"Elon's FSD with our AI will destroy the market.."-Jensen Huang

In order to train a robot, you need a lot of data, and the first version of consumer robots that we're all going to engage is a self-driving car, and we've been working on self-driving cars now for some time, and the business is some $5 billion estimated run rate this year, and that kind of puts it into perspective, and the reason for that is because in order to build a self-driving car, you need to train a mountain of data, that video data, and NVIDIA's data center business is very important to us, and so in order to build robotics, you need three computers. You need the training computer, that's DGX. You need the simulation computer for Omniverse, and NVIDIA's also in the simulation business. Computer graphics is not just for playing video games. You could use it for simulating virtual worlds for self-driving cars, and so the simulation computer is a very big business for us, and the third one, of course, is the robotics computer. All three computers are big businesses for us, and so if you just imagine just right now where the self-driving car business is, if it's already a $5 billion business for us, imagine how big it's going to be when we have 100 million new cars per year, a trillion miles driven per year. This is likely going to be one of the largest robotics industries in the world, and one of the largest computing industries. You know, the way I think about that is we have AI supercomputers in the cloud, but people would like to have an AI supercomputer personally. It's a little bit like personal computer complementing cloud computing, and so we have, in the case of the computer industry, we started in the PC, went into the cloud. In the case of AI, we start in the cloud, we're coming into PCs, into personal computers, and so if you were a developer and you would like to have your own personal cloud, you could get one of those digits and have it all to yourself, and all of NVIDIA software runs on it, which is really a miracle. It runs on something that small, and so everything that you want to do in AI, you could do on that, or you could do it in the cloud. It's up to you, and just, you know, we want people to be able to access NVIDIA AI however you like to do it. Demand is incredible, and we're racing to get Blackwells into every data center in the world. All the cloud service providers have now stood up, Grace Blackwell systems are liquid cooled and NVLink 72, and we have some 45 factories around the world building the systems necessary to deliver to them, and so it's going to be pretty exciting. I would suspect probably early this year. I would like it to be as soon as possible, but you know, of course, Blackwell is still in allocation, and we're trying to ramp up the production, but it'll be soon. AI models. Yeah. What does that look like going forward? This is Blackwell. I mean, yeah. I mean, this is literally Blackwell. The Blackwell architecture that we have in the cloud that we use for supercomputing is right here, and so the thing that you'll be able to do, of course, you'll be able to create your own AIs if you like. We have a cloud called AI.NVIDIA.com, and we put a whole bunch of AI models up there, and these AI models are all prepackaged into what we call NIMS. They're vision models, language models, animation models, physical models, digital biology models. Whatever you would like to do as a software engineer, developer, creative artist, you could go download those models and go create your own AI, and it just runs on your PC. The fundamental trend is very clear now. When you take a step back, what it says is that everybody now realizes accelerated computing is the path forward, and general-purpose computing that we have in the past, a trillion dollars worth of data centers we've already built, those data centers are going to get modernized, and so the first takeaway is that accelerated computing is now clearly the path forward. The second thing is we do things very, very different. NVIDIA, of course, our architecture is open. It's fairly pervasive, and as a result of that, you could put NVIDIA in every cloud. You could put NVIDIA in every company. You could apply NVIDIA into self-driving cars, into robotics, because it's so versatile. So I think the fact that our architecture is literally everywhere makes it the best architecture for developers, makes it the best architecture to innovate on, and it's the best architecture to build new industries like robotics and self-driving cars and others. And so, yeah, I'm quite excited about the future for accelerated computing and AI. Well, Tesla is far ahead in self-driving cars, but every single car someday will have to have autonomous capability. It's safer. It's more convenient. It's more fun to drive. And in order to do that, it is now very well known, very well understood that learning from video directly is the most effective way to train these models. We used to train based on images that are labeled. We would say, this is a car, this is a car, this is a sign, this is a road, and we would label that manually. It's incredible. And now we just put video right into the car and let the car figure it out by itself. And this technology is very similar to the technology of large language models, but it requires just an enormous training facility. And the reason for that is because there's videos, the data rate of video, the amount of data of video is so, so high. Well, the same approach that's used for learning physics, the physical world from videos that is used for self-driving cars is essentially the same AI technology used for grounding large language models to understand the world of physics. So technologies that are like Sora, which is just incredible, and other technologies, VO from Google, incredible. The ability to generate video that makes sense, that are conditioned by human prompt, that needs to learn from video. And so the next generation of AIs need to be grounded in physical AI, needs to understand the physical world. And the best way to teach these AIs how the physical world behaves is through video, just watching tons and tons and tons of video. And so the combination of this multi-modality training capability is going to really require a lot of computing demand in the years to come. Computer graphics is going to be here forever, forever. And what we've done is we've fused artificial intelligence and computer graphics together. And the images that we're generating today wouldn't be possible if not for the fact that we're using computer graphics to create beautiful pixels and then use artificial intelligence to amplify that capability. For example, out of four frames I was talking about yesterday, 33 million pixels or so in 4K, out of that 33 million pixels, 2 million pixels were computed. The other 31 million pixels were generated by AI. It understands the physical world. So for example, if I ask it a question, if I ask it to generate multiple features of a car driving down the road, it would understand the dynamics of the world, it would understand the object permanence, it would understand geometry in space, and it would create a driving scenario for the car that is plausible. The autonomous vehicle industry and the robotics industry is really important to us. And we offer three computers for them. We offer, of course, the training computer. Through DGX. The robotics computer that's inside the car or inside a robot. And now we have this new computer called Omniverse with Cosmos that is the digital twin or the playground where these robots can learn how to be robots. And so if we could accelerate the development of artificial intelligence for AVs and for robotics, it brings in a lot. It doesn't replace, it augments. And so you should collect as much world data as you can. Of course, collecting world data is very expensive. And Elon has a great advantage because, number one, his AI factory for his cars is fantastic. Has a lot of NVIDIA gear in it. His AV algorithms is incredible. It's the best in the world. And he has a very large fleet of cars on the road that allows him to collect a lot of data. And so I think he has just a phenomenal position. And he's been working on this for a long time. And so he's going to be in a great position to take advantage of it. I don't know the attitude towards AI. I know Elon's attitude towards AI. And he's very optimistic about its future. And obviously, he's working on some of the most important AI areas. XAI is working on foundation, cognitive intelligence AI. Tesla is working on autonomous vehicles. And Optimus is for humanoid robotics. These three areas of AI are the three most important areas of AI. And so I think he's working on exactly the right things. But we're a core technology company. And so we build the foundational computing platforms. We're also full stack. And so we develop the necessary algorithms, the necessary AI technologies. And then we put it out to the industry for them to adopt it and turn it into end market solutions. We're a computing platform company. It depends on use case. Which would we see first, use case? The first use case will probably be in manufacturing. You know, there's estimates tens, if not a hundred million jobs that are workers. The world is short of workers. And aging population, declining birth rates. And so I think the world needs a lot more workers. Robotics is one of the best ways for us to supplement all of that and help companies recover the lost revenues on the one hand and drive productivity, which reduces inflation for the world on the other hand. And so I think robotics is going to be very important to that. In different areas, you could have, they'll probably deploy into manufacturing first because they obviously need it most. At the limit, artificial intelligence is the single most important technology force of our time. And it's really about, we're at the beginning of that. And in the future, every single data center will be driven by AI and the type of computing that we build today. And so if you look at the world today, we're about a year and a half into the remodeling, if you will, the modernization, the reinvention of computing. And so I think that over the next several years, you're going to see the transition from the old way of doing computing, general purpose computing, there's a new way of doing computing, artificial intelligence and accelerated computing. And so we have a lot of growth to go to. The development is going gangbusters, as you can see all the different robots that are going to be around here. And the enabling technology necessary for general robotics is coming together. All the pieces are coming together. The industry still has a lot of engineering to do. If you if you look long term, you know, pick your horizon in 10 or 20 years. The number of robots that's going to be on Earth is going to be measured in probably tens and not hundreds and not potentially billions of robots. And so those those days are clearly coming. Is it going to happen in the next couple of years or next five years? Hard to say. But the development of robotics is going to be all over the world now. And we're seeing startup companies, large companies, industrial companies, consumer electronics companies all getting involved in the future of robotics. And our offering to the industry is a three computer system. And so whether they're developing the robot, training the robot, we have systems for them and clouds for them. If they're simulating the robots, we have omniverse for them. And if they want to deploy them when they're ready to deploy the robots, we have, you know, little, little, little, little computers. That basically is the computer brain of the robot that they can put inside the robot. And so we'll work with the industry across the board from the development of the robot to the deployment of the robot. And we have computer systems for them, algorithms for them, AIs for them. And we'll partner with the industry and make this future happen. Very, very quick. Which line of business grows fastest this year? Data center, gaming or other? They're all going to grow fast. I think gaming is continuing to grow. Our our autonomous vehicle business is already on its way to be a five billion dollar business this year. And so, right. And at times, right. Run rate. And so the autonomous vehicle businesses is just starting to get off the ground. And that tells you something about how we address it. And the reason for that is we get the benefit from the beginning of the development of the AIs all the way to the deployment of the cars, because a car company needs two factories, a car factory that builds the cars and an AI factory that builds the AI. I can't imagine someone who enjoys cars and building cars and building self-driving cars and figuring out where the future of cars is going to go and has broken every rule in building cars and somehow managed to have created just an amazing company. I still remember the first car I bought from him. My only question for Elon at the time was, are you going to be around to service this car if something happened to it? And he says, no, trust me. And look what happened. What an amazing achievement. I have all three versions of his cars. It just gets better and better. Not only does it get better and better, each version gets better and better. It just delights me to no end when I get an OTA in the morning, you know, and I read it and oh, wow, all these features. I didn't even pay for it. Our TCO is so good that even when the competitors chips are free, it's not cheap enough. In the last 10 years, we reduced the cost of computing by one million times. The cost of deep learning by one million times. If we could reduce the marginal cost of computing down to approximately zero, we might use it to do something insanely amazing. Large language models to literally extract all of digital human knowledge from the Internet and put it into a computer and let it go figure out what the wisdom, what the knowledge is, that idea of scraping the entire Internet and putting it in one computer and let the computer figure out what the program is, is an insane concept. But you wouldn't ever consider doing it unless the marginal cost of computing was zero. We made that breakthrough. The thing that I can't wait to do is turn the car into an A.I. itself. You know, I want to be able to talk to my car. And the moment I walk into the car, it'd be kind of nice if the car knows who I am and adjust everything accordingly. It might even just, you know, hey, Jensen, this is in a very sultry voice that that this is what your day looks like. It might even recommend what you like because it knows that that it's almost time for me to get on a conference call. Do you want me to make the call for you? You know, so so the the the A.I. inside the car is going to be pretty amazing. We're going to have an A.I. inside a car that's going to look around corners. So even if you're driving, even if you're driving, the A.I. might prevent you from an accident. And so so the A.I. will become a copilot. The A.I. could could notice that that there's a child getting out of a school bus that's parked on the side of the road. And the car should go into a more conservative posture. And the car may say, you know, Jensen, there's a there's a school bus parked on the side of the road. Some kids could be coming up and careful. And so there's all kinds of things that the A.I. could could predict on your behalf, looked around the corner. Hey, there's a motorcycle that's splitting lanes. Don't turn, you know, or don't change lanes.