Monolithic Power System Inc (MPWR) — Q1 2026 Earnings Call Transcript

Earlier today, MPS released a written commentary on our results of operations for the first quarter ended March 31, 2026. This document can be found on our website. Before we begin, I would like to remind everyone that in the course of today's presentation, we may make forward-looking statements and projections within the meaning of the Private Securities Litigation Reform Act of 1995 that involve risks and uncertainties. The risks, uncertainties and other factors that could cause actual results to differ from these forward-looking statements are identified in the safe harbor statements contained in the Q1 2026 earnings commentary and in our SEC filings, including our Form 10-K and Forms 10-Q, which can be found on our website. Our statements are made as of today, and we assume no obligation to update this information. Now I would like to turn the call over to Tony.

Thanks, Arthur. Good afternoon, and welcome to our Q1 2026 earnings call. In Q1, MPS achieved record quarterly revenue of $804 million, 7% higher than the fourth quarter of 2025 and 26% higher than the first quarter of 2025. Our quarterly performance was a result of our continued innovation, our consistent execution and the resilience of our diversified market strategy. Let me call out a few highlights from the quarter. Our communications end market grew 33% sequentially on the strength of our power solutions for optical modules and switches. The pipeline for our automotive and enterprise data end markets, including server, continue to accelerate as we won multiple new projects across customers and regions. We sampled our first high-speed interface products for DDR5 at major customers and MPS continued to grow our capacity past our original $4 billion plan with a new goal of reaching $6 billion in the near future. We continue to adjust to the fluid geopolitical and macroeconomic environment, but our diversified market strategy remains unchanged. MPS focuses on innovation and solving our customers' most challenging problems. We consistently invest in new technologies that open new end markets and applications and accelerate our transition from chips only to a full-service silicon-based solution provider. And finally, we continue to expand and diversify our global supply chain, allowing us to capture future growth opportunities, maintain supply stability and rapidly adapt to market changes as they occur. Operator, you may now open the webinar for questions.

I just want to dig a little bit into the enterprise data side of things. Can you just talk about the different trends you're seeing between the XPU side versus the server CPU side? I know you mentioned in your preamble that the backlog and visibility was improving in both. But given the strength of demand we're hearing elsewhere in the server CPU side of things, I wondered how you guys are doing there.

Both are good. Yes, Tony, you can talk to that.

Yes. I'll give a little more color. And Ross, if you recall, even last year in 2025, we had talked about CPU being a tailwind, and we continue to see that here in 2026. But if you look across enterprise data, it's increasingly hard to differentiate between AI solutions and CPU. In general, all the growth drivers are intact. We're ramping new customers, we've been ramping existing customers, and we continue to see the transition to modules. As I said, plain server has been a tailwind, and we think it will continue to be so.

And I guess the second question would be on the storage and computing side of things that seem to be a little bit better than feared in the first quarter. Talk about the tailwinds or headwinds given what's happening from a macro perspective and then potentially the difference between what you guys do on the storage side versus the computing side?

Yes. I'll start on that one, and then I'll let Michael and Rob jump in. But as you know, that segment really has two separate businesses in it. The storage side obviously has remained strong as it's indexed to a lot of the data center business. We see strength in DDR5. We see strength in HDD and SSD continuing out of last year and into Q1. On the notebook side, we're still more cautious there. As you know, there are really two dynamics: potential TAM headwinds associated with memory shortages or elasticity from memory prices. Also, we selectively play in the parts of that market that have lower margins around consumer. If we look forward to those two parts of the business going through the year, I think we're still very optimistic about storage staying strong and probably much more cautious on the notebook side.

For notebooks, we don't focus on quarter-to-quarter fluctuations as long as we develop the best solutions for power density and our customers' ease of use. The design wins we have will ramp revenue when customers need them.

Great. First, I'd like to ask about manufacturing. You noted in the press release that you passed the $4 billion target, you're now working to $6 billion capacity. Maybe you can update us as to the strategy around geographic placement of your capacity? And remind us what's going on from a technology perspective. This used to be a big focus of the various BCD iterations that you produce. Can you bring us up to speed as to what is the latest BCD generation?

I'll answer the technology question first. We are still around 60 nanometers and we may go down to 40 or 45 nanometers. As power increases, we increase power densities. It's a continuation of the work we've been doing for the last 20 years, and we believe we do it better than our competitors. Regarding the $6 billion manufacturing goal, we clearly see a lot more activity and potential in our near future. The design wins are imminent and will turn into revenues.

Maybe to add, on the $4 billion of capacity, we've talked about that being very geographically diverse, both inside and outside of China. Our strategy is to maintain supply chain diversity, and we'll continue to try to keep that balance going forward.

Congrats on the results. Maybe to start, can you just help us a little bit with the models? Any help you can give us on the guidance by segment as we think about sort of a 12% sequential growth for the June quarter, which segments should be above and below?

Let me start with what you guys are most interested in. Last time we talked to you, we said we had a 50% floor. I'll let Tony give you better news today. I'm also excited about other projects we're deeply involved in, such as building automation, audio, and robotics. These will pave the way for the next two to three years to keep us on the same growth trajectory.

I'll follow up and march through it. Josh, I know the first thing people are interested in is enterprise data, so I'll start there. We tend to be fairly conservative and wait for backlog to be in place. Late last year, we talked about 30% to 40% growth year-over-year. On the last call, we raised that to a 50% floor. The strong ordering patterns that started last year continued through Q1. At this point, we're comfortable raising that floor up to around 85% year-over-year growth. That will certainly be one of the drivers for the year for MPS. On communications, we've become increasingly excited about that end market with optical module growth and switches. Auto is a consistent story: roughly flat for the first half of the year and ramping later in the year. Storage and compute have two dynamics: we're optimistic on storage pulled by data center, but more cautious on notebook.

I'll let Tony deliver better than our last CFO.

Thank you both for all the color there. Unfortunately, when you deliver good news, you still get annoying follow-up questions. If we think about the incremental upside since last quarter, any help on how much of that's coming from CPUs versus more confidence in content or visibility into share on the AI accelerator side? Congratulations again.

That's a good try. I'm not going to give it to you.

Josh, we'll fall back to what we've been saying: the growth drivers are intact. We don't want to try to parse out between volume and content because it can be very specific.

In reality, it's very difficult to separate what's called AI from what's called servers. There's a lot of overlap. Portable AI devices and GPUs are happening, and they clearly overlap with CPU and GPU-powered solutions.

I'll add my congratulations and just a wow on the outlook. Maybe if I could ask about enterprise data: the top four cloud service providers I think just last night announced very large CapEx. It seems like you guys are clearly seeing increased order velocity. My real question is, are you able to capture all that upside? Is there anything hurting your suppliers or your ability to capitalize? Historically, you've been ready for upside and never caught short. Is that still the case?

I think that's exactly right. In AI and GPU power, performance, manufacturing capabilities, and reliability mean there are only a few players. Over the past couple of years, it's become clear MPS is one of those players. As I promised over a year ago, we continue to do well, especially on power density. We provide total monolithic power solutions: a single piece of silicon versus competitors using multiple pieces. That shows our advantage. We don't want to be the dominant supplier; we want to be part of a diversified growth strategy.

Got it. For my follow-up, I'm curious on physical AI. It's getting a lot more attention. Can you flesh out your plans for that segment? What kind of TAM have you identified? You mentioned robotics earlier. Can that be a meaningful revenue contributor next year, or when would we start to see robotics drive top line?

We see robotics activity this year, but volumes are still low and its impact on revenue is modest for now. It's very early and difficult to predict. Many companies have launched initial high-volume robots and we benefit from that. As AI adoption in robotics grows, applications will widen.

I think what you see is us running the typical MPS playbook: engaging broadly and winning designs. We can't control when customers ramp, but we can control winning sockets, and that's what's happening in 2026.

That's a very good point. Yes.

I'll offer my congratulations as well on the results and outlook. Michael, Tony, I wanted to ask on the comms segment. It was up 33% sequentially. It sounds like it could be one of the faster-growing segments in the June quarter. When I look at optical modules, 800-gig modules are more than doubling in '26. Do you think the comms segment could actually grow as fast, if not faster, than enterprise data this year given those trends?

I'll follow Tony's answer. We're not in the business of predicting market rates. We saw a lot of activity and demand for high-power-density products, especially modules. The power density in small confined areas is critical, and that's our core technology. We execute fast and capture that market.

As ordering patterns have continued to be strong and extend, we still see that through the year. It's tough to call the back half of the year, but we'd put that end market above the corporate average.

Got it. My follow-up: you guys have been sampling products for around 800 volts for a few quarters now. Can you provide feedback on how that activity is going? There's debate on whether higher-power conversion steps will be GaN- or SiC-based. If it goes GaN, will you have GaN-based solutions ready?

We are focused on silicon carbide for the 800-volt applications. I used to say I didn't believe in GaN, but last year we developed GaN for lower-voltage, lower-power segments. We began fundamental GaN development for those markets. To answer your first question, yes, we're sampling and co-developing with customers and their system-level customers. Our 800-volt product is working. The broader environment for a new 800-volt power bus in data centers has many issues to be resolved, but our application is ready. There are also other voltage segments up to 10,000 volts that require more development, and MPS will play in those segments as appropriate.

Congrats on another record quarter. A follow-up on power: there's a move to 2,000-watt GPUs. Just hoping you could touch on two of the three things that make MPS unique and differentiated to handle those power levels. That is not 2028 — it's already next year. Some color there would be great.

That's a good question. One point is we focus on monolithic integration. We pursue the most cost-effective integration and can integrate or disintegrate as needed. Integration allows us to put solutions in modules — a huge advantage — versus multiple discrete silicon pieces that make module manufacturing difficult. Second, we invested in module development since 2016 to move from silicon-only power conversion to plug-and-play solutions. We developed our own automated test and reliability systems, which are unique and necessary for high volumes and high quality. These systems are based on MPS eMotion products. Third, from a semiconductor perspective, we've used 60-nanometer and are moving to 40-nanometer, increasing power density significantly.

Great. As my follow-up, you mentioned a new product. You talked about 800-gig optical components earlier and then you mentioned you have your first high-speed interface product sampling for DDR5. When should we start to see material revenue from that business? Could that grow into several hundred million dollars over the next few years?

That's absolutely possible. From a business perspective, it's a natural way to expand our service and total addressable market. We have a good position in PMIC for memory and introduced timing drivers, timing control, temperature sensors, and now high-speed interfaces. Our engineering team has pulled this off with fewer people than some competitors, and our customers welcome another capable supplier. We sample these products. As to timing, it's hard to predict exact revenue ramps, but the market segment has welcomed another player.

Tore, to help with modeling, I wouldn't expect that to be a contributor to 2026 revenue. We highlight it as we continue to expand our footprint in that market.

Let me also extend my congratulations. I'm curious about the comms business that was a standout for the quarter. How much content do you have in these 800-gig optical modules and in rack switches? Can you put it in the context of how much you see content increasing in rack-scale solutions for accelerated compute given this beachhead in these two new applications?

I think it's more than a beachhead; we're well beyond the beach now.

We're going to stop short of giving a dollar content. In optical modules, the module-level solution is important so we look at module content as opposed to just discrete devices. For switches and rack solutions, there are multiple elements — switches, NIC cards, and other boards — that require power. The opportunity is that multiple processors in racks require power solutions, and we've been expanding across those within our Communications segment. We won't provide specific content layers for specific customers.

As my follow-up, distribution channel inventory has been running lean. Is it still lean relative to where you would normally place distribution inventory? Also, can you talk about inflationary input-cost trends and pricing actions you may need to take?

I'll take that. With respect to our distribution channel, we don't have perfect visibility, but what we have seen in 2025 into 2026 is that the channel has been very lean. That implies we're shipping to actual end-market demand. Beyond that, things look good.

For pricing, some input costs are higher. We see more activity and demand, and our goal is to keep our margin profile intact.

To add, we're not seeing broad-based inflation across the board, but there are places where input costs have risen and customers ask for expedited supply chains. In those cases, yes, we may raise prices to maintain our gross margin model.

Maybe a follow-up on gross margin. Could you share the puts and takes on the guide? It seems like positive revenue acceleration but margins stayed in that range. Any color would be helpful.

We've seen margins at the lower end recently. We continue to improve yields on modules. I don't see a major headwind, but I don't want to give false hope that margins will jump significantly. That's not the MPS approach.

Historically, we've been consistent delivering to our gross margin guidance. For the last four quarters, we've been flat at 55.5%, which is at the low end of our gross margin model that ranges from mid-50s to upper 50s. For Q2, we had the confidence to increase our gross margin guidance incrementally because we have better visibility to backlog that started in Q4 last year and continued into Q1. We do see potential headwinds in the second half, so we remain cautious in our guidance for H2.

Right. That's helpful. Maybe on robotics socket opportunities: you talked about those being up for grabs or to win this year. Are those incremental? At Analyst Day, you mentioned maybe $150 of content for a humanoid. Is that the right way to think about it, or are those opportunities expanding?

It varies by application. Humanoids are visible, but we focus on robotics that are untethered and battery-operated. Battery management plays a role there. On the compute side, GPUs and automated control units require power. We also provide sensors and actuator solutions for motion. Applications include medical assist and rehab. We sell both chips and modules, and dollar content varies widely. It's difficult to give a single number, but the trend is clear: robotics adoption will grow and automate many tasks.

Hi. First, on enterprise data: can you talk about growth on merchant solutions versus ASIC solutions this year, and what you expect with regard to content? Do you expect outsized growth in one area or the other?

We don't divide it strictly into merchant versus ASIC for the learning and inference side — they often use similar products. What matters for why we're winning is power density. Customers want smaller size and high efficiency. It doesn't really matter to us which segment they call it.

The only thing I'd add is we're comfortable raising the enterprise data floor from 50% to 85% because of what we see in backlog and extended ordering patterns. Nothing fundamental changed in our growth drivers other than improved visibility to orders in the books.

Got it. As a follow-up, on the auto segment, you mentioned flat in the first half and growth in the second half. Auto has been variable in recovery. There was some data out of China that was weaker earlier in the year. Can you give color on visibility in auto and why you think it starts to grow in the second half?

I don't focus on regional strength. We're not chasing the market. We listen to customers, prepare product and inventory, and deliver when needed. We focus on delivering better products and winning sockets.

The shape of the year for auto hasn't changed. We expect ramp later in the year based on designs we've won coming to market. We can't control customer ramp timing, but that's our current belief.

Congrats on the results. Could you talk about the rationale behind focusing on silicon carbide for 800-volt step-down while focusing on gallium nitride for lower-voltage, lower-power segments? Some peers are using GaN for higher-voltage step-down. How would that influence your competitive positioning? Also, historically you've been able to gain share in tight supply environments. Could you talk more about your supply chain management strategy and your confidence in meeting customer demand if other suppliers face capacity constraints?

It's a long answer. I used to say I didn't believe in GaN, but that was for high-power applications. Silicon carbide has been proven over decades and the material and device reliability is well established. We started focusing on SiC early and have deep know-how to use it effectively in our integrated module solutions. We're selling integrated modules, not just discrete SiC devices. Regarding GaN, we developed GaN for lower-voltage, lower-power segments starting last year. For 800-volt high-power applications, we believe SiC is currently the right approach given its maturity and reliability.

Regarding supply chain, given your history in 2021 and after COVID, how are you positioned to meet demand if other suppliers face constraints?

Throughout our history, we actively listen to customers and preemptively build inventory when needed. Our product life cycles are long, so we can carry some inventory without significant scrap risk because we expect to sell these products eventually. We try to be proactive rather than passive in supply chain management.

To be clear, nothing about our enterprise data outlook or the 85% floor is limited by supply chain constraints. We've continuously stayed ahead of supply chain challenges, so that is not an issue right now.

Thank you, operator, and thank you all for joining us on this conference call today. I look forward to speaking with you on our next call for our second quarter 2026 results. Thanks, and have a great day.