Nvidia CFO On Trillion-Dollar Future For Blackwell & Vera Rubin

We've got to talk about NVIDIA, and I'm pleased to say we are joined today by NVIDIA CFO Colette Kress joining us on the phone. And Colette, thank you so much for joining us. The quarter was strong in a word. And in particular on the call, I think one of the things that investors and analysts were excited to hear was the breakout and more detail on the CPU data center business. And that's where I wanted to begin today. And if you could talk us a little bit more through how NVIDIA is thinking about that business and the size of that opportunity. Thank you. Thank you so much. Yes, we had a great quarter talking about a lot of different things. Important to see the growth that we've seen in the AI and the different types of AI. We're in a phase right now of agentic AI, and that has really driven a significant amount of compute that we've been able to put in market. The diversifying of our customers continues to grow and all great types of folks buying our compute. But most importantly now is a focus in terms of the CPU and how important this is to the workloads as well. And we have the ability not only to just be inside Grace Blackwell, as well as what we have in terms of our systems with Vera Rubin. We now have an opportunity to also have what we call a standalone CPU and selling that. We have customers already looking at that opportunity to provide standalone along with our Vera Rubin and our Grace Blackwell together. That together, we believe for this year, can be a total of about $20 billion worth of CPUs. So when you talk about that $20 billion, is it the $20 billion as in the CPUs included with Vera Rubin and then the standalone units as well? Or is it the standalone on its own that's the $20 billion? How does that kind of work out? It's actually both. It's actually both. As you've already seen with Grace Blackwell, Grace with it, we have had standalone, and we'll see it along with those full systems as well. And then again, with Vera Rubin, we're going to see more of that as well. Colette, can you give a little more color around the standalone itself and how you all see the opportunity there? What kinds of customers are tapping into that versus the combined compute as well? You're going to see the same types of customers when they think about what they are building inside of their data centers and this important inflection that we've had for inferencing. It's going to be an important piece that they have, the CPU, enabling them for agentic types of work. You have the CPU there and available for you, and that will also be a need for some of the standalone agentic work, post-training types of pieces as well. We see this continuing within our full systems and a very big market now for just the standalone as well. I just want to touch on the through line. I think Jensen said $1 trillion by 2027. Big change, by the way, from October when he said I think it was $500 billion. I may be getting my numbers mixed up there because big numbers confuse me. But does that $20 billion add on to that trillion? I think at that point he had said it was just compute that was the $1 trillion, not networking, so I just want to get that together. It's a good question regarding the $1 trillion. We're very excited what we will see with Vera, Ruben, and what we'll see with Blackwell over the three years from 2025 all the way to 2027. That's going to be a trillion dollars. We're looking forward once we get to 2027 to really show this strong growth that we've seen over this time. That includes there's going to be situations where we also have additional types of products that we're bringing to market that will likely add to what we're seeing. When we see the part in terms of the standalone, yes, that's probably going to be an addition that even helps us. So stay tuned. Don't start now and finish because we've got plenty moving into that $1 trillion going forward. Collette, I want to ask you about the capital return too because that's sort of your purview here with you all increasing the size of the buyback authorization to $80 billion, increasing the dividend to $0.25, and you talking about the intention to return half of your free cash flow this year to shareholders. This is significant increases, right, from a penny to $0.25, pretty big boost to the dividend. Why did you all decide to make that change to increase that capital return to shareholders? Thank you. Great question. Our shareholders and the return that they earn is super important to us, and we have for many years been working on stock repurchases. We certainly do not want any type of dilution of our shareholders, but this was an opportunity given our strong free cash flow that we have that we both can work on what we need to operationalize our company using our cash to do that, but also we have been with quite a bit of investments and working in both with our suppliers and our full ecosystem has been essential, but given the strength of our growth, this gives us a great opportunity to advance what we can do in stock repurchases, but also what we can do in terms of improving the dividend for our shareholders. Now what that meant was at the very onset of what we built in terms of that dividend, it has not been the top part of where we have increased over time, and this is really looking at almost a new initiation when you see us at 25 cents per quarter, essentially a dollar a year for every single share. So we're excited to bring this. I know it's a good balance in terms of our investments as well as how we can return to shareholders. Colette, we know there are a lot of demands on your time today, so we'll let you go. Thank you so much for calling and really appreciate it. Thank you so much. Great to hear from you. Take care. So Dan Howley, interesting there, that precision from NVIDIA on the CPU number, right? Because I think the assumption on the part of the market was maybe that it was the standalone, so it's important to know that it is... It's the whole kit and caboodle. The whole kit and caboodle. I don't know what that means. If you will. I will say like bigger picture here, I've been saying in the week or so leading up to the NVIDIA earnings that the stock the day after earnings doesn't really trade on the earnings. No, not at all. Like to me, it trades over time on the huge numbers and the growth that we continue to see. And I think there's a couple of reasons behind this. One, the market is less surprised now by the consistency of the huge numbers that NVIDIA is putting up. And two, if you are an investor in AI and you're looking for where am I going to get the biggest, most eye-popping upside, they're looking at memory or they're looking maybe at some of the companies that NVIDIA is investing in, right? Coherent or momentum or whatever, you know, some of those other opportunities. And maybe that, you know, the attention is a little more there. But that said, like the numbers were big and consistent and they can, I mean, company like this side, growing sales, 85%. It's still just amazing. Yeah. And I think the memory thing is just going, we're going to see the same thing happen with memory, right? Like, you know, memory is... Eventually, yeah. Eventually. I think, what was it? SanDisk had, I mean, they broke off from, you know, Western Digital. So it's not like it just came out and, oh, it's up a thousand. I think it was up a thousand percent last year or something like that or something absolutely absurd. But yeah, I do think that when you look at NVIDIA's numbers and you say, oh, well, people just started selling it right away, even though they beat and raise. And it's like, okay, that's the expectation now is that they beat and raise, I guess. You know, they don't have that, what was it? I think it was 200%, 400% growth that we had seen at one point, right? But even if you look back at what the data center did, right? It was, I think, in the 70 billions for this quarter, 30 billions. It was up 92%. Yeah. I mean, that's crazy. The data center revenue alone. That doesn't make any sense, right? And so for a company this large, you know, and that continues to grow, it still dominates this market. It's going to continue to dominate this market. I know talk a lot about competition growing. And sure, there is competition growing, but also the market broadly is more growing. You know, people are trying to purchase more AI processors, whether that's through NVIDIA, whether that's through AMD or, you know, Amazon and Google and, you know, Cerebrus. So I think there's the idea that, you know, people see these numbers and yeah, they just kind of go, okay, cool. But it's just still blows your mind. It is remarkable. What else stood out to you? I want to move on because I know we talked to Colette for a bit, but is there anything else that stood out to you that you think is really worth flagging with NVIDIA? I know they changed the reporting structure, but like, what do you think is most pertinent? I mean, well, I'll just hit that for one second. I'm a big gamer, followed NVIDIA for so long because it was a gaming company. I mean, I used to build, I was 13 building my computers with NVIDIA cards, you know, like slapping them in there. My parents did not like that, but seeing them expensive, I mean, they're expensive at that point. They weren't that expensive, but I also may have messed up my parents' old Dell. Dude was not getting a Dell again. But so we, you know, we looked at that as, you know, the gaming company and now it's kind of moved in under edge. So it feels as though it's kind of like not exactly a gaming company anymore at all. And it's, you know, it's fully, you know, kind of embraced where we are AI and then gaming is also a cool side. So that's just me on a side. I think the other aspect was everybody was looking for some kind of clarity on China, right? I mean, Jensen had said previously that their market share there is non-existent, right? Nobody's building that into their models anyway. And they say, we are not going to recognize any revenue from China. We're not, we are not putting that into our outlook. But there was conversation after he was in China. Trump had said, look, China wants to focus on their chips. And then there was talk of, well, maybe they'll bring in some Nvidia cards. You know, there's these companies that may end up taking them. I think just that still feels very fluid at this point. I mean, I just want to know, is it happening or is it not? Will they, won't they? I don't think we're going to know necessarily. I think we're, we got to get used to that. We just count it out unless it happens. I think just shut it out. I think that's, that's some reason that made me think of, they're on a break, Ross and Rachel. I don't know why, but you know, maybe they'll see another country. I don't know.