Okay, welcome to the first edition of OK Computer for 2025. I'm Dan Nathan, joined by Gene Munster, the Managing Partner at Deepwater Asset Management. Gene, welcome.
Great to be here. I hope you had a great break. I think you and I, the last OK Computer of 2024, you and I did it together. So here we are with the first of 2025. I really appreciate you being here.
I want to just hit the semis here. I mean, right out of the gate, you know, you have a NASDAQ 100 that's up 3% on the year, you have an S&P that's up 2.25%. A lot of that performance once again is, you know, the Fateful Eight here, my man. Right out of the gates, up and running. I know that you expected a good year out of most of these names. A lot tracking to the themes of 2024, which were actually the themes of 2023. We have a little bit of a semi-charmed life cycle here, Gene. Let's hit some of the stuff, some of the stories right out of the gate today. Foxconn giving really good revenue guidance for December in particular. I guess we're going to have to wait and see what the guidance looks like, but it sparked a huge semi rally. We have the SMH up a little more than 4%. We have names like Micron up nearly 12%, ASML up, you know, 8.5%. Taiwan Semi, if it closes here, new all-time high, just broke out of a big range. Nvidia, if it closes here, new all-time high. Give me your sense of what this Foxconn guidance, specifically for December, suggests for Q1 here and why these names are rallying. Do you think it's a little too much?
The most basic level, what's going on is that this concern that we're just at the cusp of the hardware trade starting to unwind, that concern was reversed with the Foxconn guidance. Essentially, it looks like early in 2025 the outlook is going to start off on a good foot. And compare that to the middle of December when Micron reported. Now, Micron is a story that gets wrapped in as an AI-only story. The reality is that most of their business is consumer, old-school DRAM. They have an AI play with this high-bandwidth piece, but they had guided down 12% versus where expectations were for the upcoming quarter. And that, I think, in combination with some of the commentary from the Fed, really started the sell-off on the hardware and the broader AI trade late in 2024. Now we've seen a reversing and some positive comments coming out with Foxconn here. But I think what it comes down to is we're seeing encouraging data points that this hardware trade can maintain a higher growth than what many had believed. And just to put in perspective of the concern, like how do you put numbers around what this concern means, is that if you look at Nvidia on their calendar '26 numbers, it's trading in the low 20 multiple. It's probably going to grow better than 30%. Most of the street is looking for a low 20s growth rate, so kind of a one PEG on that, which for this story is evidence, I think, of investors being nervous essentially that the stocks are cheapest at the top, that kind of concept. Micron, different market cap obviously versus Nvidia, but it trades at like, even with this move, it's like 13 times next year's earnings and is going to grow revenue this year around 45% and next year for calendar '26 around 25%. So my point is that this market is really sensitive to any commentary from the hardware players about what the sustainability of this AI buildout is. And so I think that's what's going on today. And then also investors are focusing or starting to look forward to Jensen's comments related to at CES.
Yeah, no doubt. And we obviously know that he's going to be very positive and promotional. That's the way Jensen is in general. You know, the Micron one is really interesting to me, Gene. So when they reported a few weeks ago, obviously everything in the consumer electronic space was very weak, right? And so, you know, the Gen AI component of their memory just wasn't big enough, right? And when you think about it, you know, to drive let's say the performance of the stock, so here it is in one fell swoop, it's up 185% on the year and it's almost filled in that earnings gap. Now you could make the valuation argument and you could kind of be chasing things early in the year. You got a lot of time for things to work out if you're being judged on a quarterly or an annual basis here, but it feels like a little bit of a dash for trash. And I look at ASML because nearly two months ago, remember that the booking guidance that they gave for Q1 was horrible. And here's a stock, you know, today it's up 8.5%, it's up 12% of the year. I mean, listen, everything has sparked. I think a lot of the stuff that's been down on their asses is ringing a little bit. Even AMD is ringing. Intel can't get out of its own way. But so I guess my question to you, and we're going to get to the whole custom silicon bit because that's something that you've raised on numerous occasions over the last couple of months, and that was one of the reasons why as Broadcom gave their guidance, as Marvell gave their guidance, maybe Nvidia was kind of stuck in the mud. We've been hearing for months, if not quarters, how all of Nvidia's major customers, that's Google, that's Amazon, it's Microsoft, it's Meta, 40% of their sales, how they've been working with the likes of Broadcom and Marvell and others to create, you know, a second source or their own chips relative to Nvidia. So let's talk a little bit about that. You know, I mean, at this point it seems like Marvell and Broadcom very late to this kind of rally, but it looks like at least the way investors are thinking about it, this is the next leg of the trade. Does that make some sense? You saw money cycle out of Nvidia into these two names over the last couple of months.
And is kind of part of the narrative for '25 on the hardware side is going to be this custom silicon. And I mean, just to add a little bit to what you had said, is that the basic concept here is that these hyperscalers have been spending a lot of money with Nvidia, but that is helpful with a big part of what they're doing in building with AI. They need these GPUs to do that, more general intelligence type of hardware. But they also are going to increasingly need more silicon that is specific to a function. And so that's this idea of custom silicon. And so that's Marvell and Broadcom. They're going to be some of the bigger beneficiaries of that. And so there's one piece which I think is, call it a truth. We have this kind of a syntax around undeniable truth. I don't know if it quite reaches the undeniable truth level, but a truth that custom silicon is going to be at the top of investors' minds throughout all of 2025. And so I think that's good for Marvell, Broadcom. I think that's going to continue to be good because the expectation in 2026 is they're going to see more and more business from these hyperscalers. And then how do you factor in, you know, to your point, to your question, like the cutting side of this, which is them potentially cutting, them being the hyperscalers, potentially cutting orders to Nvidia around this. And I'm going to pull up a quote from the press release from Nvidia around how far, what their supply-demand equilibrium is. This was two months ago. And the press, they said that demand for Blackwell, which is the upcoming architecture, is expected to exceed supply for several quarters in fiscal '26, which is basically calendar '25, for several quarters. Now, by my math, several is three. But my point is that this whole conversation around custom silicon is also coming at a time where Nvidia is saying that they're not seeing any slowdown. And I think at least in 2025, I think both of these can exist. I think you can have Nvidia beating numbers and raising numbers, continuing around the general buildout, and I think you can have Marvell and Broadcom continuing to do well as the custom silicon starts to build momentum. It seems like it's kind of, I'm in denial a little bit, but I think the reality is at least in calendar '25, I think both of these themes can continue.
Yeah, I guess the discount, you just mentioned a little over a PEG, a PE to growth of one, right, for Nvidia. And it almost seems like investors are kind of discounting the potential for some sort of drop-off in demand, right? So Jensen's been very promotional. He suggested this kind of insane sort of demand for Blackwell. But you know, when we listen to the BG2 Pod with Brad Gerstner and Bill Gurley a few weeks ago, and you and I talked about it, they had Satya Nadella on, the CEO of Microsoft, and he was asked specifically about the supply-demand as it relates to, you know, their demand, I guess, for high-end GPUs from Nvidia. And he said they are no longer constrained. They have the chips that they want. I mean, that was a direct quote. And we'll put the podcast in there. And so I almost say to myself, okay, what's the next leg of this trade, right? And the next leg is exactly what you're speaking to. And then let's just kind of talk a little bit about some of the guidance that we heard, right? So Microsoft, there's a story out this morning, I guess it was a blog post, you know, from a Microsoft executive talking about how they're going to spend $80 billion this fiscal year on AI infrastructure. Now that could mean a whole host of things, right? It could mean chips, it could mean data center build, it could be leases, it could be power, you know, a lot of stuff going on there, right? And so that is what it is. It reminds me a little bit about what we heard of Broadcom's guidance by the end of 2027, they're going to see, I don't know, spending between 60 and 90 billion, I think that was the number. And you and I talked about it a few weeks ago, up from like 10 billion. Yeah, yeah. And you know, and this Microsoft, I think they probably what spent 60, 70 billion dollars in the last fiscal year, right? So this is a bump. But you know, who knows whether they're going to complete that or not. Which brings me to the story, and The Information is reporting this also, that Microsoft paused construction on a data center that they were building specifically for OpenAI. We know that they have this partnership with OpenAI. It seems like every week we hear a new story about some wrinkles in that partnership. They're competing on a lot of different levels. They're trying to make their own chips for their own, you know, models and the like here. So when I think about all that together and I say to myself, okay, if Jensen Huang does not, like, you know, blow the doors off at CES, you know, I start to wonder like how much legs this trade has. And as we get to, you know, Q4 earnings and really expectations for guidance and most importantly, I guess, capex by these major hyperscalers, if we're slightly disappointed by that, I'm hard-pressed to think that the names that have been benefiting the most for the last two years are going to continue to rally in this way. Does that make some sense?
Totally makes sense. I think one of the contrarian views is that these companies can keep going. And so I think like just naturally when you have runs like this, NASDAQ up 89% over two years, 33% or whatever it was in 2024, naturally you have to expect some sort of a pullback. And I expect pullbacks. We can talk about why I expect pullback, but I still think we are going higher in '25. I think Nvidia is going to continue to go higher. And part of the reason is we spend time understanding, we talk a lot about the hyperscalers and being 40%, that's a big part of Nvidia's business. But I believe there are other things, whether it's industrial AI applications, sovereignty, I think there are other things that are still just getting going on that front that are going to benefit their business. And when I think about, to the point of Jensen's comments at CES or what earnings each window that we go into, today, given we have some good commentary from Foxconn, the flavor of it is painted more positively. But in general, when you have the market has done what it has over the past two years, each window, earnings and any company events, I think investors are nervous going into this. And so any sort of blink, I think you can see pullbacks. And I think what you're going to see is maybe some commentary around in three weeks from one of the, you know, the three big hyperscalers around how much they're spending and the timing on that, you'll see a pullback. But ultimately, I just believe that AI still is in its early stages on building out the infrastructure. We can talk about where I get that belief from. But if that ends up being the case, then Nvidia is going to grow faster than what people think this year and next year. And even though it's going to be a wild ride, I think it still goes higher. And it's hard to grow at 20% when you've gone from being a $20 billion business to expected to be $150 billion next year, something like that. I mean, it's just this incredible move, but I still think that is how big the infrastructure spend is. And just one quick comment, you mentioned OpenAI and Microsoft, and you flagged this, but I'm in agreement that that relationship seems to be fraying. And I think hearing about pullbacks about what Microsoft is doing around building any sort of capacity around OpenAI, that to me feels more like a feud versus a shift in the paradigm.
Yeah. Let me just ask you on the infrastructure, like what are some other names in which you're looking at? Like is some of the like semi chip equipment guys like an AMAT, which, you know, was participating, you know, the first half of last year, right? It was all, you know, it was a free-for-all and anything related to that, right? If you're building out fabs, you know what I mean, like to kind of build more chips to go into more servers to go into more data centers, you would think that the semi equipment guys would do pretty well. You know, AMAT had a massive run and then it gave it all back. It's up a bunch today, maybe 5.5% or so. Like how are you thinking about this trade broadening out a little bit? Because especially if we're kind of in the blow-off phase, which feels like, you know, early this year we could be, if that guidance is fine and, you know, it's better than the whisper, you're going to see these names continue to go higher. And not too different than a Broadcom which had been in this range, although was up a lot prior to its earnings, had a massive breakout. Marvell did the same thing. We're seeing that right now in Nvidia. We're seeing that right now in Taiwan Semi. So if you talk about the churn, the digestion that a lot of these stories had or a lot of these stocks had while the stories remained very good, the trade broadened out a little bit. Now they're breaking out. I can't figure out whether that's just a new year sort of thing, you know what I mean, where folks are just saying, hey, listen, we had a bang-up year, maybe we took some chips off the table as we headed into the end of the year. There was definitely some volatility in December. We saw the VIX go up above 20. That had a large part to do with the Fed and what they might do. And the other thing, as I'll just say, is that, you know, you have a 10-year yield at 4.6% and you have a Fed funds rate that might not be going too much lower in 2025. This was the exact ingredients that made it really hard for some of these long-duration, high-valuation tech trades in late '21 and all of '22. And so I wonder whether macro and some of this, you know, kind of rate stuff and the dollar exposure and that sort of thing might start to play out in this space too. Because throw China in there, you know what I mean, like who knows what's going to happen in the first year of the Trump administration. We know that there are plenty of China hawks that exist in that administration or the newcoming administration.
My goal is to get this right. And I've been continuing to believe that we still got a couple years left on this. And I don't want to get lulled into the performance that we've had with some of these companies and just think that because they've been performing well, they'll continue to perform well. So all those factors that you mentioned are things that are top of mind for us at Deepwater. And I just want to mention, you know, talk about kind of like the hardware trade piece of this. And I think at the most basic level is I probably have more optimism about what is in store, the spending ability outside of the hyperscalers and the ability to keep some of these companies going. One company you mentioned power earlier, but Vertiv is another company. I mean, those the four that you had mentioned, I'm almost checklist down that checklist, Broadcom, Marvell, Nvidia, Micron, we own all those companies too. So we're still there, we're still believing that this hardware piece is going to continue. But there's a but. The but does not mean everything before it was not true. I think that is true, it will continue. There is going to be a day of reckoning around these, and that this is the nature of the hardware business. And this is why you see these big swings around all the hardware companies on the Micron news, is that investors know these are boom and bust. That has not going to change. This is not going to be subscription-based revenue from these companies. And so ultimately we're going to see some more material slowing of growth rate, and my sense is probably in year three from here. And I think that's one piece to it that could cause this to fracture. The other is related to recession. Recession will pull the plug on the AI trade. It's almost impossible for if a recession's happening for the market just to continue to rip. And then you throw in what's going on in China. My sense is that it's almost like a foregone conclusion that China's going to make a move on Taiwan. And I think most people think it's like in calendar '27 right now. So got a little bit of time there. But how do I, when I put all this together, my sense is that the commentary, the guidance from these hardware companies over the next few quarters is going to be optimistic enough, higher than where investors are thinking. We'll see what plays out, but that's where we're at right now.
I think expectations are very low on that regard, possibly, you know what I mean. And so the beating of those expectations could be a very good thing for the whole space. But before we get there, I want to talk a little bit like about these equipment manufacturers, like about other foundries like a GlobalFoundries, for instance, right? Will AMD and Intel finally join the party? You know, like some of these names that have kind of been left for dead. Dell had some very poor guidance. Super Micro did, right, on the server side. Like so it still seems pretty narrow despite, you know, having this custom silicon kind of theme play out over the last few months or so. So how are you thinking about equipment? How are you thinking about some of these server names? You know, and I'm just trying to find some places that maybe investors aren't focused on right now.
