Motley Fool Hidden Gems Investing - What’s Up With Super Micro?

Episode Date: March 3, 2024

When a stock goes parabolic, there’s more to the story than business performance. Mary Long caught up with Tim Beyers and Kirsten Guerra, advisors on The Motley Fool’s Interconnected Opportunities investing service to discuss: - Super Micro’s real business. - The original bull thesis for the data center service company. - Super Micro’s relationship with Nvidia. - What’s behind the stock’s recent performance. Stocks discussed: SMCI, NVDA, DELL, HPQ Host: Mary Long Guests: Tim Beyers, Kirsten Guerra Producer: Ricky Mulvey Engineer: Tim Sparks Learn more about your ad choices. Visit megaphone.fm/adchoices

Hello, midsize business. We see your big ambitions, but how do you achieve the wins you want when your technology is holding you back? SAP Grow is built to grow with your business, no matter its size. With AI embedded at its core, working across every system, all ready to go from day one, so you can hit the ground running. Bring it with SAP Grow. AI Cloud ERP for any size business. I'm paraphrasing here but is the late great Charlie Munger would often say that you know your results are tied almost inexorably to your ability to control your emotions in situations where you need to and if that's not you this is not a party you want to go to I'm Mary Long, and that's Tim Byers, an advisor on the Motley Fool's Interconnected

Opportunities Investing Service. Tim and his colleague, Kirsten Guerra, have been following the Supermicro story for a while now, and you may have seen the data center services stock recently went ballistic. I caught up with Tim and Kirsten to find out what this business actually does, the machinations behind the stock's performance, and what investors should do when they see a stock break the laws of gravity. So, Tim, I want to start with you because talking tech can get really technical really quickly. And before we talk about Supermicro, the company, it's maybe important to get a picture of the landscape. This company operates within the data center landscape. And so before we start talking

about the ins and outs of the business, can you give us an idea of the different types of data centers that exist and why it's important to understand that landscape before talking about Supermicro? Yeah, there's a lot of different ways to engage with a data center. Let's use a building analogy. So the data centers that most people are familiar with are the ones that are owned by the big cloud providers. And you can think of the big cloud providers. So that's AWS, GCP, Google Cloud Platform, Azure. And when you are accessing those data centers, those are wholly owned by those big tech companies, and they rent space inside of their data center. So when you are engaging in the cloud, you're using somebody else's computer. If you're engaging in AWS, you're using AWS's

computers. And those AWS computers live in a data center somewhere, which is a really large building that has lots of equipment in it, servers, networking gear, storage devices. They have tons of wires and just lots of industrial equipment to cool that facility. You could think of it as a giant factory floor that may as well be silent and frigid because you don't want computers to light on fire. That's the picture of a data center. So an AWS data center, so all of them kind of look like that. So then it becomes, okay what do they do and what do they offer so like if buildings are roughly the same but they have different purposes that's a way to think about data centers so the aws one is like the

furnished apartment you rent right out of college they give you everything you don't need to buy any furniture you don't need to do anything you just go in and you've got a place and it's all set up for you and you pay the premium for that that's aws that's gcp that's azure then there is the like equinix it's the data center that has a lot of equipment in it already that they equinix they don't always do this but there are these hybrid environments whereby and and generally they're called bare metal agreements you have lots of equipment you have lots of lots of things that are inside the data center and you want to control that equipment so this is the condo this is the condo arrangement i have a place i own it but it's still like a managed community

there's some stuff that's managed it's condo it's a townhome and then there is the co-located agreement, which is you get an apartment and you are going to bring all of your stuff to your apartment. You're going to have to furnish this. You're going to rent it, but you're getting the space, but you're paying the utilities. You are paying really everything. And Supermicro fits most specifically into that third part of it. They're the ones that furnish the apartments with the stuff that you want to buy so that you know you have the essentials in inside your apartment so you are bringing equipment you own that you have you know brought from super micros super micros selling you the bed selling you the couches all that stuff and

you bring that in into the apartment so in this data center arrangement you own the equipment and the co-locator gives you the space, power, things like that, but you really want to control the experience. And so you use Supermicro equipment because you are very interested in not handing the keys to AWS. You want to control the actual hardware experience. So let's zoom in on that equipment picture a little bit. Small company called NVIDIA has been in the news a lot lately. Supermicro rides similar- Tiny company, never heard of it. Tiny company. Okay, that's what I figured. We can do a deep dive on that at a later date. But

Supermicro rides similar tailwinds to NVIDIA, but they're allies, not necessarily competitors. And you've used a sports car analogy to visualize this partnership. Really diving into the analogies today. But Kirsten, if data centers are sports cars rather than apartments, what part of that car is Supermicro building that NVIDIA isn't? I was going to say, I love all the analogies. So if we're talking cars now, NVIDIA is the engine. And the engine is exciting. And everyone wants to talk about the latest and greatest specs. But you'd never roll up to a car dealership and just ask for an engine. It's bigger than that. You want a whole car. And if that car maybe has the fancy engine you want, but it pairs it with a restricted exhaust that builds back pressure or

poorly timed ignition system, ultimately, you've paid up for this exciting engine, but its effective power is squandered from an overall inefficient build. I hope that makes sense. I don't actually know anything about cars, but if you do, listen, I hope that helps. That sounded amazing, though. I know, I believed it. Thank you so much. So in that analogy, NVIDIA provides the engine and Supermicro is in the business of that

overall build. So Supermicro builds full rack scale server systems, which require not just compute from these NVIDIA GPUs, but also substantial memory, storage, networking capabilities. And Supermicro isn't necessarily manufacturing a lot of what they assemble. What they sell is the expertise in putting it all together. Supermicro is building these hyper-efficient systems that take full advantage of everything offered by those fancy new NVIDIA GPUs that the data centers are already paying top dollar for. So, of course, they want to get everything out of them that they can. Yeah. And just to build on that super quickly, it's not like you're buying

from a standard dealership. When you're buying from Supermicro, you're buying a tricked out vehicle that is for you. You are Vin Diesel and you want your tricked out car. Yeah, that's a great point, Tim. And I know we're kind of jumping around with the analogies here, but I do think for the layperson, they're really helpful in visualizing what these highly technical companies do. Because, okay, now that I have a grasp on what it is Supermicro makes and how it operates in this space, that's great. But there are also other companies that do the exact same thing. So why would... Again, to help a layperson understand, why would a customer choose Supermicro over a bigger name like Hewlett Packard, Cisco, Dell, etc.?

Yeah, those are the dealers. They make a car that is very easy to understand. They are not tricking out a vehicle. They are mass producing a model. And those models can be useful. And you may even put them in a data center and you might even go find a mechanic who could trick out these base models that you have bought from one of those other providers and build you an environment that suits whatever purpose you have. But very often when you are going with a co-located agreement and you're building out an infrastructure, you have a particular purpose in mind. There are probably some very specific workloads that you're trying to serve. And so in the case of Supermicro, because they do something that's much more custom, they essentially custom build what you need. and then they give you an agreement that allows you to upgrade those systems over time they're very upgradable they're also to kirsten's point before about like the vehicle being something if it's if it's inefficient in some way like if you're if it's you know the the power distribution inside

vehicle you have a bad power train or something like that i don't know much about cars either let's be honest, you just aren't getting a good ride. And in the case of Supermicro, they have some base components that are very useful. One of their biggest is that the systems that they build are hyper-converged. They're not just giving you a server, they're giving you essentially the all-in-one meal. So they're giving you everything. And so some of the vendors you mention there, Mary, they will sell servers and they might sell some networking gear and they might sell some storage equipment and they may do it through third parties and then maybe they assemble it with the help of a systems integrator. That's not what Supermicro does. They build you a

custom hyper-converged system that has all of the pieces that are built together, shipped together, and done so with some very specific technology they've created to make their systems highly power efficient, which if power is one of your main costs, as somebody who's renting out space inside a co-located facility, you would like your systems to be power efficient. That's a real draw to Supermicro as well. So it's the customization. That's one reason why people buy from Supermicro. Another reason is they have relationships with all the major suppliers. So when they custom build something for you, you know you're going to get a premium product because they have all of the premium agreements with all of the premium providers,

including NVIDIA. And third, the way they build their systems is highly efficient. In fact, so power efficient that they are essentially world-class in that area. So you do have customers who are like, we would like to be climate aware. And so in those instances, buying a highly efficient machine like what Supermicro is offering can be attractive on that basis as well. It's not the main reason I think people buy Supermicro. It's more about the customization, strong relationships across all the major chip and component suppliers, long history doing it, and then the power consumption, which saves you big money. If you look at a stock chart of Supermicro, there is a staggering, to put it mildly,

run up in recent months. We're going to get to that. But if we hop in the time machine and go back a few months before the staggering run up, what was the original bull thesis for Supermicro? In addition to the fact that this is a solid business, for many of the reasons Tim just outlined, the bull thesis was really twofold. This is a company with huge tailwinds from the exploding demands of AI. And when Tim and I were looking at this last August, we felt that was underpriced. Because last August, CEO Charles Liang started suggesting that Supermicro could get to $20 billion in revenue in just a couple of years. And that was a huge jump from the $7 billion or so in the preceding 12 months around the time he said that. So, we had to ask,

can this guy be trusted? Because that's a huge, big if true, right? So, after looking at- It was an even bigger question than that. Either he's a genius or an idiot. And which one is it? I mean, honestly. Right. So after looking into Liang's long history of issuing guidance at Supermicro, it was clear that this is someone who is quite accurate in his predictions. And if anything, he leans conservative when he does make statements like this. So we felt comfortable taking him at his word on that $20 billion projection, give or take. And when you look at the math of that kind of medium-term growth, it looked very attractively priced at the time.

So we talked about this staggering run up. Let me put some numbers behind that. In September, shares of Supermicro traded at $231. February 15th of this year, $1,004 a share. It's bounced around since then. And I'm not even going to attempt to name a number for today. We're recording this on Thursday morning. The price today could change in a few days time. The point is things have happened and things are happening. My question, Kirsten, I'll kick this back to you. what in the world is going on here and has caused this just kind of wild parabolic move? Yep. Things are happening. Anytime the market is valuing a company, there are elements at play of both the underlying business fundamentals and general market sentiment as kind of a multiplier

on that. So on one hand, you have genuinely impressive business performance from Supermicro since the time we recommended it and the year or so before. They are really seeing that revenue growth that Liang suggested. And they're seeing margin expansion from that, from operating leverage on that revenue. So that's all business fundamentals. But it's that other component that maybe is getting a little out of hand here. Market sentiment around AI is generally very bullish right now. And how could it not be? Valuation is all about predicting the future and projecting how much cash flow you think a company can generate into the future. So when you have companies like NVIDIA in very related fields that are just blowing away earnings

expectations every quarter recently with beats of 10% plus on earnings, yeah, people get a little excited. The market starts to feel like, wow, we don't even understand the potential ahead for this technology. And yeah, the market gets a little frothy and that other side of the equation in market sentiment can go a little wild. Can we also just talk quickly about the machines and how this works? What I mean by the machines is the machines of the market. When they see... When there are numbers that seem related to, in this case, Supermicro, like the belief that nvidia has just an unbelievable long-term growth story and that super micro is going to profit handsomely from it the machines all pile in at the same time

and what has ended up happening that market sentiment is essentially algorithms from institutional buyers putting way more money than has ever been put into super micro into that stock at once and that drives it to crazy levels and that's what's happened so what you're seeing if you can imagine the waves of like money that are flowing into and out of super micro on a daily basis but it was the shoreline was relatively stable i'm looking out over the horizon onto the ocean and it's it's you know i see waves but they're relatively minor it's peaceful good for a nice long walk on the beach and now i'm seeing tidal waves and what the heck is going on and it's money that's moving but if i'm a surfer and i see tidal like to a certain person that can be

appealing. And so whether it's because of earnings or machines or tailwinds or what have you, when a stock shoots up 300% in less than six months, why not just cash in on the gains and sell and walk away and enjoy it? I mean, that's a great question. And so when we offered some guidance here, unfortunately, with Inside Interconnected Opportunities, which is a notional scorecard, selling would have meant exiting the position entirely. And I tend to not believe in that. So we issued some guidance. It's a little bit in between that hold on buying it. And I still hold to that. Hold on buying it. But if you have some, it's perfectly acceptable while we wait to see just how long the growth tailwinds for this AI trend are. It's perfectly

acceptable to hold on to some. But if it had been a real money portfolio, and Kirsten can add to this here. But as we were talking about it, I think we tended to agree, if this were a real money position, and let's say we had put $10,000 into a real money portfolio and hit it gone absolutely parabolic, we would have sold some. We may have even sold more than half, but we would have kept some as well. I mean, Kirsten, you think that's fair? Yeah, that's fair. Okay. We're talking about tidal waves and wanting to surf. If I don't have a position in Supermicro, but I'm seeing this incredible run up and I'm hearing about the tailwinds and the long term

growth story of Supermicro and other companies like NVIDIA that are benefiting from the same trends. It's tempting to see all that and want to join in the fun, like to want to hop on the beach and start partying with everybody else. What's your advice to someone who sees that kind of run up and wants to join in? Yeah, I would say just be cautious. Know exactly what game you're playing if you enter here. If you know that you're an incredible surfer, then that's one thing. If you know that you're not, that's another. I would say six months ago when we recommended the company, I would have described it as more of the balance of risk-reward there rested more solidly on business fundamentals than market sentiment. And today, maybe the opposite. It's

a lot more sentiment-driven. And that's not to say that this company can't continue to surprise, but I think investors today should really have the expectations going forward of substantial volatility. And if that makes you uncomfortable, or you don't know how to surf, or you're just not someone who believes that the AI market ahead will continue to explode in ways that we can't predict, then this probably isn't the stock for you. And there are plenty of other options out there. Yeah. And I could go further. Taking your analogy there, Barry, since we're having fun with analogies if you if you're on on the beach and going to going to the party two things recognize there's always going to be another party so you can always go to another one call a friend and

find a different party or if you're going to go you know not to use the old man advertising line here but please drink responsibly i think would be the the way to to think about this and don't get overindulgent here because it's very easy to let your emotions get away from you when you are getting into into these types of situations and i think charlie munger had a i'm paraphrasing here but his the late great charlie munger would often say that you know your results are tied almost inexorably to your ability to control your emotions in situations where you need to and if that's not you this is not a party you want to go to there's not insignificant short interest in super micro it's not quite beyond meat levels about 12 of the float is short interest

in the grand scheme of things a is 12 a lot from your perspective but kind of beyond that what what are the shorts thinking what's their case i mean i have an answer for this so i'll So I'll tee us up here. And Kirsten, you can add what you want. My answer to that is the short sellers here, they have a very basic argument, Mary. They will tell you that this run and the growth that is expected as a consequence of the run-up is so enormous over the short term that the opportunity for disappointment is very, very high and they expect a pullback here. I think it's primarily driven by valuation concerns. And that does happen. There are short sellers that will do that. It's very risky because valuations can reset

incredibly fast. I tend not to be at all interested in valuation-based shorts, but there are many institutional money firms that do that. Another potential is that there's a belief among some that there's a real concern that hardware saturation is going to come much faster and the cliff will be much steeper than the market realized. In other words, you reach a tipping point where a lot of companies say like, okay, I think I've bought all the hardware I need for my AI workloads. And then you see what was extraordinary demand. If you can imagine it, you can imagine a line and then just an immediate 90 degree drop. I think that's the other thing that some of the short sellers are betting on. That is also a risk because you don't know when

that's coming. It may come, it may come, but there are no signs of that coming right now. Tim, when NVIDIA reported earnings last week, you came onto the show and you raised some concerns about Jensen Huang's abilities as a capital allocator. A hot take, I've heard you say before, is that you think Supermicro CEO is a more skilled allocator than Jensen Huang. What is it about Liang's strategy that you think is so impressive? So this is what I think. And Kirsten, correct me if I'm wrong here, but Supermicro is far less interested in buying back stock. They'll do it, but they do it very selectively. And when they've done it, it has been at much lower stock prices, which is exactly how you're supposed to do it. that's what you're supposed to do. You're supposed to buy back stock when it's cheap and you're supposed to sell your stock when it's expensive. And I think that is roughly where his head is at. He doesn't seem to be under the same pressure as most of the big tech companies to be buying back stock to do things like offset dilution. When he's buying back stock, He's buying back stock because he believes the stock is cheap.

And he's generally done pretty well with that. So we've talked about the ups, the wild ups and less wild downs of the stock over the past few months. But moving forward, if you already have shares of Supermicro and you're intending to hold for the long term, as you two have outlined throughout this, do you have any parting words of wisdom or mindset writing advice for people as we continue to watch this story? fold and anticipate whatever might come in the near future. I mean, Kirsten, this is perfectly teed up for you here.

I would just remind people that this is why we take a portfolio approach. Supermicro is one company among many. It maybe promises a lot of opportunity. Other companies do as well. And you should balance that risk across a portfolio of ideally maybe 25 or so ideas. But yeah, diversify across a lot of your best ideas, not just this one. How fantastically foolish of you. Don't be afraid to have small positions. That diversity is the number one thing you should do. So what Kirsten said is spot on. But you can always have small positions. You know, if it's a highly speculative bet, which is what Supermicro is right now,

just keep it small. Don't spend a lot. Don't put a lot of capital into it. If it goes bananas, you're going to be rewarded anyway, because it's going to go bananas. Kristen, Tim, thanks so much for walking through this with me today. Appreciate your time and all the sage wisdom and awesome context that you have provided us with today. Thanks, Mary. Thank you, Mary. 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 i'm mary long thanks for listening we'll see you tomorrow I'll see you next time.