NVIDIA Corporation (NVDA) Earnings Call Transcript & Summary

Good morning, everybody. Welcome to the Goldman Sachs Communacopia and Technology Conference. My name is Jim Schneider of Goldman Sachs, and we are really thrilled to have NVIDIA and CEO, Jensen Huang with us today. Welcome, Jensen. Thanks for being here.

Thank you. Great to be here.

Now Jensen, last year, your prediction of $3 billion to $4 trillion in AI infrastructure spending by 2030 I think, raised a lot of eyebrows in the investor community, but it seems like we're actually really rapidly progressing toward that figure right now. What technical advancements or market developments with respect to AI.

I just think we should just take a pause and acknowledge that I was right.

We're doing that. We're doing that.

Just take a pause. Take your time.

What do people miss or underappreciate? And sort of where do you stand with respect to the to.

The question is why did I know that? Isn't that right? The question is, why did I know that? And it's actually fairly simple. -- the last industrial revolution made it possible for us to power everything. And we distribute power everywhere. And then of course, the Internet made it possible so that we can find anything. That's the big idea. You plug into the wall back in the old days, Ethernet jack or a modem and now WiFi, you plug it in, you can find anything. And so it's not a small thing. It's a very big deal. And because you could find anything and because everybody wants to find everything, we have to make a part of the infrastructure because we wanted to -- we can distribute power anywhere. We want to power everything we distributed power everywhere, became part of infrastructure. And so now we can find anything, find everything. But obviously, that's not what people want. What we all want is to know everything. We want to ask anything, know everything. And that's the layer of computing that we're building now that makes it possible so that whatever you want to ask and whatever you want to know, you can. And so that's the big idea with artificial intelligence. And a computer is in the middle of doing that. The first instrument was called dynamo and the next one, obviously, called the computer, and now we have these AI factories and they produce numbers, just like the Internet produces numbers. And at some level, is that simplistic -- now the question is very simple. How did the computer industry go from that to this? And what's the implication? And there are 2 compounding problems. Two compounding challenges. The first 1 is that in the last generation of computers, it was used by humans and because you're looking things up. And you bought computers as a tool, as a terminal. And so you could find anything, find everything. And so now in this new world, and so long as you prescribe to the idea that everybody wants to ask anything, find and know everything, if that's a world that you prescribe to and that you believe in and that it could somehow be integrated into almost everything we do, then the question is how do you build that layer of computing around the world. That computer as it turns out, is a generative computer is not a retrievable computer, and I think I've explained this to you guys in the past. The last 60 years, everything was prerecorded. And we've put it on storage somewhere. And when you touch something on the phone, it goes and retrieves that piece of information. And in fact, it probably retrieved about somewhere between 3 to 10 pieces of information. And based on who you are, the cookies that are associated with you. And your previous track record and your previous preferences, a recommender system would recommend one of those pieces of information to you, but it was all prerecorded. And so that model of computing is called retrieval based. Well, if your question, if you want to ask anything and know everything, you want to actually know it, you don't want to find it, you want to know it. Then you have to generate the answer. You can't reasonably retreat that, okay? And the reason for that is because in the old days, you have this idea called a recommender system and it's based on your preferences. Well, if you want to ask anything and know everything, then the preference has to be replaced by something else, and it's called context. Right? So the query is called a prompt and the surrounding environment includes the context. -- and also your preferences. The combination of all of that, you have to generate the answer. Now basically, what that says is that a new layer of computers that's continuously generating answers based on all the queries that are coming at it, has to get built. And so that -- it's no longer a bunch of storage, it's a bunch of computers. And what does that computer look like and what's the algorithm it runs. And so it's almost like everybody has our own recommender engine. Think of it that way. Instead of having one giant exit engine for all of Meta, you now have a recommender engine, literally for everybody. And these recommendation engines are really complex because the AI models are large. It has to be smart. It compresses a lot of the world's information. And so now the second problem goes like this. It's the end of Moore's Law. And you guys -- the first time I said it, some 15 years ago, there was a lot -- a gas. I said something that hurt somebody's feelings. It just says transistors don't scale anymore, and it's not a big deal. And so if the transistors don't scale 2x per year or get half as big every other year, then the question is what's going to happen to the future as we're trying to do this other thing. This called generative AI or artificial intelligence as you would like to think about it. And so in this new world, where it's the end of Moore's Law, and we need a lot more transistors than several things has to happen. The first thing that has to happen is you still want to get 100 and 1,000x improvement every few years. And if you want to do that, then you can't just live inside the chip. You have to co-design, which is the reason why NVIDIA became a co-design company, extreme codesign company. You guys hear that all the time from us. And the second thing that you have to do is -- if you want twice as many transistors, you got to build twice as made chips. And so -- which is the reason why NVIDIA invented NB Link. Right? And so you see -- well, the first thing that we did was we used cobots and that made it possible for us to fuse multiple chips together. And then we -- that wasn't been satisfying. So we took a whole bunch of chips and we connected them together into NVLink, which is -- the big breakthrough today, if you don't have NVLink, if you can have really excellent scale up and scale up technology is really hard, scale out is hard, scale up is incredibly hard. And so if you don't have that, you're dead in the water because Moore's Law is your enemy. And these models are getting larger and larger and larger. And so -- and then the third thing that happens, the compounded result of that -- and it's something that I predicted a while back, which is the semiconductor industry is going to be X times fat larger. And the reason for that is very simple. Demand of the semiconductor industry has been growing for some time. And yet Moore's Law was a depreciating deflationary technology. And that was happening at the same time. If you don't have the benefit of deflationary technology and the demand accelerates even further than that because instead of 1 billion people using tools and we sleep, now we have hundreds of billions of agents and have to think, so it's not like a chat bot, you don't hit it 1 time, you got to think iteratively. So you compound all of this together, the semiconductor industry is going to just keep getting larger and larger, which is what we're seeing now. And so these 2 fundamental ideas that we have a new layer of computing with a new application and the end of Moore's Law. Meanwhile, people are expecting these AI models to be smarter and smart because they don't like wrong answers. Then the compounded result of that should result in a very large industry. And so anyways, -- and that's the end of my talk.

Excellent. I think one thing that investors have consistently questioned is the ROI of this technology. it's easy -- it's hard to look at the P&L today and say, oh, here's the gross margin or whatever. But coating has clearly been a killer app for the industry. I think it's fair to say we're not going back to the old way on coding for sure. As you look ahead, what other applications or tasks do you think can really move the needle on both adoption and ROI for the industry?

Everything is coding. So when you buy a cookbook, you see the recipe, that's just a an American word for coding. When you ask somebody how to do something as coding, when you codify you guys codify a business process that's coding. And so everything is coding. If you want to do something, if you want to discover the best way of doing something and then after that, repeat it over and over again without anybody deviating from that process or that methodology or that best known process. However, you guys right, there's a lot of words about it. It's all coding. And we code in a lot of different ways. We code to make onlet we code to repeatedly close our books. And and manage our supply chain, we code in order to communicate with each other in a consistent way, connecting the connecting fabric, for example, the supply chain is one giant large piece of code, not one code, but trillions of pieces of code. And so -- so everything that we do is, in fact, coding. And so it's a sensible thing. It's also the easiest thing for AI to learn. And the reason for that is because there's a right answer. And there's -- the best answer is hard to find, but the right answer is not hard to find. And so coding is an important part of it. A derivative of coding is, of course, bug finding and a derivative of that, which is a very large market, is called cybersecurity. And the reason why there's so much conversation today about cybersecurity is because the industry is getting ready to launch some products and what better way to create demand and then to create a problem. So -- who doesn't want their market to be historical about their product and line up around the corner for it. And so -- so there are responsible ways of doing it, and there's less attractive ways of doing it, but there's a lot of demand creation about cybersecurity today because new products are about to be launched. And if you can code well, so you must be able to debug well. Red teaming is finding a bug. Blue teaming is patching a bug. And so it's not a complicated concept. And -- but the fact of the matter is cybersecurity will likely be the next major use case of AI, and it's going to run continuously. And so it'll be a great new business opportunity for the labs be a fabulous opportunity for CrowdStrike, and we have a big partnership with them using open models to create red teaming and blue teaming, and have the asymmetric advantage of swarms of nemotrom models that are running continuously. We have a partner with ship with Cisco. I think today, Palantir, Cisco NVIDIA, where they're building and Cisco is going to offer an entire AI factory platform from NVIDIA. Palantir is built on top of that. That's going to be taken out to all the countries and companies around the world to help them build either proprietary AI models based on Nemotron or cybersecurity Red Team, Blue Team models based on Nemoto. And so anyways, this is just -- it's the time when -- it's the next click of AI coming out to the next market. And then, of course, there'll be future use cases. But at the core, code is very important. I think that one of the things, Jim, that I was going to mention is if you look at -- if you now -- we spoke about the industry and technology and all, but let me -- I'm here to sell some NVIDIA stock, and so I don't want this meeting the agenda to be unclear. Were the world's first and only growth value stock. I think we're in every -- people are trying to figure out which one we are. And it's like we're both. You can be both at the same time. And so why is it that we're both at the same time? And let me just -- let me do some forensics on this for you. NVIDIA is incredibly misunderstood. As large as we are, we're insanely misunderstood. And the reason for that is this, we invented the GPU. You can't uninvent the GPU. When you are NVIDIA, you can't uninvent NVIDIA. And everybody knows because they've known me for 30 years. Most of the AI researchers in the world, most of the tech -- they grew up on products I built. That's how old I am. When an AI researcher comes up and say, I use GeForce 2060 when I was 8 years old. That's not a complement they're just counting the years. And so my point is -- we have always been a GPU company. It's not who we are today, but unfortunately, that's where we started. Does it make sense. We're not unproud of it, but most people think NVIDIA builds a chip. I mean, you need airplanes to ship what we build. Each one of our system, each chip, if you will, on GPU is 2 tons. One GPU is now -- it's not $399, It's $8.5 million. That's one GPU. all connected with NVLink, 2 million parts, right, 250,000 kilowatts, that's a GPU. And we ship thousands of them I think I was just seeing the reports this morning, Grace Blackwell NVLink 72 racks. Month-to-month increase month-to-month increased 27%. And so have month-to-month increase of 27%. That compounds. That's called a high-growth value stock. All right. So number one, we went from hopper, which is about 8 -- now hoppers that enable an architecture, not a chip, and Blackwells named an architecture, not a chip. And these architectures are expanding in their scope. So Hopper was about $18,000 a or so per GPU system, and Blackwell went to about 25 and Vera Rubin is about 40%. And the reason for that is because we're offering more and more and more of the overall AI factory. We see the AI factory in our head. We're trying to use extreme codesign to overcome the challenges of Moore's Law. And between algorithms and software and system and interconnect and new technologies that we invent along the way, we create a generation every single time that's many times faster or more productive in token generation capability than the generation before. And so the first thing that we do is we're increasing our SAM of the world's CapEx. And okay? So mission number one. Not only are we growing, we're also capturing more at the same time. The second thing is it is incredible, but there are more model companies today than there was a year ago than 2 years ago, a lot more. There are a lot more frontier model makers today than there was a couple of years ago. And there's a whole bunch who are starting up right now with a whole bunch of great ideas, okay? And so my point is, we're the only company in the world that actually runs every model. We didn't use to run Gemini. We run Gemini today. Of course, Grok bot doing fantastic. I can't wait to try it and of course, metas. These are all new. These are all net new, right? And not to mention, we didn't use to run anthropic for a lot of different reasons. We just -- we didn't have the money to invest in them as a young company. Now we have more money, and we're happy to invest in them and help them. But anyhow, our share of Entropic is growing very quickly. And of course, we're delighted to see entropic and open AI in all of these labs growing. -- okay? And so the second growth that most people don't see is we're the only company that benefits from Frontier, closed, open models. We run everything -- and so there's no company in the world who is indexed to open models, which is growing incredibly fast, except for us. But -- you can't find that number anywhere. -- open router is a good place to go look at these things, but you could see the growth. The world needs both closed models and open models, and we address them all, okay? And then the third thing is -- the AI overall market is growing incredibly fast. But what we see are just a CSPs. We just see the cloud service providers. But remember, their enterprises -- some of the great names of enterprises building AI factories are, of course, Jane Street and Hudson River just about every quantitative trading company in the world is shifting into this new model of doing doing predictions. And of course, drug discovery with Lilly and Merck and BMS and many others have now created basically their robotics lab is basically lab in the loop, AI in the loop for wet labs. And so you need a supercomputer to do that. And so you have enterprises but also the regional clouds. One of the biggest challenges, and this is -- people are -- I've been talking about this and you want to make sure you understand the strategy. Upstream -- the supply chain is very challenging. And the reason for that is because, obviously, we're growing super fast. So packaging is a challenge. DRAM is a challenge, LPDDR DRAM is a challenge. Connectors are challenging. Everything is challenging, okay? Voltage regulators are challenging. Everything is challenging. Wafers are obviously challenging. -- all kinds of different challenges upstream. But remember, the supply chain goes all the way to the end downstream until somebody stands at a computer and turns on to service. And so we've been thinking about the supply chain, upstream and downstream and one of the advantages that we have because our go-to-market -- remember, we run everything. This is the power of general purpose versus specialization. Do you remember 2 years ago, 3 years ago, everybody just say specialization is better than general purpose because it's faster. Well, if I can make general purpose faster than specialization, then specialization is all good. And the reason for that -- excuse me, generalization is all good and reason for that is because generalization gives you fungibility, durability, versatility, rentability. And very importantly, today, because of capital constraints, everybody's balance sheets, investability. Finally, we have a computer, the NVIDIA compute -- it's finally a computer that could be asset-backed. And so we can use it to secure loans, and that's a very powerful capability. And frankly, the only computing stack in the world that allows you to be able to say that because nobody could right step back on it. So anyways, our market includes CSPs and of course, the regional clouds, the neo clouds. The power of the neo clouds is this. They secure land power and shelf for us that the CSPs have already exhausted. Just remember this, this is a very big deal. Many countries want to -- in many countries and regions want to secure the power and land for their own companies. And so we have neo clouds around the world from end scale. And of course, you get the core Wave and Nebia they're doing fantastically. Some new ones that are on the verge of going public or in the process of filing whether it's end scale or land or firms. You can see a whole new crop of a really, really exciting neoclouds with hundreds of billions of dollars of backlog together. And so this is -- this is a capability that allows us to secure land power and shell downstream and be diversified in our way of going to market. And so these 3 ideas, the whole AI market we serve completely all of the models we serve completely and, of course, the world's data center CapEx, we addressed a lot more of it. which is the reason why we're growing so fast.

Can I follow up on a couple of those points.

Yes.

So one on supply constraints. You talked about a couple of weeks ago in your earnings call, high confidence in delivering 70% revenue growth next year, unconstrained demand growth of over 100%. You talked about upstream and downstream. Are you more concerned about the upstream component shortages? Or are you more concerned about the downstream, which you may have a little bit less direct control over in terms of land power shell and data center availability?

Well, this is -- the downstream is where NVIDIA's advantage is incredible. And upstream, our advantage is because of our scale. We have the largest supply chain in the world. And I've been working with the supply chain and these partners now for coming about 3 decades and when I make a prediction, they come through. And so -- and they like it when there are people who are right. And so -- because they have to put a lot of money at play and because NVIDIA has a track record of actually being helpful and truthful and good at predicting these things, not to mention we can create a whole market. Don't forget NVIDIA's platform is the only one that you can take all the way to the end market by yourself. Remember, we go to market through the CSPs, but we go to market through the OEMs. Look at Dell, they're doing incredibly, almost all 100% NVIDIA. I think it probably is 100% NVIDIA. Supermicro is doing well. Lenovo is doing well. HP is doing well, and then Cisco is coming in. We have 100% of the world's enterprise go-to-market can take NVIDIA to market because we are a full stack AI factory platform. You come in, you can install the whole thing, add our software to it and be basically up in a couple of weeks and get to work. And because these things are so expensive, the economics of it is so high that if you don't make it productive as fast as possible the anxiety is really quite incredible. And so if you look at our go-to-market, we have diversity of channels. We have many ways to get into market and because of our neo clouds and because of sovereign AI capabilities and all driven by the fact that we are full stack, we have a rich ecosystem, we can reach all of these different marketplaces. And so the downstream part of it is a huge advantage for us. There is no question land power and shows a problem. And so when I hear these people talk about these big numbers, the question, and we're tracking every single gigawatt of land power shell around the world, literally everything on the planet. Let me just think about all my partners. How many neo clouds is reporting back to us, how many OEMs are reporting back to us? How many clouds are reporting back to us? How many AI native companies are reporting back to us. We're working with everybody. And so we kind of know where everything is. And it's kind of -- it's pretty amazing. We've secured a lot of it. We've secured a lot of it. How can you otherwise be upstream obviously, TSMC would never tell you. I mean, they would never say something like that, that they can somehow project how many gigawatts sold. How would they know? Because it's far away from downstream. But we're all the way upstream and downstream, which is the reason why NVIDIA is -- our position is so good.

So you feel good about the closing make gap?