough on yourself. Then you get out of bed and are attracted to the next shiny light, the next future. It's like great athletes who worry about the next point. The last point is behind them. Because I do so much of my job publicly, I say things that seem sensible or funny at the time, but you reflect and it's less funny. But you allow yourself to be pulled by the light of the future, forget the past, and keep working towards that.

Yeah. When I hear that, that's probably true about everything worth doing, right?

Yeah. I say to myself often that my first thought is how hard can it be? You get yourself into that mode. How hard could it be? Nobody's ever done it, it looks gigantic, will cost hundreds of billions of dollars, and you just go, 'Yeah, but how hard could it be?' You need to get yourself into that state of mind. Don't over simulate everything in advance. Go into a new experience thinking it's going to be perfect and fun. Then you need endurance and grit so that when setbacks happen, you can just forget about it and move on. To the extent that my assumptions about the future don't change materially, I should expect the output won't change, so I'm still going to go after it. There's a combination of human characteristics: the ability to go into an experience fresh-minded, to forget setbacks, to believe in yourself, to stay true to your belief while constantly re-evaluating. That combination is really important for resilience. I'm fortunate that life experiences gave me those characteristics. I'm always curious, always learning from everybody, humble. I'm always thinking, 'Gosh, they did that so nicely, I wonder what they're thinking.' I'm simulating and emulating almost everyone I watch.

Surprisingly, no. I would actually go the other way because I do so much of my work publicly. When I'm wrong, pretty much everybody sees it. You get humbled. Most of the things I say outside I'm fairly certain about because it's going to impact somebody else. For stuff I'm reasoning about inside a meeting, a lot of things could turn out differently. But it doesn't stop me from reasoning. I constantly reason in front of people. I show you the steps I got there so you can decide whether you believe what I said. I'm doing that all day long in meetings. I reason through it and give everybody the opportunity to intercept and disagree with a reasoning step. The nice thing is they don't have to disagree with the outcome, just the reasoning, and we can reason forward together. It's fantastic.

Yeah. Well, they knew that my first job was cleaning toilets.

Um, by the way, GeForce is still to this day our number one marketing strategy. Right. People learn about Nvidia while they're in their teenage years. Then they go to college and they know who Nvidia is. In the beginning it's playing Call of Duty or Fortnite, then they're using CUDA, then Nvidia for Blender and AutoCAD. I mentioned to a friend I'm talking with you and he said, 'Oh, they make great gaming GPUs.' Yeah, there's more to it, but people really love it. It brought a lot of joy to a lot of people. The hardware really brings these worlds to life.

Yeah, I think their perspective makes sense. I don't love AI slop myself. AI generated content increasingly looks similar. That's just not what DLSS 5 is trying to do. DLSS 5 is 3D conditioned, guided by ground truth structured data. The artist determines the geometry, and we are completely truthful to it. In every frame, it's conditioned by the textures and artistry. It enhances but doesn't change anything. DLSS 5 also lets you train your own models and even prompt it to look a different way, all consistent with the artist's intent. The impression that games will come out and we'll post-process them is wrong. DLSS is integrated with the artist. It gives them the tool of generative AI. They can decide not to use it. I think people are very sensitive to human faces and AI slop, and that's beautiful. It puts a mirror to ourselves to realize we seek imperfections. As long as it's tools that help us create worlds, it's wonderful. Gamers might also appreciate that we recently introduced skin shaders with subsurface scattering to game developers. This is just another tool they can decide to use.

Doom. Unquestionably. That was the start of 3D and turned the PC into a gaming device. Flight simulation companies were before it, but Doom had the popularity to turn the PC from an office automation tool into a personal computer for families and gamers. From a game technology perspective, I'd say Virtual Fighter. More recently, Cyberpunk 2077 is really nice for GPU accelerated graphics, fully ray traced. Personally, I'm a huge fan of Skyrim. People release mods and it becomes a different game, allowing me to replay and experience the world in a new way. We created RTX Mod, a modding tool that lets the community inject the latest technology into old games.

I think we've achieved AGI now. A claw could create a web service, an interesting little app that a few billion people use for 50 cents, and then it goes out of business. We saw many such companies during the internet era, and those websites were not more sophisticated than what open claw can generate today. Achieve virality and monetize it. I couldn't have predicted any of those companies either. You're going to get a lot of people excited with that statement. It's happening right now in China, with people teaching their claws to look for jobs and make money. I wouldn't be surprised if some social thing or digital influencer becomes an instant success. But the odds of 100,000 of those agents building Nvidia is 0%. The part that I want to make sure we all recognize is that people are worried about their jobs. The purpose of your job and the tasks and tools are related, not the same. I've been doing my job for 33 years, and the tools have changed continuously. Radiologists were supposed to go away because computer vision became superhuman, but the number of radiologists grew. The purpose of a radiologist is to diagnose disease, and because we can study scans faster, we need more radiologists. Similarly, the number of software engineers at NVIDIA will grow because their purpose is to solve problems, not just write code. Coding is now specifying, and the number of coders just went from 30 million to probably a billion. Every carpenter in the future will be a coder, also an architect. Every accountant will also be a financial analyst and adviser. Professions have been elevated. If I were a carpenter, I'd go berserk with AI. The people who are currently programmers are at the cutting edge of understanding how to communicate with agents using natural language.

Well, by the way, it's happening right now. In China, you see people teaching their claws to look for jobs and make money. I wouldn't be surprised if some digital influencer or social application becomes an instant success. But the odds of building Nvidia that way are zero. The important thing is to recognize that people are worried about their jobs. The purpose of your job and the tasks are related, not the same. I've been doing my job for 33 years, and the tools have changed continuously. Radiologists were predicted to go away because computer vision became superhuman, but the number of radiologists grew because their purpose is to diagnose disease. The same will happen with software engineers. Coding is now specifying, and the number of coders will increase to billions. The artistry of specification depends on the problem. When I give company strategies, I describe them at a level that is specific enough to be actionable, but I underspecify to enable 3,000 amazing people to make it even better. Everybody will have to learn where in the spectrum of coding they want to be. Writing a specification is coding. This artistry is the future of coding.

Exactly. When you go to Taiwan, just ask AI, 'What are Jensen's favorite restaurants in Taiwan?' and it will tell you. You're a rock star there. Maybe our paths will cross at Computex or GTC Taiwan. Do you think there are things about human nature or consciousness that are fundamentally non-computational? I believe AI will be able to recognize and understand emotions, but my chips won't feel them. The spectrum of human performance that comes from the same circumstances manifesting differently in different people—I don't think there's anything about what we're building that would suggest two computers with the same context would produce different outcomes because they felt different. The subjective experience is truly special. I was nervous talking to you. The richness of life, the fear, love, heartbreak, pain of losing loved ones—it's hard to think a computational device could do that. But there are many mysteries, and I'm open to being surprised. Scaling has created incredible miracles. Intelligence is not a mysterious word. It includes perception, understanding, reasoning, planning. Intelligence is not equal to humanity. I think intelligence is a commodity. I'm surrounded by people more intelligent than me in their fields, yet I have a role. They're more educated, but I sit in the middle orchestrating them. What is it about a dishwasher that allows them to sit in the middle of superhumans? Intelligence is a functional thing. Humanity is much bigger. Our life experience, tolerance for pain, determination are different words from intelligence. If I could give the audience one thing, it's that intelligence is not the highest form. The word we should elevate is humanity: character, compassion, generosity. Those are superhuman powers. Intelligence will be commoditized. The most important thing is not just education but the knowledge of how to use it. Don't let the democratization of intelligence cause you anxiety; be inspired.

I really don't want to die. I have a great life, a great family, really important work. This is a once-in-a-humanity experience what I'm going through. Nvidia is one of the most consequential technology companies in history, doing very important work. I take it very seriously. Of course, there are practical things like succession planning. I don't believe in succession planning. The reason isn't because I'm immortal. If you're worried about succession, break it down: the most important thing you can do today is pass on knowledge, information, insight, skills, experience as often and continuously as you can. That's why I continuously reason about everything in front of my team. Every meeting is a reasoning meeting. I pass on knowledge as fast as I can. Nothing I learn sits on my desk longer than a fraction of a second. I point it to somebody else and say, 'Get on this.' I'm constantly passing knowledge, empowering people, elevating the capability of everyone around me. I hope I die on the job, instantaneously, with no long suffering.

I've always had great confidence in the kindness, generosity, compassion, and human capacity. Sometimes I get taken advantage of, but it doesn't stop me. I start with the belief that people want to do good and help others, and I am constantly proven right, often exceeding my expectations. What gives me incredible hope is extrapolating what I see now that's possible. There are so many things we want to solve and build that are now within reach in my lifetime. You can't be not romantic about that. The fact that it's a reasonable thing to expect the end of disease, drastic reduction of pollution, traveling at the speed of light for short distances. Very soon I'll put a humanoid on a spaceship, send it out, and it will keep improving along the flight. All of my consciousness has been uploaded in the internet—my inbox, everything I've said. When the time comes, I'll send that at the speed of light to catch up with my robot. It's exciting. Understanding the biological machine is around the corner, not 10 years, probably 5. Then cracking theoretical physics and explaining consciousness—all within our reach.

Thank you, Lex. If I could say one more thing: thank you for all the interviews you do, the depth, the respect, and the research that reveals amazing people. I've enjoyed them immensely. As an innovator, creating this long-form unbelievable and captivating content is wonderful. Thank you for everything you do.