Power Integrations, Inc.

Good day, everyone, and welcome to today's Power Integrations third quarter earnings call. At this time, all participants are in a listen-only mode. Later, you will have an opportunity to ask questions during the question-and-answer session. You may register to ask a question at any time by pressing the star and 1 on your telephone keypad. Please note this call is being recorded, and I will be standing by should you need any assistance. It is now my pleasure to turn today's program over to Joe Schiffler, Director of Investor Relations.

Thank you, Brittany. Good afternoon, everyone. Thanks for joining us. With me on the call today are Balu Balakrishnan, Chairman and CEO of Power Integrations, and Sandeep Nair, our Chief Financial Officer. During this call, we will refer to financial measures not calculated according to GAAP. Non-GAAP measures exclude stock-based compensation expenses, amortization of acquisition-related intangible assets, and the tax effects of these items. A reconciliation of non-GAAP measures to our GAAP results is included in today's press release. Our discussion today, including the Q&A session, will include forward-looking statements denoted by words like will, would, believe, should, expect, outlook, forecast, estimate, anticipate, and similar expressions that look toward future events or performance. Such statements are subject to risks and uncertainties that may cause actual results to differ materially 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 12, 2024. Now I'll turn the call over to Balu.

Balu Patel Thanks, Joe, and good afternoon. Our third quarter results were on target with revenues up 9 percent sequentially to $116 million non-GAAP gross margin of 55.1%, and non-GAAP earnings of $0.40 per share. Our revenue outlook for the fourth quarter is $105 million plus or minus $5 million. The decrease from Q3 reflects a soft demand environment called out this quarter by many of our analog peers and by some of our key customers in the appliance market. Appliances account for the bulk of our consumer category, which comprises nearly 40% of our sales. Further tempering the outlook in consumer is a buildup of appliance inventory at Chinese OEMs ahead of the China government's long-awaited consumer stimulus program, which thus far appears to be having a limited impact. Overall, while the recovery is proceeding in fits and starts, we are pleased to be returning to year-over-year growth with the midpoint of our Q4 forecast up 17% compared to last year. We also expect to remain at the high end of our gross margin model through at least next year, thanks to the dollar-yen exchange rate and a favorable end market mix. While all four of our end markets should grow in 2025, we expect the strongest growth from industrials which has the highest margin of our four categories. We expect growth in multiple sub-segments of industrial, led by high power with particular strength in high-voltage DC transmission projects. Long-distance DC transmission lines deliver renewable energy efficiently to the grid and are also a key part of efforts to modernize power infrastructure to enhance energy security and meet the increasing power needs of AI data centers. In Q3, we won high-voltage DC designs with expected annual revenue value of more than $5 million, including our second multimillion-dollar design win for Tenet, a major North European grid operator. We also expect to begin delivery of our scale-to-gate drivers in early 2025 for two other major projects, phase one of Saudi Arabia's planned nine gigawatt link between Yanbu and Neom, and an undersea link connecting Japanese islands of Honshu and Hokkaido. Metering is another growth driver in the industrial category, most notably in India, where our metering business is on track to double this year as the government continues with its planned installation of 250 million smart meters over the next few years. We have earned a sizable share of that market and expect dollar content to increase over time as customers migrate to our 900 and 1,250-volt GaN products to achieve higher efficiency and reliability. While still early in its revenue ramp, automotive will contribute to our industrial revenues in 2025, leading up to an inflection in 2026. Building on our early success in China, With drivetrain emergency power supplies, we are expanding both geographically and in the range of sockets we are winning as EVs evolve to a distributed architecture, requiring multiple converters to power subsystems from the main battery instead of a 12-volt battery. We achieved an important milestone in September, becoming an approved vendor at the world's largest Tier 1 automotive supplier. On the heels of our recent qualification at a major Japanese Tier 1. We've also had key developments at two American EV customers in recent weeks. Pre-production began last week on a micro-DC-DC converter socket using our silicon carbide inner switch, while another customer awarded us an emergency power supply socket using the 900-volt GaN inner switch with production scheduled to begin in 2026. In our consumer category, notwithstanding the soft demand environment, our position in appliances remains strong, and we are leveraging our market leadership to capture incremental dollar content driven by connectivity, LED lighting, brushless DC motors, and other features being added to appliances. We're also seeing strong design activity related to new EU standby rules set to take effect next year. We are capitalizing on these opportunities with bridge switch motor drive ICs and GaN products, including InnoMux2, which enables much higher efficiency in appliances, displays, and industrial applications through the use of zero-voltage switching and the elimination of multiple DC-to-DC converters at the output. Earlier this week, we introduced a new version of InnoMux2, featuring the industry's first 1,700-volt GaN switch, marking another milestone on the technology roadmap we presented at our Investors Day in 2022. We developed our proprietary GaN process and device designs with five key attributes in mind, cost, reliability, ease of use, voltage, and power. and we are executing to plan on each of them. We concluded early on that even with its superior performance, our GAN had to be cost competitive with silicon to achieve mass adoption. Over the past six years, we have executed on an aggressive cost reduction roadmap that has our GAN on the verge of parity with the most advanced silicon MOSFETs, but with far superior performance. Reservations about reliability and ease of use are the top concerns we hear from automotive and AI data center customers that have struggled to use discrete GAN devices. We designed PowerVGAN to be inherently more reliable than other GAN technologies, and we now have over a trillion hours of field operation to prove it. Our GaN is also easy to use because of the system-level approach that has defined our product strategy for the past 30 years. Not only are discrete power devices eventually commoditized, they are also harder to use, and this is especially true of GaN because of its extreme high switching speed. All the GaN switches are incorporated in system-level products with our innovative controllers and drivers that provide comprehensive protection and optimize the performance of the switch, eliminating the trickiest aspects of designing with GaN. Engineers familiar with our silicon-based products can use GaN versions of the same products seamlessly, taking advantage of GaN's superior performance with no added design effort. In fact, this strategy has worked so well with our InnoSwitch products that we are now refreshing some high volume legacy products with GaN to enhance performance and extend the power ranges of products that our customers love to use. We also have a pipeline of new system level GaN products targeting higher power applications we don't address today, including AI data center, EV onboard chargers, telecom infrastructure, and more. These products will help us reach our goal of expanding our SAM to an estimated $8 million over the next few years. In terms of voltage, while the leading foundry technology is limited to 750 volts, our unique process and device technologies offer a clear path to higher voltages. Following our announcement earlier this week, we now have GaN devices at 750 volts, 900 volts, 1250, and 1700 volts, with more to come. Many industrial customers prefer higher voltage ratings for three-phase applications and as an insurance against surges and spikes, especially in areas with unstable AC grid voltages. In the EV market, a 1700 volt rating is essential for flyback power supplies in 800 and 1000 volt battery systems, which today can only be addressed by silicon carbide. The final track of our GaN roadmap is power, and we are approaching that from multiple directions. The output capability of our current technology continues to increase and is now in tens of kilowatts. We are also developing an all-new GaN technology capable of delivering hundreds of kilowatts, which we believe would make GaN an alternative to silicon carbide in drivetrain inverters at much lower cost and with superior performance. While the necessary breakthroughs are still a few years away, we are excited about our progress and our acquisition of Odyssey last quarter will help accelerate our development. While the full deceptive potential of GAN will unfold over many years to come, we expect wider GAN adoption to be a significant growth driver in 2025, when GAN products should account for more than 10% of sales. Recent wins include major next-generation game console, now ramping with GaN InnoSwitch ICs, a 65-watt notebook adapter for VIO, and a 140-watt multiport adapter for ASUS containing GaN InnoSwitch and HyperPFS power factor ICs. We are also seeing existing designs migrate to GaN, such as our 5G fixed wireless design in India. That program ramped this year with silicon-based products, but should generate incremental growth next year as customers add GAN to the mix, mainly to reduce the size of the power supply. To conclude, though our outlook reflects industry-wide demand softness, we expect strong year-over-year growth in Q4, while 2024 revenues will be down from the prior year due to the losses we suffered in the China cell phone market early this year. Combined sales into industrial consumer and computer categories will be up mid to high teens for 2024. We expect that momentum to carry into 2025, sustained by strong growth in GAN products and high power, a broad range of new design wins, and a return to growth in communications driven by non-cell phone applications. And now I'll turn it over to Sandeep for a review of the financials.

Thanks, Balu, and good afternoon. As usual, I will focus my remarks on the non-GAAP results which are reconciled to GAAP in our press release. Third quarter revenues were $116 million, just over the midpoint of our guidance, while our non-GAAP earnings were 40 cents per diluted share, above the level implied in our guidance as we came in slightly better on gross margin and operating expenses. While we expect sequential low revenues in Q4, we are returning to year-over-year growth, as Balu noted, and believe we are poised for strong growth in 2025. Reflecting confidence in our outlook, as well as our healthy balance sheet, our board has declared a 5% dividend increase and authorized $50 million for share repurchases, commencing in the coming days, subject to price volume thresholds prescribed by the board. Returning to the quarterly results, revenues were up 9% compared to the prior quarter. Industrial revenues grew mid-teen sequentially driven by high power, home and building automation, and metering applications. Broad-based industrial also showed sequential improvement as distribution inventories declined further in Q3 and are now at historically normal levels. The communication category was up 20% sequentially, driven mainly by the ramp of our 5G fixed wireless business in India, as well as seasonal growth in smartphone chargers. Revenue from the computer category was up 10% sequentially on strength in OEM and aftermarket notebook chargers. Consumer was flat sequentially as seasonal softness in air conditioning was offset by strength in major appliance and gaming. Distribution sell-through for consumer was less than sell-in, reflecting the trends Balu mentioned and resulting in higher channel inventory for consumer compared to the prior quarter. Overall, channel inventory increased to 8.6 weeks, up from 7.8 weeks at the end of June, with the increase driven entirely by consumer. Channel inventory for each of the other categories decreased from the prior quarter. Revenue mix for the quarter was 38% consumer 36% industrial, 14% computer, and 12% communications. Non-GAAP gross margin for the third quarter was 55.1%, up one percentage point from the prior quarter, driven mainly by higher back-end manufacturing volumes. Non-GAAP operating expenses were $43.7 million, down $500,000, sequentially, despite the addition of about a half a million dollars from Odyssey. Non-GAAP earnings for the third quarter were 40 cents per diluted share. Diluted share count for the quarter was 57 million, unchanged from the prior quarter. Cash flow from operations was 33 million dollars, while capex was 6 million dollars. Other key uses of cash were 10 million dollars for the Odyssey acquisition, and 11 million for dividends. Inventory days fell to 291 at quarter end, down 21 days from the prior quarter. Turning to the Q4 outlook, we expect revenues to be $105 million, plus or minus 5 million. Non-GAAP gross margins should be between 55 and 55.5% as compared to 55.1% in Q3. The slight sequential increase at the midpoint reflects an increased percentage of revenue from the industrial category, which is our highest margin in market. Non-GAAP operating expenses should be between $44.5 and $45 million, up modestly from the third quarter, driven mainly by headcount increases. For the full year, non-GAAP OPEX will be up just 4% from the prior year, with one percentage point of the increase driven by the Odyssey purchase. Finally, I expect our effective tax rate to remain in the low single digits in Q4, and then to increase into the mid single digits for 2025. And now, operator, let's begin the Q&A.

Thank you. At this time, if you would like to ask a question, please press the star and one on your telephone keypad. You may remove yourself from the queue at any time by pressing star two. Once again, that is star and one if you would like to ask a question. We'll take our first question from David Williams with Benchmark. Your line is now open.

Hey, good afternoon, everyone. Thanks for taking my question.

Hi, David.

And forgive me, I missed a little bit of the call here. I had some issues, but lots of things, lots of moving pieces here. But one thing I did want to ask was just around the 1,700-volt GAN announcement that you had recently, can you talk maybe just a little bit about that, how you see demand developing, and maybe just where that fits in? Is it industrial automotive? And given what would seem like large markets, how do you think that ramps?

Yeah, the 1700 volt is required to operate from 800 volt and 1000 volt battery systems in automotive applications. The cars are moving from 400 to 800 and the larger vehicles like trucks and heavy vehicles are moving to 1000 volts. And so this is a perfect fit for that market. And also in three-phase industrial applications, They prefer to have a pretty high voltage rating because these meters sit directly on very low impedance lines next to the transformer, for example. As a result, they can get significant transients from lightning and so on. So they prefer the 1700 volts. Currently, this market is served by silicon carbide. We believe GaN is a much better solution, both in terms of efficiency and cost.

Great. Thanks so much. I really appreciate the color. They're very helpful. And then just lastly, I wanted to ask maybe broader, but just thinking of the last 90 election, are there areas of your business you think will benefit or maybe be impacted by the new administration? And specifically thinking about renewables and solar applications, but also just some of the energy efficiency regulations and any other color on that would be helpful. Thank you.

You know, that's a good question. I haven't thought through all of that, but I would say that the train has left the station when it comes to electric cars and renewables and so on. The renewables are now less expensive than coal plants. So most countries are installing additional capacity in renewables for that exact reason. Of course, there's also the environmental benefit, which I think, you know, the entire world now recognizes is critical, given all the climate disasters we have had over the last few years. So I am not at all concerned that it will in any way slow down either electric cars or renewables or high-voltage DC transmission systems. And let me talk a little bit more about high-voltage DC transmissions today. The high-voltage DC transmission is much more efficient than AC transmission, and it is really quite integral to any new power station, whether it's renewable or nuclear. If you have a power station and you have to transmit that to a load, let's say AI data center, for example, you have to have a very efficient way to do that, and the DC transmission does exactly that. And then if you have offshore wind forms, the only way to get the power onshore underground is through DC transmission. AC would be way too lossy in terms of the cables. So I don't think the trend has any impact on our business. In fact, I would say in terms of generating more oil, it will have a positive impact on our high-powered business because we sell into applications like fracking. They need large motors to pump liquids into the ground, and that's a huge market for us in terms of our high-power board-level business.

Thank you. And we'll take our next question from Torres Samverg. With Stiefel Nicholas, your line is now open.

Thank you. This is Jeremy calling for Tori. Yeah, let me add my congrats on that 1,700-volt product. We've been waiting for that, and it's very nice to see. Just to follow up on that point, I guess, can you talk about maybe how much of a lead you have in high-power GAN? And you talked about the price difference almost price parity with silicon. Can you talk about maybe how it compares to silicon carbide? How much of a, I guess, advantage you have in that space? Thank you.

Thanks, Jeremy. Excellent question. Let me start with the 750 volt GAN, which primarily competes with MOSFETs. And there, we think we're getting very close to being parity with even the high-end MOSFETs within the next year or so. And at 750 volts, we are also very cost competitive with silicon or silicon carbide. Of course, if you're cost competitive with silicon, you are dramatically cost competitive with silicon carbide. And these applications include AI data centers. It includes onboard charges for electric vehicles. It also includes telecom infrastructure. All of these are in the, I would say, 1 to 20 kilowatt range, and we already have products in development. These products will be ready sometime in 2026, and that opens up a huge market for us. So with our existing 750-volt technology, we can be very competitive with silicon and, of course, much more competitive than silicon carbide, all the way up to tens of kilowatts. Now, when you go to higher voltages, like 1250 volts and 1700 volts, the real competition is silicon carbide. And we are very competitive with silicon carbide up to about 100 plus watts. Our products currently we've introduced are in the just under 100 watts, and we are very competitive there. Now, if you want to go to much higher power levels, like hundreds of kilowatts, That's where we need new technology. We are working on that for some time now. And with the acquisition of Odyssey Semiconductor, we will be able to speed up the technology because it comes with the fab. It also comes with people with the knowledge of the device. And we have made a lot of progress on that. We have a few more breakthroughs to get there. And we are now beginning to feel that we can get there in the next three to five years. And when we get there, we'll have a very competitive offering compared to silicon carbide. It will not only be lower cost than silicon carbide at those power levels, but it will also be a much higher performance than silicon carbide. And as you all know, silicon carbide requires a lot of capital investment, whereas GAN does not. And GAN inherently is going to be cost-effective long-term because of – you know, the temperatures at which they are manufactured, which is similar to silicon, and therefore it doesn't require anywhere near the amount of energy consumption that silicon carbide requires. Silicon carbide, as many of you know, requires manufacturing at 2,000 degrees, which is about half the temperature of the sun, by the way, and it takes a lot of money, a lot of energy to and a lot of capital investment to do that. And that is not the case for GaN. So we really think GaN would be the most attractive alternative to silicon carbide all the way up to several hundred kilowatts once we have this new technology up and running.

Thank you for that. That's very, very helpful. And maybe as a follow-up, you mentioned migrating some of your legacy products to GaN. Can you give us a sense of how much of your portfolio is consisting of GAN now and how much maybe you intend to convert? I know you also targeted 10% of revenue for GAN, I think it was next year. You know, as you look, it might be kind of further out, but three to five years out, how much revenue do you expect to drive from GAN? Thank you.

You know, GAN is going to replace silicon as a high voltage switch. in pretty much all of our products. Everything we are designing now is using GaN, except at very low power levels, which is really not our focus at this point. We are focusing on higher and higher power levels. For example, the AI data center product or the onboard charger product, those are all GaN products. That's the only way to get the performance you need. It's the only way to get the power levels you need. You know, the silicon just won't cut it at those power levels, you know, 10, 15 kilowatts. Even though silicon is predominantly used today, that's only because GaN technology is not high enough reliability, at least from a customer's point of view, not easy to use. That's where our technology will shine. We have much higher reliability than anybody else's GaN technology. And we also have... easier to use product because we don't sell them as discreet. We sell them as system level products. So the customers don't have to deal with all the idiosyncrasies of GAN, not only in terms of protecting the GAN, but also getting the most performance out of the GAN. So I think that the future of our company is going to be GAN. And if you were to ask me where we would be in terms of revenue, I wouldn't be surprised if we cross 100 million by 2028 in GaN products, but that's only because in some areas it takes longer for the GaN to be, the new products to be adopted. But we are seeing GaN growth in all four of our markets. You know, we actually grew GaN revenue in all markets except cell phones. where, as you know, people are going out of the box, and therefore there was a little bit of a drag in that market. But if you take out cell phones, we grew very nicely this year, and, of course, we're going to grow very nicely next year and onwards.

Yeah, Jeremy, in the next few years, Ghana is going to be really growing quite a bit, and I can see it being about 20% of our revenue in the next two to three years.

Great. Thank you very much. Appreciate that.

Thank you. And once again, that is star and one, if you would like to ask a question, we'll take our next question from Christopher Roland with Susquehanna. Your line is open.

Hey guys, thanks for the question. Um, I guess, first of all, uh, if you guys could help us kind of force rank the segments for your outlook next quarter, uh, I guess maybe versus the, uh, I think it was down nine sequential. What, what, what would be better? What would be worse? And I know it's probably a bit early for March, but we track that seasonally as flat. I wonder if you had an early expectation there for growth. Thank you.

So I think for the next quarter, as we indicated, the industrial segment, from a dollar standpoint, will be kind of flattish because the revenue is lower, will be higher as a percentage of revenue. And the other three segments will be down. So that's the directional answer we can give you at this point of time for the Q4. In Q1, it's a little early. But I think in historical terms, in Q1, typically you see the high power business, which is an industrial segment, tends to be a little lower. And cell phones are a little lower. And AC, which is an air conditioning, starts ramping up. So that's the directional idea that I can give you for Q1. But it's a little too early.

Yep. Uh, thank you, Sandy. Um, and then, uh, I, uh, just understand, I would like to understand, uh, the inventory across these end markets, um, you know, where, where it's still worse, um, overall, uh, you know, selling versus sell through, uh, how do you think that, um, kind of, uh, uh, finished, uh, across distribution for the quarter. Uh, and then lastly, um, you know, tying all into this, the appliance market, uh, I think what you were saying is, uh, finished white goods inventory was built ahead of this China stimulus, if I understand that correctly. And, uh, Yeah, just putting all this together from inventory and then sell-in, sell-through would be great. Thank you.

Yeah, as we said on our call, that the major sell-through, you know, the sell-in was higher than sell-through, and the reason was mainly in the area of consumer. If you look at the prior quarter, our consumer inventory was way below our eight weeks, but This quarter, because of the Chinese inventory build, our sell-in was greater than sell-through. It was entirely in the consumer segment. So as a result, our consumer end inventory in the channel is slightly above the eight weeks, which is normal, while as all the other area, it has come back to the normal level.

Okay, understood. And sorry, did that include... Did I get the dynamic correct on the build ahead of... white box and appliance inventory. Was that correct?

Yeah, that's correct. In the first half, we had very strong consumer business, which was kind of a false positive. Our consumer customers, which are appliance customers, they thought the second half would be very strong, especially in China, because of the incentives they were expecting the government to provide for purchase of appliances. Now, first of all, it came very late, the incentives, and also the incentives are not uniform. They are different in different parts of China, and it's not clear to us so far that there's any significant benefit or impact of those incentives on the demand. So the appliance customers were expecting a strong demand in the second half. They didn't see that. The demand continues to be weak. So they ended up with a bunch of inventory. And you have to remember, they always build the power supply first before they build the end product because they don't want the end product to wait for the power supply. And so we had a very strong consumer business in the first half. And unfortunately, in the second half, it has slowed down significantly because of the finished goods inventory of the end product at our customers. Our distribution inventory is slightly above. It's not so bad. But it was significantly below, as Sandeep pointed out, last quarter. So it was a surprise to us that the sell-through in consumer was slow. We expected a strong consumer in the second half based on what we saw in the first half.

Excellent. Thank you very much, guys.

You're welcome.

Thank you. We have no further questions in the queue. I will turn the program back over to Joe Schiffler.

All right. Thanks, everyone, for listening. There will be a replay of this call available on our investor website, which is investors.power.com. Thanks again, and good afternoon.

This does conclude today's program. Thank you for your participation. You may disconnect at any time. Have a wonderful day.