SPEAKER_00 0:00

Hey everyone, and welcome to a brand new episode of the AI Investor Podcast. We've done it. It's not exactly running a two-hour marathon, but we are two weeks in a row of a new podcast releasing on Thursday. I'm really excited about today's episode. Uh we have a the main focus of the episode is going to be we have a fund manager on today named John Rotonti. Why I wanted to talk with John is he's a specialist in areas like industrial. So we're going to actually talk about a lot of the infrastructure going in for not just the build out of data centers themselves, but also the electrical infrastructure. John, by nature, he's someone I've worked with in the past. He's more of a value investor. So, you know, you hear a lot of people on the show, they're going to be what's called growth investors looking for stocks with the highest growth profiles in front of trends. You know, John's going to have a little bit more, he's going to be a little bit more attuned to overall risk in the market. So it's a wide-ranging conversation with John. We touch on a lot of macro issues, what's going on with the market today. We talk about companies he specializes in, and he gets some really compelling pitches for companies we've mentioned on the show, but we've probably never gone very in-depth in. You know, this is areas like Eden or Quanta Services. So if those sound like new names to you, I would definitely recommend give the whole interview with John a listen. I think we cover a lot of ground, everything from uh what's happening in the market today with interest rates, all the way to what John thinks about SpaceX's IPO, which they filed their S1 uh prospectus yesterday. It looks like we're about to have OpenAI filing its S1 pretty soon as well, maybe as soon as the next couple of weeks. So we're really going to see what we talked about in the last episode, uh Cerebrus opening the window for what will be the biggest, the biggest IPL market of all time. I mean, that's just what it's going to be. SpaceX is targeting that $1.75 trillion level. So let's let's just, before we start, 99% of this episode, as I mentioned, is going to be this interview with John Rotanti. But I did want to talk about what's going on in the market. One of the reasons we're trying to do this on a weekly basis every Thursday is to follow along with the market so that we don't have these gaps between what's happening on a day-to-day basis. What we saw in the past week was Friday and Monday. That would put us back to uh, I think May, sorry, I'm doing math wrong, but we're we're filming today on May 21. So last Thursday would have been uh the 14th, uh May 15th, uh, and into Monday, May 18th, um, would have been one of the biggest rotations in two days of momentum stocks we've seen in a long time. So momentum stocks, one that I've recently appreciated, as many of the stocks in the portfolio have, uh, we we basically saw them come back to earth. A lot of names selling off 20% in the matter of uh two or three days. You know, that include names like Applied Opto Electronics, Lomentum, Intel. I had talked about on the last podcast that I was going to, in my personal portfolio, reduce some exposure to some of these momentum names with the appreciation they had and where I want my personal risk profile to be. Now, there's nothing about AI itself. Well, there was one issue. I'll get into that in a moment. But mostly the sell-off, it's it's purely macro driven, right? There are going to be times when people are selling off names in the portfolio off of fears around the overall AI space. And there's times it's just gonna be the overall macro environment. And what we're looking at right now really comes down to interest rates. We're getting a new Fed chair tomorrow, uh, Kevin Walsh, and he was nominated by President Donald Trump. Uh, largely Trump has given uh, even though the Fed is supposed to be independent, uh, he he's kind of given a directive that he wants to see lower interest rates. Now, the problem is sometimes you can say what you want to do with Fed funds rate, but the market's gonna force your hand. And what we're seeing right now is interest rates seeing significant, um, significant gains in their yield. So over 5% right now. And that's gonna make it really hard. You can't cut rates in that type of market environment. And the expectations for the market are flipping from rate cuts into actually raising rates. And again, the longer you invest, the more cycles you go through, you can see how painful often these macro cycles can be. Anyone who invested in 2022 is very familiar with this. We had a year that was very difficult, especially on stocks that rely on um momentum factors that that have higher Ps, higher growth rates, because when the economy bogs down, growth rates come down, investors race to get ahead of that, they sell many of the stocks that had appreciated the most. And um, you know, it can be a situation that's scary. So when we look at what I'm concerned with right now with the market, in terms of what I'm seeing in AI versus what I'm seeing from a macro lens, I'm much more focused on that macro lens right now. But we have seen a little bit of a bounce back. Sometimes moments like this, they just really compel you to look at how you personally handle risk. Some people can handle more risk, some people can handle less. If you're looking on Friday and Monday and getting nervous about kind of a two-day drawdown, that might be an example where you might just be getting a little bit more risk than you're going to be comfortable with. And, you know, that's where a podcast like today, where we're talking with someone in some different spaces of AI than we normally talk about, often looking at companies with different risk profiles that have maybe PEs closer to market averages, like Eden will talk about, or we'll talk about Train, or uh, we'll talk about Schneider Electric, those might be some stocks that you'd want to add to your portfolio if you still want exposure to AI, but maybe you want a different risk profile. Because again, when you own stocks that have appreciated 200%, 300%, 400%, past a thousand percent in a year, that will fundamentally change your risk profile. Looking at the, oh, I should say one of the areas where people had sold off AI stocks that was maybe not macro based in the past week since we filmed last, was we had talked about the 13F uh filing coming out from situational awareness. I'm gonna butcher the guy's name who does it. It's Leo Actionbrenner. Don't don't try and Google that. I probably said it poorly enough. You you won't get it. But he had released his 13F in the past. He had been heavily allocated into some areas that ended up being uh very red-hot areas of AI. He he was early into Intel, he was early into Bloom Energy, uh, he was early, uh relatively early into some things like Lumentum, and he had actually reduced some of his holdings and uh he he had uh some puts that concern people. Now, the reality is 13Fs, which are filings. If you've never heard 13Fs before, they're basically filings that funds have to provide on a quarterly basis that show what their holdings are. The problem is they are a stamp in time. So his 13F was filed on March 30th. We've talked about that specific date before. It was the end of last quarter. It was also the market bottom. So, you know, there there was probably some risk management at that point that there's a portfolio that's historically been filled with things like leveraged options and probably leaning a little bit more into risk management. And we did see some of the smaller companies that he had added um jump 20, 30 percent than the next day as I'd predicted. But they were generally things like um what I would consider a little bit less high quality um neo clouds, uh, companies worth around a billion dollars that are trying to pivot from things like crypto into uh you know selling compute as a service. I haven't been as inclined to add stocks like that to the portfolio. So it's not something that's probably going to change what I would recommend. Now, last, before we get to this interview, I did want to talk about the biggest news in AI this week, which is NVIDIA's earnings. Uh, the bottom line is they're excellent as always. Um, and NVIDIA fell what seems to be always after excellent earnings. I I'm recording this on Thursday, May 24th. They're down, I think, like 2%. It's not a significant drop. And NVIDIA's had a nice run up recently. The bottom line is everything looks great from their earnings. Uh, it's become clear to the wider world of Wall Street that they're gonna do more than $400 billion in top line this year. Their free cash flow across the next two years is gonna be more than $500 billion. So that's my projections. Um, that will be more than Apple, Meta, Google, and Amazon combined. Put them all together. Nvidia's free cash flow, the actual cash that they're collecting will be more. And of course, part of this is those two numbers are related, right? NVIDIA is a giant sucking noise to the cash from the other companies that are hyperscalers. Um, so you know it's interesting when you look at the broader market today. NVIDIA is down, but a lot of the smaller, again, more momentum-based companies. We could look at things like optics, we could look at some of the companies in the energy space, like Bloom Energy, they are uh often five to fifteen percent today. And that's largely based on the enthusiasm for what NVIDIA said about the AI market. You know, NVIDIA keeps talking about one trillion in demand by the end of 2027, but the other number that they're talking about is CapEx reaching three to four trillion dollars annually by the end of the decade. Right? A lot of people thought this three to four trillion number was cumulative. That's actually annual. Is that CapEx from AI will grow from 3x to 4x what it is gonna be next year by the end of the decade. Whether or not we'll reach that number, I mean, that is staggering. You know, you we're to reach that level, we're going to need something more than the cash flows from hyperscalers. You're gonna need debt, you're gonna need new companies in the game, you're gonna need enterprises getting more directly involved in AI spending. Um it is uh simply astounding measure. But for all the estimates and and um questioning on Wall Street of what is the durability of AI, this is what NVIDIA believes. They do believe it would be three to four times bigger than what will be next year by the end of the decade. And in this interview with John Rotanti, we're gonna talk about you know, some of the numbers with AI, some of the some of the long-term buying that's happening, some of how big this can be into the next decade. So once again, I I hope you enjoy this interview. I think it covers a lot of great space in terms of not just some stocks that we've maybe never covered on the show or haven't covered in depth, I should say, but also talks a lot about what this AI buildout looks like. And the bottom line, once again, this week, we've seen this kind of pullback in a lot of stocks. That's more macro oriented. That's going to be due to macro conditions. In terms of what I saw from Nvidia's conference call yesterday, it certainly looks like all systems are go, that demand continues to stay rabid, and the future continues to look brighter, especially when it comes to NVIDIA stock, than what Wall Street is giving this company cred for. So I hope you enjoy this interview with John Rotonti, and we'll be back once again next Thursday as we are moving to this weekly cadence. Um, and one more thing, if you have any questions, anything you want to discuss on the next podcast, just leave a comment, YouTube, Spotify, uh message me on Twitter, reply to me on Twitter. I'm happy to continue doing my best to cover all the areas people are interested in. So, with that, let's begin with the interview. Hey everybody, I'm happy to have John Rotante on. He's someone that I've worked with at the Motley Fool and now is a portfolio manager at Bastion Fiduciary. So, John, number one, thank you for being on. And number two, you know, maybe just share a little bit of a background about yourself, you know, maybe some of your history and also kind of what makes you different as an investor.

SPEAKER_01 13:33

Thank you for having me, Eric, for sure. So um I met you at the Motley Fool. I was at the Motley Fool as a senior analyst for almost nine years. And um while I was there, I was also for about five years the head of investor training and development. And so I helped to train all the new analysts that went through our investor development program. And then as part of that uh responsibility, I also was in charge of sort of networking on behalf of the investing team and bringing in outside experts, um, investing practitioners, but also academics like a Michael Mobison or an Oswat de Moderon to come give guest lectures to our team. And so through that, um I built up a network. And uh yeah, today I I manage an industrial and infrastructure portfolio for clients at Bastion Fiduciary. Um, and I spend my time trying to talk to my network as much as I can about investing and markets and economies and cycles and AI and everything in between.

SPEAKER_00 14:36

All right, sounds good. And John, I know you're a big thinker. If anyone wants to follow John on Twitter, I'll make sure to put his account in the show notes. He he posts a lot of thoughts out there. And one of the main areas that you're looking at is AI. As you noted, you've kind of have this focus, industrials and infrastructure. So I just wanted to let's start at the 10,000-foot view here and talk about what are the biggest questions that you're rolling around in your mind with AI. You know, that could be how it's going to impact the company, uh, companies in the economy, the economy itself. You know, what are what are the big issues right now that you find most interesting?

SPEAKER_01 15:16

Yeah, for sure. So we could go in a million different ways with this, but I'll say maybe um two or three um patterns that I'm recognizing uh across my coverage universe right now, which is, you know, not only the companies in the portfolio, but also on the watch list. Um, number one, and maybe most importantly, is AI data center customers from NVIDIA to the hyperscalers are looking for prefabricated modular integrated total systems or architectures that combine really thermal and power or electrical into one um whole system that is monitored with proprietary software. So it's not just the hardware, it's the thermal hardware, the cooling hardware, the electrical, the power, all in one because power and electrical really um are tightly integrated and they feed off of each other. And so they it's best and more efficient if they work together in one system. But then it's also the software that monitors it and then 24-7 available service teams. And so data centers are shifting from buying discrete parts from different vendors to really integrated systems from leading vendors that can provide um customized, co-designed turnkey solutions across technologies that I that, like I said, are required to work in unison, um, like cooling and power. These these converged systems provide faster build and deployment times, so faster time to power or time to token, um, better efficiency because of the way that thermal and power can be optimized when it when it's in one system. And then I think importantly, the this rapidly increasing technological complexity and customization and the converged nature of these data centers increases the franchise value of having the best global service teams in the industry. In other words, if you build the whole thing from nuts to bolts and it's all modular, it's all prefabed, it's all turnkey, um, you're not gonna go get the you're not gonna go to a third party to have it serviced. You're also gonna get that service contract. And these service contracts provide uh a very long tail of higher margin recurring revenue. Excuse me. And then as the installed base of hardware grows, this annuity stream of very profitable recurring revenue grows with it. So that's the first thing I would say that a huge shift I'm noticing to prefab modular systems.

SPEAKER_00 17:51

Can I just pause there for a second? Please. I'll just be curious. Number one, I had recently visited the Stargate complex in Abilene, and it was amazing. You go to the hotel lobby bar, and it is just almost completely overrun with people who are electrical, you know, electrical by trade, right? It is this shift and you see the value there. But as far as this trend, when you talk about that, you know, some some companies we've talked about on the show would be something like Vertiv, which I know you've been a fan of and and were an earlier investor in. What are some of what are some of the companies that when you talk about this trend, you see on kind of the leading edge, or or some of the companies that you're most interested when you when you discuss this kind of modular design around data centers?

SPEAKER_01 18:37

Verdive is is without a question um probably number one. 80% of its revenue comes from data centers, and it is the largest provider of both electrical or power and thermal or cooling equipment into the data center, inside the data center. Um inside that data center. So we own Vertive in the portfolio. Um Vertiv is competing with companies like Eaton and Siemens. I also own Eaton in the portfolio. And then across my other portfolio companies, this theme of the shift to pre-fab modular systems is um very strong across Vertiv, across Eaton, across Train Technologies, um, which is a primarily a commercial uh HVAC provider. Um and then also uh Qanta, which is an EPC company and uh engineering procurement and construction company, and GE Vernova, um, which is a a they make heavy power generation equipment mainly, um, all are talking about their ability and their unique ability, Eric. They're unique ability to provide integrated modular prefabricated systems to get these customers up and running faster.

SPEAKER_00 20:03

Okay, and uh and thanks for that. I just want to make sure sometimes you know people hear this and it's forming that connection to okay, what's what's the next level of stocks? So let's get to that second thing that's been kind of rolling around your head as well.

SPEAKER_01 20:16

Yeah, definitely. And just really quickly on those companies, like literally, if you go onto their most recent earnings calls, just go to Vertiv, then go to Eaton, then go to Train, then go to Quanta, then go to GE Vernova, and literally do a word search for modular, like it'll come up, or pre-fabricated, or pre-manufactured, it'll come up, and you can and and your listeners can read all about it. Um the second one is um almost all of my AI holdings are working pretty aggressively now to expand capacity to meet unprecedented demand. Um, and they're doing so by expanding the footprint of existing manufacturing facilities, number one, using lean and other um techniques to expand shifts at existing facilities and then making acquisitions to bring on new capacity and new facilities that way as well. Um, one of the ways you win is by having the manufacturing scale and prowess to meet demand in the age of AI, where everything is about um speed to power and speed to token right now. And then the third one, um, which is really a source of competitive advantage, is almost all of my AI holdings are also aggressively expanding their acquisition platforms right now, um, either by making more bolt-ons, more smaller bolt-on acquisitions, so increasing the number of acquisitions, or by increasing the size of their acquisitions, writing bigger checks. Um, and so what's happening is these companies, whether it be Vertiv or Eaton or Train or GE Vernova or Quanta, um they all have what they refer to um as a seat at the table. A seat at the table with NVIDIA um and a seat at the table. With the other hyperscalers that are planning these future data centers and the power and load management from the very first steps. So they have that seat at the table and they're there planning from the very early stages. And what they're doing is they're co-designing and co-engineering the future of AI with NVIDIA, with these hyperscalers. And they're planning two or three generations into the future, Eric. And so these data centers of the future are sometimes referred to as reference designs. But the point is they have they have this seat at the table. And so when NVIDIA, for example, or one of the hyperscalers says, okay, the data centers of the future, we're going to transition to 800 volt direct current, 800 uh VDC to some degree. So my companies may go out and acquire a solid state transformer company, as Eaton did, because traditional magnetic transformers can't do the step down to DC. You need solid state transformers. And so during this process, they learn, okay, two or three years from now, 800 VDC is really going to become a big thing. Let's go buy that capability today. So Eaton goes out and buys the solid state transformer company. Or they may learn during this seat at the table that some of the copper interconnects inside the server rack or the networking between server racks or the networking between data centers may be replaced with optics at some point in the future. So these companies that have the seated table go out and buy an optics company, like Amphenol did, which is another company in my portfolio. Or they may learn, as I just mentioned, that customers are demanding this prefabricated modular all-in-one solution. So as I said, they go out and they buy these modular companies, as Verde did, as Eaton did, as Train did. And so I really think that these deeply embedded relationships are such a strong source of uh intangible but durable mode because they get a glimpse into the future product roadmap and not only a glimpse, but they they're helping to co-design and co-engineer these future products. And so these embedded engineers, they're literally living in some cases at these customers' facilities. These embedded engineers can send immediate feedback to headquarters, and then headquarters can use that feedback to direct RD dollars, capex dollars, or acquisition dollars. Um and that feedback loop informs capital allocation and can drive when when well done really long-term profitable growth. And so there are only so many companies in the unique position to get this seat at the table in the first place, and then to have the balance sheet and the free cash flows to go out and fill in these puzzle pieces. So I think to an extent, the strong get stronger, so to speak, because of this constant feedback loop loop from customers and then superior capital allocation. They use that feedback loop with capital allocation to fill in those puzzle pieces. And, you know, we're seeing that from Vertiv, from at least I'm seeing that, from Vertium, from Eton, from Train, from GE Vernova, from Quanta, and down the list.

SPEAKER_00 25:26

You're looking, we we know that we're rising to the AI demand is rising, and it's gonna continue in years. Uh, the biggest question is the durability of how long this is going to last for. And what you elucidate there too is the advantages that some of these companies have, the structural advantages that make them such interesting plays. And I love the John, that's your answer right there. It's red meat to our audience. We've been talking about this 800 V uh you know cycle that's happening and and using that for what's coming with Ruben. So I think people are gonna really love that answer as well.

SPEAKER_01 26:00

Let's let's talk about real quick on the duration question, just really quickly. Sure. So um on Quantas recent call. So I I own GE Vernova as well, if I didn't mention that. And GE Vernova sells these very large natural gas turbines, which are used to um build power generation plants. Um, well, there's a wait list to get one of these things until 2029 or 2030. So right now, customers that want one or several um have to put down a large down payment up to 20% of the cost of the turbine just to get a slot reservation for 2029 or 2030. Once you have the turbine in 2030, let's say, um, it takes three years to build the power generation plant. The clock on a lot of this doesn't start, Eric, until 2033. These are such long cycle builds and such long cycle products that gives me some conviction that this is a really long duration build. Just because the clock in some cases doesn't start till 2033. That power doesn't start generating until 2033. Um and and and so, you know, on Eton's most recent call, their first quarter 2026 call, um they said that uh there's currently 32 gigawatts of total data center capacity under construction in the US today, 32 gigawatts today. 70% of that is AI data centers. Total data center backlog, total data center backlog, and this is according to Eton. Like literally, they have a seat at the table. They're talking to every customer up and down the supply chain from the utilities to the data centers. Um, total data center backlog has grown to 228 gigawatts in the US only, 228. That would take 12 years um to build that bag that backlog at current 2025 build rates. So once again, 12 years just there. Assuming all 228 of those gigs get built, the general rule of thumb is that it costs 40 billion to 50 billion per gig. So 228 gigawatts, if they all get built, it's like 9 trillion to 11 trillion of spend over the next 12 years, call it in the US alone. Um so it's really big numbers to say, I mean, really, you know, nine to trillion to eleven trillion, really big numbers, and these projects are extended over really long periods of time. One other I'll mention really quickly is you know, there's a general idea that we need to double the electrical grid in the US, double the size of it. Well, it took us 100 years to build the grid we have now. Um most research um that I've come across suggests that it'll take 20 to 25 years to double the grid. Um and so once again, 20 to 25 year build out. I'm I'm I'm I'm I'm I think my point is I'm trying to focus my portfolio on the long the longest duration trends that I can find and the largest markets that I can find.

SPEAKER_00 29:20

Yeah, and so let's maybe maybe click into there there is all this demand, this uh you know, 200 plus gigawatts. What what markets it's serving in a way is the enduring demand for intelligence. And you've had some thoughts about that that you you've posted recently. So maybe let's let's talk about that. Just AI, um putting aside all of the infrastructure that goes into it, AI as a service and what it means to the economy and and how you see the demand for intelligence being something that maybe doesn't have a limit, at least in any practical sense.

SPEAKER_01 29:58

Yeah, it's something I've been thinking about um lately, because you know, an important question is um if we're in an AI bubble, what inning are we in? Um I I I don't necessarily think that's the most important question, but it's an important question. Um and you know, Bezos recently, a few months back, did a conference in Italy, uh in Italy, I believe, um, Jeff Bezos, founder of Amazon, um, and he said, you know, it's an industrial bubble. It's not a financial bubble or a typical economic cycle bubble. It's an industrial bubble. And in that, in that presentation, Bezos said that AI will affect every company on earth. And then he he he reinforced himself and he said, literally every company on earth. And so that just got me thinking, and I'm not the only one that has surfaced this idea, but like, is there a net is there a limit to demand for intelligence? Is there some natural end to this cycle? Um and I'm I'm not sure. You know, I think I think intelligence is is quite possibly the demand for intelligence is quite possibly unlimited. And if you think about AI as um, you know, it is an it is an age defining technology. I think it was a partner at um Sequoia Um said that it's the most important technology of the of the next 50 years. And I believe that. That's kind of how I think about it. I think that you know, AI um in combination with our um vast resource of very cheap natural gas here, um, so if you take this this AI plus our our res our energy resources here is going to be the technology that that um could potentially, and I think at this point, this is what I believe, but I could be wrong and I could change my mind, but could potentially usher in an industrial revolution. And so it revolution, I really mean the word revolution. Revolutions don't, you know, they don't end in three years or four years or five years. Revolutions, you know, last decades. And so that's kind of where my thinking is right now. You know, agentic AI uses 1,000 times more tokens than a simple LLM query. Um Bernstein put that note out earlier this week. Um, so I do think that token use usage is going to explode higher. Um and and you know, I think that there's a decade-long, decades-long build to this industrial revolution, which um, you know, AI is the enabling technology of it, and then it will, you know, it will include not only build out of the electrical grid, which I touched on a little bit, and these data centers, and and all of the new businesses and new discoveries we haven't made yet, right, that come from AI, but then also physical AI, robotics, um, and then and then space. And you know, I think Elon Musk and and and both Jensen Huang have said that sort of physical AI could be a $50 trillion opportunity. That number is so big that I can't like, I can't tell you if it's right or wrong, right? But it it just it's a really big opportunity, I think is the point. Um and so yeah, I think I think it could be the technology that ushers in an industrial revolution.

SPEAKER_00 33:29

Let's maybe pull back from I I would like to get some, I I love these high-level questions, so maybe we'll circle back, but I wanted to talk about two, with what's happening in the market today. We've seen, as we've talked about a lot on this show, it's it's felt very frothy, especially coming out of this run where you had the semiconductor index, I believe, uh closing in the green for 18 straight days. The previous record was nine straight days. We've suddenly seen a little bit of a reversion this week as we're filming this on uh May 19th, largely driven by rising interest rates. So what what are you watching on a macro level this year? Does regime change at the Fed impact how you're thinking? What what are the areas that you're looking at? And also I would ask, how do you feel investing at this moment um with with where stocks are trading relative to this kind of backdrop, but with with rates and midterms, a lot of uncertainty in the market? I feel a lot of that uncertainty, Eric.

SPEAKER_01 34:29

Um, and so I deal with it. Um I have 30%, roughly 30% cash in the portfolio. That's that's one way that I deal with it. Um as as far as what am I focused on, I think most importantly I'm focused on uh the price of oil only because um historically every time that the price of oil has doubled in a very short amount of time, we've gone into a recession. Um now the price of oil is not up 100% just yet, but I think it's up like 80% since before um the conflict in Iran started and the Strait of Hormuz closed. And so we're we're bumping up against that 100% mark. Um and like I said, that pretty reliably puts us into a recession. And the reason is because um, as you know, about 70% of the US economy is um consumer spending. And when prices at the pump go up, people feel pinched and they have to pull back spending elsewhere. That's that's largely why. Um another thing I'm really focused on is um regime change from Kevin Walsh, who's gonna be um taking over in the next few days, I think. Um we we don't really know how that regime change is going to play out, though. I mean, I've read all the stuff on Warsh, and we know that he thinks that um we can lower rates um without stoking inflation because he thinks AI is so disinflationary. Like I understand that. Um, but uh right now it looks like it's going to be hard for him to lower rates. And if you look at futures markets and where the two years, the Fed has historically followed the two-year um treasury, and the two-year treasury is now higher than the Fed funds rate. And so that would suggest that we need to actually raise rates. Um I don't know if he's gonna raise rates, but I don't think his initial plan of lowering rates is gonna be as straightforward and and easy um as when he first proposed it, once again, because the conflict in Iran and tariffs and maybe a few other things um have sort of stoked inflation a bit again and and rates are rising. And so I I don't know what that regime change is going to look like. Um I will say the last three last three Fed chairs that took over, um so going, yeah, POW, but then going back two more, their first move was actually a rate increase. Um I don't know what that means, but just historically putting that data out, that data point out there, their first move as new Fed chair was to actually raise rates. So we'll see what happens with worse. But there's some uncertainty there. Um and you know, if he is so intent on lowering rates, which is obviously what the White House wants as well, um are they gonna let inflation run hotter? Are they gonna be more comfortable with inflation running hotter? And so there's some uncertainty there. The f you know, watching the 10-year um inching up closer to 4.7% and then 5%, 5% is this like psychological level where I think a lot of people go into their models and increase their discount rates and or increase their required rate of return, their hurdle rate. And obviously, when you do that, it it um decreases the present value of free cash flow. And when you decrease the present value of free cash flow, you decrease intrinsic values. And so um I think 5% is this psychological level where stocks could maybe sell off. And so oil prices want to see what happens with Warsh once he really takes over that share, and then the 10-year approaching 5% are sort of the macro themes that I'm watching. Um really quickly, since we're on macro, you know, I talked about how right now my base case is sort of that AI is going to usher in this industrial revolution and this infrastructure build out could last decades. That's sort of my base case right now. That in no way means that I don't think we could have a typical like economic recession in between now and then. You know, um we're maybe do. It's been like 18 years since we've had a real economic cycle-driven recession. I don't count um the COVID pandemic because that was like a hopefully once in a hundred year like you know, event. Um, but we haven't had a real economic recession since 2008. Um, so like, you know, it's now also a data point is markets don't die of old age. Um bull markets don't die of old age. Um, but you know, if that oil price continues to sneak up and if that five-year continues to sneak up, and um I just think maybe some things are starting to stack um that that could you know could put us into some sort of you know downturn or recession at some point in the future. Um, and so I definitely think that we could have this decade-long infrastructure build-out, but also have like typical economic cycles happening in between.

SPEAKER_00 39:34

And maybe one more macro question, then I'll get into a few more long-term um ones. But I I was curious too. We just had the summit with President Trump and uh uh Qi from China. There is starting to become this idea of not just AI as a technology, but also uh one of the leading policy areas for the world. And we have this almost emerging Cold War with kind of a Chinese tech stock and a US tech stack. Um, we saw uh Jensen Wang, he he he caught he caught the plane in Alaska because I think Trump uh overruled his advisors at the last minute, and that led to a a bit of a rally in NVIDIA stock. So I was just curious how you think about this kind of emerging Cold War and what implications that might have within um, you know, not only the stocks in AI, but where the trend heads in the next five plus years.

SPEAKER_01 40:29

Yeah, so I've definitely um I definitely think there is this emerging tech cold war between the US and China. I mean, I can see it when I read everything and when I watch the news and things like that. Um I'm not an expert on US-China relations, but what I think it means, Eric, for from it from an investor standpoint and a tech investor standpoint, an AI investor standpoint, is um, you know, I think it's good. I think it's a tailwind. Um, because I I do think that there is this tech war between the US and China. And I do think that um we eat you know, both sides have some advantages, which we can discuss. But I I think most importantly is both sides sort of see it as existential in nature, and as you said, a matter of policy because it's a matter of national security. Um and so when things reach the level of national security and policy, um, you know, countries can throw a lot of support and money behind it, um, and and laws in the form of policy. And so I I think it's a tailwind to the sort of broader AI infrastructure um investing ecosystem. Um, you know, the US has a clear um indisputed lead in semiconductor design intellectual property. Um China, however, has an indisputable lead in the amount of available power generation. Uh China is also uh they have deployed far more robots than we have um so far. And uh China has more manufacturing capacity, and they are better manufacturers than we are at this point. And so um there are advantages to both sides. I don't know, I don't know um how it's going to play out, but I think it's an it's a tailwind for investors.

SPEAKER_00 42:30

Well, let's let's double-click on that because one thing that you've been talking a lot about in your tweets, and you've called it as far back as 2022, is this need for building stuff, actual manufactured goods and a return to that. And that's part of this, that AI can be a renaissance in areas of manufacturing, especially for uh, you know, many of these Western companies. So, what what are you seeing in terms of we've kind of had this wave of AI where we've got Chat GPT and then reasoning and agentickets all on the model side. And there's probably a coming wave. You you teed up $50 trillion in physical AI. We we see actual stuff with the task tires themselves, obviously. But what are you most excited about in this next kind of hardware evolution of the trend?

SPEAKER_01 43:18

Yeah, so you know, you threw thank you for for pointing out those those posts from back in 2022. Um so you pointed out, you know, that sort of like high-level $50 trillion physical AI. Um, but if we just come back, come down a little closer to Earth, because that's like so up there, um, Eton, once again, in their most recent earnings call, says there's currently a backlog of 3.3 trillion in mega projects across the US. That's just the US um announced announced mega projects. So that's a that's a big number. You know, going back to that grid, um, building out this grid, you know, that's we're talking about like a four. I've read anything from like three to you know five trillion dollars to get the grid where we need it in the next you know 10 to 15 years. And so these are really, really big numbers um extended over a long period of time. Yeah, you know, I just think that I think I think that so in 2020, as you know, um 20 thereabouts, um the market was myopyally focused on software as a service. And you know, going back to 2020, and I've I've posted about this as well, um, I saw issues with the software as a service business model um and their and their and their management and their governance. Um you know, most importantly is it the business models were based on stock-based compensation, largely based on stock base, or at least partly based on stock-based compensation. The actual business models were based on this concept, but the concept only worked when stock prices were going up. And so early on, I was worried about um what happens when stock prices start falling and that model breaks down. And we're sort of seeing that now. Um But the other the other issue, you know, going back to 2020, was a lot of the applications, software applications that were sold as a service um weren't critical or indispensable to the world or to the US. They weren't an asset to the US. You know, you you hear all the sort of cliche, it's a dog walking app. And um, they were not that that's not important. I love dogs, I had three dogs at one point in time, but like it's not indispensable. And so, you know, and so you you got to a point where you had like two dozen, you know, human resources apps, all doing the exact same thing with a slightly different user interface, UI, and they were just the same. And the same thing with marketing, and the same thing with legal, and the same thing with accounting, and and on and on. Then you, you know, deliveries, food delivery services, and none of that was really like indispensable or critical to national security or to our way of life. So that was my second issue with SaaS at the time. And then at the same time, Mark Andreessen put out this white paper, which I recommend everyone reads, and it's like we need to build, or it's time to build. And it was so significant to me when it, and this was in 2020, when Andreessen put this out because this is the same dude that wrote back in 2014, I think, software is eating the world. Right? And like he's more connected than almost anybody, and I think he's just brilliant. And so in 2020, I went down this, I read that paper from him that the US is like incapable of building. Like we couldn't build like raincoats like during COVID. It was embarrassing, it was embarrassing. Yeah. Like smocks, like, you know, for hospitals and stuff. And and so I went down a rabbit hole and I and and a deep rabbit hole, and I haven't come out yet. And basically what I what I learned was that the US underinvested, US and other parts of the world, but underinvested in infrastructure basically writ large across the board, from housing to energy, um to you know advanced manufacturing, whether it's tech, whether it's pharmaceuticals, um, life sciences, and things like that. And uh yeah, and so when when when Cale Smith, my boss, and the founder of Bastion Foodshare, reached out to me and we started talking about me coming on board, he asked me what type of he wanted some thematic theme to the portfolio. That was the only guardrail he gave me. Um and so when when he asked me what thematic theme I wanted, I just said industrials and infrastructure, because I think it's a um it's a long duration opportunity.

SPEAKER_00 47:48

So let's let's look at two, you know, you've you've been mentioning this time to build. We we kind of have an exciting era coming up. We're about to see a $2 trillion, I I I can't even say $2 trillion IPO, uh, you know, 20 times the size of Facebook for something like SpaceX, but we're also probably going to see things like Andreal, yeah, which is a company that's building real things for the defense industry um coming up this year. So are are there any industries or companies that you're pretty excited about that you see that maybe they've been on the backlog, they've been pre, um, you know, they haven't been a public company, they've been private that that you're interested in or that you think is going to broaden this opportunity set beyond what investors have been able to buy so far in these kind of AI adjacent companies.

SPEAKER_01 48:41

Yeah. So I talk about this with my team so much, Eric. Honest to God, true story, that I was just like, we're a small team, seven full-time employees. And I know you're very familiar with Slack. Um, well, I was posting about how bullish I am on space in our general team channel. Because I couldn't help myself. And it's like every day I'm writing something new about space. This is like going back now, I don't know, like six months maybe. Um maybe four months. And so finally, my boss, Cale Smith, who I love, he's like, let's move this to another channel. Let's just do like, let's just do the John Musings channel or something.

SPEAKER_00 49:20

John Space channel.

SPEAKER_01 49:21

Yeah. And so, yeah, I mean, um to answer your question, um, I'm very excited about the SpaceX IPO. Um, I don't think that I'm coming at this from a two million is such a big number. I get that. But I I don't think that I'm coming at this from a position of FOMO. I really don't. Like, you know, like I said, like I never got on the SaaS train for better or worse. I mean, I mean, Eric, when I was at the Motley Fool, every analyst um managed a model portfolio and something that we called Fool IQ. And we measured how good of stock pickers we were and everything over time. And uh that started in 2014 when I started at the full. And so I had eight and a half years of managing this model portfolio at the Motley Fool. I never put one SaaS company in it. Never. So that's just an example of like I think that I'm aware of like FOMO and I think I do a good job of like controlling my emotions. SAS was never something I got excited about. Web3, never something I got excited about. Virtual reality, never something I got excited about. Um, so I'm I'm usually pretty good when it comes to like you know keeping my senses and my emotions, but with space, I cannot help myself. Like I I can't, like I I am very excited about the opportunity. Um, and so yes, SpaceX and Andereel are the two companies that I'm going to dive into first. Um and you know, I the way I see space is most important. If I had, if I had, if there was like one takeaway of space for me, it's that it's that the smartest best builders in the world, in my opinion, are just going to will this into existence. So, you know, like who's doing it? Elon Musk, Jeff Bezos, Palmer Lucky. Like, these are the smartest best builders in the world, maybe that the world has ever seen. And I just think they're gonna will this into existence. I really believe that because they have the capital, they have the following, they have the brains, and they have the capability. A lot of people have ideas, but these people can bring those ideas to fruition better than anyone I've ever studied. Um, and so I just think they're gonna will it into existence. Um, and I think that there's also a lot of economic reasons to do things in space. Um, and the way I see space today is the US before the Industrial Revolution. Um, the the first original industrial revolution. And so, you know, today we have factories on Earth. I think we're gonna have factories in space. Token factories, so AI data centers, but also factories that make parts for the other infrastructure that we have in space. You know, whether it's three 3D printed, manufactured by robots, additive manufacturing, I don't know, but I think we're gonna have factories in space. You know, today we have mining operations on Earth. I think we're gonna have mining operations in space for rare earths and minerals and all of that stuff. Uh obviously defense. You know, we have massive defense installations and weapon systems on Earth. I think we're gonna have that in space. Lasers, you know, whatever it is, not only for defensive purposes, but also to take on asteroids and things like that. I think that so we have transportation networks on Earth. I think we're gonna have, you know, landing pads, you know, whether it's on the moon and Mars or wherever, maybe somewhere in between floating. I I I'm not an astrophysicist, but I think we're gonna have these transportation networks in space. Um and and and you know, so everything that I not everything, but a lot of the infrastructure that we have on Earth, I think we're going to um eventually try to build in space. And space is a big place, and you know, it's not gonna happen overnight. I don't even know if it'll happen. I don't even know if I'll see a lot of it in my lifetime. But you know, I think this could be like a 50 or 100 year build-out of the infrastructure of space.

SPEAKER_00 53:20

Yeah, it's uh deeply fascinating. So when you're looking at SpaceX, and you know, I know they still need to file all their documents, and there's still a lot of unknowns. But what in the near term, what are you most interested in? Is it going to be Starlink? Is it this next evolution for data centers? I know you talked about that. Are you a believer that data centers in space will eventually um be something very material? I know we're about to start getting a ton of SpaceX questions. So when I have the opportunity to talk to smart people, I just want to, you know, get their take on uh what they're most excited about uh of all these parts with SpaceX, right? Because there are all these moving pieces.

SPEAKER_01 53:59

I think Starlink in the in the you know intermediate shorter to intermediate term is the strongest part of the business and maybe the most important part of the business. Um, you know, Ron Barron, who's obviously a huge SpaceX investor, um I think he's got one of his funds like 20 or 30 percent SpaceX, literally. Um you know, he says it's gonna be the internet for everyone on earth. And I I don't know if that's true or not, but I think that's the most exciting part of the business right now. But the most just at a higher level, the most expi exciting part for me is is Elon Musk. I just there's not a lot that he can't do. I'll be the first to admit, maybe it doesn't all happen on the timelines that he sets publicly, but also like he's also done a lot of things that I never thought I'd ever see in my lifetime. Um, and so you know, this is just a bet on Elon Musk. I've talked on some other podcasts and in my in my letters to investors and stuff. The most important thing that I um focus on when investing money on behalf of my clients is the management teams. And so um that's the main reason I'm invested. I mean, I'm interested in SpaceX, it's just Elon Musk.

SPEAKER_00 55:15

Gotcha. And and you know, John, too, let me know if there's anything that we haven't covered that that you think is really important. But I I would be curious too. You you've talked about a lot of the infrastructure. We had in a recent episode talked about many of the actual construction firms, things like Mcore that are building out this. Is that an area that you've been interested in? Um, because it it does lead to exceptional management teams really delivering exceptional businesses. So I just wanted to get your take on that uh quickly if you have it ready. Absolutely, man.

SPEAKER_01 55:45

I'm so glad you asked. So um I think, you know, as of like yesterday when I looked, Quanta is our largest holding in the portfolio. Um so Quanta is the largest engineering, procurement, and construction firm in the US focused on utilities, primarily transmission and just distribution or TND. Um, but it also works on a variety of other projects like natural gas pipelines, LNG facilities, solar and wind farms, and now data center construction. Um and even and even it's through a joint venture, it's getting into the construction of um NAT gas power generation plants. Um it has vertically integrated Quanta. Um the ticker there is PWR. Um it has vertically integrated and now Bernstein estimates, it it it may it manufactures 20% of the transformers in the US. This is really important because transformers are uh a main bottleneck. There's lots of bottlenecks. Um quanta solves for a lot of these bottlenecks, which I'll get into, but one of them is transformers, which can have like a three to five year um wait time. Um and so they now manufacture 20% of the of the transformers, sort of medium voltage and lower in the US. Um it's now a vertically integrated solutions provider to energy and tech companies. Um, utilities, as I said earlier, are being asked to double the size of their business because there is a general idea that we need to double the size of the grid in the US. And Qanta is a direct beneficiary because they have built over 50% of the long-distance transmission lines in the US and 15% of all the TND transmission and distribution in the US. Um, so they have they have massive market share and very long-term uh relationships. I talked about management CE, their CEO, uh uh Earl Austin, they call him Duke. Um, he's exceptional. Um, he says flat out that the electrical grid build out is a decade plus. Um and that's be and you know, they need we need to do this grid, we need to build out this grid, not only um because of like extreme weather events, and because it's already maxed to the load, we're already getting an increase in blackouts, but it's also everything we talked about, Eric. It's it's to support AI, reindustrialization, decarbonization, robotics, and then the broader electrification of everything. Um so, you know, like I said, to double the size of the grid, you know, you're you're talking about 2 trillion to 5 trillion over a over a 20-year period. So it's a very large, very long duration uh build, and that's where I focus my attention. So um what I love most about uh Quanta is their customers, a lot of their customers are utilities, and utilities have regulated programmatic spend, basically, and so they grow um alongside utilities. And every utility that I've looked at, maybe a dozen, um, have all increased their five-year capex numbers by like 15 to like 40 percent. So utilities are growing, they're asked to double in size, we need to double the grid, and that's programmatic regulated spend, and so they grow alongside those utility customers. Um and so their business is sort of like supported or buoyed by this, by this like floor of annual spend from utilities, and that makes the business um pretty predictable. They recently had an investor day. Um, they gave five, so this is sort of like my investment thesis, the numbers part of it. So they gave uh five-year guidance, and they just had this maybe a month ago, to at least double EPS in five years. And so they said EPS will grow by a 15 to 20 percent cager. But Duke went out of his way to say that's very prudent and that they think they can stack above that. When he says stack above that, he means he means he thinks they can beat that number. Um, I think they're gonna beat it by a lot, like a lot. So I think it's a pretty low ball guide. In the 10 years, Eric, from 2015 through 2025, Qanta's adjusted EPS grew at a 25% kger, 25% kegger. They just they just guided to a 15 to 20% keger for the next five years, but in the last 10 years, they grew at a 25% kger. And that was during a period where electricity demand or low growth was flat. 0%, basically, thereabouts, around 0%. Now load growth in the US is expected to grow around 3% over the next five years. So this is a step change in a demand tailwind. And so I think they're gonna be able to grow their EPS by at least 25% over the next five years, not the 15 to 20% they said. Um, and I think they're gonna raise that medium-term guidance, you know, in the next, let's say, one and a half to two years. The stock looks rich. Um, it's it's up about 100% in the last 12 months, but that's not nearly as violent, as you know, as some of these other like memory stocks and optical stocks and bloom energies of the world that are up like 800% or a thousand percent or more in the last 12 months. So it's only up 100% in the last 12 months. Um, trades at a next 12-month PE of 50, which is rich. It's rich for my blood. But I think it's justified because, like I said, the business is predictable because their regulated customers spend programmatically. Um, and because this is an extended super cycle that I think could last 20 or 30 years as we get the grid to where we need it in this country. And when you pay a really high multiple, ultimately you're making a bet on the duration of the per share growth. And so I, you know, I do think that that this could be a very long duration per share grower. And so that's that's that's one of the places I'm focusing my attention.

SPEAKER_00 1:01:49

Well, and you look as well too, yes, it it a lot of these stocks are priced rich, especially relative to their historical norms. But in an environment like right now, where um we have this macro overhang about rates, you know, this these are always companies too. People can add to their watch list. And if if we have some of these sell-offs, it looks like Quanta's down uh eight or 10%. I'm just eyeballing it across the past week. That might be something, you know, investors bump up to their list, especially if they have a higher allocation to some of those areas you mentioned, like memory and optics that we know are are going to be extremely volatile. Uh, how about this? One question. I know on these shows it is there's so much cool stuff with AI and mind-blowing things. Uh we 100% on this show share your review, John, that this will be the era defining technology. But I am curious, you're a very thoughtful person and you are very serious about your risk. I'm just curious from you, what kind of keeps you up at night? What if something were to go wrong in the next 18 months, kind of with this thesis around these companies, what is the area that you'd be most concerned about? Is it is it something like um a broader recession that takes um the ability to fund this next phase out? What are you worried about that it may not stop the eventual future of AI being the most exciting technology, but it could make these kinds of ups and downs a lot more accentuated or or uh introduce a lot more risk than we're currently planning on.

SPEAKER_01 1:03:26

Yeah, I think I think an economic recession is kind of the biggest risk. So, you know, the two numbers that I probably follow most closely or the two metrics would be hyperscaler capex. Um, you know, those hyperscalers are um sort of you know five companies or whatever you want to call it, uh single-handedly supporting a lot of this AI trade. There's no question about that. And, you know, they continue to increase their capex at um just sort of like mind-blowing rates. And, you know, I think you know next they're gonna do like 750 billion in capex this year. Um next year, you know, 2027 estimates are for somewhere between like 800 billion and a trillion in one year, in one year. Um from five companies. I mean, it's it's it's it's mind blowing. It hurts your hair to even think about it. But that's the first thing I watch is now though now they cannot continue to increase the rate of capex um as much as they have been indefinitely. So the first thing I think we're gonna see at some point is the rate of increases um become smaller. So, you know, rather than increasing capex 60% a year, they increase capex 20% a year and then 10% a year. So capex continues to grow on an absolute dollar basis, but the rate of increase slows. And then you know, at some point we may even see capex drop. Um but if that happens in an economic recession, I still think the number like let's say CapEx drops from 700 750 billion to 600 billion, that's still a lot of spend. Now, it may not be enough to placate the AI crowd in the short term, and I do think a lot of stocks will sell off, but I think that's a buying opportunity if that happens. Um but you know, 600 billion hypothetically is still a huge number. And if you look at the history of CapEx from the hyperscalers, um going back to the dot-com bubble, or before then, um, CapEx almost increases every year. Almost every year. And when it does fall, it's not but it's not that severe, that significant. And so, you know, I think we're gonna continue to see very, very large numbers from the hyperscalers. Like I said, those five companies are very important to this AI ecosystem. They're basically holding it up themselves. Um, But if I'm wrong about that, that's a big risk. Um, the other thing that I watch closely is the shortages. So it's it's you know, shortages in memory, shortages in optics, shortages in CPUs now, um, shortages in labor. Um, and so as long as we're talking about the markets and investors are talking about shortages, I think we're okay. Once you know that flips and we're um we're no longer in a uh supply constrained, you know, I think we could have some some mini bubbles uh along the way. Um and so those are probably the the two biggest risks I'm considering is um how long do the shortages last and then how long can the capex stay growing at the rate that it is. Um and then you know, my portfolio at the portfolio level is you know average PE of 30. I think the PEs are justified based on high returns on invested capital, um, recurring revenue business models as far as their services business, higher margin, um, and you know what I think are very long duration builds, and so long runways of per share growth or long profit cycles. So I think that the the PEs are justified. And even though they may be trading at 30 times today, they're growing quick enough that that drops to 20 times in like two or three years. And so if you use out-year numbers, they look much more reasonable. But yeah, there's some valuation risk to my portfolio as well.

SPEAKER_00 1:07:35

And um, well, number one, I just wanted to thank you for being on, John. This has been a great uh discussion. I just wanted to see, too, you know, are there any other areas we haven't covered today that but before we end this, you're like, I just really want to make this one point or anything um before before I let you go.

SPEAKER_01 1:07:54

I'll just, you know, one of the things I think about is how do you invest in AI um when so many stocks have gone hyperbolic, right? And so I talked about Quanta, which hasn't gone quite as hyperbolic, but maybe, you know, two others is really quickly Eaton. So we mentioned it several times on today's show, but um, it's the largest supplier of electrical equipment in North America with a share of around 30%. Um and you know, back in 2025, Bank of America put out reports saying it was the third largest vendor of electrical products going into the data center. So that's behind Vertiv and Schneider. Um they they sell things like transformers and uninterrupted power supplies and switch gear and circuit breakers, bus bars, liquid cooling, and on and on. Um Eaton also has an aerospace business that sells fuel pumps, hydraulic pumps, pneumatic systems. These parts have an extremely high cost of failure because they're in in planes, um, and they're generally spec in for the life of the aircraft, which is measured in decades. And 40% of its aerospace sales are recurring aftermarket replacement parts. And the way that works is the FAA mandates that aircraft parts are replaced after so many flight hours. And so what really happens is these planes, um commercial aircraft, basically get stripped down every like seven to ten years, and everything in them gets replaced. Um, and so 40% of their aerospace business is these aftermarket replacement parts. Um the aerospace business is driven by a 10 plus year backlog for commercial aircraft and increasing defense budgets around the world. Today, 90% of Eaton sales come from electrical and aerospace and defense, but it has an and 10% come from this e-mobility business, which is like losing money or not making any money. Um, so they've they've announced they're gonna spin that off. And so, really soon, like I think Q1 of 2027, 100% of their sales will come from higher margin, faster growing um electrical and aerospace and defense. Um I think only about 20 to 25 percent of their of Eaton's total sales come from data centers today, but that no number's growing very rapidly, Eric. So here's a stat. It's data center orders were up 240% year over year in Q1 2026. Um, and Eaton's products are are helping to power and cool Nvidia's latest generation Vera Rubin AI server racks. So they are in Vera Rubin. Their orders are growing 240%. Um it's a good company, it's a good company, but you know what is it? Like the stock is up 12% in the last year, it's up 12% in the last two years, and it's only up 150% in the last five years. And so it's it like sort of has been glossed over in a way, even though it's in Vera Rubin and its orders are growing 200%. Um it's only up 12% in the last two years. So, you know, that's something it's trading at next 12-month PE of 27, um, which is low for the like AI ecosystem. And and like I said, that that drops to 20 in two or three years. And then, you know, one more quickly, sort of the same story, train technologies, ticker TT, we own it as well. Um, like Eaton, it's a broad industrial technology company. I think some people sort of reflexively think of train as a residential H VAC company. Um, but you know, train is all train's residential business is only 15 to 20 percent of the overall company. Um it also so so let's say 15% residential. It also owns Thermo King, which is the largest refrigerated trucking or refrigerated trailer business in the US. That's 15% of their business. So 70% of their business is commercial HVAC. And for that, they sell into 14 different verticals, only one of which is data centers. Um, but its data center business is growing very rapidly, both organically and through acquisitions. And so it's um it's a they call them applied book. Applied are these very complicated custom HVAC systems as opposed to simple systems. So they're they're customized, they're very complicated. So it's applied bookings in America's in their America segment, which is where their AI data center revenue is generated, grew 160% in Q1 2026. And that was the third consecutive quarter that bookings have been over 100%. So another sort of no one really thinks of train as an AI company, but its AI business is going growing really, really, really quickly. They recently made an acquisition of a company, they made a liquid cooling acquisition, um, but then they also made an acquisition of a company that makes modular, back to that prefabricated modular theme. It makes modular chiller farms or chiller plants, and it's called Stellar Energy. And right now, Stellar is doing $350 million in revenue. And um the CEO of Train, Dave Regnery, says this will be a $1 billion business in two to three years. Um, so it's just growing really, really quickly. A third of its business is um higher margin services, and those services are equal to eight to ten times the original cost of the equipment. So it's a very long annuity of service higher margin services revenue. Um same thing. The stock is up 4% in the last year. Um, it's up only 150% in the last five years. Once again, like not even comparable to some of the other AI names. And the PE is in that 30% range again. But once again, you know, it drops down to 20 or so on a few years, few years out. So that's kind of how I avoid the mania. I'm looking at companies like Train, like Eaton, like Quanta, which haven't gone hyperbolic.

SPEAKER_00 1:13:54

Yeah, I really like that. And you know, as as I noted earlier, there are a lot of people listening to this podcast that were earlier in a lot of these names that had major movements elucidating these areas. You know, you look at Verdev, that one has gone relatively parabolic.

SPEAKER_01 1:14:10

But you know, and and I put and I and I pulled back a little, Eric. So I own it, and I don't I hope I never sell that thing. I'm just I'm you know, they say never fall in love with a stock, but I sort of love that company. Um but yeah, you know, it when when something goes up 250 or 300 percent in 12 months, I I trim it, I trim it back for um, you know, risk mitigation purposes.

SPEAKER_00 1:14:31

Yep. And so these are some alternatives, similarly positioned companies, but ones that I haven't seen, you know, this this extreme run up. So, you know, if you are looking to b balance your portfolio, people are probably getting a little bit of a gut check in this past week with with the market. You know, it's it's in moments like this you're gonna be able to assess how you handle risk. And one way to be able to reduce your risk is, you know, to spread out to some of these names like an Eatin or a train, as you mentioned. So, John, I just wanted to thank you for coming on. It's it's been almost an hour. I know you have to get going. Um, so thank you so much for coming on today. Uh, we really appreciated all the thoughts that you've shared.

SPEAKER_01 1:15:11

Eric, thank you so much for having me. Um, I wish you and your show the best of luck.

SPEAKER_00 1:15:15

All right, thank you. And with that, we'll close off this episode of the AI Investor Podcast. We will be back next Thursday.

SPEAKER_01 1:15:23

The AI Investor Podcast is for educational purposes only and should not be considered investment advice.