The Next BIG Thing In AI – Data Center Cooling

[00:00:00] Speaker 1: McKinsey says $6.7 trillion will be spent on data infrastructure through 2030. Cooling accounts for 30 to 40% of a data center's total energy cost. Ergo, data center cooling might be worth a lot of money. Today we're going to look at five stocks plus a bonus stock that experts say may benefit from the explosive demand for data center cooling. Let's start with the most popular name on this list, which is Vertiv. So when you're investing in a particular theme, you always want to consider the exposure that you're getting. What are you actually investing in? We often refer to this as invest in everything with Google. So you'll see pundits out there say that Google's an investment in everything from robotics to data centers. But what Google actually is an investment in is ad spending, of course, though cloud is certainly formidable for them and growing. So when we look at Vertiv, their estimated contributions back in 2020 from thermal management were 20%. And that's relevant because in 2023, they made a cooling acquisition. And you can see back then 70% of their revenues came from data centers. So we can do the math really simply here. Back in 2020, they had thermal management revenues of around $860 million. Well, fast forward three years later when they made this acquisition, you can see that contributions from thermal management have now jumped to 30%. So again, you can do the math. In 2023, they had $2 billion coming from that particular segment. But what's very notable about this slide is when they talk about the thermal market outlook, liquid versus air. So you can see here that liquid cooling is not only growing a whole lot faster than air, but it's also displacing air cooling. So then you would want to pay attention to companies that only offered air cooling because they would essentially be having their solutions cannibalized by liquid cooling. Though, Vertiv tells us that there's a coordination between these two systems so that they can coexist and benefit data center providers' need cooling. Now, disruptive tech investor Spear put out a great presentation that I've referenced here. So if you look at data center total addressable market, what they say outside the rack. So our largest investment is inside the rack with NVIDIA. We've been invested in NVIDIA for close to a decade now. But outside the rack, there's some interesting things going on. One of those is thermal management, which they expect to grow at 40%. So you need to be careful with all these market estimates and things like that. When you start to cross-validate them, it's really a rabbit hole. But if thermal management was worth $5 billion in 2023, then that should imply that Vertiv with $2 billion in revenues from that particular segment should hold a 40% market share, right? Well, no. Spear says that actually the top three players control around 45% of this market. And they point to some names here that we can start to explore. You have, of course, Vertiv. You have Schneider Electric, Eaton, Legrand, and Envent. And what's very notable about this slide would be the percentage of total revenues that data center represents for each of these large players. Say you look at a firm like Envent, well, it's under 20%, or at least it was back then. We're going to take a look at those updated numbers. But going back to the point I made about making sure you're getting exposure to what you think you're getting exposure to, the problem what you'll have is that a firm might have 20% exposure to data center that's growing like mad, but the rest of their business is stale. So the blended growth rate that you get is mediocre at best. And, of course, these numbers are changing over time. You see Del Oro Group here with their Q1 2025 data center physical infrastructure quarterly report. So now we're talking about the physical infrastructure, which would include things like liquid cooling. And they note here, liquid cooling's growing like mad, right? That's what Vertiv said. So that matches what we can expect there. They also said power distribution remains one of the fastest growing segments. It's growing even faster than liquid cooling. And they mention here how Schneider Electric and Vertiv are virtually tied for global market share. Again, not for liquid cooling, right? But for data center physical infrastructure. Well, that's useful then. So let's take a look at Schneider Electric. Here you see this is their last quarter. They say, well, our revenues were up 8% organic. And then you start to look at their three segments here. Well, products, which is nearly half of their revenues, only grew 2% organic. Systems grew 17%. They say most notably data center. Well, is data center only confined to that segment? Well, we don't know. They don't tell us. Then you have software and services. What contributions are being made to each segment by data center? So we can look further in their earnings release. We see this other table here breaking down the business into a total energy management and total industrial automation with energy management being the vast majority. Well, it still doesn't tell us what we're looking for. It's interesting to see North America energy management growing strongest out of all these segmentations. But we're told by Grok that 24% of Schneider Electric's revenues last year came from data center. They point to an investor deck, and I browsed the heck out of it. I couldn't see this number. But presumably, Grok is more intelligent than I am. So we can cross-check that against Speer saying it's 20%. That would have been back then. So let's just slice it right down the middle and say 22% of their revenues are coming from data center. And since we're told earlier this year that their market share is nearly the same as Vertiv, then that's another data point. Now, when we move on to looking at a firm like Eaton, they actually tell us very clearly data centers and distributed IT, whatever that is, made up 17% of their total sales. Now we have a number we can work with. This chart is really cool. Look, they're showing us all the breakdown of segments here. And for each segment, they have whether you're going to see a decline, a small increase, or a double-digit increase. That's the green arrow pointing straight upwards for each segment. So then you can see, well, how these other segments might offset that growth in data center revenues. So whilst data center might be quite lucrative at the moment because you can command whatever prices you want, there's such a strong demand for this hardware, that's only going to last so long. And of course, these other segments will drag down the profitability of data center. I think another good example would be a firm that one of our premium subscribers raised in our discord group. Our discord community is great. It's full of intelligent people that support us financially and create a real hive mind when it comes to figuring these things out. Someone raised Munter's group. Again, what's interesting to note here, as you can see in the center chart, it shows food tech, which is growing almost as fast as data center. It's relatively interesting segment there. They have air tech. First of all, that's not growing. Second, look at this other chart on the right. It shows profitability. Look at how that majority segment of air tech is dragging down the profitability of other segments. It leads to the entire group seeing a drop in profitability. In other words, all the great stuff they're doing in data center might very well be muted. Again, on price here, when demand is this strong, everyone's going to enjoy pricing leverage. And when you think about the companies that you want to invest in for the long term, larger firms, we always invest in leaders whenever possible, are going to be able to better compete on price when it comes down to that. And, of course, your largest customers have the most to gain from moving in-house. Looking back at this press release here, it says, well, did Amazon crush Vertiv stock tanks over 6% after Amazon's bold cooling move? Well, Amazon had developed some cooling technology in partnership with NVIDIA. It took them just 11 months from design to production. They're able to move very fast. Not that sophisticated of a problem they're trying to solve, apparently. And a Bloomberg analyst calculated that 10% of Vertiv's revenues come from liquid cooling. That's useful to know, right? And they calculate that Amazon might be their biggest customer. Well, since then, that no longer seems to be a concern to the market. But it's certainly something that you need to pay attention to, especially when it comes to customer concentration rate. So, Amazon is indeed a customer that accounts for more than 10%. You should see that in their financials. Brings us to a company called Invent. And I want to be careful here about how we're classifying the term thermal management. So, you see this slide from Invent talks about how they divested a thermal management segment. And when you read into the details there, whilst they still have liquid cooling capabilities, that segment they divested actually has to do more with heating than cooling. When they used that, I think they got $1.7 billion. They sold that segment to a private equity firm. They said they're going to use it to make more acquisitions. Well, that's one thing that you're seeing a lot of in this space. When we look at the six firms we've talked about today and estimated their data center revenues, you see Schneider at the top there with $9.8 billion. Then you have Vertiv at $8. I think Vertiv is expecting $10 billion this year. So, some nice, strong growth there. Eaton at $4 billion. Legrand at $2.4 billion. Invent at $0.76 billion. And Munters at $0.45. And then when you look at growth, this is the midpoint of guidance for total revenues. In some cases, it's only provided organically. Remember, you can't be disrupting anything unless you're seeing revenue growth. And as I said, the more pure play you are, the more likely you're going to see strong growth coming from data center. And I plugged this question into Grok just to get a comparison, sort of a sanity check on our O numbers. And you see here we came pretty close across the board except for Eaton. And that actually seems to be a mistake Grok made because very clearly we looked at this slide that said 17% of their total sales come from data centers. However, now that I'm looking at it, it's interesting because this slide says data centers and distributed IT. So, Grok has said, well, 17% comes from those two. 10% comes from data center. Maybe Grok's correct. But you see that this becomes very difficult when you start going down the rabbit hole figuring out, first of all, what exposure they have to data centers in general, but also exposure to cooling. That's what we want the most exposure to. That's what we're most excited about, right? And you also need to pay attention to organic growth versus growth made by acquiring other firms. As you see here for Legrand, they made three acquisitions in the first half of 2025 for data centers. They're listed out here. Just remember, I think it's upwards of two thirds of all acquisitions don't realize the synergies that the acquiring company expects. So, there seems to be this frenzy to acquire as much of these smaller firms as possible. They're essentially acquiring growth, right? But you'll see in the case of some numbers I looked at, they'll say, well, organic growth was 10%, but total growth was 5%. Well, what happened there? Well, they acquired something that took away from that organic growth. So, pay attention to that. So, just some thoughts here. AI is the most impactful technology we'll ever see. That's one argument. AI is the most hype technology we'll ever see. That's the other argument. Guess what? Both can be perfectly true at the same time. I think Sam Altman was making that point recently. Companies need a plan for when their cash cows subside. And this always happens. You look at the massive amount of money being poured into these AI startups. A lot of those are going to crash and burn. And the demand for hardware and everything that you see will eventually subside and reach some sort of stable point, right? Where you have maintenance or replacement, things like that. But there's a massive boom happening right now, no doubt about it. And, you know, these companies we pointed to that are quite diversified, the more diversified they are, the more resilient they're going to be. And the other side of that sword is the less growth they're going to see as other segments offset the fastest growing segment. So, what we've found during this research is the best segment breakdowns you're likely to get from any of these companies are sporadic or they just say data center. This is a very tough segment to analyze. So, pick your player and start to monitor whatever key metrics they offer up that are relevant to the exposure that you're trying to get. Invest in companies, not stocks. Noted that Vertiv is a very volatile stock. You have all the instant analysts out there doing astrology for men on charts. That isn't how you become a successful investor. That's nothing but noise. That said, out of all of these, Vertiv really seems like, well, it is definitely the purest play bet. And it isn't overvalued. The last time we looked at Vertiv, they had quite a low simple valuation ratio. That's just market cap divided by annualized revenues. In this case, we've calculated that again. Comes in around five. Our catalog average is seven. So, even at, I think it's close to all-time highs, it's still not what we would consider to be overvalued, at least based on revenue metrics. It's also extremely volatile in an ex-spac, which is a red flag in itself. But this company definitely seems to be executing. So, there are other ways to play the boom in data center infrastructure. And I'm going to leave you with a video that talks about some other stocks that you can get exposure to for what we call inside the rack. So, give that a watch. Next, thanks so much for taking the time to watch this video today.