It's still good morning, everybody. Welcome to the KeyBank Technology Leadership Forum. My name's Brandon Nispel. I cover IT hardware among many other sectors for KeyBank. Thanks all for being here. This is a 25-minute fireside chat. We have Michael Staiger, Senior Vice President of Corporate Development with Super Micro . Michael, thank you for being here.
Hey, thanks for having us. Thank you to KeyBank, and thank you for your coverage. I just want to highlight, please refer to our website regarding cautionary statements, regarding forward-looking statements. This is a standard legal disclaimer.
Let's just start. You guys reported earnings last week. Why don't you give us a quick recap of the fourth quarter?
I mean, without getting into too many details, we think we had a pretty good quarter. We're within range. We grew for the full year. We grew 47%. We have lots of new customers coming into the fold. We talked about where we grew one customer to a scale customer in 2024, four in 2025, and we expect to add two to four more in fiscal 2026.
Got it. One of the things that's come up, and I think it's sort of unclear what it exactly is, but a lot of answers to the questions were around Data Center Building Block Solutions. Can you help us understand what that is for Super Micro?
Yeah, from that perspective, you really have to look at what's going on in the industry, what's happening. AI is obviously on everyone's minds. Everyone's hearing about hyperscaler CapEx, and hyperscalers are building out like crazy, building out like mad. We have a whole host of customers that are following suit: NeoClouds, Storage, and Enterprises. From that perspective, the hyperscalers have the manpower, the staff internally to develop systems across the full scale to offer these services. Many customers don't have that capability. We're moving beyond just supplying the system or the rack to the customer, to the full suite of solutions for the customer so that they can get up and running in fast time so they can offer the services to those customers.
Those customers on a DCBBS, whether it's retrofitting air-cooled to liquid for a data center that might not be ready and adding different componentry, different services, integration services, support, cabling, and get full suites because those customers don't have those capabilities. We're providing them for them as opposed to just being in a competitive situation where it's rack to rack, where there's a lot of noise on that front, where some of the larger customers are just trying to build as fast as they can. Give me as many racks as you can. Like I said, these other customers are like, can you please help us get those racks on the floor?
You sort of alluded to it, but how does this different, sort of change how you guys operate and do business? Is this a big strategic lift, or is it sort of a little bit more simple from an operational standpoint?
I wouldn't call it a big lift. What I would call it is that we've always been focused on performance, density, systems design, engineering, et cetera, and getting the best, lowest cost per compute per watt for a customer. This just happens to be in our DNA. Innovating on a platform, as we move forward, these platforms are getting more complicated. Our ability to innovate and scale those innovations is starting to expand. It's expanding. From a competitive perspective, our focus is innovation. We're a serial innovator. If you think forward, what will things look like? There are some announcements today about the RTX 6000 for inference. The market's focused on the AI platforms. Language models are getting built out. If you think about the context windows, they're too small. As they expand, the training models are going to get bigger.
At the same time, you'll have more users on the inferencing side. We look at it in a holistic approach where we will deliver the best, most complete training platform, if you want to call it, with our partners. At the same time, we'll have innovation that goes around that on the inferencing side. We have a full set of solutions for the customer to optimize for the application. Those applications are starting to expand in use. We heard this morning at your conference a couple of users where one or two could peg one of the services by running a couple of models. That much capacity by one user gives you an idea that the core AI systems are not built out yet. The training models aren't built out yet. There's a whole ecosystem, and we're focused on building to that ecosystem.
Can you talk about sort of the higher value-add services within DCBBS, what is it?
DCBBS.
DCBBS, thank you. Like software services, how do you sort of implement more higher value-added services? Talk, you know, weave in there how you expect that to translate into like gross margins, profitability for you guys.
From our perspective, like I said, as we look at a customer and the customer's focused on trying to get up and running, we are trying to put together the pieces that they need so that they don't have to rely on multiple vendors to bring that data center up. It is a complete suite for the data center. In many instances, we're working with other ISVs, like take storage, for example, and we'll do a special purpose-built backend, whether it's flash-based or disk-based storage system with the partner. If we can deliver that at the same time with the platform itself, along with some of the cooling elements, whether it's a sidecar, whether it's the rear door heat exchange, et cetera, whatever that customer needs, we'll put that on the floor for them in record time and encapsulate services on that, so ongoing support.
When we do that, we won't be in a position where it's rack for rack pricing competition. That will lift our margins up to the ranges that we talked about in the longer term.
Got it. Let's talk about revenue growth. Super Micro's been, you know, I think an industry leader, outpacing the industry from a growth standpoint. Just help us unpack sort of where you think you are from a market share standpoint, how you've grown over the years, and ultimately taken market share.
If you look at the historical element there, about $3.5 million in 2021 to $5.2 million to $7.2 million to $15 million to $22 million, and then we got it to $33 million. Clearly, there's something there that is a huge value. The growth rates are strong because we're innovating for the customers. We're delivering the best platform for those customers, and we're giving them the best cost per compute per watt. The extension of that will be to wrap around not just the core of that, but we'll wrap around more product into that and continue to expand our growth. We feel like we're at early stages, and on a comparative basis, what we're doing is innovation as opposed to, like, can you just deliver a rack to us? People are trying to be fast followers.
Those innovations and the generations in our partners' product sets, which keep changing year to year, we're staying in pace with that. We have the capability of delivering it. We do the design, we do the engineering, we're the manufacturing. All these are in-house. We control all these elements. We're able to flex faster. Not only that, but the learnings from the prior platform will bring them to the next platform. People are focused on the reference architectures from our partners. We're focused on not just the reference architecture, making that work the best, but we're focused on the reference architecture. How can we take the reference architecture and advance it so it benefits our customer so that we have a differentiated product and so that we carry that product forward? Right now we're in a land and expand kind of mode, and we've proven our mettle.
By the fact that we've had large-scale customers that are relying on us for very important builds as we move forward, the enterprises and the smaller customers are like, wait a minute, Super Micro has these capabilities, and I know some of the other players out there aren't innovating, so I can get the best solution and best for the application that I want to deploy so I get the best bang for my buck and deliver to my end customers. We're trying to help our customers do that, and it all comes around serial innovation, being at the forefront and controlling all these elements and keeping them in-house. I think if you look at it that way and you look long term, we could potentially be the largest supplier in industry, and that's the long-term goal of our CEO founder.
Got it. You alluded to it. The company recently guided to greater than $33 billion in revenue in fiscal 2026. It's over a 50% growth rate and acceleration year- over- year. Can you help us understand sort of what you guys are seeing and what data points you can share that can help us get some confidence in that growth reacceleration?
I think the historical pattern of what we've done should give people confidence that the growth is there. The second element of that is if we had one scale customer in 2024, now we have four in 2025. We said that we expect those customers to grow with us into 2026, and two to four more would come into the fold at least in 2026. That gives you the framework of the expanse of what we're doing. That's a pretty good understanding without being Wall Street specific of who, what, why, and when. We feel pretty confident that at least $33 milliion is a great number and an indication of where we're growing.
Help me bridge from a modeling standpoint because you also guided through your fiscal first quarter 2026. There's about 10% growth at the midpoint. Is this a stair function higher throughout the year? Should we just step it up right away in 2Q? How do you want us to think about like this?
We didn't provide any specific linearity with respect to how the street would model the 2026. What I can say is that with new customers coming into the fold, with the fact that we have NeoClouds, there are more that are emerging, the fact that Sovereigns are emerging, those builds, and then you parse that back with supply, timing, build, et cetera. If they all come in at the same time, the numbers could move faster, so it's a timing element. When we think about our business, obviously we have to guide to the quarters from a street perspective, but we really think about it in terms of long term and the visibility that we have with the markets. The at least $33 million, how that works out is amazing to be seen. You've seen variance in the quarter to quarter.
Larger customers have a greater impact on things moving around. We talked about where we had a new scale customer or potential scale customer come back to us and say, hey, look, we need to re-architect this design to fit this one particular data center, which kind of altered the flow of revenues in the prior quarter. Those things can flex the business model, makes it a little bit of a challenge, but if you think more along the lines of where's the long-term direction going of one, the industry, two, the fact that we're the leading technology innovators in that industry, where we will be in 2026, 2027, 2028 is going to be in a different place than it is today. I think it'll be a much, much larger.
Sovereign seems like maybe the biggest incremental opportunity because you have some NeoClouds as customers today. How do you sort of characterize what the demand outlook or opportunity is around Sovereign? You did announce a deal for up to $20 billion for Databoard. Help us understand sort of what the Sovereign opportunity is.
As we've indicated in the past, we have numerous, numerous entities approaching us. If you think about that customer set and what they're trying to do inter-country to leverage AI to be more efficient, whether it's at the government level or through their, you know, whatever industries they own, whether it's telecom, they've been coming to us. When they take a look at Super Micro , they're not looking at Super Micro like how many racks can you give me today, right? They're looking at us and saying, hey, can you carry me through the product transitions that we've heard about into the future and support us? Can you help us set up an architected data center, which rolls back to the DCBBS? These customers also as well are fairly well funded as being government entities, right? It's a little bit of a deeper partnership.
We've seen we have business in some of these particular pockets, and we expect some of those to turn on to be much larger. It's an opportunity set that our partners are talking about. We're in the process of realizing those opportunities and developing products around to really help them out, which rolls back to DCBBS. There'll be more product elements to DCBBS that will be announced later on in the year. I think it's all coming together, and it looks pretty promising from our perspective, which rolls us back to we think that in the two to four additional large-scale customers, it's likely that there'll be a Sovereign entity in that element.
That should build over time, you know, outside of fiscal 2026 into 2027 and 2028 as other Sovereign entities sort of develop their AI strategies.
You would think about where we are right now. It's like land and expand. Obviously, there's a lot of competition, but at the same time, land and expand and innovate at the same time. That's an arc of pretty, takes a high degree of intellect to do that at the engineering level. Roughly half the force is engineers, and they're focused on multiple different SKUs and products to bring this together, to take advantage of our partners' technology so that we can serve the customer base. The customer base is going to be much larger in the future. The current customers that we have, the repeat nature of the customer where we're engaged with them is very high. We're pretty comfortable on a long-term scale, like where we're going relative to some of the things that some of the sell-siders might be saying. Not you, not you, I'm just saying.
Let's shift gears. Let's talk about gross margins a little bit. You guys just reported about a 10%, just shy of 10% gross margins. You know, the company continues to sort of put out expectations to return to a 15%, I think to 17% gross margin. Data Center Building Block , part of that solution. What have been sort of the factors that have got us from the traditional level of gross margins back to today? How do you see that playing out sort of in the future to get back to the normalized rate?
From that perspective, you know, there's been a tremendous rush to try to put something on the floor, right? You know, everyone's saying and claiming that they can do X, Y, Z, or they can deliver racks. We think over time that those claims maybe tail off a little bit. We're focused more on the full suite for the customer and doing the best for the customer. We're a customer-focused company. We're an innovation-focused company. From those perspectives, the innovation element, if we can bring an advancement in any one of these reference architectures that's significant and unique, that's an instantaneous margin enhancement profile. We think we'll be able to keep gaining ground in that area. In the short term, there's been some pressure to satisfy some customers and customer builds. We think that'll ease over time.
We're not saying it's going to ease, you know, immediately, but the focus of the company is to get back to these levels. I think even Charles mentioned like, hey, if we wanted to, we could grow faster, but there were lower margin levels. No, the focus is to have a very profitable business as we move forward and serve our customer base and expand our footprint. That's a land and expand mode right now, where we're at right now.
What are you seeing from your main competitors? Are they becoming more aggressive from a pricing standpoint? How would you sort of characterize the competitive environment, and then the impact on gross margins?
The competitive environment has always been aggressive. Everyone knows that it's a competitive environment. What I think the disconnect is that they're not valuing the innovation element that is rapidly happening under the covers. If you have to support legacy products, legacy applications, legacy software, you don't control the engineering, you don't control the design, you don't control the production. There's a disadvantage in that and in the long-term element of that business. This platform business, AI business is going to evolve. It is evolving, and we are very well positioned to take advantage of that switching change.
You're clearly trying to make some progress on gross margins operationally with DCBBS. One of the arguments that we've made is margins are sort of structurally moving lower, and it's almost just like a mathematical equation where the bill of materials in the overall server rack contains a higher cost of a GPU, and so margins sort of just structurally move lower. What do you think of that type of thesis?
I mean, it's a thesis that feels sort of evident in the numbers today. What is missing is what can we do to wrap around this core platform that brings that margin level to where we're more comfortable operating. Right now, as we garner and grab the confidence of these particular customers as we move forward, our builds, our reliability, our systems designs are the best. When we add into the other products, we have already seen this by learnings. We think that we'll be in an even better position downstream. Those factors will kind of wash themselves out.
Got it. The technology is changing so fast too. I mean, NVIDIA comes up with a new GPU every year, it seems. How do you think about sort of the company's inventory position in terms of, you know, you've had some inventory write-downs. How do you think the inventory position is? How does that sort of speed of technology change play a role?
I'll hit on the speed of technology first because in the speed of technology and the changes, literally like at quarter to quarter, builds are, they're not all the same. Right? The product is like what we're delivering might not always be the same. You know, it's easy. The platitude is like you're making a widget. The widget costs X and, you know, P x Q. Right? It's way more complicated than that. We're managing through that. We're managing through the inventory element as well so that we have the right balance for the customer at the time because we don't want to be stuck with, you know, parts that we can't use. At the same time, we're trying to be expansive to be able to offer the solutions to a broad base of customers.
If you look forward and we think that there'll be some inference platform that becomes more standardized in a volume production, we have the volume capacity to be very cost-efficient, manage that inventory for that particular product, and have that product line be super successful, whereas people might not even have that on the floor available yet. We'll wrap that around with DCBBS. It's a complicated challenge, but we've been dealing with this for many, many years and been very proficient at it. We have our eye on this and we are expecting to manage that successfully into the future.
Any questions from the audience?
Yes, we talked about cooling solutions as a special purpose built when you work with the ISVs. Overall, how is your cooling solutions being adopted by your customers? Is there anything specific you can tell us about kind of how it evolves? Is it like any vertical kind of specialties with those?
All right. Yeah, from that perspective, on the liquid cooling front, we believe we continue to lead the march there. Every system that we look at that needs liquid cooling, literally it will have some sort of change element to it. We're constantly, we'll have DLC2 that's, I believe, we made an announcement with the HGX platform where we'll be able to bring a much more efficient cooling element to the platform. There's no impedance on the dielectric fluid that we're going to be using that will increase the confidence rate of customers that are not willing to do liquid cooling because they're nervous about some element of reliability. We have super reliable systems. We're making them even better. We continue to innovate, like I said before, and the liquid cooling element of our solution is encapsulated in basically what the core of the platform is.
We're core compute in the data center, right? We'll expand upon that, but the liquid cooling is an element of the core compute. When a customer comes to us, it's like, well, you know, give us the compute platform and we're going to go get the cooling somewhere else. They want it all wrapped up and together and integrated. Our ability to innovate and integrate the liquid cooling solution using DLC2 and an advancement where we're taking all the heat out of the system, whereas before we were just taking the heat out of the chips, is more efficient. A customer can get more watts on the floor, whether it's a retrofit air-cooled, liquid-cooled solution, or whether it's just pure liquid-cooled. We're constantly making advancements on this front.
Got it. Another question.
If you had two to four or more large-scale companies.
That's a great question because we ran the business the prior year. You know, we got to $22 billion, roughly $2 billion capital base. You know, we have $5.2 billion in cash. We have access to $1.79 billion from an air credit facility. You know, $7 billion roughly of liquidity should support future growth. We are very mindful. Charles is very mindful of dilution and we're going to manage that quite effectively. Part of that was in the convert. $2.3 billion raise. We bought some shares in to limit the impact. The cap call, the conversion premium is like $81.78 to limit the dilution. We have a strong eye on that and we have plenty of capital to address the needs of the customers as we move forward to support the 2023 and beyond.
Got it. Any other questions from the audience? With one, we talked a lot about sort of the NeoClouds expansion of that customer set, right, and Sovereign . How do you think about the enterprise as a big channel for you?
The enterprise is a focus. We mentioned that on the call. The enterprise customers are fantastic. They tend to be really nice repeat customers. The ones that are technology forward happen to be our favorites because we can build to their needs. Like Tesla's an enterprise customer. There are lots of others under the covers. We have a lot of the enterprise customers that are using the NeoClouds services to test out. They know under the covers that Super Micro is powering a lot of these things. By virtue of our success with some of our NeoClouds customers and some of the other enterprise customers, we're well known. We're getting more opportunities, chances to bid. We're scaling up the go-to-market strategy for the enterprise to offer those customers the solutions that we have. We give them the greatest choice. One solution is no longer like a serial compute.
It's super commoditized. As we move forward, it'll be a little bit different. Each industry vertical has a different kind of use case. I think our largest partner suggested that. Whether it's to be digital twins, whether it be in the bio industry, drug discovery, there are lots of different use cases. We're optimizing for each one of those. We think there's a lot of opportunity. Those customers will need the DCBBS support that we talked about. At the same time, they don't quite have the scale purchasing power that some of the handful of scale customers will have. The pricing environment should balance itself out.
Margin, more margin creative.
We should be margin creative. Yeah, absolutely.
We're just about out of time. Michael, thank you very much for being here. Everybody, thanks for your participation.
Thank you.