Prepared remarks
Good day, and thank you for standing by. Welcome to Monolithic Power Systems Inc. Second Quarter Earnings Conference Call. At this time, all participants are in a listen-only mode. After the speakers' presentation, there will be a question and answer session. To ask a question during the session, you will need to press star 1 then 1 on your telephone. You will then hear an automated message advising that your hand is raised. Please be advised that today's conference is being recorded. Now I would like to turn the conference over to Arthur Lee to read the Safe Harbor statement. Please go ahead.
Earlier today, MPS released a written commentary on the results of its operations for the second quarter ended 06/30/2026. 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 2000 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 Q2 2026 earnings commentary and in our SEC filings, including our Forms 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 Q2 2026 earnings call. In Q2, MPS achieved record quarterly revenue of $981 million, 22% higher than the first quarter of 2026 and 48% higher than the second quarter of 2025. Our performance was a result of our continued innovation, our consistency, and the resilience of our diversified market strategy. Let me take a few moments to call out a few of the highlights from the quarter. All end markets grew sequentially with enterprise data growing 45% as we continue to see strong broad-based ordering patterns. We extended our capacity goals significantly beyond $6 billion to support future revenue growth and our transformation into a full solution provider. We received initial orders for high-speed DDR5 memory; we expect to grow our SAM into next year. We began sampling high-voltage AC-to-DC products for 800-volt data center architectures to expand beyond our current AI and server core power solutions. And finally, in our automotive market, so far this year, we have shipped products for over 1,500 new sockets as we increase our footprint in both ADAS as well as in other applications within the vehicle. Overall, while we continue to adjust for the fluid geopolitical and macroeconomic environment, 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 chip-only to a full-service silicon solution-based provider. And finally, we continuously 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. Before moving to Q&A, I am also pleased to announce that our Board of Directors has authorized an additional $500 million for stock repurchases, increasing our total current authorization to $1 billion. Operator, you may now open the webinar for questions.
Questions and answers
Ladies and gentlemen, to ask a question at this time, you will need to press star 1 then 1. Please stand by while we compile the question roster. Our first question is coming from the line of Rick Schafer with Oppenheimer. Your line is now open.
Thank you, and congrats on the beat and raise. I have to ask about capacity, Tony. Communications was up 80% this quarter; I think last quarter it was up about 50% if memory serves. That is a pretty big step up. I'm curious: is that mostly transceiver power, or are you seeing meaningful contribution now from the other sockets like switches, DPUs, SmartNICs? Does that kick in? Is that helping drive that? And then any color you can give on how the second half sets up? I mean, does that momentum continue to grow into Q3?
Yeah. We see a lot of growth and a lot of demand in modules and chips, and we see it from multiple customers. Tony, do you want to go into the detail?
Yeah. I'll add a little more color. Rick, you do see contribution from both optical modules solutions as well as from what we generally bucket as switches. As I have mentioned previously, that is a large bucket that includes power solutions not only for top-of-rack switches, but for DPUs, NIC cards, and other things in the rack. So you are starting to see that grow as well. Optical is still the larger portion just because it has had a longer runway since last year, but you are seeing growth from both of those really driving the communications end market.
Okay. As a quick follow-up, I believe you are shipping 48-volt vertical power modules to a couple of customers now. Could you update us on that? Do you expect to add any more this year? Maybe a sense of what your expectations are for 48-volt vertical power modules in terms of the mix versus VR this year or next year? How would you couch it?
It is more than a couple of customers. We see new customers coming online. So we have started to ship to multiple customers and we see that increasing.
I think over the long term, as power requirements continue to increase across our end markets, we've talked about the fact that modules and solutions will increasingly be an important part of our business. So I think you will see that trend continue over the next couple of years.
Thanks, you guys.
Our next question is coming from the line of Joshua Buchalter with TD Cowen. Your line is now open.
Hey, guys. Thank you for taking my question, and let me echo the congrats on the fantastic results. Maybe to start: you gained nearly $120 million in the data segment in one quarter, which is pretty astounding. Can you walk through the drivers of that upside and growth? And importantly, you had the inventory dynamics play out a couple of years ago. Can you speak to your confidence that there is no inventory building here and the overall visibility in that segment? Thank you.
Thanks. Remember the last couple of years with all the shortages in the industry. We pulled through then, and we will continue to. There's no reason not to believe that we will make it happen even though it is very difficult. We will continue to execute.
If you look at the underlying growth drivers for that particular end market, they really have not changed from what we've talked about. We've talked about ramping existing customers, ramping new customers, seeing module content increase, platform refreshes that drive content, and CPU. We saw all of that very strong in Q2. Regarding sustainability, there are a couple of indicators. One is channel inventory, which remains very low, so we believe that is continuing to sell through. Based on what we can see, we are comfortable raising the floor for that particular end market from 85% for the year to 130% for the year.
Enterprise data center business is relatively new for us; we started to see significant business about two to three years ago. We do not have concentrated customers; we engage large and small customers. Revenue is happening broadly, and we expect that to continue into the next year.
Thank you both for all the color. I will take the hint and not ask about AI servers again. That said, it is great to see the initial orders for the DDR5, high-speed interface controller. You mentioned that could be SAM-expansive. Can you help with how expansive it could be? It seems like a new capability for Monolithic. Are there other applications beyond memory controllers for this technology?
Yes. This is new to us. This is a very high-speed technology into the gigahertz range and requires very precise analog circuitry. We can expand the technology into other communications applications once we have established the know-how. These are true fundamental know-hows. We want to get this product to launch first, then migrate the technology to other applications.
In terms of the second part of your question, which was how expansive it could be, I would still say we are very much in the early innings.
We want to prove ourselves in this particular market. It is too early for us to call any kind of revenue ramp on our side. We are just signaling that we continue to run our playbook and look for new sockets that can expand overall SAM over time.
Historically, when we've mentioned something new, it has taken time to turn into revenue. That pattern is likely to hold here as well.
Thank you both, and congratulations again.
Our next question is coming from the line of Tore Svanberg with Stifel. Your line is now open.
Yes. Thank you. Let me echo congratulations for another record quarter. How should we think about the segments for Q3? It sounds like all markets are growing right now, but on a relative basis, could you give us some color on each segment into Q3?
I think the consumer segment is the one where we are kind of still lagging because we've focused efforts elsewhere. That does not mean we give it up; we will continue in a very diversified way to grow our business. MPS is transitioning from a chip company to a semiconductor-based solution provider with our solutions. Other segments, including automotive, continue to grow. The communications side will continue to grow. Industrials are kind of lagging a little bit, and that does not mean we are defocusing.
The story will feel similar: you are certainly being led by data-centric businesses, with enterprise data and communications leading the pack. Industrials could grow a bit, but this year is more of a design-win year where we pile up additional sockets that will turn into revenue in the future. Areas we are cautious on are consumer and the notebook side of storage and compute, which we expect to remain cautious in the second half.
That is great color. As my follow-up, you mentioned you are now sampling the 800-volt solution. When should we expect to see some revenues from Monolithic Power there? Are these products based on GaN or silicon carbide, or a combination?
In the past I did not believe in GaN, and I have been proven wrong to an extent; we cannot ignore GaN now. Last year we developed our own GaN and have a working device. For the 800-volt solution, we rely on our silicon carbide devices. Revenue timing will depend on when the data center transition happens; when that transition happens, we will have revenues. In terms of wins, we know about as much as you do at this point.
Thank you very much. Congrats again.
Our next question is coming from the line of William Stein with Truist Securities. Your line is now open.
Great. Thanks for taking my question. I want to add my congrats to the fantastic results and outlook. I'm wondering if you could talk about whether pricing meaningfully influenced the sequential growth or the outlook in Q3?
Yes. If you mean whether we increased prices to drive the revenue, we do not. MPS never gouges prices when the supply chain is tight. We want to build a consistent model and execute consistently. When supply chain is tight and prices rise generally, we do not take advantage by raising prices; we operate in a consistent way, and our customers appreciate that. So the growth is not due to price increases; it is all product-driven.
To add, our approach to pricing is consistent. We have raised some prices, but primarily in three areas: where input costs have gone up to protect margins, when customers ask for expedites which affect our supply chain, and when customers ask for specific supply chains outside of China that can be more expensive. We maintain a consistent approach as we've discussed previously.
That is really helpful. Michael, I want to shift for a second to some things closer to your interests, like robotics, humanoid robotics, and home or building automation. Can you talk about your traction in those emerging areas? They are of interest to us.
We have a million-square-foot building installing our building controls. The hardware is mostly done and with some minor revisions; the software is key — the ease of use and how we implement it. By the end of this year, we should complete everything. There are multiple potential customers waiting for us to install in their buildings. To my surprise, this market is about $40 billion to $50 billion, and MPS has the key products and technology; we are building the software. Regarding robotics, especially in China there are many companies focused on entertainment robots; we have design-in solutions in the U.S. and China, and they are using MPS solutions. The next question is where robots will be used. In our own factory we use automation that is not humanoid robots; we use our own products to improve modules, testing, and reliability testing. We use our motion controls and robotics for these purposes, and our customers and suppliers will use those solutions and sell them elsewhere.
Thank you.
Our next question is coming from the line of Quinn Bolton with Needham and Company. Your line is now open.
Hey, guys. I will offer my congratulations as well. Tony or Michael, I wanted to come back to the optical transceiver part of the business, since I think that is the biggest part of communications. Can you discuss what you are seeing on the competitive landscape? Is competition mostly PMICs, or mostly discrete DC-to-DC converters from folks like TI or Analog Devices? Can you give any sense of what you think your share might be for power management within those optical transceivers? Also, I have a follow-up on capacity.
That is an important topic. Over the last few years, we've developed power modules — very high power density encapsulated modules that integrate inductors and capacitors and provide a total power solution in a very small form factor. We have some of the highest power density ICs and we fully integrate them in these modules. I will say that we are among the highest power-density companies in the world now. As long as we provide the best power density we will win many sockets, including in optical modules.
On share, we will not discuss specific customers, but the overall market still has room for us to grow. Net of both TAM growth and share gains, this can be a primary growth driver in the communications segment going forward. We cannot quote an exact percentage of share at this time.
Okay. But you certainly still see share gain opportunities in that market?
Yes. At some customers we see additional opportunity to gain share in some sockets. Net of TAM and share, there is still a substantial growth opportunity ahead of us.
Great. Coming back to capacity support: in the past, when you went from $2 billion to $4 billion, you talked about incremental capacity largely coming from outside China. As you have now built capacity up to and beyond $6 billion, can you give us some sense of the geographic split of that capacity? Is it fairly balanced between China and non-China, or does it lean one way or the other?
It is not settled down yet. We can do it depending on customer requirements. At this time, wherever they need product, it goes there. We have built a very balanced approach, but it is difficult to call an exact percentage right now.
When you call out a total number, some of the detail gets lost. We've increasingly focused not just on the foundry side but also the back-end part of the business because modules and solutions will become increasingly important. The back-end process is more complicated, so as we bring new partners on in a geographically balanced way, it applies to both front-end and back-end.
Understood. Thank you.
Especially our module assembly: it is more difficult, with a 3D effect and more complexity than phone assembly. It requires experiments and know-how to get these modules right, and we are expanding to find the equipment and capability to make it happen.
Thank you.
Our next question is coming from the line of Joseph Quatrochi with Wells Fargo. Your line is now open.
Thanks for taking the questions. Could you give an update on how you are thinking about automotive demand in the second half of this year? You talked about 1,500 new design sockets year to date. How should we think about the ramp-up revenue from those new wins?
The year is playing out pretty much as expected. We thought the first half would be flat with the second half ramping up, and we feel comfortable with that second-half ramp. On a year-over-year basis, we are still thinking the end market can be growing mid-teens. The ramp that gives us additional confidence is that it is very broadly based and not isolated to one or two customers. The diversification in automotive is increasing: the 1,500 sockets are not just ADAS, but additional sockets across the vehicle.
We are focused on zonal architectures, 48-volt systems, battery-side solutions, LiDAR — these are emerging requirements in automotive that we expect will be installed and become popular in the next few years, supporting MPS revenue growth.
Thanks for that. As a follow-up, last quarter you talked about plans to enter the RCD market and start sampling with customers. Any update on how that is going?
We are sampling and still developing. It is new to us, but we are confident it will turn into revenue. We have some revenue now, but the business is still in development.
From a modeling perspective, it is not going to be a needle mover in 2026 in terms of revenue.
Thank you.
Our next question is coming from the line of Christopher Caso with Wolfe Research. Your line is now open.
Thank you. First question: an update on where CPU server power stands right now. You have gained a lot of share over the years in that market which is heating up because of the Gen 6 CPU. How impactful has that been to the enterprise data segment, and where do you see that going into the end of the year and into next year?
One of the things we've been saying for close to a year is that CPU has been a growth driver for us in enterprise data. It has been part of the overall growth story and should continue to be one. If Gen 6 AI drives further CPU growth, that will be a tailwind. It is difficult to separate pure AI sales from CPU sales, but we are broadly indexed across both x86 and ARM players, so we think we can participate regardless of who wins in that race.
On the CPU side, what is our target market share? I have said in past earnings calls that if it is lower than 30% I would call it a failure. I think we are comfortable with our position now and we are past that threshold. We will continue to push forward.
As a follow-up, looking out two-plus years, enterprise data has been your fastest-growing end market. Do you expect that to continue to be the case, and does that make it more difficult to diversify the business?
Good question. We never focus on any single market segment. We provide foundational solutions — picks and shovels — and let market demand decide where growth happens. Sometimes automotive is big, sometimes consumer is big. We focus on fundamental development and doing the best products, and over the long term we will win. A few years ago when AI demand went sideways, other businesses like automotive grew tremendously. We aim to provide consistent growth for investors.
Even in Q2, outside of enterprise data, the rest of the businesses saw double-digit growth. So while enterprise data may be growing fastest, we are still seeing substantial growth outside of ED.
Thank you.
Our next question is coming from the line of Kelsey Chia with Citi. Your line is now open.
Hi. Based on the strong and broad-based ordering patterns, how much visibility do you have in the enterprise data end market? Can you provide any color as to how the 2027 outlook could look based on design wins, visibility, expected product ramps, or incremental revenue opportunities?
I'll start. 2027 is a bit far away; we are still trying to land 2026. Regarding visibility, the longer-term ordering patterns that began late last year have maintained. Our book-to-bill this cycle was well above 1, so we have more than a quarter of visibility like we had midway through 2025, but not necessarily all the way to the end of 2027. Underlying structural growth drivers haven't changed: best-in-class current density and modules, more sockets, and continued wins. While it is too early to put a number on 2027, the structural case for continued growth is sound.
True demand is determined by the market. We do not want our customers to end up with wasteful inventory, and we also do not want to end up with inventory we cannot sell. We balance that. We do not force NCNR orders into customers; that would damage long-term relationships. Although orders are strong, we do not manage the business by overordering; we act according to customer demand.
Got it. Also, with such a strong revenue outlook, can you help us think about gross margins and OpEx trajectory? It seems you are running somewhat below your long-term OpEx guidance right now.
Hi, Kelsey. This is Robert. I'll add to what Michael said. The strong order levels that continued into Q2 gave us the ability to incrementally expand our guidance on gross margin, again slightly, but it is there. We are at the low end of where we want to be, but we are feeling comfortable about where we are for Q3.
And, Kelsey, the last part of your question was a little bit on OpEx.
We have not changed our thesis around investing for the future. When you get to higher revenue levels, it is difficult to maintain the same level of spending relative to revenue, so you see leverage and operating margin expansion. You've seen healthy expansion in operating margin over the past couple of quarters.
Got it. Thank you.
Our last question is coming from the line of Sebastian Naji with William Blair. Your line is now open.
Thanks for taking the questions, and congrats on a great quarter. It's great to see momentum across the business. AI data center is clearly a strong tailwind for enterprise data and communications, and even a bit in storage and compute. If I start to add up all these segments, you are looking at data center exposure that is approaching roughly 50%. Is that the right way to think about it across your different end markets?
You are picking the right components. The optical module business and switch business can be pulled to the data center, and enterprise data is clearly data-center-focused. Within storage, there are three big sub-buckets: DDR5, HDD, and SSD. HDD and SSD are becoming increasingly focused on data center over time. We have not broken out storage and compute versus notebook here, but you are picking the right components that could ride healthy data center demand.
Okay. Quick follow-up on incremental supply capacity you are bringing online: is any of that on newer node processes like 40-nanometer, or is it mostly on existing 65-nanometer solutions?
I think it is mostly existing nodes. We've been talking about increasing capacity for some time. This is not only about silicon and wafers; we are building more systems, which changes the complexity beyond silicon alone. Since 2017 we've matured our procurement and qualification processes. Moving to system assembly requires new equipment and qualifications — UL and other processes — which is a different model. We've become more mature, but there is a lot more room to improve and increase capacity.
Got it. Great. Thank you so much.
There are no further questions in the queue at this time. I will now turn the call back over to Tony Balow for any closing comments.
All right. Thank you, operator. Thank you all for joining us for the conference call today. We look forward to speaking with all of you on our next call for our third quarter 2026 results. Thank you again and have a nice day.
This concludes today's conference call. Thank you for your participation. You may now disconnect.