Motley Fool Hidden Gems Investing - AI Gets Star Power

Episode Date: January 23, 2025

...as if it didn’t have enough already! (00:14) Asit Sharma and Mary Long discuss: - The new venture to build out American AI infrastructure. - How 20 data centers get a $500 billion price tag. - GE Aerospace’s razor-and-blades business model. Then, (19:15), Seth Jayson joins to walk through why the rooftop solar industry doesn’t look so sunny. Companies mentioned: MSFT, ORCL, NVDA, GE, ENPH, SEDG To become a premium Motley Fool member, go to www.fool.com/signup Host: Mary Long Guests: Asit Sharma, Seth Jayson Engineer: Rick Engdahl Learn more about your ad choices. Visit megaphone.fm/adchoices

Stop wasting your nights on a mattress that doesn't get you. Experience the most comfortable mattress in the world. The Sleep Number Smart Bed. At the touch of a button, you can personalize your comfort. Choose firmer or softer. Adjust cooler to warmer. And right now, save up to $2,500 during our massive Labor Day event. Hurry into your local Sleep Number store today.

Because we have your number. we're headed up to the skies and the stars you're listening to motley fool money i'm mary long joined today by the one the only asit sharva asit thanks for being here on this lovely thursday morning mary i am excited for this conversation as am i we're going to kick things off today because there's a new partnership in town. I'm going to use that word because I don't fully know what else to call it. And it's allegedly worth about $500 billion. Put another way, that's half a trillion dollars. So a lot of money. Yesterday, President Trump announced a new AI venture that brings together some big names, Oracle, SoftBank, and OpenAI being three of them.

Much has been made about who's in and who's out of this project. I got a lot of questions actually about the project itself. So let's start there, Asit. What is Stargate and why are these three companies the ones that are at the helm of whatever this really is? So Stargate, Mary, is nominally a joint venture, some type of partnership that is aimed at building out AI infrastructure, nominally data centers. There is a $500 billion price tag associated with this build out that we think will consume some capital expenditure between a few companies over the next five years. Other than that, the details are sparse. The actual announcement was handled by OpenAI, which was and is a key player in the AI landscape and is going to be a key player

in this project. But as you mentioned, there isn't really a detailed roadmap for exactly what this joint venture is supposed to achieve. So maybe the best way to talk about this is for you to throw your questions at me, and I will try to answer them. So in the announcement that OpenAI put out about this venture, they noted that SoftBank has financial responsibility. OpenAI is going to oversee operations. But Oracle was a big mention in this rollout. Do you have any sense of how they're going to fit in to this venture? So Oracle is probably going to be one of the leads when we think about building out those data centers and the know-how that's involved with bringing together servers, softwares,

networking, hardware. Oracle's very good at this. Of course, they were the preeminent database company for many years, famously thought that the cloud wasn't going to be a big deal, and then basically reinvented an approach to cloud computing and now present themselves as a very good competitor to Amazon Web Services, Microsoft Azure, et cetera. And Larry Ellison himself, the chairman of Oracle, is a visionary. So I think he's a good point person and Oracle is a good point company to have as an infrastructure partner in this deal. And I guess I did gloss over what SoftBank is doing. So SoftBank Group,

This is the big investment company, Japanese, helmed by the one and the only, the truly one and only Masayoshi Son, who is a big venture capitalist and has been on the scene for many years. He has a tendency to invest in companies very early on, Mary, and he's all about scaling unit economics. Son has had a number of successes, but he's also had some prominent failures in the venture capital world. But I would say overall, he's a respected partner here. And then, of course, we have OpenAI itself, which is, as you mentioned, supposed to be overseeing operations. OpenAI is somewhat capital constrained itself. It is the company that is developing large language models, and it gets most of its coin just now from Microsoft, which we'll touch on in just a bit.

The number that you hear a lot when we talk about Stargate just over the past, like, what, 24 hours has been this $500 billion number. But it's going to be about $100 billion that's put in upfront. That $500 billion is going to theoretically be invested over the course of four years. How did this number come to be? Because I've heard that part of the plan is to build out 20 data centers. That number has been specific. Just do some back of the napkin math, $500 billion divided by 20, that's $25 billion per data center. There's got to be more that's a part of this plan than just building out those 20 data centers, right? Yes. And the fun thing here is that there are any number of ways we could make that

$500 billion or imagine it playing out. So one of the ways is to take that number that you just came up, very nice back of napkin math, I think, and add in some GPUs. Just look at Elon Musk, his side project to develop an AI supercomputer thinks in terms of 100,000 GPUs. So if you're looking at NVIDIA hardware, now we're talking in terms of $3 billion or $6 billion when we're looking at 100,000 GPUs in a big data center or 200,000. That doesn't include associated server costs. So it's not just the GPUs. So you could quickly tack on, I think, several billion dollars to each of those data centers if you start to increase the compute capacity. But that still doesn't get you up to $500 billion. Where's all that money going? So part of it could be going

to OpenAI, which tellingly with this new deal also sort of announced along with Microsoft that they would not be an exclusive partner with Microsoft going forward. Microsoft is still going to fund OpenAI. They will still rely on Microsoft to provide cloud computing, but now they're free to be more of a venture partner with Oracle, which has really great and fast data centers. They're building out some $16 to $17 billion worth of data centers each year now. So we see lots of moving parts and pieces. And I think, finally, there is something of the venture capital sovereign government ethos going on here, which I mean, throw a big number out and see if it sticks. This is what Masayoshi Son is good at. I note that MGX, which is a sort

of a sovereign government fund from the Middle East, is also an equity partner. These types of ambitious projects sometimes gain momentum just by throwing out a huge number, even if the principals haven't worked out the exact feed of money into the project. We've focused thus far in the conversation on a lot of the private sector players, but importantly, it was President Trump who announced a lot about this. That $500 billion number that we're talking about, it does not include funding from the U.S. government. So then where exactly does the USG fit into this project? What kind of support from the federal government, what does support from the federal government actually look like in Stargate? Yeah, it's speed, Mary. So under the

previous administration, the Biden administration, there was a lot of government investment into semiconductor technology. That piece, as you pointed out, is obviously missing here. But what the Trump administration brings and what the US federal government will bring is a faster process to build out. There'll be less regulatory scrutiny on any of this. There will be, I think, streamlining of permitting, maybe less attention to environmental impacts. So everything that the previous administration were sticklers on, whether you agree with that or not, is going to be pushed a little bit to the side here so that these data centers can get built out as soon as possible. And of course, President Trump mentioned sort of the ongoing geopolitical tussle with China to be

preeminent in artificial intelligence. It's not just a corporate thing or a business thing or a tech thing. It's for the US, it's a national security interest. So you have that element as well. So they will make this whole thing flow pretty quickly. The three big names that we've talked about thus far that are involved in this. We've got Oracle, SoftBank, and OpenAI. But there are other companies that are also listed within the OpenAI press release, not mentioned within the OpenAI press release, but that you flagged is MGX, some equity from the Middle East featured there. But other business partnerships that are involved here are also ARM, Microsoft, and NVIDIA. They're named as key initial technology partners by OpenAI. If you're an investor in ARM,

Microsoft, or NVIDIA, is there anything not to like about a potentially $500 billion deal with other massive names in tech and the U.S. government? Well, if you're an investor in ARM or ARM, it's sort of Jekyll and Hyde. Yesterday, we saw the stock of ARM shoot up because they licensed chip technology. And today, I think investors woke up and said, wait a minute, if SoftBank is a funding equity partner of this deal, and we know that SoftBank isn't quite the beast on its balance sheet that it used to be, where are they going to get some of their funds? Well, they might sell some of that huge stake they have in arm to raise billions for this project, which would dilute ARM shareholders.

So that's something that you may want to evaluate if you own shares of ARM. For Microsoft, I mean, it's good in a way. We heard Satya Nadella, even though Microsoft isn't one of the funding partners say, hey, I'm good for $80 billion, my $80 billion this year. So Microsoft this year was projected to spend in capital expenditure about $60 to $65 billion. Anyway, they've added another $15 billion. That's good because Microsoft will have a yield on their CapEx investment when we're all spending more and more time using these large language models. And I think NVIDIA, it's generally positive for them as well, because as I mentioned before, part of the data center power comes from how much compute you pack into it. They are still the major player here for

the highest value computation. Much has been made about Elon Musk's reaction to this and his responses to Sam Altman on X. I want to focus on another big name that stays a little bit more out of the spotlight than Sam Altman and Elon Musk, Dario Amadei. What's his reaction been to all this. He's notably not included in this venture. Amodei is asking where the deets are. He's late to the party. This seems a little vague and amorphous to me. And I don't think that's sour grapes. I think you and I have talked about Sam Altman in the context of Amodei and how different their personalities are, how different their backgrounds are. I think this is just a rational builder of large language models who sees the need for this to get built out,

wanting to ask, okay, what's the roadmap here? How does this get expressed? He has the same questions that you have, Mary. I still can't get to the $500 billion number. So I think his was a rationalistic question. I will note that he said, look, overall, we probably do need to be investing on this scale. But I'll say personally, Mary, I think that $500 billion is going to get invested anyway without this deal or no from various players. I got one more question for you on this before we move on to our next topic of the day. You're a student of literature, Asit. I'm a student of literature. We're both self-professed words people. What do you make of this name? Where exactly do the stars come in in this Stargate situation? I don't think the stars are aligning

here. So when I heard Stargate, it made me think of things like Space Force, which is not a great name for our ambitions to be a military force in space. It brought up Heaven's Gate, in my mind, which wasn't that like a big budget failure at the box office. It seemed like Star Wars, so not quite Star Wars either. Just sort of a rapid mishmash of concepts. And the idea of a gate is really fun in the semiconductor industry, but not so much as a metaphor. Like, wouldn't you want the path of least resistance. So I'm going to grade this one, since we are students of literature. Actually, let me ask you first, what's your grade on this name? Well, I'm going to give it a low, I mean, you know, on the one hand, the star kind of,

it gets you excited about the future. It whips up some hope. But I have to say, especially from a political perspective, gate doesn't have a great track record. I'm referencing Watergate, typically not an awesome history to be attached to. So anyway, we will take the Stargate and use that as a nice segue into GE Aerospace, which reported earnings earlier this morning, shares up about 9% after dropping fourth quarter results last time I checked. This was GE Aerospace's first year as a standalone company, and the picture looks pretty rosy. Just going to throw out some top line and bottom line results here. We've got revenue for the commercial engines and services unit that grew 19% year over year. Total orders increased 46%, reaching $15.5

billion. Adjusted earnings per share for the quarter up over 100%. Planning to hike the dividend by 30% and repurchase $7 billion in shares. Asit, what sticks out to you? Mary, I think the themes that management has been talking up for more than a year now are just coming into play. And that's really what stands out to me. The industry itself is sort of supply constrained now. There's so much demand for new commercial airplanes and new military airplanes, and there's only so much production that can be output. And we've had supply chain kinks going on all of last year. So this is something where it seems on the surface of it, you would think, okay, building planes, it's so hard. How fast can that industry grow? But there are

estimates that it can grow anywhere between 8% and 13% for the next several years. So GE is benefiting from that. That really leaped out in the numbers to me today. I'm glad you brought up that growth because already three out of four commercial flights are powered by GE engines. So that growth point is kind of important because that's a bigger number than I would have honestly expected. There aren't, to be fair, many others that play in this business. There's Boeing, there's Airbus, but how does GE Aerospace fit into the broader jet engine landscape? So as you mentioned, Mary, it's one of the few companies that can make jet engines that satisfy a few demands. One, that they should be faultless if they are kept in good working condition. Two,

that they can be economical, they can provide fuel efficiency as this industry keeps expanding and the costs keep rising. So GE Aerospace, through a joint venture with a French company called Safran, the joint venture is called CFM, produces these specialized jet engines and only has a handful of competitors, as you mentioned. So we have to say it's sort of a dominant force in this industry. But don't forget, it also has a defense component as well. It supplies to the U.S. defense industry and some other global purchasers as well. You've written that GE aerospace is, I'm quoting you here, Asad, quote, the quintessential razor and blade model in the aerospace industry. If engines are GE's razors, what exactly are its blades?

So its blades are simply maintenance services and spare parts. So you sell the engine, but that engine has to be aloft for thousands and thousands of hours. In fact, modern jet engines now can last up to 20 years. So while the company does make a lot of money selling a single jet engine, what it's really going to capitalize on is a revenue stream for 15 to 20 years of helping keep that engine in good working order. So we compare that to a razor blade versus the razor. You buy the razor once, you have to keep buying the blades. I mentioned at the top of this segment that this was GE Aerospace's first year as a standalone company. Once upon a time, it was only a portion of the larger General Electric conglomerate.

Now we've got three separate companies that trade on the New York Stock Exchange. GE Aerospace has retained the GE ticker symbol and the CEO, Larry Culp. But you've also got Vernova, which is the energy segment of the business. And you've got GE Healthcare, whose specialty is self-explanatory in the name. Remind us why Larry Culp spun off these three companies into separate entities in the first place. There was a time when GE was held up as the model American conglomerate because it had so many industrial companies. And it also had this huge financing arm, GE Capital. And they were so great at managing earnings expectations to the penny, used to be the phrase of how Jack Welch, the then CEO of GE, managed investor expectations. What happened along the

is that Welch over-prioritized financial management. The various industrial businesses under GE themselves lost their ambition and were just cobbled up into this big hole that started suffering from pension obligations, from mismanagement of its financial arm. And we had just a train wreck of a stock. And what Larry Culp has done is to bring value to shareholders by separating these businesses out, letting them run on their own, giving them ambition again. He's also just brought so much clarity to the investment thesis in each one of these and split out or spun off the underperforming parts of GE, sold off divisions that weren't going to affect the bottom line.

So he came in as someone who had a vision to pull out what was important of the company and leave the non-important parts, the parts that were obscuring performance or dragging it down behind. Asit Sharma, always a pleasure to talk to you. I feel like often when we get together, we come up with like side hustles that we could be good at. And if there's any takeaway from today, it's that you and I both might be in the business of helping to name newly formed government partnerships better than the agencies themselves.

We are going to work on that idea, Mary, and we are going to have ourselves a very focused revenue stream a la GE Aerospace. Appreciate it. When Asit's not talking stocks with us on the show, he's got a whole other day job, searching for quality companies that can beat the market for long-term investors. Asit works on our flagship service, Stock Advisor, in addition to a number of other premium Motley Fool services. If you're interested in more analysis from Asit, two stock picks each month, access to Stock Advisor's full scorecard of companies,

and more, visit www.fool.com slash sign up. There will also be a link in the show notes. Okay, up next, the outlook for solar stocks is looking a little cloudy. Fool analyst Seth Jason joins me for a look at Enphase and the existential crises facing the rooftop solar industry. Seth, we're talking about a solar energy stock that's been on quite the ride. Before we dive into kind of the business of Enphase, can you give us an overview of the science of solar energy, how it works, and where exactly in that process, Enphase, the business, fits in? Well, I don't want to get too sciencey because i'm not a scientist but i can get uh i can get you the basics which is that when

you know the sun pours all of this uh light onto your roof uh it gets kind of hot unless you have some solar panels in the way and then it can turn that into some electricity right but it turns it into dc electricity direct current electricity when you plug stuff in your wall a lot of people might know this but some may not you plug stuff into your wall you're using alternating current AC electricity. So different kind of current, different voltages. The job of an inverter in a solar setup is to change that DC electricity into AC electricity. And there are tons of different kinds of these, different scales. You can imagine a utility is going to need enormous inverters for those solar farms. And even rooftop systems used to use, well, still probably still

use string inverters, which is, you know, an inverter that might handle, say, a group of panels, five or eight panels or something like that. Enphase's business was to put a smaller microinverter underneath or attached to each panel. And the idea was that as shade is on maybe a portion of the roof or the panels are kind of varying, despite the fact that they're the same, they're varying in output instead of an entire string or a bank of panels having to put out a lower amount of electricity. Because of that, the microinverters handle each panel on their own and thereby are designed to give you the most from your system as well as hopefully last longer because you've got each inverter doing like a little bit less work underneath one panel. So

that is kind of the business in a nutshell, the inverter piece of the business. Yeah. So there's that inverter piece of the business and you know where I'm going with this. Enphase also has a battery business and EV chargers. So how substantial are those offerings to the larger Enphase business model? Well, the EV charger business, that's not a great business for anybody, but you might be a little better off if you're integrating it with an entire system for home solar. But I've been involved with EV charging stocks in the past. And so I know from experience looking at their financials, it's not a great business. It's a bit of a commodity product a lot of differentiation but what is a better business for Enphase is that backup battery

business and so that is you know putting in enough power to say last six eight hours in case of a power blackout but more recently that was kind of the backup battery biz until a few years ago more recently the idea is to use batteries that can be attached to smart systems that will allow you to tap them and fill them at certain points and then use them at certain points in order to try to take advantage of differing electricity rates if that is the case in your system. If you're in certain places, the rates can change. Certain plans, the rates can change. The problem with the backup batteries as a way of trying to fix some of the net metering changes, which will probably be our next topic, is the backup batteries, we'll just say for the right now,

they can add like 50% to 100% to the cost of a system. In other words, if it's going to cost you 10 grand to put a solar system on your roof, a backup battery can add $6,000 to $10,000 to that pretty quickly. Renewable energy broadly is a sector that I get pretty excited about. I hear about this solar technology here, you describe it, hear what Enphase is doing. And I think, okay, this is pretty cool stuff. Sounds like awesome to me. And yet Enphase shares have been on a steady trend down, down, down since late 2022. What's behind that drop? Why aren't investors feeling the same kind of excitement that I feel just hearing you talk through this company? Well, in one word, regulation and regulatory changes. The reason that it used to be an okay

deal in some jurisdictions and maybe still in others, but especially in places like California, where there's a lot of sun during the day, is that they had a system where you got a credit for the full amount of energy that you put back into the grid during the afternoon when those solar panels were really pumping out a lot of electricity you got a credit on your bill for the retail price so if you were paying i'll just say 18 cents a kilowatt hour you got a credit for 18 cents per kilowatt hour that you put back in and california and many other states now have are changing have changed in case of california and are considering changing in many other states to a regime where instead of getting that retail price credit, you only get a credit for the cost

of the electricity that the utility would have paid that they didn't have to deliver. And that may in some cases be two-thirds to half of what that retail credit used to be. And so that has the effect of really reducing the payback that you got every year or month from putting electricity back into the grid. And that completely changed the dynamics of financing solar rooftop solar systems. And at the same time, we saw mortgage rates go up and a lot of loan rates, consumer loan rates also went up. And so attaching solar to new homes got more expensive, attaching it to existing homes got more expensive. We're still talking about systems that in some states might cost $12,000, $20,000 for a house. As we were talking about this company,

before we started recording, you'd mentioned to me that you'd sold Enphase and that this rooftop solar has a bad and often negative payback. Is what you just explained, is that why you wound up selling Enphase or is there kind of a way that this company might be able to overcome the problems that you just mentioned? Well, they'll keep selling those inverters. The level their sales will be is is the real question the solar industry in general really is in trouble in places where it was formerly doing great and a lot of that is like we were talking about if you want to use a tool out there that's easily accessible you can use project sunroof from google to kind of grab your house and get an estimate of what your payback is on a solar installation i actually just before the

podcast during my preparation grabbed like a house in southern california that had a good south-facing roof. I picked it off. I zoomed into Google Maps and picked it off the map. So it was a perfect house for a solar installation. No shade on the roof or anything. It wasn't that big a house. So they said the upfront cost of a 2.5 kilowatt installation would be about $10,000 and that it would cost about $24,000 over 20 years to use the electricity from this, plus the electricity you still need to buy. You wouldn't be replacing all of your electrical use all year round. This was their estimate. Even after subtracting a $3,000 state and federal incentive, your 20-year cost with solar was going to be $31,000. Without solar, just paying

your electrical bills was going to be $44,000. So over 20 years, you were going to save $13,270. So if somebody gave you a check for $13,000 right now, you'd say that's awesome. If somebody gives you a check for 1 20th of that you're not as excited and if you take the net present value of that at a four percent discount rate it's actually less than seven thousand dollars so it's just not as good a deal as it used to be and the end phases revenues have dropped back to where they were a few years ago because of this situation and it's like i said batteries are expensive so it's much more difficult to sell a solar system and say aha but if you use this battery you can bank that extra energy and then not have to buy it. That sounds great, except you might be paying

another five, six, seven, $10,000 for the batteries. So the economics for rooftop solar just don't work as well. It's different for utility scale solar, which is what we're seeing still expanding quite a bit in the United States, even places like Texas, but rooftop solar is in trouble and is going to remain that way. So it sounds like the rooftop solar industry has some existential crises that they've got to parse out. But if we look at Enphase and one of their competitors, SolarEdge, I see an interesting split, right? SolarEdge has burned cash for the past seven quarters. Meanwhile, Enphase, despite facing these very real, again, I'll call them existential crises that we've talked about, their free cash flow has been a little wobbly,

but consistently positive. So you've got these two companies playing in the same space. What's driving that kind of split in their management styles and their results? Well, Enphase has had pretty good free cash flow production. And so I don't want to get into what SolarEdge is doing differently. But Enphase, they were just doing a better job of converting their sales into actual cash. But I like to look, if you're an investor, so if you're only looking at free cash flow kind of from the outside it'll look great but of course right now it's actually dropping quite a bit for the last trailing 12 months i see free cash flow according to my spreadsheets here of like 336 million and that is down from let's see the prior year almost 600 million so that is cut in half

which is what you'd expect but a deeper way of looking at free cash flow especially for growing companies like these i like to have another bar on my charts that is free cash flow subtracting the amount of money that they spend on stock buybacks. Because as an outside investor, those stock buybacks, especially in these fast growing companies, tend to be just to soak up equity that is delivered to employees as compensation, right? So the free cash flow seems nice, but when you look at how much of it is basically just converted straight to compensation, the picture is a little bit different. In that case, I see for last, say for fiscal year that ended 12-23 and FASE had free cash flow of $586 million. But if you subtract

the stock buybacks, you are left with the $55 million in free cash flow, which isn't as great. And through the trailing 12 months, that free cash flow figure, once you net out stock buybacks, is actually a negative 56 million. So the free cashflow picture there isn't as great as it might look at first blush and other investors may disagree with netting out that cost, but that's just one of the ways I do it.

When I make a spreadsheet, I have about four different ways of measuring free cashflow because some of them are more applicable to some companies and some are better applied to others. As always, people on the program may have interests in the stocks you talk about and The Motley Fool may have formal recommendations for or against, so don't buy or sell stocks based solely on what you hear. All personal finance content follows Motley Fool editorial standards and are not approved by advertisers.

The Motley Fool only picks products that it would personally recommend to friends like you. For Asit Sharma and Seth Jason, I'm Mary Long. Thanks for listening. We'll see you tomorrow.