Super Micro Computer, Inc. (SMCIP) June 2, 2026 Conference Transcript

We've got Michael Stager, who's the Senior Vice President of Corporate Development. Michael, thank you so much for taking the time today.

Thanks for having us.

Good to see you. We also have Krishna Shankar, who's the VP of Finance, and we have Ian Tully, who's the head of IR in the audience. So thanks, guys. Thanks for coming today. I've got lots of questions. I'm going to start with a question focused on customers. So, Michael, when I think about Supermicro, you guys have done phenomenally well with the Tier 2 CSPs and large data center customers. As you look out to expand the business over the next couple of years, how do you see that customer set expanding? Where do you see expansion? And I guess a question that I get from clients is, if you're doing so well with this customer set, what are the pluses and minuses of trying to get a different customer set? Why even go for that?

Yeah, that's a nice introduction. I would have to say that if you would just step back and look at the foundation of the organization and the company and where we came from, it was all enterprise customers. And if we're just talking about the customer conversation, which is important, you know, what is an enterprise, you know, is our NeoCloud's enterprises, et cetera. We grew up in the x86 world where we delivered more power and more compute per footprint for customers back in the day when those elements weren't as important as they are today. So going back to where we have been more successful as of late is technology forward customers that want to maximize the value of the platforms that they're providing to their customers, either in a service provider environment or enterprises that are serving their enterprise customers. And so we're focused on the application optimization of that. And to your point about customers, of course, AI is top of mind, how to leverage those platforms. And, you know, NVIDIA has been leading the charge. But we have in that grouping platforms with AMD, we have platforms with Intel, ARM, others. And those will be application optimized for those customer sets. And when you think about why go after all these others versus focusing on one market, if we serve the highest element of the stack, the compute platform and the integrated system in the form of DCPBS, which we'll talk about, there'll be more customers that will be attracted to what we're bringing to the market. and it'll help fuel growth in a market that's expanding quite rapidly right now.

Got it. So Michael, when we think about the AI server space, what generally the discussion ends up being is the trade-off of revenue versus margins. So help us understand, how does management trade that off? And when you look at the opportunities that you have, is it better for Supermicro to try and gain more market share or is it the time to focus on margins or both? Just give us your thought and management's thought on this.

So our first thought is to serve customers, right? From that perspective, we have a lot of customers coming to us asking for solutions. And so the tradeoff over solving a customer problem where we can leverage that platform for other customers downstream, there may be some margin give up there, so to speak, as we develop that platform. That's kind of going under the cover. So we're very focused on the margin element and optimizing the solution for the customer and bringing more to them. And we've always done that. And now we have more shots on goal to do that. So the consequence will likely be, quote, unquote, market share. I know the market likes to look at, hey, who's gaining share, who's winning share. But right now, it really does appear that there is an expansion of opportunity. At the same time, differentiation is going to accelerate, and we're going to bring those products in an expanded format to the marketplace. So, you know, it's a balance. I know it's not like, you know, we like one versus the other, but it's a balance, and we're trying to drive the growth of the company. And, again, if the size of the market is anything what our partners talk about, $2 to $4 trillion, you know, two, three years out, and we say 10% of that market. You know, it's a $200 billion revenue opportunity, and we're closing on 40, as we speak. So we have plenty of room to grow, and we have plenty of opportunity sets here to expand the margin underneath of this.

Yeah, it sounds like a very exciting time to be in this market. But one of the things that I've noticed is over the last two years, if you look at the competitor set, right, it's not just the OEMs that one thinks of, like Dell and HPE. You've got ODMs. You've got, you know, EMS companies. Everybody seems to be doing a data center rack for NVIDIA. So talk to us about Supermicro's competitive advantages that you see. Like, where is, like, you know, you were a first mover when ChatGPD came out. You were, you know, first to market doing AI servers. But talk to us about your enduring competitive advantage.

Well, at the core of the organization, it's always been bringing engineering skill set, scale, and manufacturing, if you want to call it that, to the customer and solutioning the customer. So anyone can potentially assemble something, but can they integrate it? Can they make it work as a system? Do they have the management software? Do they have the servicing capabilities? Do they have the full stack? And with DCPBS, we're very focused on bringing those solutions to the customers above and beyond just selling a server. In the older days, the transaction was a server. We looked at it as a server company. And now racks, so the backward view is like, oh, someone can build the rack. Great, but can you get 1,000 of these racks to be integrated and work together? Can you get the reliability of those systems to the point where if you're a service provider, your uptime is $99.99, and if another assembler provides less of a figure, a lesser uptime, it costs them money. So we're delivering tokenomics for these customers, and we're differentiating rapidly, and with other and different platforms. So, again, the market's expanding. There's a lot of players in the food chain, so to speak, on the supply side, et cetera, to make this happen. And it seems like the whole entire industry is focused on building out, quote-unquote, AI or leveraging this technology to the benefit of, you know, a business to mankind, to GDP, et cetera.

Got it. So, Michael, what clients are focused on now is the expansion of AI into enterprise and specifically using AI for inferencing as opposed to model building, right? So, again, coming back to the fact that Supermicro has been phenomenally successful with, you know, the tier two CSPs, the large model builders. But when we think about, you know, the enterprise space, and I'm talking about, like, say, medium, mid-tier enterprise, or at some point even small, medium business, Do you think that Supermicro offers something that its competitors can't? And is that a space that you want to even look at?

So, again, it kind of goes to solutions. So if you have an off-the-shelf solution, even in an x86 type of world, or an off-the-shelf solution for inference, right? One size does not fit all from an application standpoint from a customer. And we have a large OEM customer that has literally, we have 20 different system builds off of x86 to optimize for all the verticals. And those systems are running in the S&P 500 all over the place. We have customers that have gone to find solutions in the retail space where they need rugged eyes. Whatever it is, we're able to put that together. And the strength of the company would be on the manufacturing side, on the design side. So we can design something for them. And we're fairly nimble to put them together for those particular customers. And we keep them. And they tend to grow with us. And a perfect example, I don't know if I should be naming names, it would be Twitter, where we provided them x86 platforms. And as the technology advanced from 140 characters to video, et cetera, And as the organization grew with their capabilities, we've been able to deliver more and more differentiated or optimizations for them to benefit the organization. So it's all about solving customer problems. Customers have plenty of problems. AI is expanding. There's applications that have not yet been full-scale delivered that are yet to be, that will be the driver of more systems and more use cases. So we look to supply those types of customers.

You know, as a follow-up to that, a question we keep getting from clients is when they look at Dell and HPE, the thought is that they have lots of enterprise customers. They're well entrenched in that customer set, and they provide services that are tailored to enterprises. Talk to us about the services that you can provide to these type of enterprise clients. Is it true to say that you don't have that level of service, or are people mistaken in that? And can you talk to us about your level of service?

Yeah. So on that front, we stood up some of the largest clusters. We have some of the largest suppliers or service providers. There's more coming, more coming on stream. They need support. We're there for them. We're there for our broader basic customers. I think if you look at the traditional IDC kind of Gartner share of the overall quote-unquote market, you know, we're, I hate to use the word number two, but we've grown up from that perspective. The organization is expanding to meet the opportunities in front of us to support those customers. And so our service capabilities are improving in real time and go-to markets improving in real time. I mean, yeah, for sure, our competitors have, you know, some brand names and some history and some legacy that helps in some cases and hurts in a lot of other cases. And if we're talking about innovation, bar none, like customers know by the virtue of us serving some of the most advanced customers out there, that we have systems and we can solve their problems and they're going to come to us and we'll support them, you know, in any way we can and develop that.

Got it. Just to round out the discussion on customer types, you've talked about sovereigns as being another class of customers. I think you had an announcement with Data Vault a couple of years ago. So how is that customer set evolving? And do you think that they have different needs that Supermicro can address?

So from a sovereign perspective, and I can't speak for every nation, I mean, there's plenty of them, the anticipation and the design plans. I think we just made a recent announcement with connection with Spain, Catalonia, as an example, something that we've done also in Korea. So there's lots of activity there. And I would imagine that they're a little more, a little slower governments, if you want to call it that. But at the same time, I think that they're executing some of their initial deployments through some of the neoclouds. So you're seeing neoclouds that are forming to serve the particular areas. and the sovereigns will want localized data control, et cetera. They have a different set of circumstances, but still will want a highly tuned, efficient, fully enabled system, so a total solution. So it kind of fits into the DCBBS mold that we're presenting to them to solution those customers. And so I think the activity there is good. It's phenomenal, and there's lots more to do in that particular segment. So, yeah, that's what I would say about that.

So for those of us who don't understand this term DCBBS, can you just kind of help explain what that is and how do customers order it? Like would a data center operator say, give me three DCBBS and, you know, how does it work?

So it's tough on the Excel line item for you. So I get that, right? Like, what is it? So, you know, recently Computex, we showed a pretty broad swath of the different components that go into the data center. And so what we're trying to do for customers, you know, and because of a customer saying, hey, I need a couple of servers, right? Well, the conversation is like, well, has gone to I need a couple of racks, and we've noticed that over time that a couple of racks, they're having a difficult time integrating those racks and connecting it through the networking stack or through the storage stack, quite frankly, to power cooling. So what we're doing is designing or helping them design a full suite of solutions to fit the data center. Not every data center is the same. Not every power configuration is the same. And some have different constraints and requirements. So when they start talking about platform, we're rolling out AI, we're on the Blackwell platform, we want to be on this platform, what can you do for us to enable this? So it's enablement technology and data center building blocks, so we're building data centers, so it's no longer like you come to us for a part, which would be easy for you to model if it was just a part, right? So on that standpoint, we talked about the margin element. So it also ties to availability and readiness of the customer. So we've had customers that have had issues with power at the site or there was some networking components that weren't at the site. And the networking element held up the verification of a system so that, you know, revenue recognition would occur. So there's all kinds of things. This is very complicated to bring all these parts and components together for customers to do it. And we've said this before in the past, the hyperscalers have teams of engineers that do these things for different platforms across their footprint. And we all notice that they're using different platforms for different problems, right? So the broader market will ultimately go that route, and DCBDS is us to enable those customers to do that. And we've stated that it's a margin accretive from our perspective. It's also a way to solution the customer so that we have the competitive element in our favor. We are doing the engineering and the manufacturing for the most part here in the USA, working with our partners in the USA and addressing these customers and providing these solutions. And it's working. The growth rates of the company are excellent. And, you know, we're working on getting to a higher margin run rate as we exit the calendar year and into 27.

So you brought up a couple of things that I want to touch on. First on component availability, right? So are there still parts that you're having problems getting and is that impacting revenues? And you've given a certain guidance for the year. Like how should we think about the impact of these component shortages on your revenues?

Well, there's a lot of components that go into a rack. A lot of components that go into a data center, so it's pretty difficult to isolate which ones. I think prior to our earnings report, you know, the street was concerned about memory and memory availability and the margin compression that we see. And obviously through the reported results, we showed that that was not a constraint for us at that point in time. So we're working through it. We're working through it. We have a pretty good idea of what our customers want. We're a major supplier. We're a major buyer at the same time. We're working through the supply chain with all our vendors to alleviate any problems, if you recall, during the COVID period. Not that we want to talk about COVID because we're tired of it, but the ability to interchange components and parts really helps us in our design. So I think the supply element we'll just work through, as we've shown in real time, to deliver to customers. So it's probably always going to be something that we have to manage, and we're pretty good at it. We've been doing it for 32 years, going on 33 years.

I want to talk about manufacturing capacity. So are you at the point, or Supermicro at the point now, that you can deliver 6,000 racks per month, and how many of those can be liquid-cooled? And when we look at your footprint, what goes into management's decision to add more capacity? Because I think your existing capacity can support a lot more revenue than what you're guiding for this year.

Yeah, so Charles has always been foot forward on making sure that he had the capacity to serve customers and customers at scale. So now we're a material scale player, you know, that splits 3,000 liquid, 3,000 air cooled. There's still a pretty strong market for air-cooled. I mean, most of the data centers out there are air-cooled. Through DCBBS, by the way, we can retrofit and improve the cooling or the thermal footprint for our customers, which would give them more performance so they get more for their dollar out of their existing purchase. So there's some interesting benefits there. But the capacity element, you know, through automation, through optimization, without having to build more, I'm sure we can lift those numbers as we move forward. That all said, we've made some announcements, and we have been expanding capacity. And so the view that we have on a long-term basis of the volumes that we think that we'll be able to serve is there, and we're putting that capacity in place, and the CapEx required is fairly light from our perspective. that we've managed through it, and it puts us in a good position when we do come to multiples of scale customers. They know that we can support them, and we can support them for many, many years to come.

Got it. Let's talk about margins. So margin performance last quarter was good. There was some one-time items, but you're guiding for better margins even this quarter. I mean, 8%. Talk to us about how you see margins progressing. I mean, should investors think that there's a path to double-digit margins this year, or just how aggressive should investors be in thinking about margin progress?

So the end markets and the environment is very dynamic. So I think we all know that. From a margin perspective, we've been discussing or talking about the path to double digits. And on that front, we've shown that we can get there. It depends on the customer and product mix. So some customers might have less differentiation in the build. But the solutioning element and the diversification of the customer base will be very helpful. It'll be helpful to be able to deliver new innovation in the form of AI CPU systems that are built on x86 or ARM-type platforms where our competitors might not be there yet, and they might not be yet there to incorporate the elements that would stand those systems up in the solar solution. So when we do those things in the backdrop of large NVIDIA-type deployments on balance, and we will get to a much higher margin of revenue runway, and we're kind of looking at a step up over time as we exit counter. As you know, our year is over June 30th, fiscal year 27. We haven't provided guidance specifically on that yet. Obviously, we do at the end of the year. But the cadence of the business suggests that there's just a lot more to do with customers and a lot more product to bring to the market that could be margin-enhancing and we should be well ahead of competitors.

Got it. I have a lot more questions on this, and maybe we'll come back to this. But I want to ask you, over the last year and a half, You know, there have been several headlines that we believe have been distracting, if not, I've heard the brand of Supermicro. I want to ask you, how does leadership at Supermicro hold itself accountable for the actions and operations of the company? And what measures is management putting in so that it can address investor concerns?

So we've done quite a few things under the covers to address, I wouldn't say concerns, but to improve the strength of the organization. Certainly some of the headlines haven't been fantastic, but we're making clear steps and we've articulated those steps over time to the market as we've gone through. So I think there's not a lot to add today in that respect, but we're conscious on continuing to continue to improve processes. At the same time, customers realize that the technology that we're bringing to the market is superior. And, you know, we'll balance the two and we'll overcome any of the obstacles that potentially have occurred in the past.

So I want to touch on two things. One is on the recent export compliance issues. I know that Supermicro is having an internal investigation. When should we expect any findings? Is there a timeline for that?

So from that perspective, we didn't provide a timeline because obviously it's independent. I would imagine that that would be wrapping up in reasonable short order. And when we get the results, we will share them with the public.

And then the second thing I want to address is you had some issues with internal controls that you were addressing. Talk to us about how many issues were there, how many have you addressed, and again, how How should investors think about the time length?

So I think the one thing you have to take a step back on is the explosive growth of the organization and the changes from the processes with internal and these are system level upgrades to SAP, et cetera, linking them to inventory management, MRP, et cetera. So CRM systems. So we're going through those growing pains as we speak and all of those will need to be fully tested and will need a period of proven success. And we're tracking to do that. It's a very, it's a large effort as we want to get that aligned and move forward and to support the next leg of growth that we see coming. So that's where we're at there.

So Michael, this is a working capital intensive business, right? How does management make that trade-off of how much credit to give to customers from a days of receivable standpoint, how much inventory to have, and when you come to market for more funds, how is that decision made? And how do you see working capital and free cash flow trending over the next couple of years?

So I think if you take a step back and look at the products and you take a look at what we're going to bring to the market and take a look at the margin profile and that we have said or articulated that the customer base should expand. The DCBBS element will contribute. There will be more differentiated product across different zones, different marketplaces. Those things should compress the working capital cycle and alleviate those things, so the cycle time. So we've chewed through some really large contracts in the past from time to time, and those cycle times have been out of the 90-day period would show that kind of strain, but they'll always catch up, so to speak. So we're conscious of that, and I think the margin improvements and the diversification of the customer side should help that. All that said, okay, the pace of what customers are trying to do and the programs that they have in place, if they still stay at this pace, I mean, there could be some pushes and pulls on the timing of getting that working capital cycle compressed to the point where it's optimal and we're buying shares back like we've done in the past. So that's what I would say there.

So when we were talking earlier, you mentioned a couple of names like AMD, beyond NVIDIA. There are other companies with accelerators. Is Supermicro qualified to build the Helios rack, the MI450s that are coming up? Or do you expect Supermicro to get qualified? And do you think that you're going to be building racks for AMD as well?

So they're a fantastic partner. We work very well with them, very closely with them, and I am sure that we will be, we're already delivering AMD-based products that we'll be providing those for our customers and supporting them as we move forward. So I don't think there's a question there. Will we be doing it?

It's a matter of like, when will we be doing it when customers and the products that line up for delivery? so there's kind of a strange situation happening in the industry where if you look at non-gpu servers or non-gpu server racks i mean a lot of manufacturing companies are saying that their margins are better on those versus ai gpu servers maybe because of competition so when you think about supermicro's focus would supermicro ever consider maybe increasing your percent of revenue from non-accelerated racks, which I think you said 80% of your revenue was from AI-related or GPU-related servers. Talk to us about that. I mean, is that a focus for the company?

It's an absolute focus. I mean, again, historically, like we grew up in an x86 world, right, supplying customers. And if you take a look at customer behavior, 10% of the customers are looking for the, just roughly, right, the best performance. And they chew that performance up and they'll buy the next set as soon as it comes out. And then there's 10% of the customer base that is, you know, the lowest price. So you balance those out, that's 20% of the market that's like ideal for potentially for Supermicro, right? So we've been busy with the AI side, but those are enterprise customers. The neoclouds are serving the enterprise. So indirectly, there's an enterprise component. There's a heavy dose of CPUs in the AI systems themselves. So, you know, people aren't really calculating that. But as you move forward and you look at serial-based agentic behavior, We will have systems optimized for those customers with our partners, and we will deliver a lot of volume in that space. And, again, you mentioned competition. This highly optimized and flexible manufacturing environment is kind of contra to a large, you know, we're stamping out a million units and we can't change our line for anyone. So I think the future looks really bright for this, for us. And we're very focused on bringing the brand because we've been so successful with these higher-end systems to the rest of the market. So it's 100% a focus of the organization.

So a lot of component suppliers are asking for long-term contracts. I mean, the memory guys, the HDD guys. What does Supermicro think about this? Because if you have such high revenue growth, would it make sense for you to kind of guarantee some supply over many years?

Also, on that front, we don't typically talk about what our contracts are on the supply side, but we've been very effective at managing that and working with our partners. I'm sure they would love us to sign five years up 100%, right? So I don't know if we want to do that. So we've been very, I would say frugal, but very focused on that element. We kind of have an idea of what we need at a base level, and so we're trying to keep that in balance in real time. And, you know, I think it's an art as opposed to a science. It's partially a slight competitive advantage. Actually, not slight, a pretty big competitive advantage to be able to manage the supply side.

When I look at Supermicro's revenues, I think most of your revenues today are from the U.S. or from Asia. You also have some revenues, I guess, from Europe. How do you see the European market? And when you think about investments, and I'll ask you about CapEx and OPEX, Do you think that there's more investments to be done to grow your presence in Europe?

We have a good presence in Europe. We have some really good customers that are on stream there that are big and healthy. We see more. As you said earlier, we signed some things in Europe, southern Europe, that look pretty compelling. And I think there's lots of opportunity there. So we're not lacking opportunity in any zone in the world. we're tended to respond where demand is and work as fast as we can to help those customers.

So, look, we've got about a minute left. Talk to investors about what they should expect from Supermicro over the next year. What are they missing about the Supermicro story? What's the takeaway?

I think the takeaway is that we have a fully fleshed out product offering. It's differentiated. We're solving customer problems. We're forefront on the AI side and empowering, whether it's CPU-based, agentic, or it's AI learning, so to speak, or the machine learning. Or not the machine learning, but those elements, we have all those. We're looking at moving margins higher. We think there's a couple hundred billion dollar opportunity for the organization itself. And if you math out the valuation and the fact that we're trying to deliver shareholder value for the long term and build a much, much durable, larger organization, We've been doing that under the covers as we speak and material larger than we were the year before, the year before, the year before. And ironing all those things out and becoming a major, we're already a major player, but a much, much larger player.

Okay. I think we've covered a lot of topics. Thanks for having us.