POWI - NASDAQ

Hello, and welcome to the Power Integrations Q3 Earnings Call. [Operator Instructions] I would now like to turn the call over to Joe Shiffler, Director of Investor Relations. Please go ahead.

Thank you, Hayden. Good morning, everyone. Thanks for joining us. With me on the call today are our CEO, Jen Lloyd; and Interim CFO, Eric Verity. We're doing a premarket earnings release this morning since we'll be traveling later today to Chicago, where we'll be attending the Stifel Midwest 1x1 Conference tomorrow. We look forward to seeing some of you there. Later this quarter, we'll also be attending the UBS Technology and AI Conference in Arizona on December 3 and the Virtual Northland Securities Growth Conference on December 16. Our discussion today will include forward-looking statements denoted by words like will, expect, should, outlook, forecast and similar expressions look towards 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 and in our most recent Form 10-K filed with the SEC on February 7, 2025. During this call, we will refer to financial measures not calculated according to GAAP. Non-GAAP measures in the third quarter exclude stock-based compensation expenses, amortization of acquisition-related intangible assets, expenses associated with an employment litigation matter and the tax effects of these items. A reconciliation of non-GAAP measures to our GAAP results is included in today's press release. 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'll turn it over to Jen.

Thank you, Joe, and good morning, everyone. I'm going to cover 3 topics in my remarks today. First, I'll review current business trends and the Q3 results. Next, I'll expand on the opportunity for Power Integrations in data center following the announcement last month of our collaboration with NVIDIA on their 800-volt DC power architecture. And finally, I'll offer some thoughts on my first 100 days in the CEO role and priorities for the months ahead. Starting with recent trends, we said on the Q2 call that we had seen a slowdown in orders in July with bookings down about 20% compared to the monthly run rate of the first half of the year. The lower run rate continued through the third quarter, accompanied by a weaker distribution sell-through. Appliances are by far the largest driver of the slowdown with orders down about 40% in Q3 compared to the first half. Appliances make up the bulk of our consumer category, which accounted for about 40% of our sales in the first half. Throughout the year, we have called out the sensitivity of white goods and other appliances to tariffs owing to their high dollar value and their steel content. We've talked about the unusually strong growth in our appliance business in the first half and we've highlighted commentary from the largest U.S. appliance OEM regarding what they've called "extensive preloading" of imports from Asia in the first half. This was a key topic again on their Q3 earnings call last week. All of this is to say that the softness we're seeing in the second half is not a surprise. Appliances are a great business for us and typically generate a steady and fairly predictable revenue stream. But tariffs have severely disrupted that industry, adding to the difficulties caused by stagnant home sales in the U.S. and China's weak housing market. And because it's such an important part of our business, we are seeing volatility in our revenues. We expect fourth quarter revenues of $100 million to $105 million with the consumer category driving a large portion of the decrease compared to the third quarter. We also expect industrial to be sequentially lower, directionally consistent with recent Q4 seasonality. But our industrial business continues to be strong with revenues up nearly 20% for the first 3 quarters of 2025. That growth is coming from a broad range of applications where we are capitalizing on big picture trends like electrification and grid modernization, encompassing renewables, energy storage, high-voltage DC transmission and smart meters. Our high-power gate driver business sits squarely in line with these trends and continues to gain momentum with revenues up more than 30% year-to-date. In Q3, we built on our already strong position in the growing Indian rail business, adding a major new customer with our first design win at one of India's largest suppliers of systems for electric locomotives. We also won our largest design yet with our scale EV automotive driver boards at a major German manufacturer of drive systems for heavy vehicles. In low power, we continued our progress in passenger cars with 6 more design wins in Q3, adding to the 40-plus EV models now on the road using our products. We continue to win a robust share of inverter emergency power supplies and are using that foothold to go after other high-voltage sockets like auxiliary power supplies for battery management and onboard charging. We see strong interest in our GaN-based solutions helping to drive the market to higher power micro DC to DC converter architectures. We have a strong pipeline of design activity in these applications and expect a healthy revenue ramp over the next several years. Eric will cover the finer details of the quarterly numbers, but I do want to highlight our cash generation and return to stockholders. We generated $30 million in cash from operations in Q3 and are on track for more than $80 million in free cash flow this year. We naturally expect free cash flow and free cash flow margins to rise as revenues recover and that confidence is reflected in our cash returns. Including our fourth quarter dividend, we will return nearly $150 million to stockholders this year through buybacks and dividends. Our Board has also declared a $0.005 per share dividend increase effective in Q1 of 2026. Turning now to data center. At last month's OCP Global Summit, we published a paper demonstrating the advantages of our 1,250 and 1,700-volt GaN technologies in 800-volt DC AI data centers. We also announced our collaboration with NVIDIA to help realize the potential of the new architecture to improve efficiency, use less copper and reduce the amount of data center space consumed by power infrastructure. The white paper is available on our website and I encourage you to take a look at it. In short, our proprietary 1,250-volt GaN accommodates an 800-volt input in a conventional power supply topology, while standard 650-volt GaN requires stacking of multiple devices, compromising power density and reliability while adding complexity. Another alternative silicon carbide can handle 800 volts but has significant limitations in terms of power density due to its slower switching speed. The paper also explains why our 1,700-volt InnoMuX2 is an excellent fit for the auxiliary power socket in the 800-volt architecture. The white paper includes reliability data comparing PowiGaN to other GaN technologies. Reliability has been an obstacle to GaN adoption in the data center as well as the automotive market and the fact that we're seeing traction in both these markets speaks to the superior reliability of our unique GaN technology. In fact, one of the key attributes of our technology that NVIDIA and others in the data center ecosystem have found attractive is the fact that it is automotive qualified and already shipping into the automotive market. While we're excited about the 800-volt opportunity, GaN can also bring significant improvements in power density to existing AI data center architectures, which are expected to remain prevalent for years to come. By the end of this year, we expect to deliver early samples of our system-level GaN product for rack-level AC to DC converters with production release planned for late 2026. And now I'll conclude with a few thoughts on my first 100 days in the CEO role. As I said on our call last quarter, just after I joined, that I was excited about our unique technologies and the depth of our expertise in high-voltage processes, packaging and systems. I could also see that the need for innovative high-voltage technology is growing because of the global trends that we've talked about, grid modernization, electrification, decarbonization and of course, AI. A 100 days in, I'm just as excited about the opportunities ahead of us and developing a clearer picture of the steps we need to take to best capitalize on them. As I said last quarter, our core power supply business is back on a growth trajectory with a mix moving toward higher-margin industrial applications. The growth in our high-power and automotive businesses shows that our products and expertise has significant value in those markets, while our collaboration with NVIDIA validates the unique capabilities of our GaN technology. We continue to receive encouraging feedback in our conversations with other key participants in the AI ecosystem. I'm confident we have a lot of what we need in terms of technology and engineering talent, though it's clear to me that we need to adapt our organization and our processes to increase the ROI on our R&D spending and better match the needs of the markets that we expect to drive our longer-term growth. Data center, auto and high power have different requirements and a different geographic footprint than the mass market power supply business and we'll be taking steps in the months ahead to better align our R&D and go-to-market resources with those markets. And while we need to reallocate some resources, I don't believe we need to spend more to accomplish what we need to do. We have important hires to make, including some at the senior level, but are limiting hiring to critical needs and I'm pushing the team to tighten up on OpEx and capital spending. Our top priority is to drive shareholder value by growing our cash flow. While revenue growth is really the key to that, disciplined spending will enable us to expand cash flow margins faster as we grow our revenues. So it's something I'm emphasizing as we plan for 2026. And now for a review of the financial details, I'll turn it over to Eric Verity. Eric has been with Power Integrations for more than 15 years, serving as Senior Director of Finance for most of that time and we're very pleased to have him step into the interim CFO role. Eric?

Thanks, Jen, and good morning, everyone. I'll focus my remarks on the non-GAAP results, which are reconciled to GAAP in our press release. Third quarter revenues were up 3% sequentially to $119 million. Looking at the sequential changes, Industrial was up high single digits on strength and traction in high-voltage DC transmission in our high-power business as well as growth in metering and automotive. Communications was up high single digits, driven by strength in cell phones due in part to a design win that we announced earlier in the year for a GaN accessory charger recently launched by a major device OEM. The computer category was up mid-single digits, driven by tablets and aftermarket chargers. Consumer revenues were down mid-single digits, driven by softness in major appliances as well as seasonality in air conditioning, offset by strength in gaming. Revenue mix for the quarter was 42% industrial, 34% consumer, 13% computer and 11% communications. Non-GAAP gross margin for the third quarter was 55.1%, in line with our guidance and down 70 basis points from the prior quarter, driven by higher input costs flowing through our inventory as well as smaller benefit from the dollar and exchange rate. Non-GAAP operating expenses were $47.4 million, in line with our guidance and up sequentially due mainly to higher legal expenses. The non-GAAP effective tax rate was 2%, resulting in non-GAAP earnings of $0.36 per diluted share. Diluted share count was 56.2 million, down about 200,000 from the prior quarter, driven by repurchases. Inventories on the balance sheet fell by 18 days to 278 days. As Jen noted, we saw lower distribution sell-through in the quarter, which resulted in higher channel inventory of 9.8 weeks at quarter end. Sell-through has exceeded sell-in thus far in the fourth quarter, drawing down a significant portion of the channel inventory that accumulated in Q3. Cash flow from operations was $30 million for the quarter, while CapEx was $6 million. We used $42 million for the buybacks during the quarter, repurchasing 919,000 shares and completing our buyback authorization. We also returned $11.8 million during the quarter in the form of dividends. As Jen noted, the Board has increased the dividend by $0.005 to $0.215 per share effective in the first quarter of 2026. Turning to the Q4 outlook. We expect revenues of $100 million to $105 million. We expect significantly lower consumer revenues driven by the softness in appliances as well as somewhat lower industrial revenues. At the midpoint of the Q4 range, full year revenue growth would be about 6%. We expect non-GAAP gross margin for the fourth quarter to be between 53.5% and 54%. The decrease in Q3 reflects a less favorable end market mix with appliances and industrial driving the sequential revenue decline. Lower back-end production volumes will also contribute along with the increase in the yen versus the dollar that took place in September of last year. As a reminder, at our current level of inventory, changes in the yen-dollar exchange rate take roughly a year to affect our gross margin. We expect gross margin to rebound from the Q4 level in the first half of 2026 as mix swings back toward industrial and appliances and the impact of the yen moves back in a favorable direction. The yen has weakened considerably against the dollar of late, which should provide further support for our gross margin towards the end of 2026. Non-GAAP operating expenses for Q4 should be around $47 million, down slightly from Q3. The effective tax rate for the fourth quarter should be around 3% before rising to high single digits in 2026, driven by a lower exemption for overseas income, a provision of the 2017 tax reform legislation. Finally, I expect share count to come down by 400,000 to 500,000 shares compared to Q3, bringing our share count below 56 million. On a split adjusted basis that's significantly below the share count at the time of our IPO in 1997. And now operator, let's begin the Q&A session.

[Operator Instructions] Your first question comes from the line of Tore Svanberg with Stifel.

Yes. Jen, I was hoping you could talk a little bit more about the consumer business directionally here. Obviously, there was some pull-ins into the first half that are now being digested in the second half. But it does sound like you expect consumer to bounce back in the first half of next year, at least based on Eric's gross margin comments there. So help us understand some of the dynamics there. And maybe also you could include what the -- what this would mean for the channel inventory, whether it's going to be back sort of to that 8-week level as you exit the year?

Sure. Yes. So first, maybe let me talk about the decline that we saw what we're expecting and then maybe talk a little bit about slightly longer term. So we knew that the appliance decline was happening. We knew that Q4 was going to be sequentially lower. It was difficult to forecast just because of the lack of visibility. And the inventory situation there is really finished goods that were shipped into the U.S. and we have very limited visibility to that. But what we did see at our distributors is they did bulk up in Q3. So as you said, the sell-through ended up being somewhat soft. But we are already seeing that channel inventory coming down right now where we are in the fourth quarter. So we are expecting that to bounce back. We're just not 100% sure where the timing is going to be when that comes back. But we have heard from, for example, Whirlpool said in 2026, they're expecting that to normalize as that preloaded inventory clears out at the end of this year. So we are expecting our consumer business to get back to growth in 2026. Just it's a little bit hard to predict the timing of that. I don't know if Joe or Eric wants to add to that.

Yes. We did see a significant sell-through in October to take down that inventory that you are mentioning. And we do see it normalizing next year and we're expecting moderate growth in appliances for 2026.

One more point on that, Tore. The consumer business typically has some positive seasonality in the first half of the year because the air conditioning builds are going on for the summer. So that should -- that piece of the business should grow sequentially in Q1. The bigger question mark, obviously, is around major appliances, which is the biggest component of the consumer category. And as Jen noted, the world expects the preloaded inventory to be largely cleared out by the end of this year. The bigger question really for 2026 is just what happens with consumer demand for appliances. As you know, housing has been a challenge, certainly in China, but also in the U.S. There's not a lot of turnover in existing homes, which is a pretty big driver of major appliance sales. So with rates coming down, that could potentially help with demand for major appliances.

Maybe I'll add one last comment on that is that we still do see a great future for appliances. It's a great business for us. And I just wanted to reemphasize some of the growth drivers for that are really efficiency standards and the GaN adoption, which means more dollar content. So we do think there's going to be growth in units. And yes, on top of these macro and cyclical factors, the growth drivers are there.

Very good. That's very helpful. As my follow-up, I had a sort of longer-term question. And I think, Jen, you mentioned a little bit of this on the call where it does sound like data center, automotive and high power are going to be a big focus for the company. So I'm just wondering, does that mean you're going to change a little bit how you go to market, how you're structured internally? Obviously, today, you have the 4 main end markets and you've got tons of applications within each one. But yes, just wondering if that's going to cause a reorg and sort of the focus being more on data center, auto and high power?

Yes. Yes, maybe 2 comments there. The first one is you're correct, we are going to be focusing more on those markets, both in terms of our R&D investment, but also in terms of our go-to-market approach. And we've already taken some steps realigning our project spend to accelerate some of the developments that are in those areas. But I did want to comment that we still have a very strong core business and we will still be investing to drive that business. We're just being -- we are going to be pivoting more towards the data center, automotive and high power.

Your next question comes from the line of Christopher Rolland with Susquehanna.

I guess, first, if we could maybe talk about next quarter and how you see things playing out in terms of strength or weakness between comms, computer, consumer and industrial, that would be very helpful for us.

You mean Q4, Chris?

Yes. Yes.

Yes. I think the -- as we indicated in the script, consumer, we expect to be down pretty significantly after the accumulation of the channel inventory in Q3 that took place when the sell-through there didn't quite match what the distributors were buying for. And this is very consistent with what we've heard from Whirlpool about the pull-ins that happened in the first half, shipments coming and being preloaded from Asia. So there's clearly one more quarter there of inventory burn in the finished goods and that needs to happen. So consumer makes up the biggest part of the decline. Industrial also down sequentially. That's really just kind of a function of some seasonality in parts of the market like tools, some of the electrified or battery-powered lawn equipment and other tools that have a seasonal aspect to them. Also, some of the other parts of the business, high power has just some kind of normal lumpiness in order patterns. These are big project-driven. It's a project-driven business. So the timing of orders in high power and also metering, which is driven by government tenders in India. So it's really just a timing of orders thing there. But as Jen said, the industrial business is still doing very well. So those 2 are really going to drive the sequential decline. I think computer and comms are probably closer to flat, maybe slightly down, but the bulk of the decline comes from consumer and industrial.

And then Jennifer, maybe a data center question for you. So as I understand it, you're doing -- I believe it would be the main power conversion in the PSU for data center AI power supplies. And I think originally, this is silicon, I believe most think this is going to move to silicon carbide. Of course, you have this unique high-voltage GaN product. And so it does seem like maybe there could be a debate here, silicon carbide versus high-power GaN. How are your engagements going with the PSU OEMs? How do you guys ultimately view share shaking out? And do you think you will be the primary here or the backup here? It seems like GaN would have some cost advantages over [ SEC ]. So I'm curious if you have any prognostications as to how share shakes out between these 2 technologies longer term.

Okay. Let me try to address that. Maybe I can address that by talking about where we think the opportunities are for GaN. I can talk a little bit about silicon carbide. I think it will be difficult to say how the share is going to shake out. There are, as you know, a lot of players going after this market. But what we talked about recently is about what we think is the future opportunity as the data centers move to higher voltage like 800-volt DC. So the first opportunity there really for Powis and the aux supplies. And that's an application we already address in existing data center architectures. But in the 800-volt DC architecture, that really requires a 1,700-volt switch. And that is where the other option would be silicon carbide, but we think the 1,700-volt GaN provides some advantages. So that remains to be seen, but we believe that the power density achieved by the GaN will be stronger and make that a better choice. There's also opportunity for POWI in the 800-volt DC to DC conversion and that's where we think the 1,250-volt technology will come in and we talked about advantages there. That technology is shipping into other markets, but now we're working to build products for it -- sorry, for the 800-volt data centers. And we're engaged there with NVIDIA, but others as well at the architectural level to build products that will best suit their specs and just add that that product we expect to be released in 2027. We also think there's other opportunities, the high-power AC to DC converter that sits at the front end of the data center. We have gate driver boards there that are a good fit for that. And we have drivers for the silicon carbide modules that will be used there. So that's a place where silicon carbide will show up. So I think it depends socket to socket, whether you're going to see GaN or silicon carbide, but we believe that the GaN -- the high-voltage GaN will prevail in the aux supplies and the main DC to DC conversion.

And Jennifer, do you have wins at the power supply OEMs or through the supply chain? Or is it too early given the 2027?

We do have wins in the OEMs, yes, with the aux supply.

Your next question comes from the line of Ross Seymore with Deutsche Bank.

Why don't I just stick with the long-term question first, which is, Jen, what do you view to be the TAM opportunity for where POWI is playing in the AI data center side of things? And roughly speaking, what's the kind of time to revenue? What sort of slope are you looking at? And it was great that you guys got added to the collaboration list, definitely a positive, but you are just 1 of 13 other companies. So it seems like it is going to be a competitive field.

Yes, definitely. There are quite a few involved. And as far as the TAM, I think we think it's -- for us, it's too early to know. I mean, I think it's too early for everybody to know how fast the 800-volt DC market will take off. So it is really hard to estimate the size of the market. What we focused on is looking at what our content would be in the AI server rack. So we feel like at this point that our content is probably somewhere around 1,000, but higher in the 800-volt DC. So that's a little bit about -- you can size based on what our rack content is. As far as time to revenue, I mean, we have products today that can serve the existing data center market. The content will go up, as I said, as it shifts to the 800-volt DC architecture, but meaningful revenue generation is going to be a few years out for us. So I think 2027 is when we'll be releasing the first products that can go into the main supply of the data center architecture.

One more point, Ross. That list of 14, there are a lot of different sockets in play here and not every one of those 14 players is going after every one of those sockets. So in the 800-volt architecture, the 2 sockets that we're best positioned for are the auxiliary power supply with the InnoMuX2 products and the main converter, the 800-volt to either 12 or 54-volt socket. And we think you really need high-voltage GaN for that socket. And not everybody on that list has high-voltage GaN. So it's not that we're competing against 14 other companies for these sockets. Everybody is going after different pieces of that market. So -- and then just to add one more thing. We're talking here about the 800-volt opportunity. But the bigger piece of the AI data center market for the -- at least the near future is still going to be the existing architectures where you have rack-level AC/DC converters. We have a product that's going to be sampling, as Jen mentioned in her prepared remarks, going to be sampling here before the end of this year and then be ready for release in the latter part of 2026. We've got a good lineup of customers interested in those early samples. And that's a product we can start to generate revenue sooner than the 800-volt opportunity.

And I guess the near-term question I have, I guess, it will be more on the consumer side. But just thinking about the channel inventory side of things, it seems like you guys are burning a ton in the December quarter. As that normalizes, do you expect -- how big of a tailwind do you expect, I guess, the first half of the year in the consumer business? So whether you want to talk about what the revenues would be without the inventory burn in the fourth quarter guide or the size of that revenue on kind of a normalized consumer quarterly run rate? Whatever is the easiest framework. I'm just trying to figure out how much pain you're taking now and when it bounces back to normal, what does that really mean?

Yes. Well, the -- we were at, I think, 7.6 weeks of inventory coming out of the third quarter. We added a couple of weeks here during the third quarter -- sorry, coming out of the second quarter, we were at 7.6, added a couple of weeks largely in the consumer category. Based on what we're seeing so far through October, it looks like we'll burn off most, if not all, of the inventory that accumulated during Q3. Where that lands us in terms of weeks exactly, it's hard to say. It kind of depends on the denominator a little bit, but we should be in a much cleaner position as we start the first quarter. And then from there, it really just depends on end demand in the appliance category as to what happens with consumer growth. As I mentioned earlier, the air conditioning part of the business typically trends up in the first half. Major appliances, really more of a question mark. Tariffs not only kind of disrupted order patterns with the pull-ins, but also there's a little bit of demand destruction aspect to them as well because you get -- it affects pricing for consumers. And there's -- some of the inflation data earlier this year showed some pretty significant increases in appliance prices. So a lot of variables there. But what seems pretty clear is the preloaded inventory should be cleared out by the end of this year, at least that's according to Whirlpool and our own channel inventory should be in better shape as we exit the year. So from there, it will be a question of demand.

Your next question comes from the line of David Williams with The Benchmark Company.

Maybe first, if we're thinking about the PC market and potential adjacent opportunities there to expand the business, how do you think about maybe more of the -- on the PC side and compute, just where you think some additional opportunities could be for you guys?

So let me clarify the question. Are you talking about the data center, the server or more in the...

Yes. No. No, outside of the server, more on the PC, just more on the compute side, the more mainstream consumer-based type products.

Yes. David, I think the real opportunity in that -- in the PC market is really in GaN penetration in notebooks. I mean, that's the key opportunity for us. And that's an area we've been making kind of steady progress. There hasn't really been a mass move yet by the PC OEMs towards GaN, but there have been some. We've had some good design wins and notebooks become a pretty meaningful part of our consumer category over the last couple of years. So I think the story in PC for us is really just how quickly does GaN get adopted over the next few years. We have a lot of design activity going on. And it's really just a question of how quickly the PC OEMs who haven't gone towards GaN yet want to do that.

Great. And then maybe just on the automotive side, you mentioned some nice design wins there this quarter on top of the 40 that are already on. Can you talk maybe about the traction you're seeing there, what those opportunities look like? And do you see that as a large -- or I guess, how would you size the magnitude of that potential opportunity going forward?

Yes, I'll talk about the design win that we were referencing was for heavy vehicle win. So let me describe that a little bit. Basically, that was a win with a systems company that sells to vehicle manufacturers. So kind of like a Tier 1 for passenger cars. And we believe that that design is for a mining vehicle. So the unit opportunity there, it's much smaller than what you'd see for passenger vehicles, maybe 15:1, but the content is higher. So our content there is probably about 10x with current products. And that's where we're selling gate drivers for the traction inverters in addition to power supply chips. So we think that's a good area where we can see more wins. But it is a bit fragmented of a business, so difficult to grow rapidly, but we do expect it to be part of the mix in our auto business over time.

Yes. And then on the passenger side, which is obviously going to be the bigger part of the automotive business for us. We talked a little bit about it in the scripts. It's an area we're seeing a lot of success. The emergency power supply in the inverter is an application we're doing extremely well in, winning most of the opportunities that we go after. We just -- we have a very elegant, very effective solution for that with our automotive qualified InnoSwitch products. And that's a socket that we're using as a foothold in the automotive space. And it's getting us on the group vendor list for a lot of these OEMs, getting us access to more sockets. The architectures in EVs are evolving in a way that's very favorable for us. More power supply sockets are being built into these evolving EV architectures, auxiliary power supplies for some of the subsystems. We mentioned micro DC/DC converters, which are small power supplies that allow some of these ancillary systems to run more efficiently, handling things like over-the-air updates and video surveillance that the cars are doing when they're not being driven. Those kinds of functions all need power and they all need efficiency because you don't want to be draining the battery while your car is in what you might call standby mode. So a lot of opportunity in automotive. The SAM long term, of course, is going to depend on EV adoption. But it's going to be a very large -- continue to be a very large SAM for us and we're having a lot of success there.

[Operator Instructions] There are no further questions at this time. I will now hand it back to Joe Shiffler for closing remarks.

All right. Thanks, Hayden. Thanks, everyone, for joining. There will be a replay of this call available on our website, investors.power.com. We look forward to seeing some of you tomorrow in Chicago, and thanks again for listening.