I will now hand the conference over to Joe Shiffler, Director of Investor Relations. Please go ahead.
Thanks, Ben. Good afternoon, everyone. Thanks for joining us. With me on the call are Jen Lloyd, our CEO, and our CFO, Nancy Erba. After Jen and Nancy's prepared remarks, we will open it up for questions. Slides accompanying today's earnings release and conference call can be found on our investor website at investors.power.com. Our discussion today will include forward-looking statements denoted by words like will, expect, should, outlook, forecast, and similar expressions that look toward future events or performance. Such statements are subject to risks that may cause actual results to differ from those projected or implied. Such risks are discussed in today's press release, in our most recent annual report on Form 10-K, and in subsequent quarterly reports on Form 10-Q. During this call, we will refer to financial measures not calculated according to GAAP.
Non-GAAP income statement measures exclude stock-based compensation expenses, amortization of acquisition-related intangible assets, accruals for a judgment in a legal matter, a restructuring charge recognized in the first quarter of 2026, and the tax effects of these items. A reconciliation of non-GAAP measures to our GAAP results is included in today's press release and in the accompanying slides. This call is the property of Power Integrations, and any recording or rebroadcast is expressly prohibited without the written consent of Power Integrations. Now I will turn it over to Jen.
Thank you, Joe, and thanks everyone for joining us. We delivered solid Q3 results with revenue up 5% sequentially, led by strong growth from GaN products. We beat our outlook on gross margin and expenses, reflecting our continued emphasis on profitability. Our non-GAAP operating margin was up more than six points year-over-year in Q3, while non-GAAP EPS rose by 28%. These results demonstrate that we are executing well across our end markets. Industrial demand remains healthy, with particular strength in energy infrastructure and rail. In fact, we currently have more demand than we can satisfy in our gate driver business due to certain supply constraints. In personal electronics and computing, we saw strong sequential growth in Q3, driven by new GaN design wins and seasonal ramps.
While memory shortages are a factor in these markets, the impact on our business is mitigated by our customer mix and new design wins. In the consumer market, which makes up a third of our sales, demand continues to be soft due largely to stagnant housing markets and higher prices for appliances, driven by tariffs and rising raw material costs. Against this uneven backdrop, we are on track to achieve our key financial and operational priorities for 2026. These include revenue growth, improved profitability in the markets we serve today, advancement of our GaN technology roadmap, and progress in our strategic pivot to the data center, EV, and energy markets to drive faster growth in the years ahead.
I am excited about our progress in each of these areas, and I will provide an update on each, starting with our core business, where growth is being led by penetration of GaN across our end markets. We are on track for a record year of GaN revenue and had another strong quarter of design wins as customers look to our GaN products to improve efficiency and reliability. That includes our first design wins with TOPSwitchGaN, which we introduced earlier this year, winning three major appliance designs at top-tier OEMs. Other wins include a recently announced smart home product for a major U.S. OEM and a TV design for a major Korean OEM using a GaN chipset addressing both AC to DC conversion and power factor correction. I will turn to data center, where our high-voltage GaN continues to open doors at customers designing 800 V systems.
At last month's OCP Summit, we met with hyperscalers, XPU suppliers, power supply and server ODMs, and suppliers of energy infrastructure. Our discussions reinforced the demand for technologies that maximize power density and efficiency in 800 V data centers. Engagements span applications from grid to rack, ranging from solid-state transformers and renewable energy to high-power sockets in the rack and power sidecar, as well as a diverse range of AUX power sockets throughout the data center. For high-power sockets, customers continue to show strong interest in our 1250 V GaN to optimize power density, both by collapsing stacked power stages to reduce board area and by taking advantage of GaN's high switching frequency to shrink magnetic components. We are developing a range of high-voltage GaN products for the main power path, reflecting customer interest in a variety of integration levels and package configurations.
These products are expected to sample to customers in 2027 and then roll out in the 2028, 2029 timeframe. Auxiliary power represents our most immediate opportunity to penetrate 800 V architectures. We are seeing strong growth this year in AUX power for existing server architectures, including five new design wins in Q3, and we have four design-ins for 800 V sockets scheduled to begin production in 2027. Our opportunity pipeline has more than tripled since the start of the year and now includes a wide variety of 800 V designs, including AUX supplies for compute trays, IT racks and sidecars, as well as network switches, fan power modules, and solid-state circuit breakers.
In conjunction with the OCP Summit, we announced a collaboration with Ideal Power on a solution for solid-state circuit breakers. The collaboration addresses the need for innovative approaches to circuit protection in high-power, high-voltage data center environments where mechanical circuit breakers are too slow to protect the valuable rack equipment. Existing electronic breakers introduce losses into the system, both from the protection device itself and the surrounding control electronics.
As rack power levels rise to hundreds of kilowatts and higher, any drain on system efficiency is a significant problem. An added challenge is that electronic breakers must be powered, and in 800 V systems, the power source is usually the 800 V bus itself. That means the AUX supply must be capable of handling an 800 V input. So in the architecture we're developing with Ideal Power, our 1700 V InnoMux handles the 800 V with ample margin. We power the intelligence that controls the Ideal protection switch, which in turn protects the power path to the compute load.
The result is an efficient, self-powered breaker that meets the needs of 800 V data centers, and we're excited to partner with Ideal Power on this innovative solution. Customer engagements for data center involve not just rack-level power conversion and protection, but also solutions for beyond the walls of the data center. Attach rates of battery storage systems to data centers are projected to rise rapidly in the years ahead, and we are well-positioned to capitalize with our successful high-power gate drivers. One of our top high-power customers is now supplying battery storage inverters to a top U.S. hyperscaler, with plans to ramp with multiple new customers in 2027. We added to our position in battery storage in Q3 with a high-value design win at a new customer, a major U.S. supplier of energy infrastructure.
We are also expanding our customer engagements on solid-state transformers being developed for later phases of the 800 V transition. Our high-power gate drivers are an attractive solution for driving switching modules in silicon carbide-based SSTs, while GaN-based SSTs are an emerging application for our high-voltage GaN technology. Finally, turning to automotive, we won 10 new designs in Q3, including an inverter emergency power supply for a top German automaker, and our pipeline has increased by more than 40% since the start of the year. We believe the outlook for the EV market is promising, as higher gas prices have renewed consumer interest and the pricing delta between EVs and ICE vehicles continues to shrink. OEMs are moving to distributed power architectures in which micro DC-to-DC converters will power subsystems from the main battery instead of the lead-acid batteries.
One of our Chinese automotive customers recently announced its intention to eliminate the 12 V battery, and we expect other OEMs to follow suit. Meanwhile, rising battery voltages are driving interest in our higher-voltage GaN, not only for low-power sockets, but also higher-power DC-to-DC converters and onboard charging. Together, these two architectural trends have the potential to dramatically increase the number of high-voltage sockets per vehicle, driving our available dollar content from single-digit dollars to nearly $100 over the next four to five years and significantly expanding our SAM. In closing, we're excited about the opportunities ahead of us in data center, energy infrastructure, and automotive. We expect these markets to drive rapid growth in our SAM, with a goal to double SAM by 2031. Our strategy for driving revenue growth in these markets is taking shape, and I'm confident in our team's ability to execute the strategy.
I'm delighted to announce a new addition to the team, Tony Stratakos, who will join us next week as our Chief Technology Officer and SVP of Engineering. Tony is a noted thought leader in power semiconductors, having founded Volterra, served as CTO at Maxim Integrated, and later founded a company specializing in utility-scale battery storage systems. I'm pleased to have Tony joining the team, and I look forward to introducing our executive team and providing an in-depth discussion of our strategy at an Investor Day in March 2027. Now I'll turn it over to Nancy for a review of the financial highlights.
Thanks, Jen, and good afternoon. Our Q3 results demonstrate continued progress on revenue growth, margin expansion, and cash generation. We reduced balance sheet inventory for the third consecutive quarter and announced an increase in our quarterly dividend. These results reflected disciplined execution across the company. Channel inventory increased during the quarter, reflecting both softer demand in certain markets and higher customer buffer stock requirements, which I'll discuss in more detail. Revenue for Q3 was $125.5 million, up 6% from a year ago and 5% sequentially. The communications and computer categories each grew more than 20% sequentially on a combination of seasonal strength and recent GaN design wins. The industrial category continues to show healthy growth, with a 6% sequential increase in Q3 and 13% growth year-to-date compared to last year. That strength, led by high power, has been fairly broad-based within the industrial category.
Consumer revenue was down 7% from the prior quarter, driven mainly by seasonally lower revenue from air conditioning. As Jen noted, we have seen continued softness in appliances as that market continues to feel the effects of higher retail prices and weak housing markets, especially in the U.S. and China. As previewed in my introduction, channel inventory rose by 4.5 weeks during the quarter to 11.8 weeks. Three factors drove this increase. The first is softness in appliances, where we expect sell-through to take one to two quarters to improve. The second is project delays in metering, particularly in India. We are starting to see signs of improving sell-through in metering in Q4 already and expect project launches to continue ramping through 2027. The third is channel partners building inventory to support end customers' higher buffer stock requirements.
Some replenishment was expected as we entered Q3 with below normal channel inventory of 7.3 weeks, but the increase was higher than anticipated. Both our channel partners and end customers tell that amid broader supply chain constraints, they are holding additional buffer stock. Channel inventory, therefore, is likely to remain above our eight-week target while these conditions persist. We will continue to monitor sell-through and customer requirements closely as we manage shipments. Turning to gross margin, non-GAAP gross margin was 55.2% for the quarter, above the high end of our outlook and up slightly from the prior quarter, benefiting from improved mix and higher manufacturing volumes. Non-GAAP operating expenses in Q3 were $42.4 million, down $2.8 million from the prior quarter. Compared to the third quarter of last year, non-GAAP operating expenses were down more than 10%.
We remain focused on disciplined expense management while protecting priority investments in GaN and our strategic growth markets. Our objective is to maintain ongoing profitability improvements while advancing the technology and customer programs that support longer-term growth. Non-GAAP operating margin for Q3 was 21.4%, up more than four points from the prior quarter and more than seven points higher than the full year 2025 figure. This improvement reflects the operating leverage in our business model that comes with revenue growth and disciplined expense management. Non-GAAP earnings for the quarter were $0.46 per diluted share, up 24% sequentially and 28% from a year ago. Turning to the balance sheet and cash flow, cash flow from operations was $43 million for the quarter, while CapEx was $6 million, resulting in free cash flow of $37 million.
Lower inventory was a significant contributor, falling by more than $7 million during the quarter. Days on hand fell by 19 days to 246 days at quarter end. Cash and investments on the balance sheet increased by $25 million during the quarter to $288 million. Reflecting the strength of our balance sheet and confidence in our ability to grow cash flow going forward, our Board of directors has raised our quarterly dividend to $0.22 per share beginning in Q1 2027. I'll now review the fourth quarter outlook. We expect revenue to be between $118 million and $126 million. At the midpoint, that would be an increase of 18% from the fourth quarter of last year and bring our full year growth rate to 7%.
Our outlook reflects continued industrial strength and contributions from recent design wins, balanced against softer consumer demand, elevated channel inventory, and supply constraints, including packaging and PCBs. I expect non-GAAP gross margin for Q4 to be in the range of 55%, ±0.5 point. Non-GAAP operating expenses for Q4 should be in a range of $43 million, ±$500,000. At the midpoint, that would be a slight increase from Q3, but once again, lower than a year ago. We are on track for a full year reduction of about 4% as we continue to drive expense discipline while protecting priority investments in our strategic growth markets and advancing our technology roadmap. Finally, I expect non-GAAP operating margin for Q4 to be between 19% and 21%. In summary, Q3 showed our ability to improve profitability and generate strong cash flow.
We maintain measured in our near-term expectations for market growth while encouraged by our design win momentum and longer-term opportunities. Our focus is on disciplined execution, investing selectively in our strategic priorities, and translating growth into sustainable earnings and cash flow to drive shareholder value. We look forward to sharing more about our strategy and target business model at our upcoming Investor Day on March 8, 2027, in New York. Joe will be sharing more details in the coming weeks. Ben, let's begin the Q&A session.
We will now begin the question-and-answer session. Please limit yourself to one question and one follow-up. If you would like to ask a question, please press star one to raise your hand. To withdraw your question, press star one again. We ask that you pick up your handset when asking a question to allow for optimum sound quality. If you are muted locally, please remember to unmute your device. Please stand by while we compile the Q&A roster. Your first question from the line of David Williams of Needham & Co. Your line is open. Please go ahead.
Hey. Good afternoon, and thanks for taking my question. Maybe first, you had talked about at OCP, it seemed like a lot of enthusiasm around the higher voltage power architectures. Was there anything there that changed maybe your view? It sounds like you're getting the design wins, maybe better opportunities, but do you think there's anything there that pulls things forward or makes you maybe a little more optimistic about that 800 V opportunity?
Yeah. Thank you, David, for the question. Yeah, I mean, what we've talked about before and we continue to be really happy about and our progress on, first of all, we're engaging across the ecosystem. We are seeing the AUX designs growing in terms of our pipeline and our revenue across a broad range of power. As I mentioned in the prepared remarks, we are now seeing design-ins specifically for 800 V there. So we're happy about that, and we're continuing to demonstrate good performance in available sockets. As we've also talked about, the main path architecture for us, we're really focused on that opportunity with our 1250 V GaN, so a little bit further out in time from the current 650 V designs that are going on.
I think what was exciting at OCP is really seeing the expansion of the partnerships beyond what's inside the data center to some of the new things that we're seeing on the racks, as well as what's happening outside the data center. As we discussed, we're pretty excited about our high-power business's ability to address some of those opportunities. So yes, I think we're seeing a broader opportunity for ourselves.
Great. Thanks for the color there. Maybe some of the inventory dynamics, it seems, I think it is more of your exposure there on maybe on the consumer side, but that is kind of a different story than what we have heard from a lot of your peers. Is there any other dynamics that we should be thinking about there? I mean, obviously, you pointed to multiple things that drove that, but do you still feel like the demand is there and there is no real softening on the broader demand environment? Does it feel like that is the case, or is there other things maybe that we should be thinking about driving that inventory? Thanks.
If I talk about consumer first, we are seeing softness there, and it is typically a softer quarter with air conditioning, as you are well aware, in terms of that cyclicality. So that inventory buildup in China, and we are looking at it really region by region and by channel partner, as I talked about in the past. That buildup will take a couple of quarters to work through. If I look at India and the metering specifically, that is really a delay in project starts within India, which have now started to loosen up, and we have already started to see that in Q4. So I feel very good about that higher inventory in India moving through, starting in Q4, but really into next year as those larger projects kick off and get that moving within the channel.
The third area is one on the supply side, which is actually new. As you can imagine, we have been having very frequent discussions with channel partners and our end customers as we saw this build up. Certain customers, end customers, are requiring these channel partners to hold additional inventory as they are seeing other supply constraints. So that piece is the piece that will likely take a little bit longer to work through, and it is really going to depend on when those end customers allow that buffer stock level to decrease. But that demand is continuing to be healthy, and as Jen said, we are seeing more and more wins that are going to drive continued revenue growth. Unfortunately, it turned into a blip this quarter in terms of that higher level of weeks on hand.
Great. But it does not feel like it is potentially a golden screw type scenario where you are building inventory waiting for other critical components in the supply chain?
I think some customers may be.
Okay.
But on a broader base, I think no, we have multiple reasons for why we saw the buildup.
Thanks so much. Appreciate the help.
Your next question from the line of Tore Svanberg of Stifel. Your line is open. Please go ahead.
Yes, thank you. Maybe following up on that last topic. I am just trying to understand, Nancy, is it fair to say that the 4.5 weeks is sort of equally distributed within those three reasons? Assuming that supply will remain tight for the industry, as you sort of go through 2027, would you expect the channel inventory to be a little bit more elevated than normal? Would your target be kind of more like nine weeks versus eight? Any color there?
I think it is fairly distributed. China with appliances, we saw that tick up later in the quarter, so it seemed to be moving fairly well in the first part of the quarter and then ticked up toward the end. That I would expect to move through in the next couple of quarters, as I said. I think the metering will move through in the next, call it two to three quarters as well, as those projects get underway. The one that is a little bit more challenging to predict is those that are increasing buffer stock. I think for Q4, the level of weeks on hand will remain higher than we have targeted.
I think for next year, it is potentially a scenario where it sits at nine to 10 weeks until we see the reduction in buffer stock, but we will have to just keep watching it and stay in connection with our end customers, not only our channel partners, but their customers, to really understand when they start to see that buffer stock requirement reduce.
Got it. But it is fair to say that it is probably not going to go above 11.8 weeks, or at least you wouldn't let it go above that level at this point?
Yeah. We certainly will do everything we can not to have that happen, yes.
Very good. As a follow-up for you, Jen, I do recognize 1250 V, your GaN technology is a little further out. You also didn't mention 1700 V on the call. Just curious, especially given OCP and so on and so forth, what's the level of interest in both of these very unique GaN technologies that you're now sampling? Thank you.
Yeah. I thought I did mention 1700 V. Maybe I didn't, but a couple of the wins that we referenced were using the 1700 V technology, just to clarify that. I think the level of interest is high. I mean, what I'm really hearing, again, and it's back to just the criticality of power in these systems, is that the interest is high in technologies that will improve the efficiency, the power density, the board area, which translates to power density. The challenge with the uniqueness is the goals around sourcing, right? So right now, our 1250 V is very unique, and so we're working with our customers on it, but we will have to align and meet some of the challenges that they'll see with that as a unique technology.
Very good. So it's fair to say that the design wins for both 1250 V and 1700 V, as far as materializing the revenue, we're looking at 2028, 2029 timeframe?
No, no, 2027 for the 1700 V.
Okay.
Yeah, the 1700 V will be in 2027, and it's AUX.
Got it.
Right?
Right.
Then, main will be out in 2028, 2029.
Perfect. Very good. Thank you.
Your next question from the line of Christopher Rolland of Susquehanna. Your line is open. Please go ahead.
Hi, guys. Thanks for the question. I guess mine is around your supply chain, your foundry partners. They're a little bit different from others. Some might consider a little bit more unconventional. This has benefited you in the past when others have been kind of super tight, expanding lead times, et cetera. I guess, first of all, have you seen this as an opportunity? Is this something that you talk with your partners about, maybe more availability of supply or lower lead times? Secondly, do you think this has anything to do with the channel build, maybe the channel getting these parts because they are available in front of either expanding lead times or price hikes like they've seen from other partners?
Yeah. I think that is a potential. I want to see Q4 and Q1 start to improve and see things move through, but that availability certainly has benefited us historically. As Jen mentioned, in certain areas, in particular in high power, we are actually tight ourselves and are asking customers to provide longer visibility to their forecast because we are seeing some level of shortages in PCBs as well as in packaging needs. We are starting to experience a little bit of that as well on the high power side. On low power, certainly the availability is a benefit to us to be able to meet customers' needs, and we are working closely with our end customers to make sure that they understand and that we can see that visibility to make sure that that continues to be an advantage for us.
Excellent. I think you preempted my second question. You may have been referring to the silicon gate drivers. You said they were supply constrained. I guess you might have just mentioned those chokeholds, but if there is anything else to discuss there. Then also just the demand that you are seeing for this, is it substantially above or is it really just these supply chain factors?
We have right now more demand in our high power business than we can supply, which is good, right? It is building that backlog. We are doing everything we can and working with our supply chain to expedite and bring in the, and I will use PCBs as an example, to try and bring in more in advance so that we can meet that demand.
Is that from an upside in demand or your demand expectation, or is it just—
Yeah.
—are you highlighting the supply chain issues?
No, it's an increase in the demand, which is great. As Jen mentioned, that's one of the areas that we are hearing a lot of interest from our customers and from new customers in terms of our capabilities in high power is certainly a large area of focus in the industry right now.
Thank you so much.
If you would like to ask a question, please press star one to raise your hand. Your next question comes from a line of David Williams of Needham & Co. Your line is open. Please go ahead.
Thanks for taking the follow-up. Jen, just want to ask, I know in the past you've talked about your portfolio being a bit more narrow, and you would like to expand that a bit. Of course, you have those initiatives internally to build that organically. But if you look out and think about inorganic opportunities, how do you rate the environment today? Are there opportunities for you? When could we see potentially you being more acquisitive in the future, if that is where you would like to go? Thanks.
Well, we're probably not going to pre-announce anything here on the call. That's a joke. But we are, as we have mentioned, actively evaluating different technologies and are looking at the ecosystem around us, and whether that's through partnerships or potential M&A, are looking to expand our product portfolio. Our first priority is to do that organically, and I think Jen will talk a bit about that with Tony joining. But we are absolutely looking at other ways that we can potentially use an inorganic opportunity to drive scale.
Yeah, I'll just add to that, David. We are looking at what we can do organically, and I think you'll hear more as time goes on. We've talked about in automotive, expanding our BOM content, and some of that will be through some new topologies. Same thing in data center. We're seeing some things adjacent to our traditional space. So, stay tuned on that. Yes. Thanks for asking that.
Great. Thanks for that. Just maybe lastly, if I could, on the Ideal Power partnership or relationship there, can you talk to us a little bit about how that is set up, and will that be a combined technology, combined packaging, or is it a transfer of IP? Just any color around what that relationship could look like. Thank you.
Sure. Yeah, that is really a co-development, I would say. What is exciting about it is that we are seeing that the two companies coming together, our technologies play really well together. There are advantages in combining those technologies. It is really early days. We are co-developing, we are working on some reference platforms and demonstrating the technology to customers. We are talking to a lot of customers. It is very early days. We are excited about that work.
Thanks again. I appreciate the help.
Yep.
There are no further questions at this time. I will now hand the call back to Jen Lloyd for closing remarks.
All right. Thank you again for joining us today. I want to thank our employees at Power Integrations around the world for their hard work, innovation, and commitment, as well as our customers and partners for their continued collaboration and support. Thanks to the global trends like AI and electrification, power is becoming a critical constraint, driving the need for step function improvements in efficiency and power density. Our high voltage technology and system expertise are well-suited to meet these challenges, and we continue to execute on our technology and product roadmaps to create a long-term value for our stockholders. Thank you for your continued support, and we look forward to updating you again next quarter.