Penguin Solutions, Inc. (PENG) Earnings Call Transcript & Summary

June 29, 2023

NASDAQ US Information Technology Semiconductors and Semiconductor Equipment earnings 63 min

Earnings Call Speaker Segments

Suzanne Schmidt

executive
#1

Thank you, operator. Good afternoon, everyone, and thanks for being with us on today's earnings conference call and webcast, where we'll go over SGH's results for the Third Quarter of fiscal 2023. With us this afternoon are Mark Adams, our Chief Executive Officer; Jack Pacheco, our Chief Operating Officer; and Ken Rizvi, our Chief Financial Officer. The slide presentation and press release that accompany this call are available in the Investor Relations section of our website, and we'd encourage you to visit that site during the quarter for the latest information about the company. I'd also like to remind everyone to review the note on forward-looking statements that appears in the press release and the earnings call presentation. As a reminder, during this conference call the company will make projections and forward-looking statements, including statements regarding its growth trajectory and its 2023 financial outlook. Such forward-looking statements rest on current beliefs and assumptions, are not guarantees of future performance, and are subject to risks and uncertainties, including, without limitation, the risks and uncertainties described in the press release and the earnings call presentation filed today as well as in the company's most recent annual and quarterly reports. These forward-looking statements speak only as of the date they are made, and except as required by applicable law, we assume no responsibility to publicly update or revise any forward-looking statements. In addition, we will discuss both GAAP and non-GAAP financial measures. Non-GAAP measures should not be viewed in isolation from, as a substitute for, or as superior to our GAAP results. We encourage you to consider all measures when analyzing our performance. A reconciliation of the GAAP to non-GAAP measures is included in today's press release. And with that, let me turn the call over to Mark Adams, CEO. Mark?

Mark Adams

executive
#2

Suzanne, thank you, and I appreciate everyone being with us for the Fiscal 2023 Third Quarter Call. During the last three years, SGH's transformation has advanced considerably. Our top line has grown, gross margins have risen, we've posted strong earnings per share, and we've reinforced the balance sheet through positive operating cash flow. Earlier this month, we announced an agreement to divest a majority stake in our Brazil Consumer Memory Business, which allows SGH to move our business toward high performance, high availability solutions for enterprise customers. Even though macroeconomic headwinds continue, we believe our position is strong for long-term success, powered by secular tailwinds in AI, machine learning and data analytics. Although Ken will go through the financial details more thoroughly, I'd like to highlight a few key points from how we performed in the third quarter. In what is still a highly difficult worldwide economic climate, our team delivered solid results for Q3. Non-GAAP gross margins came in at 28%, which is 230 basis points higher than the same quarter a year earlier, and non-GAAP diluted earnings per share were $0.66, with sales reaching $383 million. We produced robust operating cash flow of roughly $41 million during the quarter and finished Q3 with a healthy balance sheet, including all-time-high cash and cash equivalents of $401 million. Let me now walk through each of our business lines. Beginning with IPS, made up of our Penguin Computing and Stratus Technologies brands, we handle design, manufacturing, deployment and managed services for high-performance computing across the data center, cloud and edge. Across an expanding range of industries, AI, machine learning and data analytics are turning into foundational technologies for enterprise success. With over 25 years of experience deploying HPC Systems and Solutions, Penguin Computing is, in our view, well positioned at the forefront of the generative AI revolution. Some of the largest and most powerful supercomputers in the world have been designed, built and managed by our team, including oversight of more than 50,000 NVIDIA GPUs in total for AI training at scale. The way we work with customers centers on building solutions specifically suited to their needs. We design a system that brings together technologies like compute, storage, and networking, tuned to each customer's particular HPC and AI workload demands. Our ClusterWare software, developed by our skilled team, delivers an intelligent management suite that covers node provisioning, image customization, and cluster monitoring, and it also acts as a foundation for other software components. What sets Penguin Computing apart from rivals, in our view, is that we stay with customers from the design stage through integration and on to deployment and administration of large HPC and AI clusters. On top of that, we can deliver solutions on-premise, in the cloud, and at the edge. During Q3, IPS brought in $171 million in sales, accounting for 45% of SGH's overall revenue—a sign of the transformation underway. Excluding Stratus Technologies, IPS sales climbed 31% versus the same quarter a year earlier; with Stratus factored in, the gain was 79%. Service revenue, the bulk of which IPS generates, made up 15% of total SGA revenue for the third quarter, reflecting the value we add for customers. Our services span everything from point-in-time work such as design and implementation to managed services delivered over the longer term. All quarter long, we continued launching new technologies and capabilities into the market. Penguin introduced a new release of its flagship skilled ClusterWare software platform, a strong management and monitoring tool that lets customers grow their own HPC clusters. This latest release offers more scalability, improved performance, and greater ease of use. Stratus rolled out the next generation of its fault-tolerant computing platforms, which let customers run business-critical applications—from transaction and payment processing to industrial automation—at the core, in the cloud, and at the edge. In the quarter, Meta announced that the second phase build-out of its AI research SuperCluster, known as RSC, was complete. As our implementation partner, Penguin collaborated with Meta to design the AI optimized infrastructure powering their RSE, enhancing overall cluster management and implementing the system. The Navy DoD Supercomputing Resource Centers recently added a supercomputer named Nautilus — a Penguin True HBC machine featuring 48 GPU nodes. Final testing on this HPC system wrapped up in April, and it enables scientists and researchers to run sophisticated applications that uncover weather patterns and ocean modeling, producing predictive insights regarding the Earth's climate and sustainability. We are also announcing today that Dave Laurello, who joined as CEO of Stratus at the time of the acquisition, is moving into a wider role as President of IPS, taking over from Thierry Pellegrino. We feel Dave's extensive background in building and leading technology teams while serving Fortune 100 enterprises will assist us in scaling customer engagements along with overall operations. Prior to joining Stratus, Dave served in executive roles at Lucent and Digital Equipment Corporation. While interest from customers in HBC-AI opportunities has grown stronger, we acknowledge that we are still in the early phases of generative AI and machine learning. As we mentioned on previous calls, IPS sales are likely to be uneven because of deployment timing choices and the patterns of large customer installations. Longer term, with Dave at the helm, we believe we are well positioned to take advantage of these emerging market trends. Now on to Memory. Our Memory Solutions group is made up of 2 businesses at present: Specialty Memory, which concentrates on specialty memory applications for the enterprise, and our Brazil module business, which mainly serves the consumer market. Memory brought in $148 million during the third quarter, accounting for 39% of total SGH sales, and came in roughly level with the second quarter. We've built a lengthy record of supporting enterprise clients across networking telecom and industrial end markets. Going forward, a major priority is to apply that expertise toward driving innovation in memory subsystems for AI and machine learning. For instance, our CXL products removed the latency and performance bottlenecks that existed among GPUs, CPUs and memory, as well as HPC applications; design activity and customer interest in our CXL lineup are strong, and the wide variety of CXL architectures customers require aligns directly with our strengths. Beyond our CXL development work, we keep deploying cutting-edge test processes so enterprise customers get high quality and reliability. Our zero-failure-rate, or Zefr, memory offering is drawing considerable interest from major customers, since it substantially cuts field failures and helps maximize system utilization. In the quarter, we began shipping DDR5 Zefr memory models, and paired with our DDR4 Zefr memory models, these give SMART modular key differentiation along with high availability memory products. Thanks to the specialty character of our value-add business model, our memory business keeps showing relative stability across the present memory cycle. For the third quarter, that translated to operating margins of about 8% in our memory business. Turning now to our LED Solutions Group—operating under the Cree LED brand, it produces application-optimized LEDs for specialty lighting, video screens, gaming displays, horticulture, and outdoor and architectural lighting. LED Solutions reached $64 million in the third quarter of fiscal 2023, representing 17% of total SGH sales, and was up 15% sequentially from what may have marked the bottom in the second quarter. While customers continue working down inventory levels, customer design activity is picking up. Typically, as we progress through these cycles, we see that as a precursor to improving demand. While the LED market's cyclical upturn is still at an early stage, we believe the business will deliver a better fiscal 2024, and we feel good about that. Cree continues to be viewed as a front-runner when it comes to innovation driven by customer needs. During the third quarter, LED Magazine presented Brightstar awards to three of Cree's newest LED product launches, citing their innovative design. In the LED light sources category, the XLAMP Element G and Pro9 LEDs earned recognition, and the Photophyll select LEDs were recognized in the horticulture SSL and control systems category. Through sustained investment in cutting-edge LED technology, Cree is able to give its customers innovative new LEDs so they can build outstanding system-level solutions. In our view, Cree has been placed in a strong position amid these difficult conditions by bringing together best-in-class IP technology, an outsourced manufacturing approach that requires little capital, and tight cost discipline; early signals of better customer demand are appearing, and we anticipate a slight rise in revenue for the fourth quarter. With that, I'll turn things over to Ken, who will walk through our third quarter financial results in greater detail along with our outlook for the next quarter. Ken?

Ken Rizvi

executive
#3

Thank you, Mark. My comments will center on the non-GAAP figures, and you can find the reconciliation to GAAP in the tables of our earnings release; let me now move to the results for our fiscal third quarter of 2023. SGH posted total revenues of $383 million, while non-GAAP gross margins landed at 28%, the midpoint of the guidance we provided. For the third quarter, non-GAAP diluted earnings per share came to $0.66. At the start of this fiscal year, we started giving a breakdown of total revenue between products and services. As a reminder, services revenue encompasses both longer-term services and point-in-time services like logistics and implementation. Overall services revenue in Q3 reached $58 million, compared with $48 million in the same quarter a year earlier, aided by the addition of Stratus, a company we acquired at the beginning of this fiscal year. Product revenues amounted to $326 million. Revenue for the third quarter, broken out by business unit, was: IPS at $171 million. Memory at $148 million, and LED at $64 million; that works out to a sales mix of roughly 45% IPS, 39% memory and 17% LED. SGH's non-GAAP gross margin in Q3 was 28%, an increase from 25.7% in the year ago quarter, mainly on the strength of IPS. Non-GAAP operating expenses in the third quarter were $70.9 million, lower than the $72.5 million recorded in the second quarter of 2023. The sequential decline in operating expenses stemmed chiefly from ongoing cost containment efforts along with reduced bonus accruals. Third quarter 2023 operating expenses got a boost from $0.4 million of financial credits in Brazil, which was less than the $1.4 million in the second quarter of 2023 and also below the $3.3 million seen in the year ago quarter; we expect this credit to deliver only minimal benefit in the fourth quarter of 2023. Non-GAAP diluted earnings per share for the third quarter of 2023 came in at $0.66 per share, versus $0.87 per share in the year ago quarter. Our non-GAAP EPS for the third quarter reflects a tax reduction of about $6.5 million, or $0.13 per diluted share, since we expect to use additional prior year net operating losses to offset U.S. taxable income for fiscal 2023, which also helps lower our cash taxes for the year. Adjusted EBITDA for the third quarter of 2023 was $45 million, or 12% of sales, compared with $64 million, or 14% of sales, in the year ago quarter. Now for balance sheet highlights. On working capital, net accounts receivables were $244 million, versus $229 million last quarter. Days sales outstanding stood at 42 days, 6 days more than the last quarter, mainly reflecting the timing of invoicing and collections for IPS. Inventory at the close of the third quarter was $226 million, down from $294 million at the end of the prior quarter. That inventory decline was driven mainly by lower IPS and memory inventories during the third quarter. Inventory turns came in at 7.6x for the third quarter, compared with 6.3x the quarter before. As we've done historically, net accounts receivables, sales outstanding and inventory turnover are computed using gross sales and cost of goods sold, and for the third quarter those figures were $526 million and $428 million, respectively. Just to remind everyone, gross versus net revenue differs because of our logistics services, which we account for on an agent basis — in other words, we recognize only the net profit from logistics services as revenue. Cash and cash equivalents reached a record $401 million at the close of the third quarter, an increase of $25 million from $376 million at the end of the previous quarter. Cash flows from operating activities in the third quarter came to $41 million, versus $101 million in the prior quarter. For those of you following capital expenditures and depreciation: capex was $13.3 million in the third quarter, while depreciation was $9.5 million. Earlier this month, we announced an agreement to divest the majority stake in SMART Modular Brazil. This 81% divestiture of the commodity module business in Brazil lets SGH concentrate on our strategy of delivering high performance, high availability solutions to our enterprise customers. The proposed purchase price puts 100% of SMART Brazil and an initial enterprise value at $205 million, with the 81% disposition valued at $166 million. We'll get an upfront cash payment of roughly $138 million, subject to certain customary adjustments, plus a deferred cash payment of roughly $28 million payable 18 months after closing. The transaction should close by the end of calendar year 2023, subject to regulatory approvals and satisfaction of customary closing conditions. The parties also agreed to a put call feature covering the remaining 19%, exercisable across 3 exercise windows between 2027 and 2029. This put call values 100% of the Brazil business at 7.5x its fiscal year net income. For instance, if the put call is exercised in 2027, the reference net income would be fiscal 2026, and SGH would receive 19% of that value at the time of, and subject to, the exercise of the put or call. As a reference point, revenue for the Brazil business totaled $153 million through the first 3 quarters of 2023, and approximately $39 million in the third fiscal quarter. Given current run rates for the business, at closing we'd expect this divestiture to be immediately accretive to our non-GAAP gross margins by over 200 basis points, and also neutral to slightly positive for our non-GAAP earnings per share. Now let me turn to our fourth quarter fiscal 2023 guidance, which includes Brazil. We expect revenues for the fourth quarter of 2023 to be approximately $375 million at the midpoint, plus or minus $25 million. When we look at the fourth quarter, our guidance rests on a few key assumptions. For IPS, which tends to be more project-driven and can shift with the timing of hardware sales, we're expecting sequential revenues to decline slightly at the midpoint. For memory, SMART Brazil included, we see total revenues holding roughly flat on a sequential basis. And for LED, we anticipate a modest revenue increase in the fourth quarter. On a GAAP basis, gross margin for the fourth quarter should come in around 26% at the midpoint, give or take 1%. Non-GAAP gross margin for the fourth quarter is projected at roughly 28% at the midpoint, plus or minus 1%. Non-GAAP operating expenses for the fourth quarter should land near $71 million, plus or minus $2 million. GAAP diluted earnings per share for the fourth quarter is estimated at about $0.02, plus or minus $0.15. On a non-GAAP basis—leaving out share-based compensation expense, intangible asset amortization expense, debt discount and other adjustments—we anticipate diluted earnings per share of roughly $0.45, plus or minus $0.15. Our GAAP diluted share count for the fourth quarter should be about 54.5 million shares, based on our current stock price. The non-GAAP diluted share count is expected to be roughly 52.5 million shares, since it reflects the benefit of our convertible note cap costs. Cash capital expenditures for the fourth quarter should fall within a range of $15 million to $20 million. Our non-GAAP taxes for the fourth quarter are expected to be in the 11% range, as we keep benefiting from our U.S. net operating loss carryforwards; that said, looking ahead to fiscal 2024, we will have drawn down the majority of our available U.S. tax attributes, and our non-GAAP effective tax rate should rise into the low to mid-20% range. We remain prudent in how we run our operations as we work through a challenging environment, while we also keep investing in our long-term growth. For more detail, please see the non-GAAP financial information section and the reconciliation of GAAP to non-GAAP measures and tables in our earnings release; with that, let me hand it back to Mark for a few remarks before Q&A.

Mark Adams

executive
#4

Thanks, Ken. Even though we're still at an early point with AI, machine learning, and data analytics deployment, we believe SGH is positioned well to grow across the coming 3 to 5 years. So we keep transforming into a company centered on delivering high-availability, high-performance enterprise solutions that assist our customers in solving for what lies ahead; we're not just gaining from the correct secular tailwinds in terms of end markets, we also have an operating model that has held up well through the cycle. Our move toward enterprise solutions plus those secular tailwinds, together with this discipline in how we run the company, leave us optimistic about what's ahead and about our capacity to deliver long-term value to our shareholders. With that, operator, we're now ready for Q&A.

Kevin Cassidy

analyst
#5

There are, of course, probably plenty of questions right now about the IPS business given all the interest in AI. Could you describe how your RFQ funnel, or perhaps your opportunity funnel, has expanded since, say, the start of the year? And what's your approach to it? I suppose, how do you decide which customers to go after?

Mark Adams

executive
#6

Kevin, it's Mark here. Well, obviously, activity levels are clearly ticking upward; and the second part of what you asked is genuinely both an opportunity and a challenge for us, since our model works somewhat differently. What we seek are customer relationships that truly want us in a trusted adviser capacity, spanning right from the outset of our engagement on designing systems and solutions through to implementing them and then managing them after deployment, so to speak, after installation. So the criteria for how we do this, given just how much activity there is, really come down to what fits best in terms of a partnership model with the customer and what role we're able to play. That's the way we're viewing it, because as you can imagine, if it's purely a hardware play, that's a game we don't want to be in. And we're trying to pick the kinds of relationships we engage in so they're the ones that lean a bit more toward full service.

Kevin Cassidy

analyst
#7

And perhaps I could just piggyback on that: what kind of interest are you seeing across different vertical markets? Is it broader than cloud alone, or would you say the interest spans many areas?

Mark Adams

executive
#8

Yes. That's interesting too, since it has really taken off when it comes to expansion across verticals, so to speak. Of course, the cloud hyperscalers were somewhat early to market on some of this. And I'll return to that shortly, because it isn't only the big names. There's a group I'd describe as Tier 2, Tier 3 cloud providers that are building out data centers for custom-type application environments, which clearly is very interesting to us; but I think financial and health care are 2 that are very interesting for us. Oil and gas remains one that's very interesting to watch. And I'd say one we keep looking at is simply education and research, and while that was sort of the original market segment for HPC to expand across the years, AI is creating a new definition of what that opportunity looks like when it comes to science and research and so forth. And so I know that sounds like a lot, but what I mean is everyone is examining how they can enable AI and their enterprise. So it has been an interesting year to watch customer interest expand horizontally.

Tom O'Malley

analyst
#9

I’d like to touch on the AI topic once more—clearly it matters a great deal right now—but is it fair to say you have some exposure to “AI” within Specialty Memory, and then some via Penguin? If you had to give a single percentage for how you see AI as a share of total revenue today, could you size that for us?

Mark Adams

executive
#10

That’s a harder one for us to pin down. In my script, I mentioned that our deployment of GPUs at scale is roughly more than 50,000 GPUs, and if you extrapolate that across the industry, you can see that kind of scale. It’s difficult because I have to stretch definitions when we discuss what counts as an AI system: does it cover the networking piece, the storage piece, the service piece? Or is it only the platform? So we don’t truly have a solid metric, and I simply don’t want to mislead you. But the indicator I just gave—50,000 NVIDIA GPUs under management—should signal that we intend to keep growing and investing in that area. Ken, though, will add a little more.

Ken Rizvi

executive
#11

That's a fair question. I think if you take a look at it, and as Mark said, it's hard to make a call on a one-for-one basis. But when I consider systems that either rely heavily on GPUs or, in Jack's business, the specialty memory systems headed into the data center for these generative AI applications and similar workloads, my guess is that figure for our fiscal '23 is likely running above $250 million in aggregate, and could exceed $300 million when we view it on a blended basis.

Tom O'Malley

analyst
#12

So what you're saying is that for fiscal '23, the AI exposure comes in above $300 million?

Ken Rizvi

executive
#13

I'd put it as more than $250 million, getting close to that $300 million.

Tom O'Malley

analyst
#14

Mark, you also mentioned that looking back, Q2 might have been the low point for the LED business. Could you discuss what you've observed regarding shifting trends there? Clearly, China had some difficult periods. But could you just comment on what has improved?

Mark Adams

executive
#15

Certainly. For us, the main indicator has been customer design activity, which serves as a good leading index — despite some ongoing channel burn, that customer design activity is positive and rising. And look, I don't want to over-celebrate from the bottom. But revenue quarter-over-quarter, 3, 2 was up about 15%, and that happened while there was still additional channel burn. So we're not — it's not as if we're stuffing the channel with inventory to hit revenue growth. We achieved solid revenue growth from the bottom at a time when the channel is still burning off some revenues; so overall, I'd just say that, combined with the design activity, it's quite favorable. Another point I'd make is that this LED market situation isn't germane or specific to Cree. In fact, I'd encourage you to benchmark Cree's performance against the competitive landscape. Cree's done a great job; and it's simply been a tough business, very similar to the memory environment, Tom, but Cree has done a great job navigating some really tough headwinds during the macroeconomic environment they face. And I'm just really impressed with how they've kind of navigated and gotten to this today.

Tom O'Malley

analyst
#16

Helpful. And if I might squeeze in one more question. Normally, there's pretty pronounced seasonality in your IPS business heading into the November quarter. You mentioned in your slides for IPS that you've just finished phase two with Meta on the Supercluster. Is there any reason November seasonality would look different from what you've historically seen there, given that some of the programs have wound down?

Mark Adams

executive
#17

Well, that's somewhat of a loaded question, but I'll do my best to answer — though it's a hard one, because it comes down to when these deployments happen. When you're growing the way we have, we've managed to navigate much of this, but the timing of certain deployments — not only the upfront budgeting on our customers' side, or even with the large-scale customers and how they're allocating budgets. But also, once the orders come in, when they get installed and how they get installed, plus supply chain challenges — I mean, supply chain has broadly improved. But if you look at certain key technologies, the networking component of these installations, the availability of leading-edge GPUs in the industry. I just can't promise that everything will be the same as a prior year given all the factors I just went through — lumpiness, concentration, deployment timing, capital budgets, supply chain. We're long-term believers. We think we're in a really good place, but I want to stop short of calling quarters that far in advance.

Sidney Ho

analyst
#18

It's really encouraging to hear how strong demand has been on the AI side for you folks; so I have a couple of quick questions. First, are you seeing demand for those AI servers pull spending away from traditional services — in other words, it comes out of one fixed budget, though that might not be hitting you guys? And my second question is whether you could update us on your Penguin on Demand solutions, which would be great. But what appetite are customers showing for the consumption model as opposed to a traditional CapEx model, especially when it comes to the HPC and AI market?

Mark Adams

executive
#19

Could you repeat — restate the first part of that question?

Sidney Ho

analyst
#20

Yeah. The first thing is just attempting to determine that everyone is spending far more money on AI. And naturally, you're seeing a significant increase in the demand stream. But do you view that as — essentially originating from a fixed IT budget, with some of the money being drawn away from conventional spending on conventional servers?

Mark Adams

executive
#21

I can't really provide you with a definitive answer on that. What I would say is this. During fiscal '23, we observed money being reallocated when there's commercial success in deployment and the production side of the house for these kinds of systems. Regarding next year's budget process — and clearly, our fiscal year is off cycle relative to the budget process of our biggest enterprise customers — I think if I had to guess, I'd say yes, probably some budgets would shift toward this. But you're talking about different kinds of customer environments: in the UltraScale hyperscale customers, they obviously have a great deal of infrastructure they must build out — not only AI, but simply pure cloud resource capabilities. And with some of the newer verticals, I think the math adds up; and we've got — let's keep in mind that we aren't quite out of the woods yet on the overall macroeconomic headwinds. So I'm guessing there will be money that needs to be shifted appropriately, but it's such a broad question that I want to caution my answer just a little bit. And depending on the demand piece for us, that's something where we're kind of shifting our model on demand slightly. This year, in fiscal '23, we announced a partnership with Google Cloud as an HP partnership model with them. And so while we do have cycles being used on our Pod Penguin-On-Demand platform, our strategy really is to leverage some of the infrastructure capabilities through our Google partnership. In addition to that, leverage our control plane software, the skilled cloud software platform that lets people run their workloads in a hybrid environment — both on-prem and in the cloud — and really just focus on their workload optimization, not necessarily carrying where the workload is being processed.

Sidney Ho

analyst
#22

Could I maybe come back with one more question, about the sale of the Brazil memory business? You said EPS would be neutral to slightly positive. Can you walk through what that assumption covers in terms of how the divestiture proceeds are deployed? And a bit related to that, now that the Brazil memory business is sold so you can concentrate on enterprise solutions, how are you thinking about the other businesses that aren't enterprise-driven? Maybe you could speak to your M&A strategy from here? That would be great.

Ken Rizvi

executive
#23

So, looking at accretion dilution — first off, in the formal part of the transcript we said we'd expect gross margin accretion to come in above 200 basis points. And based on the current run rates, the transaction ought to be roughly neutral to slightly positive for our non-GAAP EPS; that doesn't count the net proceeds we'd get. So it's purely based on where the Brazil business stands today and its run rate. If we turn to the net proceeds, I believe that was your second question. When it comes to how we view this and where we'll put it to work, we've said there are really, I'd say, 3 to 4 vectors for thinking about capital allocation. First and foremost is to keep investing in the business. I think Mark emphasized in his formal comments the continued investment in the IPS segment. There are plenty of opportunities, as you can imagine, given our platform and our capacity to grow over the longer term in that segment. #2 is to examine M&A and inorganic opportunities like Stratus, which we closed earlier this fiscal year. Those have been terrific additions to our overall business line. #3 is returning capital back to shareholders. Looking back over roughly the past year, share buybacks have put about $58 million back in investors' hands. Beyond that, I'd note that priority 4 involves working down our gross debt. Net debt is at a very sensible level, but we're examining whether we might trim gross debt as well. So when it comes to the net proceeds from Brazil, once that deal is completed, those are the options we'd be considering.

Rajvindra Gill

analyst
#24

I've got another question about the AI business — the IPS business. You said it was up 31% year-over-year, excluding Stratus. When we think about the total market size for this opportunity, can we approach it in a bottoms-up way? How are you valuing the overall market opportunity? Are you valuing it by potential projects across certain verticals you're targeting? You mentioned conceptually that some of these verticals you're focused on are financial services and health care. Any thoughts on the long term regarding how you identify and quantify the market's size and how that feeds into your pipeline of business you want to pursue?

Mark Adams

executive
#25

Sure. I'll start things off, and Ken can jump in if he wants to add anything. Raj, I appreciate the question. The way I'd frame it is that when you look at AI, there's a sizable market opportunity out there. One lens we use is this: some of those big enterprises are going to roll out the technologies and systems themselves, and they really won't have much need or appetite for the upfront outside services and capabilities that we offer. As I've said previously — and a comparison of our gross margins against our competitors' makes it clear — that's simply a game we're not going to play. It just isn't a good game for us to be in. That isn't our strength. So I prefer to view us as a slice of the AI market in which we serve not only as the system designer and developer but also in the capacity of a service and trusted adviser, consulting-style model — full service, end-to-end. That's a narrow slice, a smaller portion of the larger pie, since big companies exist that want to deploy using their own in-house capabilities. Frankly, some of our customers will eventually want us to take on more of that, or they'll develop those capabilities themselves, and we need to keep adding new customers in that manner. That is just how this industry works—things shift constantly. There is a great deal happening on that front, and we feel good about where our business is headed, though I want to be candid: we do not represent the entire AI market. Discipline will guide us, since ultimately I have no desire to push us into a hardware-only operation; such a business would neither deliver strong profitability for us nor let us compete as effectively as we do under our core trusted adviser approach.

Rajvindra Gill

analyst
#26

Got it. I think there's real value in trying to understand how your model is different from, say, other companies' approaches. On the competitive advantage question — and again, you mentioned that your focus is on smaller, more customized solutions — where do you think Penguin Computing's genuine edge lies relative to the other players? And could you tell us who you compete against directly? Has that changed at all, given the way AI is being deployed within enterprises — are you seeing the competitive landscape shift? Any thoughts on what gives Penguin Computing its advantage? Why are you winning? And when you do lose, what are the reasons you might lose?

Mark Adams

executive
#27

Yes. I'll do my best on that. If I overlooked something, feel free to ask again about whatever I didn't cover, but this is our internal view. Picture the pyramid of engagements and customer types we're talking about: at the top are the big enterprises that want to put AI into production; the middle tier may also be fairly large companies, though their spending isn't necessarily at the level of a Tier 1 project—so you could think of it as a middle-tier project. Then, naturally, there's what I'd describe as early-stage technology deployment, though really it's more about testing applications and getting their feet wet. If you think about it as three separate tiers, we look at every one of those opportunities and how we play in them. On the differentiation question — you brought up competitors — the primary ones we run into are Dell, HP, and Super Micro. Each of them is a big hardware vendor, and their margins in this business right now are much thinner than ours. I'm not going to sit here and claim our models are inherently superior; it's just a matter of what we excel at and where we put our attention. Naturally, those three aren't the only players, but they illustrate the kind of competition we face. And when a company attempts to build its own in-house AI setups and treats it purely as a hardware contest, fine — that's their business model, and that's the margin structure they're built to succeed with. As for us, our strength lies in what we do well, and it's backed by 25 years of experience putting these kinds of systems in place. And we're not discussing a single server plus some off-the-shelf software program. Just yesterday, I visited — we have one of the first-ever installations of liquid immersion technology for a major customer of ours, and it sits in our lab. What sets us apart is that we sit at the leading edge of this technology—we're learning it as we go, which lets us bring it to market and engineer it for environments we already understand, since more than 25 years have built up a great deal of organizational know-how. So be it on the design side or the actual deployment side—consider data centers and how complex it is to link enormous amounts of compute power with memory, with storage, with networking, and to ensure the proper power infrastructure is in place. That is not easy. It's probably an understatement. Ken mentioned it, and I also covered it in my prepared remarks: we've spent 25 years doing exactly this kind of work, and in that time we've delivered a number of the largest AI deployments anywhere in the world so far. Bring all of that together and you get our added value. Sure, we know how to build these products and systems, but the real contribution is designing them around a customer's environment, getting them deployed, and running them.

Rajvindra Gill

analyst
#28

Thanks for that. Ken, one last thing on the model. You mentioned IPS excluding Stratus was up 31%. Doing the quick arithmetic, Stratus comes out at about $45 million a quarter. Is that roughly the run rate you have in mind — $45 million per quarter? And how would you describe the growth drivers in the Stratus business?

Ken Rizvi

executive
#29

Yeah. For context, that business was closer to 41% last quarter; this quarter, you're right, it's nearer that 45% to 46% range. That figure shifts somewhat between quarters, largely because of certain hardware rollouts, especially those tied to Stratus. Still, when we examine it, where has that business been expanding? Growth has come primarily at the edge, where they offer a leading solution for high availability and fault-tolerant applications. So what does that mean, and why does it matter? If you need to run an application that cannot be down for more than 5 minutes a year, you turn to Stratus. So, those applications might sit in oil refining, in oil and gas, in retail, on offshore rigs, or at water treatment facilities — but wherever a company needs a business-critical compute system running those workloads, that's the reason customers turn to Stratus, and that's exactly where this business has been expanding across the past year.

Mark Adams

executive
#30

And on the follow-on point, Raji, I’d like to add just one more remark if that’s all right. We have consistently said that sure AI and the on-premise HPC environment are where a great deal of the early activity will occur. Yet over the long term, our belief is that AI will be deployed more widely across the enterprise and certainly at the edge. What Ken just outlined will create an opportunity for us to broaden the AI and enterprise customer relationships that we already have.

Mark Lipacis

analyst
#31

Mark, regarding the IPS business, you called it lumpy. Could you walk us through the process you anticipate going through to move that toward being less lumpy? And I'm curious whether each project you take on winds up turning into some sort of recurring service or management stream across a multiyear span, and then those simply pile up one after another with every project you complete. Is that how it works? Or does it look more like becoming the in-house provider?

Mark Adams

executive
#32

That's a terrific question. It really underscores the differentiation I believe our capabilities are delivering, and the byproduct of that ought to help reduce some of that lumpiness over the long term. Yet, as you noted, thanks to our strategy, our aspiration is to grow these services over time so they become increasingly recurring and predictable. And keep in mind, we're at an early stage; we're 2 years into this, guys. I mean — and when you look at how dramatically the services have grown, we're excited about that. But our focus is this: if we choose the right engagements with our customers, we'll keep building that over time; and at present, given how substantially we have been growing, yes, services are up, but we can only sustain that through continued execution. That's why, when someone asked me earlier on this call how we differentiate ourselves, well, it's because of our capabilities and our skill sets — we're going to lean toward selecting customers who value that, rather than hardware-only transactions that don't provide that stability, since those tend to be one-offs. Building that is going to take time, guys. I mean, if you consider where we've come from, I mean, with the divestiture of Brazil, we now have a business that is a compute story with gross margin percentages dramatically up, and operating income performs pretty stellar through a hard, difficult cycle. I mean the memory business has been — this is the best it's been in 15 years in the memory business, generating cash and investing in this new transformation; and we're long-term company builders, guys. We're not trying to run off people on a quarter-by-quarter basis. We are excited about this opportunity. The specialty Memory business really plays well into these enterprise solutions for AI and machine learning and data analytics, because the biggest bottleneck right now in high-performance compute AI is the memory connectivity. And I think SMART Modular is a leader; so if you put this all together, we have a pretty bright future; it might not be perfect every quarter because of the lumpiness. Still, as we build things out and bring our customer relationships into better alignment with a full engagement model—one that includes solutions and services over the long term—I believe we can smooth out some of that lumpiness and the concentration problems we've experienced before.

Mark Lipacis

analyst
#33

Understood; and I realize you're just a couple of years into this. I mean, do you have something like the baseline of services and management, and would you be willing to break that out?

Mark Adams

executive
#34

I believe Ken has discussed it, and I'll let him clarify regarding what percentage of our top line revenue it represents. Now keep in mind, I think this — I'd better not go there, but services have risen by a multiple — a big multiple of 4 to 5x since we began here 3 years ago.

Ken Rizvi

executive
#35

Yes. I think if you look at roughly the past 9 months, services are up around 3 year-to-date — so comparing the first 9 months with the same period a year earlier, they're up about 60% year-over-year. Within that, you have point-in-time services such as logistics, design, and implementation. And then there are the ongoing services that we've also been growing over time.

Mark Adams

executive
#36

Mark, one more point — since you raised that question. For us, and this is an area where I'm deeply committed to working alongside the team, the discipline will be to avoid those one-off hardware bids, keep doing what you're doing, stick to your knitting and play to your strengths; and we can certainly design excellent systems. We've demonstrated that. But the real question is whether the customer values the overall offering, and that's something we'll keep focusing on as we continue to expand and invest in IPS.

Brian Chin

analyst
#37

Perhaps we should revisit Brazil. I suppose there's a question of whether Brazil ought to be reported within discontinued operations. More significantly, though, I'm somewhat surprised at how small the EPS effect is once Brazil is removed. Essentially, Brazil was running at a modest loss, and specialty memory operating margins appear to sit somewhere in the low to mid-teens. Thinking across the cycle, though, what did operating margins in Brazil look like when that business was running at peak revenue levels, perhaps 4 to 5 quarters back?

Mark Adams

executive
#38

Right. So if you examine the Brazil business and its position today, Brian, we attempted to lay out the first 3 quarters, along with the run rate as of Q3, which fell in that $39 million range of revenues. So it isn't a surprise in the sense that the margin profile — the gross margin profile for that business — is lower than it was, say, a year ago. Yet this has consistently been a business carrying the lowest margins, or the lower end of the margin profile, for SGH on a historical view, and well under the corporate average. Even at peak periods, it has sat well under the corporate average. Where we've performed better on both the gross margin side and the operating margin side is in the specialty segment, where we aim at enterprise solutions and a higher value-added offering for our customer base. I'm not sure whether that answers the question, but that's the state of the business, okay.

Brian Chin

analyst
#39

All right, that's enough. Fair enough. So then, perhaps to get a bit direct and to the point about IPS: I suppose, generally speaking, are you beginning to rebuild the funnel in IPS? And instead of asking about a quarter, I'll just go ahead and ask about the whole year — so how are you thinking about IPS growth rates in fiscal '24, given current visibility?

Ken Rizvi

executive
#40

Well, I think — look, we're seeing a great deal of activity right now, Brian. It's difficult for all the factors I mentioned at the start of the call, meaning the timing of budgets and deployments and how we're working on that. We are bullish long term. We're not ready to forecast the year at this point, because given the timing of our fiscal year and the budgeting process that our customers — but the budgets are kind of being worked on. We've got tremendous activity. There are some supply chain issues that we're working through, as it relates to how that would impact future deployment. Obviously, the broader macro trends. I'm not sure we're out of the woods on the macroeconomic environment. So we're trying to be careful just not to overstate where things are. We're excited about our long term, obviously, and we'll kind of give you more color as we get closer to our customers' budgeting process and future Q1 and Q2 activity.

Mark Adams

executive
#41

Brian, I see you raised the question of the disops we'll record — since this deal was made public during our Q4, we'll assess it as we move toward the close of Q4, determining whether or not the Brazil business will qualify as discontinued operations available for sale, and we'll make that assessment as we exit Q4.

Brian Chin

analyst
#42

All right. And then perhaps just one final thing, to go back to the conversation about AI — when you consider how long your sales cycle runs, I suppose your sales design and implementation cycle too, plus those potential component supply limits, would something like the middle of next year be a reasonable point to perhaps expect part of this uptick in AI engagement to turn into revenue?

Ken Rizvi

executive
#43

I believe that's fair as things stand today, and I don't mean to overhedge, Brian, but I think it's fair given what we're observing regarding some of the core technology lead times and manufacturing ramps and a few of the newer launches in the industry. I'm certain you track the industry as closely as we do, and there have been some technology announcements and early customer shipping in certain areas and some of the market leaders in GPUs, for instance, and I would expect that in the middle of next year, we begin to see somewhat more favorable conditions for supply availability.

Mark Adams

executive
#44

Thank you all for being with us on this call. Our outlook on the business remains quite favorable. When you consider the past 3 years, we believe we are considerably further along than we would have expected the transformation to be when we were contemplating it back in August of '20. There are numerous positive developments taking place at the company; we are enthusiastic for both our own team and our Brazil team regarding the paths we each have ahead of us as we head in different directions, while continuing to support one another throughout the process. Additionally, we are encouraged by the secular terrains in AI, machine learning and data analytics that will drive our future. And we are grateful that you took part in today's call.

Operator

operator
#45

That brings us to the end of the SGH Third Quarter Fiscal 2023 Earnings Call. We appreciate everyone joining us. You can go ahead and disconnect your lines now.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Penguin Solutions, Inc. transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.

Get the API View API docs →

This call discussed

For developers and AI pipelines

Programmatic access to Penguin Solutions, Inc. earnings transcripts and 132,000+ others is available through the EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments, full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.

More tools for stock research

OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Penguin Solutions, Inc. or any ticker on your watchlist.

StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.