Digging into the bottlenecks of AI

0:01Tyler Crowe:Diving deep into the AI supply chain today on Motley Fool Hidden Gems Investing.

0:09Tyler Crowe:Welcome to Motley Fool Hidden Gems Investing. I'm your host Tyler Crowe and today I'm joined by longtime Fool contributors Jon Quast and Matt Frankel. So we wanted to start the discussion today about AI spending basically and some of the kind of nooks and crannies that a lot of that is falling into it. And to do that, we want to start with the big news event of the day, and that was Oracle's earnings. They reported earnings early today, and the stock is down. Pre-market was down about 11 % right around the time that we are taping. And the numbers were pretty good, at least from a quarterly earnings perspective.

0:42Tyler Crowe:But I think the biggest reason we saw this massive decline in the shares was because CapEx spending is going way, way up. They spent much more than they had expected. They were expecting$50 billion for this entire fiscal 2026, but they ended up spending $55. And their plan for 2027 was to add more to that and do about$70 billion in capital spending. Now, there's a lot of reasons for that. The remaining performance obligations, basically, however hyperscalers are spending has gone way up. And we'll get into that. So guys, is one of the things I think about with Oracle compared to like the other hyperscalers is this is not like their bread and butter business.

1:24Tyler Crowe:This is something they've been really getting into. And is this seeing these kind of large, large numbers? Is this like a risky approach? Is this like the way Oracle should be attacking this?

1:35Jon Quast:You're right that the earnings report was very strong, but the earnings are not the full story. If they were, the stock would probably be up today. Both revenue and earnings grew by 20 % or more. They beat expectations. Cloud revenue grew by 47%. The RPO number that you mentioned, it is simultaneously the company's biggest risk factor and the biggest bull case, really. I mean, having a$638 billion backlog would be a dream of most companies. But when it's all tied to one customer, it does seem to be diversifying a little bit. It's now estimated that quote, over 50 % of that backlog comes from OpenAI, as opposed to virtually all of it not that long ago.

2:15Jon Quast:The number that you just mentioned, it increased$85 billion sequentially. And so that's not OpenAI deals. Prepaid AI contracts, it's worth mentioning now total$75 billion of that. So that kind of helps the need to raise capital, but there's still a pretty big gap there. So having said all that, the company does have a massive CapEx need. $56 billion in fiscal 2026, That's up from$21 billion a year ago. Negative free cash flow for the first time in a long time and probably for the foreseeable future, if we're being honest. It could certainly pay off if OpenAI and the others can afford to pay for all of their committed spending.

2:51Jon Quast:But it's still a risk factor in the near term. Almost$130 billion in debt, and that's growing fast. And anytime you're adding debt quickly, you're adding risk.

3:00Matt Frankel:To your point, Matt, there's really no other way to do this, to be fair. when the space is moving as fast as it is and you've got to spend a ton of money to build something that's very expensive up front, of course you're gonna take on a lot of debt and you're going to take it on very quickly. So I wanna be fair to Oracle on the one hand and also I wanna acknowledge that it is a long game that the company is playing. This is to the RPO number specifically, okay, 638 billion, yes, But over half of that is more than three years out. So it's not expecting this money anytime soon. Now, that's a little bit risky for sure.

3:42Matt Frankel:Do I think that AI data centers are going to still be a big thing in three years? Yes. Do I think that all of that money is necessarily in the bag right now for Oracle? No, not necessarily. That said, I mean, OpenAI, much of what it is doing is the Stargate initiative. And there's a lot of companies involved in that. Also, the White House is supportive. So it's risky, but I wouldn't say that it's, you know, humpty-dumpty risky.

4:11Tyler Crowe:To your point with the idea of it being risky or not or something like that, one of the things that is, it does seem like these companies are really dependent on, Matt, you said it was over 50 % of it is at OpenAI. I'm assuming a lot of it is anthropic. We've seen Anthropics, the OpenAIs, these other AI companies in the world, they have been signing some very large checks related to spending on compute and stuff like that. Part of the reason when we discussed the S1 for SpaceX, I think it was last week or a couple weeks ago, we were talking about Anthropic doing, I think it was$1.2 billion per month in basically renting compute from SpaceX.

4:49Tyler Crowe:And a lot of what we're seeing here with these RPOs are similar deals. These are big, big numbers and a lot of money that they have to even cough up kind of up front. And that kind of gets me to like one of the things we I've also seen OpenAI just confidentially submitted their S1. So they're also planning on going public relatively soon. Anthropic already did it. We're expecting that S1. When we see these numbers coming out from Oracle, I'm starting to wonder, do you think that the reason these companies are going public right now is because the bills are starting to come due for some of these very large companies?

5:23Jon Quast:Yeah, and I mean, don't be fooled just because they recently both did funding rounds. Anthropic very recently raised$65 billion. It's really rare for a company to have a big, you know, series F or G or whatever it was for this, and then do an IPO almost immediately afterwards. And it's because they know they'll burn through this capital very quickly and are going to need more. I mean, fortunately for them, investors seem very happy to open their wallets right now and give them whatever they need. We're seeing this with demand for the SpaceX IPO. It's just a question of whether that's sustainable.

5:55Jon Quast:In a year from now, if they wanted to do a big offering, would they still be able to do that? No. So they're rushing while it's a hot IPO market. So the rush to IPO, especially if the market gives them trillion dollar valuations, which is expected, is driven by their massive capital needs. After all, it's really a lot easier to sell an IPO to investors than it is a follow on offering as Oracle has found out.

6:19Matt Frankel:Here's the thing with IPOs. you definitely want to go public when the timing is attractive and most attractive. And you look at the valuations in the stock market right now, I'd say that they're pretty generous. I'd say it's a pretty attractive time to go public. And I think that is why, more than anything, we're seeing three, perhaps the three biggest IPOs of all time are going to happen right in a cluster when you talk about Anthropic, OpenAI, and also SpaceX. I think that that speaks more to what is going on. The market has a very, very favorable, it's a very favorable time to go public. It is smart to take the most money that you can when it is available, right?

7:04Matt Frankel:When we had 0 % interest rates, it was smart to borrow money because there's really no, there's no interest payment involved. So yeah, why wouldn't you borrow? So taking money when it is very attractive is smart. I think that is more the motivation for going public right now rather than where are we going to get the money to pay. I think that financially they're in better shape than it's not a desperation move. It's a this is a really good time to do it move.

7:32Tyler Crowe:To that, we're talking about the IPOs. I think within the past 48 hours, I have seen news articles or headlines basically going from the SpaceX IPO is two times oversubscribed, which means there's twice as much demand as the shares that should be going available. And then I think it was 24 hours later, it was four times oversubscribed. Who knows how true those are? It might be just a little bit of leak to kind of drum up interest in the IPO. It's hard to necessarily confirm that. But, you know, I've got to imagine this is, we're recording this on Thursday. SpaceX goes public tomorrow. I'm sure we're going to see a couple more with basically this rabid demand for these IPOs.

8:13Tyler Crowe:And I would not be too, too shocked if we saw something similar with the OpenAI and the Anthropic IPOs in the coming months. But I think a lot of it might depend on what we see with this upcoming one. Now, we've done a lot of discussion on kind of the big companies, AI CapEx spending, who's going to be spending it. So coming up after the break, we want to dig down into some of the nooks and the crannies of who's making the money in these areas. What are some of the bottlenecks and maybe some of the smaller companies that people might not be paying attention to?

8:48Jon Quast:Support for the show comes from Fundrise. For the past 70 years, there's been a room in finance most people couldn't enter. A room where you could have invested in some of the biggest names in tech, companies like Airbnb and Uber before their multi-billion dollar IPOs. I'm talking about venture capital, but unless you had millions of dollars and the right connections, the door to the venture capital room was closed. You were stuck waiting in line with everyone else. But recently, Fundrise took a sledgehammer to that door when they launched their venture capital product and made it available to anyone with a minimum investment of just$10.

9:21Jon Quast:Fundrise says their mission is to give everyone the chance to invest in the best tech and AI companies before they go public. Visit Fundrise.com slash fool to check out Fundrise's venture portfolio and start investing in minutes. All investments involve risk, including the potential loss of principal. Past performance is not indicative of future results. This is a paid advertisement.

9:43Tyler Crowe:As is the case with just about any investment trends, there are some of the obvious winners. We were talking about SpaceX. We were talking about Oracle, some of the big hyperscalers that we did in the previous segment. But with every sort of industry like this, there's always like the hidden winners. We dive deep down into the various parts of the supply chain and you can find bottlenecks and constraints in companies that tend to like dominate small niche parts of their industry that most people may not be thinking about, but can be incredible wealth creators for a lot of investors. So guys, we've been talking about AI infrastructure build-out.

10:20Tyler Crowe:I've lost count of how many times over the past six, eight months here doing this podcast. And what I want you guys to do is, instead of thinking about the big companies here, think about the bottlenecks. What are some of the constraints that you have been seeing as the AI infrastructure gets built out? And what are some of the opportunities? Because I know that when people are listening to this podcast, it is obviously investing, it's stock picking. Where are you seeing some bottlenecks, unique industry perspectives, and perhaps some companies? John, when you pitched this to me, you said chip production and kind of like the real deep cut parts of the industry.

11:00Matt Frankel:Yeah, I'm looking definitely at the fact that chip production is increasingly a national security issue. And there's a big push to bring all that production domestically, not to mention that there are more chips needed just in a general sense. And so we're seeing all of these companies projecting, hey, we're going to build this, we're going to increase this production, that production. Talk about Tesla's TeraFab. We can mock it a little bit, but they're saying because they're saying they want it to be the biggest facility on Earth, essentially. But is it going to be probably a big facility? Yeah, I'd say so.

11:35Matt Frankel:Is it going to pump out a lot of chips? Yeah, definitely. And so what is interesting to me about that is, you know, perhaps the margins for the chip makers could go down. But I think that it creates an opportunity. It is creating an opportunity for a company such as PDF Solutions. And that's ticker PDFS. This is not PDF from Adobe, the portable document file. This is process diagnostic framework. And so this company, essentially what it does is it is looking for defects on the chips. And then as it detects them coming up with a way that we can now fix the process so that we don't produce a ton of chips that have the same defect.

12:21Matt Frankel:The value proposition here for the chip makers is as these chips become more complex, more expensive to make, now it makes more and more sense to have a company like PDF Solutions coming in and finding the flaws and so that the processes can be fixed as they go. And so this is creating incredible business opportunity for PDF Solutions. Revenue has more than doubled over the last five years. One of the things I like about it here is that this isn't a recent all of a sudden skyrocket. This is more of a steady line up and to the right for this company. The margin profile is improving as it scales.

12:58Matt Frankel:And so you look at it, yeah, 40 times forward earnings, that looks expensive. But at the same time, I believe its opportunity is only getting bigger. And I believe that its margin inflection is just now starting. So maybe not as expensive as it seems at first glance.

13:12Tyler Crowe:I want to ask a question about this because when I think there's actually this interesting little fiefdom in like the semiconductor industry where, you know, I think of like the suppliers, you know, ASML Holdings. People have heard of this. This is the one that makes those lithiographic machines. They basically have a monopoly on it. And there's other like random processes within the chip making process. Like that there's either one company that makes 80 % of them, or there's two companies and they're always fighting like 40, 60%. I think of like KLA Corp and Lamb Research. When for PDF solutions, are they a dominant market player in their respective field or is this kind of one of the more fragmented parts of that industry?

13:56Matt Frankel:I would say that there are more dominant players in PDF solutions. And there's a lot of sometimes overlap between companies. One company might have a broader portfolio. One might have, so I mean, not always are they directly competing on all aspects with its competitors, sometimes just in one area. And so there are many companies out there doing, I don't know if many is the right word, but there are multiple out there. And so I'm not necessarily saying PDF Solutions is a screaming buy. Definitely one in this area, I would say this area is one that is definitely a fertile hunting ground for potential opportunities because you look at the chip makers, maybe the margins do pull back at that level, but it makes sense that more and more this kind of a company is gonna be in hot demand.

14:41Matt Frankel:And so that's where I'm looking for opportunities.

14:43Tyler Crowe:Matt, you really kind of whetted my appetite because the pitch that you had said it was energy. And then I kind of saw a little bit of your notes. I was like, well, this is a very interesting take on energy. So what do you got?

14:55Jon Quast:Yeah, so I mean, John's absolutely right that the chip making is a big constraint when it comes to building out AI infrastructure. I mean, that's why Micron hit a trillion dollar market cap recently, which was not on my bingo card for 2026. But all the chips in the world, it doesn't matter if you can't power them. So I'm closely watching the energy bottleneck. Data center projects are being delayed left and right. And power constraints are the main reason for that. It's not likely to become any less of a problem in 2027 and beyond. Companies getting ahead of it and actively working to solve the power problem could be in a great position.

15:27Jon Quast:But having said that, a lot of energy sector companies, as you know, are very richly valued or are speculative in nature. I'm thinking the small nuclear startups, for example. One company that is kind of a hidden player in this is Prologis. It's not a hidden company. Prologis is ticker symbol PLD. They're the largest real estate owner in, I think, the world. They've been quietly getting into the data center space with a particular focus on power capacity. So they explicitly state in their proxy filing that power availability remains a key constraint and securing capacity gives them a meaningful advantage over their peers.

16:05Jon Quast:So they have long established relationships with utility companies. In some of their markets, they are the number one customer to big established utilities. And that's a tough relationship to replicate. They're increasingly looking at on-site power generation rather than waiting for utility upgrades, gas turbines, fuel cells, solar, nuclear options, battery storage. They've already deployed over one gigawatt of installed battery and storage, solar and battery storage capacity across their data center portfolio. And I mean, the sites with secured power capacity right now for data center development are trading at pre, they're selling for five to 10 times just conventional industrial land.

16:44Jon Quast:And it's because of this, the power constraint problem. And they have 3 ,000 development, development ready acres in their portfolio right now. And it's, it's a real big competitive advantage that's being underappreciated.

16:54Tyler Crowe:This might be a little bit more of a challenging one, but one of the things, one of the things we're also seeing too, is not just access to power, but also access to water. That's kind of a little bit more of a tough nut to crack here. Does Prologis kind of have similar relationships with water utilities that they do with the powers that this shouldn't be a problem? They do.

17:15Jon Quast:And most of that land I mentioned, it's located near water sources. So their land's concentrated in coastal areas. The Northern Virginia data center market, they actually have a pretty substantial presence there. But no, water capacity is kind of a different nut to crack. And I wonder if after, you know, there are more solutions in place for the energy problem, which is honestly the bigger of the two right now. I'm wondering if water isn't going to be one that we're talking about on this show in six months.

17:44Tyler Crowe:We've said it on the shows before that it seems to be like the supply chain or bottleneck whack-a-mole of, oh, geez, this is the problem now. And before you know it, next time it's going to be insurance or some like weird switchboard piece of equipment that nobody else can find. When the numbers that we see for this industry is getting put up there, Oracle's like, oh, we're going to do 20 more. Google or Alphabet says we're going to do 25 more just because stuff costs more, not because we're building more. You see things like this, these infrastructure bottlenecks, these kind of, you know, constraint companies, these ones that basically hold the keys to the kingdom here are going to be increasingly important.

18:19Tyler Crowe:Probably going to be able to exert an immense amount of price and power. So maybe this will be a nice little recurring segment that we do as we think about this going forward. Coming up next, we're going to hit the nail back.

18:32Jon Quast:In a world full of noise, long-term thinking stands out. On the Capital Ideas podcast, Capital Group leaders explore the decisions that matter most in investing, leadership, and life. It's a rare look inside a firm that's been helping people pursue their financial goals for more than 90 years. Listen to the Capital Ideas podcast from Capital Group, published by Capital Client Group, Inc.

18:54Tyler Crowe:Hey, everyone. Just a quick reminder. If you want to get your questions in and have us answer them on air, email us at podcasts at fool.com. That's podcasts with an S at fool.com. All three rules as always, keep it foolish, keep it short enough I can read on air, and don't try to avoid anything that involves personalized advice. We have to try to keep things relatively generic. So this first one comes in from Neil, which based on the question, I think he's based in Europe. So here was the question. I was fascinated by your episode about European stocks and was baffled as to the huge difference in valuations on the NASDAQ and New York Stock Exchange, basically all the American exchanges, compared to other similar components in the UK and Europe.

19:34Tyler Crowe:There was a very long list of companies he was giving examples to, but I want to kind of conduce down to a couple of the big ones here. It's like, for example, Danish wind farm engineer Vestas Wind Simper. It's one of the largest in the world. I think it's basically a duopoly with them and GE Vernova. and Danish semiconductor manufacturer supplier, which is ASML Holding, mentioned it earlier. Excuse me, Dutch, not Danish. Again, this is a problem that we have for European companies. Stocks barely know where the countries are. His argument is basically they would be worth a lot more if they were, you know, given standing in the American market.