Motley Fool Hidden Gems Investing - The Infrastructure Behind the AI Revolution

Episode Date: February 19, 2025

If you’ve got a network that can’t go down, you call Arista Networks, a company building the infrastructure for the AI revolution. (00:21) David Meier and Ricky Mulvey discuss: - Why Microsoft and Meta rely on Arista Networks. - How Arista CEO, Jayshree Ullal, is managing Wall Street expectations. - The downfall of dating app Bumble. Then, (18:45) Anthony Schiavone joins Ricky to discuss Mid America Apartments, and why some housing costs are swinging back in favor of renters. Companies discussed: ANET, MSFT, META, BMBL, MAA, AVB Host: Ricky Mulvey Guests: David Meier, Anthony Schiavone Engineers: Dan Boyd, Rick Engdahl Learn more about your ad choices. Visit megaphone.fm/adchoices

You gotta try breakfast at A&W. You gotta try breakfast at A&W. And what better way than with a delicious Pret Organic Coffee, starting at just $1 all day, every day, now until December 31st. You gotta try breakfast at A&W. At participating A&W locations in Ontario. You got the trend right, but why was your stock a loser? You're listening to Motley Full Money. I'm Ricky Mulvey. Joining me today, he's not a loser. He's David Meyer. He joins us right now.

appreciate you being here, man. Thanks for having me. We got some earnings. We got some interesting earnings to dive into with Arista networks and Bumble. Let's start with Arista because this is for your picks and shovels AI plays. This is certainly one of them. And it's somehow disappointed investors with its earnings. I gave you a little preview of what it does before we get into the specifics. I want you to translate for me what exactly this business does. this is the description, and then you get to translate it into English. They are an industry leader in data-driven cloud-to-cloud networking for large AI data center campus and routing environments, end quote. David Meyer, what the heck does that mean? That means they take data

from one place and make it go to another place so someone else can do something with it. It's literally that simple. It's really a matter of having the right equipment to gather the data, do any processing, move it around, share it between different networks, in order for basically any company to get the most out of its data. Right now, we live in a cloud world. There are data centers all over the country. Some are public in terms of they're owned and then rented. Some are private, such as ones by Microsoft or Meta. They're their own data centers. But there's so much data moving through them that all those networks have to run seamlessly. Think about it. You and I panic if our internet goes down for 30 seconds. Imagine if you had

something that was really important. I don't know. All your business processes go down for 30 seconds because of the data center you're using. So they're the pipes that help data centers talk to each other and get information transferring. You get a little benefit of that when you're operating more large language models that require a lot more data transferring between each other. That's exactly right. And it's especially even more important now with those large language models being trained. Arista shareholders, if you're an investor in this company, if you're looking at this company, one of the things you're really banking on is Arista playing nice with its largest customers. That's Microsoft and Meta,

which collectively make up more than a third of Arista's revenue. Those two customers. Now, I want to put on my evil consulting hat. I have gone into Microsoft and Meta, and I'm looking at this money we're sending to Arista, and I say, why can't we just do this for ourselves? So, what does Arista do for these big tech giants that they don't want to do for themselves. So, what is Microsoft's core competency? Building software. It's software, right? So, what is Meta's core competency? Getting your attention. Correct. This is not in their area of expertise. This is a situation where they have decided to buy

versus build. Now, that being said, a company like Meta actually works very, very closely with Arista networks, to make sure that the hardware that they buy is actually somewhat optimized for the jobs that Meta expects it to do. So, it's not that Meta or Microsoft couldn't do this on their own. It's more of, they don't want to do this on their own because Arista does it better. But they still work with Arista to make sure that the hardware and software perform as they need for their data centers. And these are companies with a lot of data moving between data centers, so they want to outsource the piping, if you will. Unbelievable amounts of data moving. Now, when I lived in Ashburn, Virginia, just to give you an idea, at peak loads, I believe

on the order of 30% to 35% of the world's data, the world's data, moved through my backyard. Wow. Well, let's get into the business results for this quarter. Now that we have that lovely setup revenue, it's up about 20% to $7 billion for the full year and margins are also improving. This is wild. David Arista achieved about a 50%, five, 0% operating margin for the full year. And also we talked about this on the show yesterday with Jason Moser, a net promoter score of 87, which I learned moves from negative 100 to positive 100. So that's really high. What stood out to you from the quarter? It's really simple for me. It's beat and raise. For the quarter, revenue came in higher than expected for the fourth quarter of 2024.

Guidance came in ahead of analysts' expectations for the first quarter of 2025. They actually increased their guidance for the full year of 2025. In December, they said they were going to do between 15% and 17% growth. Management came into their conference call and said, nope, we think we're going to do 17%. They've essentially raised their full-year guidance. They're not expecting any drop-offs in terms of margin. What you can deduce from the fact that they're growing as fast as they are and achieving the margins that they are for what they provide, as well as their net promoter score they are making customers happy that is sure i'm sure of that based on those that data is you're setting as you're listening to this segment set that

information aside in your brain when we get to the next company we talk about arista has a software product it's called network data lake and this is a meaningful part of the business it did a billion dollars for the year and that basically david i tried watching a demo on this product and here's what it seemed to me is I am a podcaster. I'm not a tech person monitors all of your cloud data for security and patterns into my untrained eye. If you're a business, okay. You have an AI system to monitor all of the data for network security and spot patterns. You know, this kind of sounds like we don't need to go to a restaurant for Palantir when we have a network data lake at home, what's going on with the software. So that's almost correct. Palantir is actually

can work with, we'll call it, any type of data. Whereas, the Arista NetDL, which is the acronym for its network data lake, that really focuses on what's called the state of the network. Basically, what signals are they gathering to say, is our network running efficiently? Is there anything that's any trouble brewing out there? Meaning, a workload is going up, or, oh my goodness, it looks like something's getting ready to fail. What's interesting is, within their NetDL, they have what's called an autonomous virtual assistant. So, it's their AI assistant or agent that continuously monitors all these workloads and workflows and then proactively says, hey, to someone on the network operating side says, hey, there could be a problem here,

you may want to do something about it. Or it says, hey, this customer has actually been asking for more of the network. Maybe we should give it more resources. So, the way to think about it is what Arista is providing here is a very, very small niche of what Palantir can do. But it's, again, very important to a customer like Microsoft or Meta. They want to make sure their network never goes down, always stays efficient, because efficiency equals cost, and going down would mean no revenue. That's the purpose of the network data lake, and it's a very important piece of what Arista sells in terms of its platform. We talked about the positive sides of this business, but it seems Wall Street's a little unhappy. CEO Jayshree Ullal is trying to get

ahead of it in the earnings call, saying, quote, while I do appreciate the exuberant support from our analyst community on our momentum, I would encourage you to pay attention to our stated guidance. We live in a dynamic world of changes, most of which have resulted in positive outcomes for Arista." What's she trying to tell the analysts here? Yeah, this is a little interesting. If we go back to 2023, the original projection for the 2024 for revenue growth was 10% to 12%. So, what did they deliver in 2024? Almost 20%. There can be expectations that you're sandbagging. And again, let's go back to December. The company says, we're going to do 15% to 17% revenue growth, and then ups that up to a solid 17%.

there is definitely some managing of expectations here. Like, don't expect this to go to 34% type of a thing. It's just natural. And that's the way markets and analysts tend to work. Once you see things happen, a pattern develops that you outperform, the expectation is you're going to continue to outperform. But the CEO is definitely getting out ahead of that by saying, no, I'm serious, right? Stick with what I'm saying, not with what you want me to want to hear from me. Stick with what I'm saying, but also I might be sandbagging because we live in a dynamic world that often has positive outcomes for our business. Well, there's also the point that I might be wrong and I could be wrong on the negative side this time, right? And revenue might

not appear. Uh, and that would be really bad if expectations are too high. And that could happen In a few ways, you think about a company like Arista, if companies spend less money on these LLMs, less data moving through the pipes, that could have a negative impact for Arista. This is one of those companies that I have looked at from afar for a few years now. Now, you're just talking to Ricky, David. You're just talking to me. One of the stocks I own in my retirement portfolio is ASML, which builds the machines that builds the machines that build the most highly advanced chips in the world. This is one of those companies that I own. Just sock it away. Don't worry about the little earnings blips that we've been talking

about here for the past 10-ish minutes. I'm starting to think, does Arista belong in that same basket for me? I don't own the stock, but it is on my watch list. I think that Arista, given, again, what we've been seeing, which is incredible growth, high margins for the products and services that it sells, and very happy customers. So, I don't think anything's going to come tumbling down. But if we look at some data within the most recent conference call, one of the reasons I think the stock is actually down today is over a concern that Meta moved from 21% of sales in 2023, so 21% of all of Arista's sales in 2023 to only 15% in 2024. That implies that they pulled back on their spending.

One scenario might be, if big customers, the Microsofts, the Metas, the other hyperscalers, pull back on their spending, maybe because they found an alternative or maybe because they're negotiating price concessions, could be for a variety of reasons. That's the negative surprise that would actually be very bad. But again, go back to who else is providing such great hardware, such great value to shareholders, and keeping customers happy? It is Arista. There's no doubt about that. So, this seems more like maybe a blip in the road than a serious problem forming. All right. Let's move on to our next story. David, let's mansplain a female-forward dating app how about we do that bumble that sounds good it is where where women make the first move

reported as well and this is a falling knife that seems to be continuing to fall this was a company that at its peak was worth more than eight billion dollars now it's well under a billion when i look at these dating apps and the underperformance in some ways it mystifies me because i know people get tired of them i know people are hesitant to pay but also these companies have created some of the most powerful dopamine delivery software applications in existence. I hesitate to think of ones that deliver more dopamine, maybe your gambling apps, but where did this relationship between long-term shareholders and Bumble go wrong? That is such a great question. So, I looked back at the revenue for the past five years, and 2021 was the peak.

They grew their top line almost 32%, but that's when revenue started slowing. So, what happened following 2021, right? We have the pandemic that hit. There's lots of positive investor sentiment about, hey, this is the way the world is moving, right? We can't go out. We can't do things. We can't see people. So this is the mechanism for people to get together. But growth has decelerated from that point forward. And when you have high expectations and growth slows down, investor sentiment sours. So at the beginning of 2021,

it's forward enterprise value to sales ratio. So this is the projection of where they think revenue is going to be during the next year was around 12% to 13%. That's extremely high. That is a company that is essentially doing nothing wrong. But that's because investors expected lots of growth. Today, that same ratio is 1.7%. The reason is because investors are not expecting hardly any growth. It is simply that, for probably a multitude of reasons, They have not been able to stay on the growth path that they were on into the pandemic and coming out of the pandemic. That multiple getting cut by about 8X. It's tough to keep up the engagement. A recent Forbes poll of dating app users showed that

folks on the apps are spending about, we'll call it, a little under an hour per day. 10 years ago, it was 100 minutes daily. We went from over an hour and a half to under an hour. Logically, it's hard to really grow, I think, from 100 minutes per day on these dating apps. Then you're really getting into the three, four-hour range. You're watching Lord of the Rings movies for the same amount of time you're spending on these dating apps, David. Yes. One thing we should note here, because even though I agree directionally, would want more engagement on an app. If we think about it, I'm sure, even though I have never used this app, being married, I'm sure that the actual experience got more efficient. Meaning,

the developers of these apps could take the data that they were gathering from them via all that engagement and just making the process more efficient. So, naturally, I would actually expect right? The time on the app to go down because hopefully you were getting better at getting to your end state, which was going on a date or finding someone that you wanted to have a relationship. So, I'm not necessarily as concerned with the number of minutes going down, but what I am concerned about is the declines in the revenue per paying user. That's a proxy for people are not getting the value out of the app that they thought they were going to get. And this is a company where for basically both of their apps, they've grown the number of paying

users they have year over year. However, the average amount that those users are willing to pay has dropped since then. So it's basically, you're swimming a little bit upstream there if you're looking at this company. Yep. You also got a founder story, Whitney Wolf Hurd. She is returning to the executive chair. What does she need to do on this tour to turn investor sentiment around? Yeah, that's a great question. So let's go right to the source. This is what she said she's getting ready to do in the conference call. So to quote, as we execute this transition, basically her coming back as CEO, my focus is on the following key areas, the deep love and understanding of our

product that only a founder can bring, re-inspiring the unique magic of the Bumble brand, and operating with purpose, efficiency, and excellence across the entire company with a particular focus on technology. So, frankly, that seems like quite a bit of high-level jargon, more than a serious plan. But I will also say, at this point, it's very early in the transition. So, I can understand sort of, hey, the message being, look, this is my baby. I know all about it. I love it. I'm going to bring that magic and have it permeate throughout the organization. Not only that, but things have gone wrong with the Bumble brand and I'm going to turn that around. And oh, by the way, she praised the previous CEO for these things, which is operating efficiency, cost control,

things like that. And basically she says, we're going to keep doing that. However, the proof is going to be in the pudding. We'll see what she says and then compare it to the actual outputs over time. Let's leave it there. David Meyer, thanks for being here. Appreciate your time and your insight. Thank you so much, Ricky. All right, so you just heard about a stock where investors got really excited about a trend, but then got ahead of their skis in terms of expectations. Up next, Motley Fool senior analyst Anthony Chavone joins me to talk about a REIT that is trying to come back from a similar phenomenon. So Anthony, there was a great migration to the Sunbelt a few years ago,

but maybe some institutional investors and retail investors over-calculated. And this is a trend we've been talking about for a long time on Motley Fool Money, and it kind of went like this. Offices closed during the pandemic. A lot of remote workers wanted to go somewhere sunnier, more outdoorsy. Why not go to a place like Nashville or even where I'm at, Denver, Colorado? Real estate developers followed, but maybe they over-delivered. For example, now looking at Redfin data, the average rent in Nashville has fallen almost 30% 3-0 since January 2023. Denver fell about 4% in the same time.

I would honestly expect, just seeing rents on Redfin right now for me, I would expect it to fall even more. Let's get into the investing side of this, though. There's a company I know you follow that I've looked at before as well, MidAmerica Apartments. It really likes this region. How has this story played out for MAA? The major theme in the Sun Belt, where Mid-America owns most of its apartments, has always been about supply. If you look at the past decade-plus apartment construction in the Sun Belt, it's always been higher than the rest of the country, but the demand has always been

higher, too. That's usually led to good performance for Mid-America throughout the market cycle. And I mean, real estate, it's always been a cyclical industry. But if we look at the last five years especially, the development cycle has been pretty intense. So, if you go back to 2020, we obviously had the pandemic, and that delayed a lot of construction of new apartment buildings. And then also, at the same time, like you just mentioned, Ricky, remote workers relocated to warmer markets, which accelerated the migration and demand into the Sunbelt. So, then you move into 2021, 2022, we had that limited new supply because of COVID, and then that coincided with growing tenant demand. And so, you had that supply and demand imbalance of the rapid rent

growth for landlords and ultimately incentivized new development. And with interest rates at historical lows, builders decided to build, and they decided to build a lot. Plus, the construction that was delayed during COVID, well, that was also being delivered. So, that kind of brings us to the supply place today where supply is currently outpacing demand in the Sunbelt. And for Mid America, that meant a very challenging 2024. Pricing for new move-in tenants was down about 6%. Net operating income was down about 2%. And then earnings fell to low single-digit percentage. It's been a pretty difficult year for Mid America, but I do think the supply and demand fundamentals start to shift back into their favor over the next few years. I think we'll

touch on that in a little bit. And if you're a renter like me, this is great news. I am cheering this. I'm ready for rents to come down a little bit and to have more bargaining power is someone who's renting mid America. One of their strategies is that they're looking for apartments, not in the like hottest possible spot. They're usually looking about five to 10 miles outside of that. So it's people that want to be in the area, but look for a little bit more of an affordable place, but still, still nice. They still got, you know, washers and dryers in unit dishwashers, that kind of thing you can go to the pool and meet new friends at a mid-america complex mid-america made three acquisitions in 2024 here's what i found interesting these were in properties that

were about two-thirds full and the story they've been telling is that they're in super high demand areas and they are buying these spots that are one-third empty to flip the fraction that seems pretty low anthony for these high demand areas well yes the occupancy rate is definitely low but there's a good reason for it. Over the last year or so, Mid-America's strategy has been to buy newly constructed multifamily properties that are still in the lease-up phase. After a building is constructed, it usually takes anywhere between six and 18 months to fully lease an apartment building to get up to 90% or higher occupancy. That's why the current occupancy rates and what they're acquiring is so low. The reason why they're specifically targeting those properties

in lease-up is because, one, they're brand-new, high-quality assets. And two, it's very hard for the developer to refinance that property now that interest rates are a lot higher. So, these are properties that probably started construction two to three years ago when interest rates were zero and rent growth was going strong. But now that that equation has changed, most of those developers, they aren't able to get that permanent finance they need to continue owning the property. So folks like MidAmerica can come in with lower borrowing rates and purchase the properties at a pretty attractive return. So these developers are basically saying, we can't get the cash flow that we assumed based

on this rent growth and these low interest rates. So we're going to cash out now and sell to MidAmerica Apartments, which has a ton of cash on the books, and they can come in and we can end the investment here. Outgoing CEO of MidAmerica, Eric Bolton, told investors that while there is a large amount of oversupply, that's going to level off and demand's going to catch up in 2026, 2027. I mean, I don't know. I think the remote work trend is starting to shift a little bit. We're seeing that certainly at a federal level. I think some companies are also walking back remote work a little bit more and more. And I think that's a pretty key part of this thesis with more people moving into the Sunbelt. I don't know. What say you?

Yeah. Well, on the earnings call, Eric Bolton said that the tide is starting to turn when it comes to supply and demand fundamentals. And I know he's talking his book, but I think he's got a pretty strong argument because last year, Mid-America had a 50-year high of new apartment supply in its markets, a 50-year high. So the fact that NOI, net operating income, and earnings only fell slightly in 2024, despite that massive supply wave, I think that just kind of demonstrates how strong tenant demand has been. Last year, Mid-America's occupancy rate was still almost 96%. Their resident turnover was near an historic low. If you look at their lease pricing for both new and renewal leases, rents only declined by less than 1% on a blended basis. The results have

definitely moderated. There's no question about that. But the demand side of the equation is still really strong. Looking forward on the earnings call, Eric Bolden also mentioned that new construction starts dropped by 50% in 2024. That's largely due to interest rates, lower rents, and higher construction costs. As long as the economy stays in good shape, I think the supply and demand fundamentals in 2026 and beyond look pretty good for Mid-America. I think rent growth has a potential to reaccelerate them. Bad news for renters as we look ahead. Maybe if you're listening and you're looking to rent a place, looking for an 18-month or 24-month option could be in your interest. There's another multifamily REIT I wanted to talk to you about,

and that's Avalon Bay. This is one I know you've checked out more than me. This is a little bit more geographically diversified, a little bit more suburban than MAA. And one of the things in their earnings presentation they showed that I thought was interesting was how much cheaper it is to rent versus own, especially in their established markets that you're looking at the East Coast. And in the Sunbelt, still, it's about $700-ish per month cheaper to rent an apartment than buy one. In the established markets, that's $2,200. So, renting is $2,200 cheaper than buying per month in these established markets. That's a number salad. But what does this trend mean for Avalon Bay? What are you seeing here? Yeah, the affordability spread between renting

and owning a home is something pretty much all public REITs have called out on their earnings calls so far, almost regardless of geography. But since it's so much cheaper to rent than to own, the resident turnover rates are near historic lows for most public apartment REITs. So, fewer tenants are leaving apartments to purchase a home because interest rates are so high. And that's important because it usually leads to lower costs because you don't need to repair or remodel the unit if the tenant is still living in it. And additionally, you have fewer units that are sitting vacant and waiting to be released. For Avalon Bay, I think the affordability gap really helps on the expense side of things more than anything. I think just the combination of

relative affordability and the fact that there's also not a ton of new supply in Avalon Bay's more coastal, more suburban markets relative to the Sunbelt, I think that's a big reason why they were able to grow earnings at a decent rate this year. Then Avalon Bay is making a little bit of a different bet than MAA is. They're looking towards the suburbs more. What's the bet on the suburbs that Avalon Bay is making? About 73% of Avalon Bay's portfolio is in suburban markets on the East and West Coast. They actually plan to get that up to 80% over time. There's a few reasons why Avalon Bay targets the suburbs. The first is, and this might be surprising, but there's typically less new supply in suburban markets than urban

markets because the entitlement process in the suburbs tends to be a lot more difficult because those local jurisdictions don't really want new rental housing to grow in their markets. And secondly, a lot of Avalon-based tenants are higher-income tenants who are renting by choice, which also can lead to lower turnover costs and lower remodeling costs over the long term. And as we talk about these two REITs, both of them pay a little more than a 3% dividend. that's about what you can get from the Schwab dividend ETF, SCHD. For investors that are thinking about income, that are looking at these REITs, what expectations should they have? I think of the shorter term. I think the returns that these companies could generate are a little

bit higher just because of the fact that they're leaning in a lot. Well, we didn't really mention that they're leaning into development a lot right now, which I think is interesting. Ahead of that stronger 2026, 2028 period that they expect when rent growth is supposed to be higher. So, like, Avalon Bay, for example, in 2025, they expect to have $3.5 billion worth of construction, which is 50% higher than where they ended 2024, ahead of that increased earnings growth period that they expect. So, I think in the shorter term, I think these companies can provide pretty good returns. But over the longer term, I wouldn't expect anything more than, say, 10%, 12% per year.

I think that's a reasonable expectation for real estate. But the key fact is that the risk associated with these investments, in my opinion, is a lot lower, because people are always going to need somewhere to live. So I think that's the intriguing part about investing in a REIT, is the lower volatility, the lower risk associated with it. Better risk-adjusted returns, I should say. And I think, yeah, the housing shortage is going to go on for a long time. A lot of people that locked in those near 0% mortgage rates are probably going to want

to hang on to those as long as they possibly can. I want to get back to the story we told at the beginning as we wrap things up. So we talked about the building boom back in 2021 as investors got really excited about the Sunbelt. And they were right on the trend. A lot of people moved out to Denver. A lot of people moved out to Nashville. But what happened is investors were right about the trend, but they over-indexed. When we look at REITs, we think about the price to FFO, funds from operation. This is the REIT version of your price to earnings multiple, the price tag for the stock. What happened was that price tag shot up for mid-America apartments. If you were excited about this trend when it was heating up back in late 2021, you haven't totally

lost, but the stock is down and you've basically just collected a dividend of 3% for the past four or so years. I'm wondering if this story could foreshadow anything else in the real estate market. I'm seeing a similar trend for data centers right now. Investors may be right about the trend. But what happens if they're over-indexing, and then a few years from now, expectations cool down a little bit? You've studied real estate a whole lot more than I have, but what do you think about this story? Could that be an unfair comparison? Well, usually with real estate, when you see an industry, a property type that has such strong rental growth, like we're seeing in data centers right now, that typically leads to overbuilding,

which eventually brings rental rates down to a more reasonable level, and so sort of evens out. And that's happened for pretty much every property type you can think of, apartments, industrial, office space. The cycles all might be a little bit shorter or longer, depending on the property type, but it's generally the same. I think manufactured housing is probably the only property type that really hasn't overbuilt. In the case of data centers, I'm not so sure, because if you think about the power requirements that a data center needs in order to operate, That is such a huge cost for anybody developing a data center. And developing these data centers takes so long to build. So, I'm not so sure if it's going to be the same as all the other property types, but

it's definitely going to be something interesting to watch. Anthony Chauvin, appreciate you being here. Thank you for your time and your insight. Thanks, Ricky. As always, people on the program may have interests in the stocks they 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. I'm Ricky Mulvey. Thanks for listening. We'll be back tomorrow.