Good morning, everybody, and thank you for joining us today in the UBS conference. Today we have with us Power Integrations, and we have Radu, VP of Technology, and we have Sandeep, the CFO of the company. Thank you very much, and welcome here.
Thank you.
Thank you. Good. So, you know, Power Integrations is a, it's an analog company, but it will be good if you guys can maybe introduce which products do you have and which markets do you serve? How this company fits in the analog landscape.
So, we're part of the semiconductor sector, and within that, part of the analog semiconductor group. And within that, we're a specialist in power conversion semiconductors. We make the chips that power your laptops, that charge your cell phones, make the washing machine run, make also wind turbines operate and locomotives in railway. What is important to understand is that our products support all three components of decarbonization. The first being renewable energies. Second, and most important in my opinion, the elimination of waste through better efficiency, and also transmission of electricity from the point of generation to the point of use in high voltage DC transmission lines, for example. How we do that is by having developed over 30 years, significant expertise in high voltage semiconductors.
When our peers talk about high voltage in the analog sector, they talk about 20 volts, 40 volts. For us, high voltage starts at 600 volts and goes to 1 kV, 1.7 kV, and so on. So that is our expertise. The reason we can contribute significantly to decarbonization is because, as the International Energy Association has published, in order to reach the Paris Agreement target by 2050, two-thirds of that decarbonization needs are due to energy waste, which can be mitigated by better efficiency and to the use of renewables, and our products support both of them.
Okay. Thank you very much. A pretty comprehensive overview. Thank you. So, just, you know, coming a bit more to what's going on now in the markets. You reported Q3 results a few weeks ago and guided a significant sequential decline for Q4. Can you recap for us and talk about how do you see this Q4 quarter, and also, how do you see a bit more beyond what's happening this time?
Yeah. So I think what happened was a little unusual. And we've been talking about, you know, the turnaround coming by the fourth quarter, and we were getting some good orders earlier in the year. But what happened in the third quarter is we had a large cancellation last minute from an OEM, which also had an impact in the fourth quarter. Now, it was not just that, but things like the consumer and industrial segment has also weakened quite a bit. Now, this gyration is not unique to us because, you know, we see these downturns and upturns earlier, and also because we make power supplies for OEM, we are a little further away from the end, and there's an extra layer of inventory. And that's why that caused us a little bit.
We are also exposed to appliances, you know, which, as you know, with the whole housing market in China, the higher interest rates, all have had a little bit of exposure. If you look in the semi space, the companies that have really held up pretty well is in automotive, even though they are now seeing a little bit of an adjustment. We are just in the initial stages of getting into automotive because the main play for us is in the EVs, not in the traditional car, because as Radu mentioned, we are in high voltage application.
Good. Good. Yeah, definitely consumer, it's been a softer demand on the market for analog, so, makes sense. So you also thought that Q1 for 2023 will be the bottom, but we are seeing another lag now. What did you see over the course of the year on top of what you just mentioned, that changed your outlook? And maybe if you can help us also trying to understand how 2024 will shape up in the, more in the maybe long-term perspective.
We actually did think the Q1 was, and in fact, we thought we'll have sequential increase all throughout the year, and we did in the first two quarters. And this was further substantiated that the orders that we were starting to see in the March, April, May were very strong. But a few quarters, a few months later, it was like things just fell off a cliff. And again, we saw the slowness in demand in industrial as well as in consumer. That has impacted us quite a bit. As far as 2024, it's, you know, it's a little hard because it's not clear. We have announced a fairly low quarter in Q4, and I think it'll take a quarter or so to come back.
One of the things about POWI is we see this stuff happening a little earlier than others, and the reason is the power supplies are made earlier than the final product. Nobody wants to wait. As I said, you know, when the things were great, I said, "That's not sustainable," when we were having, during the COVID time, very steep increases. I also say this is not sustainable, because this level of normal down cycles are like four quarters, now we are looking like, like six or seven quarters. But I think we're gonna come out of it, and we're gonna come out of it strong. If you look at even in the area like our appliance area, we are this year going to be below 2019 levels.
So we have looked at the different gyrations and say, if the COVID didn't happen, and if this thing happened, what's our normal run rate should be? And it should be around $150 million a quarter. Now, I don't believe we will get that in 2024, but we'll definitely get it in 2025. But I think during 2024, we should gradually go back to about $135 million. I really think the channel inventory is going to get adjusted, the demand has been pushed down, and then hopefully, if the interest rates normalize a bit, that'll help with the housing market. And the best guess we have is on the second half, that we should see a better second half of 2024.
Clear. You mentioned about channel inventory. How do you see that evolving? Are you under shipping channel? Do you have a sense for what's going on?
So we were, even during the whole COVID period, when there was, you know, people trying to get as much inventory to, for shortages as you know, we tried to restrict how much we could ship, but people were trying to get as much as they can. And as you know, what people did during this time is, if their normal inventory levels were 60, they started changing it to 90 and 100 days to catch the. We tried to arbitrate, but it was very hard. We tried not to overship. Right now, we are sitting at about 11.5 weeks of inventory, which is above our normal of 8 weeks. We believe this should get normalized in the next 2 quarters.
We prefer to keep the inventory at our level because we keep it in wafer form, and we are able to ship to the customers. We've had such good relationship with our foundry partners that in 2021, when we had a 44% uptake, we were able to supply at a very short time.
Good. Good, good. Thank you. Now shifting gears to a bit of a more long-term perspective of the company. During your Analyst Day, you announced plans to double your serviceable market from $4 billion to $8 billion over the next, next several years. Obviously, we have this bump on the road, but can you maybe explain to us which are those markets? Which are those products, more or less, what's the plan to increase the, the SAM, gradually?
So, the growth of the company over the next period that you mentioned will focus primarily on new high growth in higher power areas. Automotive is definitely a focus for us, that will add significant SAM, data centers, and all these markets will be supported by products that are already in design. Other areas are high power, and by the way, our high power business has never gone down during this downturn. It's the one component of our business that continued to grow steadily, and that's the business that supports the renewable energy markets, the windmill applications, the solar utility grid, the solar systems, and so on. So high power continues to be another strong growth area that will add significant SAM.
We're also expanding in additional products that use new materials and new technologies. We do not separately quote numbers and business growth in terms of technologies, but gallium nitride, as an example, will continue to be a very significant component of the growth that will add significant SAM, given the higher power levels it can reach.
The one thing that I'd like to add is, SAM expansion is an integral part of POWI's strategy to grow its business to a double- digit. We were $1.5 billion, went to $4 billion, and as Radu indicated, we're gonna take this to $8 billion. Coming from automotive, another big area is motor control, the high power area, and GaN. And that's why you'll see the stream of announcements we've been doing in GaN, and the announcements that have been made in motor control, and we have a lot more products coming out there. But the key part for our company, the SAM expansion, is an integral part of our growth story.
Clear. You both guys spoke about gallium nitride, and that's a very hot topic now in semi. So can you maybe mention, because you have been doing gallium nitride for quite a long time, that there's some recent announcements. What is the portfolio of gallium nitride now, the roadmap? Can you elaborate on that, please?
Thank you. First, I'd like to explain the fact that all our technologies are proprietary. We are a fabless company, but we do not use foundry processes. And the reason is, in high voltage technologies, you need to control end-to-end your process technology. Even more importantly, in power devices, the design of the device and the process that's used to make that device are 100% intertwined, and you can simply not afford to use a black box technology that a foundry provides, unlike an integrated circuit, which can be designed without knowing what the process is.
All our technologies are proprietary, where we don't own any fabs, so from that standpoint, we're fabless, but we have full end-to-end control of our proprietary technologies, and we have them run in partner fabs, mostly in Japan, that run these processes for us. Silicon technology is still the vast majority of what we make today, both in terms of integrated circuits, controller, very advanced con-- power management, and power conversion controllers, but also power devices. Our motor control product line uses our vertical, as proprietary silicon MOSFET, which is extremely well-suited for motor control applications. In the same vein, we continued with the development of the next generation technology, which is gallium nitride.
We also looked at silicon carbide a few years ago, but we decided against silicon carbide at the time, and we can talk more about how they compare. And gallium nitride that we have developed is superior in many ways to other gallium nitride technologies that are available today. And one specific area where it shines is the ability to go to higher and higher voltages. And we recently announced the 1,250-volt GaN technology and product. The product line that uses this technology has just started. The first product will be released soon, and more higher power devices at this voltage level are planned for the next couple of years, so you'll hear more about that as we go forward.
This is just the, the next step in our evolution of our GaN technology towards higher voltages. At 750 volts, 900 volts, and 1,250 volts, we are already competing favorably with silicon carbide. As we move forward to even higher voltages, that overlap will increase, and GaN will always win. Whenever it can overlap with silicon carbide and serve the same market, the same segment, GaN will win because of better performance, and most importantly, much lower cost. The holy grail will be when GaN will be able to achieve both the very high voltages and the high powers that today are the domain of silicon carbide, for example, in electric vehicle drivetrains.
That is a little bit further away because it requires breakthroughs in the GaN technology, but we think it's feasible and we will be looking at that.
Well, yeah, that's... It's going to be difficult technically, because, you know, gallium nitride, it's a horizontal device, right? It requires more die, more material for transporting the same amount of current. Therefore, you have less wafer. There's some cost associated with that. Can you maybe help us understand how gallium nitride can overtake silicon carbide? Give a bit of background there.
Well, like I said, in the applications where the lateral GaN can compete favorably with silicon carbide, that's already a done deal, at 1,250 volts, and all the voltages below, as well at 1,700 volts, at a certain power level, probably on the order of 10 kW or less. Current GaN technology, which is lateral, as you mentioned, can compete favorably with silicon carbide. Obviously, the answer to your question lies within the question, that you have to change that in order to be able to really compete with silicon carbide at very high power levels.
That's a challenge. That's a-
It's a challenge, and it will require some breakthroughs, yes.
Very interesting. So, how do you see the competitive landscape for gallium nitride currently, with so many companies trying to play in the field?
Well, indeed, the field has gotten more crowded. Five years ago, we were the first company that commercially shipped GaN-based products. We've been developing the technology for about 12 years. The difference between us and others that sell GaN today is first and foremost, we have full end-to-end control of our technology. We have developed it from the basic substrate through the processing and the special steps that are needed for gallium nitride power FETs, and we have qualified it and qualified at the system level. Because we're Power Integrations, we're not selling discrete devices. We're selling system- in- a- package. So a lot of the expertise in the company is in system level, at the system level.
That's very important for a new generation of materials technologies, such as GaN, because we do not have to let the customer find the problems, the reliability problems that need to be solved or system-level, short-circuit, withstand capability, and so on. We do that in the company. We have a very, very advanced, system-level applications group, both, for new product qualification as well as, field, helping our customers with the system-level, design of our products. And they are the customer zero for us. So we never release a product, and certainly not a new technology platform, without, without running it at the system level for a long period of time, through all the tests that go beyond JEDEC, and so on, inside the company.
So that sets us apart from a vendor that only does GaN. The other important difference, again, due to the fact that we have our present technology is proprietary, we do not use a foundry, and we have our own GaN. And as I mentioned earlier, it's very critical for power devices to have full control of both the process technology and the design, a device design, in order to fully harvest the benefits of the new material.
Clear. Thank you. So we asked the question about the cycle. Now, it's the second question that comes up in every session here. It's about China. I think it's around 60% of the revenues last quarter that came from China, for you guys. But China is investing large amounts of money in new capacity, in semiconductors, both in foundry, fabless, IDM. And gallium nitride is one of the fields, and power semiconductor is one of the fields they're investing. For instance, we have the example of Innoscience, a success that they came up with the 8-inch gallium nitride wafer, whenever I was betting against that, and it happened, right? So how do you see that competition coming from there? Do you think they may take some of this market share that you have in China,
Well, the, there's definitely very strong competition coming from China. Innoscience, to their credit, have done a fantastic job in investing in enormous capacities. And they have a technology that works, and they are definitely a competitor, especially in China, as being a Chinese company. The difference is that a lot of the products that are made in China are made for China, domestic consumption, as well as for export. Typically, what we see is that our business in China for export. I should say, our customers in China that build for export prefer to use Power Integrations because of the brand name, the history of reliability and quality, service, and so on.
The other point is that we don't really have GaN competitors. We are not selling GaN, we are selling system-in-a-package products that are integrated, as our company name indicates. And yes, some certain components of our product can be made by other companies, GaN by Innoscience, and so on. Controllers can be made by On-Bright or BYD or other Chinese companies. But when it comes to high reliability, highest performance, usually people, including in China, prefer Power Integrations.
I mean, the component count from our product versus theirs, there's a big difference, and that's why the reliability is absolutely at a different level. And that's why even there, for exports, as Radu indicated, they prefer to use our product. You know, we've been in this business a long time now. Innoscience is well-funded, but many Chinese company, what happens is come and go. They get government funding, they do cost plus, and that model is not a sustainable model. And that's something we have seen a lot of companies come over the last decade and try to come after what I would call the low end of it. Now we compete, but we don't compete for the last dollar.
What we try to do is make sure we are playing so that we can keep them at bay, and but yet we are able to service what I call the high-value market.
Clear. Okay, thank you. Just as a reminder, you guys can pose questions through the app. We still have five minutes, so... But, there's also, you mentioned automotive a couple of times as one of the new big markets that you are going after. What are you doing as a company in the auto space, and what's the roadmap to service those-
I mean, it's an early stage, and I'm going to let Radu tell you the different areas. But as we had talked about analytics, we were not in the traditional car, but with high voltage, we have a tremendous application. And in fact, we have announced, you know, product in the emergency power supply where we are winning a lot, and the recent GaN announcement will allow us to get, in the future, in onboard charging and other areas. Radu, would you like to add to that?
Yeah. So, when people talk about the electric cars, the primary thought goes to the powertrain, the motor drive. There's a multitude of other power conversion applications inside an electric car today. There's an emergency power supply, there's onboard chargers, there's the need to eliminate the lead-acid battery and power everything in, including the cockpit systems and appliances from the high-voltage battery. So those present a plurality of opportunities for us because we're specialists in high-voltage power conversion. So we had our first products in automotive address those lower power applications within electric cars, emergency power supplies. We have cars today running on the roads that use our product in emergency power supplies. The onboard charger products are in design.
In order to address the very high power motor drive, application, that will require, different, technology, which, will require some breakthroughs, as we discussed. Today, that is served primarily by silicon, with silicon carbide making some inroads because of, you know, better efficiency. But the Holy Grail will be, something that will meet the silicon carbide performance at a much lower cost, because we remain convinced that silicon carbide is, it has some fundamental limitations in terms of ability to reduce cost. You mentioned 8-inch for, you know, Innoscience. And there's a lot of companies that, invested in 8-inch capacity for silicon carbide. What is, however, important to know that there's a fundamental difference between, wide bandgap, material technologies and silicon technologies.
In silicon, in CMOS, bigger wafers is always better because all the process steps are made on the same wafer in batches that can take advantage of the larger number of wafer. In wide bandgap materials, the epitaxial step and the substrate itself are a huge part of the cost, and these, unfortunately, do not scale with the size of the wafer. The cost per square inch of GaN epi or silicon carbide epi is the same whether you populate a reactor platter with 5 8-inch wafers or 14 6-inch wafers. Actually, it's cheaper if you use smaller wafers, 'cause you can cover a circle with more efficiently with smaller circles. Yes, the part of the process that follows in the fab can benefit from higher, bigger wafers because it's CMOS-like.
But overall, the cost, the cost reduction that bigger wafers bring about is not comparable to what has been traditionally the case in CMOS.
That brings the question, because we have so many big players investing large amounts of money. I think probably most of the money invested in silicon, in power semiconductors now is going to wide bandgap, and probably the largest part of that going to silicon carbide, new fabs, people vertically integrating, like the case of onsemi buying GTAT. So that contrasts a bit with the perspective that, you know, they have like a floor for the cost, that they will not be able to break, that gallium nitride can break through and just become more affordable.
Yes, that is the, like I said, the Holy Grail. So, just like the Holy Grail, people are still searching for it, so, we'll see what, what the future brings.
Okay. So another question. Europe recently passed some rules requiring electronic products sold without chargers. Any insight on that? This is, by the way, the last question.
Yeah, you know, it does. The negative side is it does reduce the number of chargers, but what happens is now the consumer has a charger. It's gone out of the box in a couple of cases. The consumer has a choice. They want higher power level, they want multi-port. So as a result of which, the value proposition goes up. So even though the number of chargers come down, the consumer who wants to use it wants to get a better charger than a cheaper charger. I've always said this: Do you ever share your charger with your spouse? The answer is no. So everybody will want their own charger, and they'll want a value, so it becomes like an appliance, not like a throwaway piece. So I think for us, it's going to be a positive.
Overall in our business, if you look forward, and even in our Analyst Day, we talked about, this communication revenue as a percentage of revenue will come down, but we will increase in the other areas, like Radu talked about, automotive, motor control, data center. But I think the value of what we sell in communication will be at a higher level.
Clear. Thank you very much. We know we are running out of time, but it was a great conversation, and yeah.
Thank you for taking the time.
Thank you very much.
Thank you.
Thank you.