Good morning. I'm Joe Shiffler, Director of IR here at Power Integrations, and it's my pleasure to welcome everybody here in person and online to our Analyst Day 2022. Here's a look at the agenda we have for today. By the way, this is available on the Investor Day tab, the Analyst Day tab on our investor website, which is investors.power.com, along with a PDF version of all of today's presentations. Balu's gonna get us started here in just a moment, followed by Mike and Radu. Then we're gonna have a brief Q&A session with our first three presenters, followed by a short break. After the break, three additional presentations, followed by another Q&A session with all of today's presenters as well as two other members of our management team that are here.
Then we'll break for lunch, and we have some product demonstrations next door. So before we get started, let me just mention that during today's event, we will make forward-looking statements based on current information that is subject to rapid and even abrupt change. Due to risks and uncertainties associated with the company's business, actual results could differ materially from those projected or implied. These risks and uncertainties are explained in the slide you see here and under the caption Risk Factors in our most recent 10-K. We will be presenting non-GAAP financial information today. Reconciliations to GAAP results are available in the financial perspective section of the PDF version of our slides and in our historical financial tables, both of which are available online at investors.power.com. With that, I'm gonna turn it over to Balu.
Well, good morning. Thanks for joining us here and also online. If you're new to our company, this will be a fantastic introduction to the company. If you've been following us for a while, I am confident you will also learn a lot of new things as we talk about our approach to our business and our multidimensional innovation that has made us successful over the last 25 years as a public company. Also the attractive opportunities that'll take us for the next 25 years. Probably most important, you'll get to meet my team. We'll have Mike Matthews and Radu Barsan who'll talk about products and technology, and that'll be followed by Doug Bailey, who will talk about SAM expansion and the key market opportunities. Sandeep will provide the financial perspective. We also have two additional members here, Sunny Gupta and Yang Chiah Yee.
They'll be here to answer any questions in the Q&A session. They are the latest additions to our team. All of us will be here for lunch, and we'll also be here at the end when we conclude the session. Our overriding goal today is to explain how we managed to outperform the analog industry by almost 2x over the last 20 years, and how we're gonna sustain that going forward in the future. Let's take a look at our track record. One thing to note here is the gap in outperformance has widened over time. In fact, the biggest years of outperformance is in the recent years. That's direct result of increased pace of innovation and also SAM expansion over the past two decades. Past decade, I should say.
Now, we have done this while competing against vertically integrated large companies with huge R&D budgets and sales forces, and also against startups, well-funded startups, all claiming to have something new or better. I believe there are two reasons for it. One is focus. We do nothing else but high-voltage power conversion. For the large diversified analog companies, it's at most a niche market. That's all we do, and that's why we are so good at it. The second one is that, we look at everything at a system level. Our approach is very different. We are the only company who does that in the high voltage area. Now that brings me to the first of the four key takeaways that we would like to communicate to you. High-voltage power conversion is a highly specialized corner of semiconductors.
It's hard to do unless you focus on it. PI is the only pure play in high-voltage semiconductors. Now I have to clarify that because many people know about power management, which is a very large category with a lot of companies competing in it, but that's very different from high-voltage power conversion. In fact, most of the companies who are in power management have very little to do with high-voltage power conversion. Again, it may be confusing because they also talk about high voltage. When they talk about high voltage, they are talking about 12 V and 48 V. You know, these are DC to DC converters. I'm sure a lot of you know about that. When we say high voltage, we really mean high voltage. Almost all of our products are rated at 700 V or higher.
You may wonder, why do you need 700 V? It turns out that if you plug into a wall outlet or a grid, you have to have a minimum of 700 V. In fact, if you go into some grid applications, you need a lot more than that. Our gate drivers, for example, go to several thousand volts. It is so different from everything else semiconductor companies do, that very few understand the challenges that come with this high voltage space, especially if you're trying to do something at the system level. Let me give you an example. If you're plugged into the wall socket or connected to the grid, you have all sorts of real-world issues you have to deal with. You have to deal with lightning strikes. We have to deal with line transients.
I mean, if you're in a country like India, the voltage could go all the way up to 300 V, 350 V. We have specialized ways of handling that within our chip. That's why we are so successful in areas where you need that kind of protection. The other thing about high voltage is you have to meet safety with your product in every country you ship into, and the safety rules are a little bit different for each country. Similarly, you also have to meet EMI or electromagnetic interference regulations in each country. In the U.S. it's called FCC. There are other standards outside of the U.S. that you have to meet, and each customer has their own requirements, things like ESD, for example. High voltage is hard, and making it reliable long-term is much harder.
I'll talk more about that. How do our competitors deal with this? Well, they deal with it by not dealing with it, essentially. They separate the high voltage component from the other circuitry and accept all of the compromises that come with that. Even more important, they push all of the problems, the system problems, to the customer. The customer has to deal with all the issues I talked about. Now we take a different approach. We integrate for a better outcome for the customer. It makes the design much simpler, easy to manufacture the product. It makes the designs reliable because you've far fewer components. The form factor gets smaller. Most importantly, you get a much higher efficiency in your product. This is a comprehensive approach to integration. The reason we are able to do that is that we control everything in our company.
We have our own high voltage process. We have our own high voltage device technology. We have extensive knowledge of the systems. In fact, all of our products, and I'll say most of our products, use our own control schemes. They're very, very unique. Nobody else uses them. We have a lot of applications know-how, and we have developed this over the last 30-some years. For someone else to start from scratch, it'll take a long time. The third item is package technology. For those in power industry, know packaging is very critical, but it's even more critical in high voltage. It's one thing to make an IC meet a high voltage spec. It's totally different to provide that high voltage specification in terms of long-term reliability, and packaging has a lot to do with it.
Packaging impacts long-term reliability of the high voltage product, and that's why we do it. We do our own packaging. Lastly, we make the designs even easier for customers by providing a design software, which is PI Expert design software. They don't really have to understand our product. They don't have to understand the entire system. They can put in the specs and it'll give you a complete schematic. It'll give you a design of all the components. It'll even give you a BOM list that you can buy directly from a distributor. Lately, some of our products have a software component associated with it. For example, we have a single chip solution for USB PD that has a USB PD protocol stack built into the firmware. Recently, we have introduced software to drive BLDC motors.
As you know, we have BridgeSwitch, a product that allows you to drive BLDC motors very efficiently. To make it easier for customers, we are now providing both single phase and three-phase BLDC software. Now, all of this technology is well protected by a lot of patents, a lot of know-how, and I think many of you who've followed us a long time know that how defensible our patents are. Let's talk about innovation. When we talk about innovation, we are not talking about incremental steps. We would rather swing for the fences. We seek disruptive innovations that not only advance the state of the art, but gives customers benefits that they never thought was possible. As many of you know, we have hit a number of home runs.
I don't think it's an exaggeration to say that PI is responsible for every major breakthrough in the power conversion IC technology over the last three decades. Not just evolutionary. We're not just adding features or shrinking the die. We do that as well. But each new product is a significant advance in the state of the art, culminating in a multi-year R&D effort. Take InnoSwitch, for example. It's the fastest ramping product in the history of our company, and I would argue it's the most revolutionary product in the history of the industry. Using IC package to provide a safety isolated data link so that we can not only integrate the high voltage side, we can also integrate the low voltage side, and in doing so, can provide even higher efficiency and much higher level of integration.
Mike will talk a lot more about it and show you how unique this product is. The most telling part is after a decade of introduction, there is nothing comparable in the industry to this day. This is a synergy of the process technology, the control technique, and packaging all combined into one. It's only possible with system-level approach. Now, like all other breakthroughs we've achieved, InnoSwitch took a long time to bring to market. 5+ years of R&D and many millions of dollars of investment. Revolutionary technologies are always hard, and high voltage makes it that much harder. But in our history, the payoff justified the effort. Each of these products have a very long sustained life of revenues and profit generation. Now let's take TOPSwitch, for example. I think some of you will remember TOPSwitch.
Now, we've had a few generations of refinements, but fundamentally, it's the same product we introduced in 1994. That's three years before we went public. In fact, it took us public. In 1998, we introduced TinySwitch, and in 2001, we introduced LinkSwitch. These three product families last year accounted for nearly 50% of our revenue. Now just think about that. These products are 20-25 years old, and they constituted nearly 50% of revenue last year. Even more remarkable, these products grew 20% last year after that many years. Now, most of the companies won't be talking about old products like this, not in Analyst Day, going back this far. Why do we do this?
It shows how different our business is, how different our markets are, and how differently we think about our products and about markets. Now, what does this mean? If you have long life products, it directly translates to R&D leverage. We can address a broad and diverse set of end markets with a single product or single product family, and all the while focused on one functionality, which is power conversion. Because all products that require power require a power supply. The same product can go into all of our end markets. You compare that, a typical analog company of our size, they have to have hundreds of application-specific ICs to cover the same end markets because they're different for each and every application. Whereas our products can go into appliances, chargers, industrial applications, automotive.
In fact, the automotive product that we've just released is actually InnoSwitch. We're able to leverage our R&D efforts and focus our R&D efforts and leverage them across many markets. Okay, the second key takeaway is that our SAM is expanding and also diversifying. If you go back 12 years ago, 2010, we had about $300 million in revenue, and we had a SAM of $1.5 billion. It comprised of AC/DC products, these are power supply products, in the 0-50 W range. Now we made it our top priority to increase the SAM over the past decade so that our growth is not limited by SAM.
What we realized was we don't have to go outside of our core competence to do this because there are so many opportunities in the high voltage power conversion space, and that's expanding all the time. We now are at $4 billion, almost 3x, and we expect to go to $8 billion in the next five years. Thanks to new products and also secular megatrends like IoT, like electrification, of transportation, for example, and decarbonization. All of these are expanding opportunities, and we'll talk more about that. There are two ways we are expanding our SAM. One is these secular megatrends I talked about. As you know, appliances are getting smarter and more efficient. They have more features like connectivity, like LED lighting, and there is a huge transition going on to BLDC motors, which are much more efficient than the current AC motors.
All of this is content increase for us. Let's take IoT and home and building automation. Many, many opportunities for us. Every output, a wall outlet is going to have a USB output. This needs to be very small, very reliable, very efficient, and it's permanently connected to the grid. You can't unplug it, which means it has to have extremely low standby consumption. Just think about it, there's gonna be billions of these connected to the grid. If you don't pay attention to standby consumption, you'll have to build hundreds of power stations just for the standby consumption of these devices. Perfect fit for our technology. Most of these are in standby most of the time. You know, whether it's your smoke sensor or your door lock, they're all in standby most of the time. You hardly use it.
They're connected to the grid 24/7. Let's take chargers. You know, we think of chargers as this throwaway material, you know, commodities. Not true anymore. Chargers become a very sophisticated appliance. It not only can charge your battery fast these days, in minutes in some cases, but it can actually communicate with the device and adjust the output in very fine steps to optimize not only the charging time, but also the lifetime of the battery. There are multiple port chargers, and then of course, it has USB PD protocol implemented inside. So it's got software inside now. Probably the biggest transformation, technological transformation of all is electrification, specifically electric vehicles. Electric vehicles have so many sockets for high voltage, and it's increasing every day. I mean, even we are surprised.
You'll notice when we talk about the content, our content has increased significantly just in the last six months. The last one, which is by far the biggest, is decarbonization. The only way you can achieve this is with technology, and high voltage plays a huge role in that. We'll talk more about that as well. Now the second way we are expanding our SAM is through our products. We introduced a product line called HiperPFS product line that expanded our power range from 50 W-500 W. Now with the GaN technology, we can extend that much further, and that opens up a lot of new markets like data centers, like communication infrastructure, and so on. Of course, InnoSwitch, because of the level of integration, our content increases significantly.
It also provides isolation, which means we replace the optocoupler, and we provide much higher performance because it's much faster. BridgeSwitch for BLDC motors, we talked about that already. Dramatic expansion in our content in appliances. Of course, we acquired a company about 10 years ago that got us into the gate driver market. That allows us to address the renewables, traction, which is basically locomotion and trams, and high voltage DC transmission. Now you may think, wow, this sounds like a little bit different business. You will notice it's not. There's so much commonality with everything we do. We use the same package, we use the same FluxLink, and we'll talk about that. Mike will talk about that later on. Let's take a different look at how we get from $4 billion to $8 billion.
Of course, AC/DC market is expanding. This is the incremental, you know, change in AC/DC market over the next five years. This comes from higher power, it comes from more content. Then you have motor drive. Now we already have a large SAM and motor drive with the current product, BridgeSwitch, but we're gonna expand the power level, so we can go to higher power motor drives, and GaN will help us even further. Automotive, almost $1 billion worth of added SAM in the next five years. We are a perfect fit for so many sockets in automotive, not just power supplies, but also gate drivers. Thanks to GaN, as I mentioned, we can now get into data centers and communications infrastructure. Lastly, more gate driver business.
You may say, "Wait a minute, that looks really small." Well, that's only because we have some already in our existing SAM, and also that gate driver SAM will expand significantly beyond 2027 when renewables become a major market for this product. It will, you know, as other countries, all the countries start installing solar and wind power, we'll see a significant increase. By the way, a portion of that is in automotive as well. Automotive has gate drivers. Now let's take a look at it in terms of our large market segments that we talk about, which is industrial, consumer, computing, and communications. One thing you'll notice right away, the largest growth in SAM comes from industrial, automotive, and appliance markets. To a lesser extent, computing, and to a much lesser extent, communications, because that market is saturated.
We'll talk more about that. Doug will explain all the details. That means our mix is gonna change. In fact, Sandeep is gonna talk about our mix, how it's gonna change over the next five years, and this gives you a hint. Now let's go to the next one. I mentioned GaN a couple of times in the context of SAM expansion, but let's talk more about this. We believe GaN will, as a high voltage switch technology, dominate the power conversion market below 1,200 V. It is much better than silicon when it comes to efficiency, and we also believe our GaN, PowiGaN, will reach cost parity with silicon very shortly in the two to three years. we have by far the best GaN transistor in the world, highly differentiated from other technologies offered by other companies.
Yet we don't sell a single GaN transistor on its own. Why? A discrete solution with GaN is still a discrete solution, and we talked about why discrete solutions are not as attractive, and Mike will go more into that. PowiGaN is purpose-built for our products, and as we always do, we integrate the switch with the rest of the system and offer a system-level product. In fact, we do such a good job of it, we were selling GaN for a long time without even telling customers there's GaN inside. The only difference is the performance difference they see. Of course, a couple of customers thought there's something strange going on here. This can't be silicon. Later on, it became public, as many of you know. That's how easy to use our products are.
It doesn't matter whether it's silicon, GaN, or silicon carbide, the outside circuits looks the same, and it's as easy to use. Now, we are not zealots. We will use the best technology for any given application, and we do believe silicon and silicon carbide have a place in our product future. However, GaN will be the dominant technology. We have a large number of exciting GaN products in the pipeline that'll address much wider power and much wider voltage ranges. Again, you'll hear a lot more about that. We are already the leader in GaN, both in terms of revenue and breadth of products. We have also a large number of designs in the pipeline, approximately 400 designs going on right now. It covers an increasingly wide range of applications, including H&BA, which is home and building automation, and other industrial applications.
The fourth key takeaway, and I touched on this earlier, the high-voltage semiconductors are very crucial for a low carbon future. This is a graph from IEA, or International Energy Agency, and it shows how to reduce carbon dioxide emissions from the current trajectory to a sustainable scenario, which is in line with the Paris Agreement. The thing to note there is more than 2/3 of that savings in carbon emissions comes from two areas. One is efficiency, the other one renewables. Efficiency and renewables. Right in our spot. The high voltage is in every link from power generation to power transmission and usage of that power, ultimately in your appliances. As we look forward to the low carbon future, what we're talking about is carbon free renewable energy generation. We are talking about high voltage DC transmission.
The current technology is AC transmission, but DC transmission is much more efficient, but it requires inverters on both sides of the transmission line. If you don't know what an inverter is, Mike will explain that to you. Of course, we are talking about efficient consumption, including EVs. For those who have followed us for a long time, you know that we've been in the energy efficiency business since the 1990s, way before anyone has heard of ESG. We are widely recognized for our role in making electronics more efficient through awards, through listing in stock indices, and also by policymakers who look at us for what the technology can offer. I'm very pleased to say today that by our estimation, 95% of revenue in 2021 came from products that qualify as EU Taxonomy Regulation eligible. 95%.
The reason is our products not only save energy being more efficient, but they also go into carbon reducing technologies like renewables and high voltage DC transmission and so on. Okay, now we set the stage for what we're going to hear today, and Mike will be coming on very shortly to talk about our product development philosophy and how the system integration sets us apart from our competitors. Along the day or along the morning today, you'll also hear from few of our customers as well. We will start with first one of those videos, and this customer's name is Tridonic.
We here at Tridonic, we're producing LED components, be that drivers, be that LEDs, for the lighting industry. Tridonic has always been focused on power savings and energy efficiency. Over the last years, we've been growing from a magnetics company into an electronics company and now providing the best possible LED solutions, which was always in the focus of saving energy. PI is our strategic partner in terms of getting power supplies into our LED drivers.
Our relationship between Tridonic and Power Integrations started more than 10 years ago and was focused on really from the beginning, on trust and reliability.
Over the last year, we actually strengthened the partnership with PI significantly. You guys been helping us big times in building up and also continue to serve our customers.
It's not just a pure buying and selling. It's much more talking about the future needs.
We are not jumping from one supplier to another supplier. We do value long-term partnerships, good collaboration, and we do also see that into the future.
They're really asking us, where is the problem, and what do you need as a solution? Finding the right solution.
We found a good and a very valued partner with Power Integrations.
We made the right choice in this partnership.
Good morning. My name is Mike Matthews. I am the VP of Product Development here at Power Integrations. Been with the company for about 30 years. Started off as a technical support guy in Europe, and then moved over to the U.S., and I've been in my current role for about 10 years. I'm here to talk to you today about what we call the power of integration and our product development philosophy. As Balu already said, we really focus on highly differentiated products. Products that are gonna drive the market and are gonna revolutionize the market rather than evolutionary products. In so doing, we've developed a large portfolio of IP.
We have well over 2,000 patents in the history of the company, which, because we're a pure-play focused company, is a very large number of patents, probably the biggest in the industry in our particular area. In generating all of that IP, of course, we've developed a great deal of know-how at a system level, and that's very important, as Balu has already talked about. Again, I'll show you some examples of how that's become very beneficial to us. We're very proud to say that, as Balu mentioned, long before energy efficiency was even a topic, we were driving, we were pioneering the area of energy efficiency in power conversion, and we continue to do that to this day. A great deal of our IP is focused on energy efficiency. The way we generate our IP is what we call multidimensional.
We are a semiconductor company at the core, but a great deal of our IP is focused on things other than semiconductors. We have developments in high voltage process and our controllers. Those are semiconductors. We also have a great deal of IP in packaging technology. For example, topologies of the power supply. A lot to do with IP of components we don't even sell, and we offer those as a free license to customers who are using our chips. We have a great deal of system knowledge, and it benefits our customers in the end because we provide therefore a system solution. As we've already mentioned, it's highly defensible as well. We had on occasion had to defend our IP. We've been 100% successful in every case.
We develop all this IP in order to put a ring fence around it in what we think of as IP safe countries, where there's less leakage of IP away from us. Before I go any further, you're gonna hear a lot about power converters today. I wanna just take a chance to describe what a power converter is. There are two main types that we will talk about. One is this one shown here, AC/DC, or we sometimes just call them power supplies, AC/DC converters. The job of that power supply is to take AC from the wall outlet, as shown here, going into a washing machine in this case, but it could be many different appliances in the house or in the industry. It converts that high voltage AC down to low voltage DC, which is isolated.
It's very crucial that it's isolated because the customer can come in contact with that low voltage on the output, and it has to be safe. They can't get electrocuted from that. That's a fundamental requirement of the AC/DC converter. I think a lot of us, to the extent that we ever think about power supplies, think of something like this. Plug in the wall adapter. This is an AC/DC power supply as well. This is actually a relatively small percentage of our business. This is only about 25% of the business. The vast majority of AC/DC converters are in all of the other products that plug into the wall. Many in this room and certainly in your houses and homes, you'll find whether it's a washing machine, dishwasher, air conditioners, electricity meters, things you wouldn't even think about, have AC/DC converters in them.
Now into this market, we provide ICs. The one shown here on this board, this is a zoom-in of the board with our logo on it. That's our InnoSwitch product. We sell a bunch of other different ICs into the same essential application. This is a chip business for us. The other type of converter I want to mention, because again, you'll hear a bit about this, is kind of the exact opposite of the first one. It's a DC to AC. A number of renewable applications produce DC as the output.
Here's one example of a solar farm, and you need what is called an inverter to change that DC back to AC to put it onto the grid, and then eventually that is transformed down and comes back to an AC/DC, which is the first converter I mentioned, to provide power to the different products in the home. These inverters require isolation, ideal for our high voltage technology, and energy efficiency is also very critical in that application. Both of these high voltage converter spaces are ideally suited to our technology. These are the terms we're gonna be using, and I hope this is helpful just so you can see what we're talking about. Now, I want to start off with an example, a real-world example of what we think is revolutionary integration, revolutionary innovation in power conversion.
I'm sorry to put up what looks like a complicated diagram. This is an actual circuit schematic, circuit diagram of an AC/DC power supply. On the left-hand side is the AC coming from the wall. On the left-hand side is the AC, on the right-hand side is the customer load. In this case, we're showing a laptop converter, a laptop power supply. This is a real-world application as well. This isn't some academic study. This is an actual production power supply. This is all of those shapes on that circuit diagram are electrical components that have to be procured by customers and have to be placed onto the circuit board in the manufacture of the power supply.
Now, just to put it in context, I'm gonna highlight a few areas of this power supply. This particular power supply uses a gallium nitride from a competitor, a gallium nitride provider. All of the other circuit blocks that I'm going to highlight on this diagram are external to the gallium nitride transistor. This is what we'd call a state-of-the-art discrete power supply. The integration that we bring, and I'll use the example our InnoSwitch product, is every one of those colored boxes that have specific functionality in the operation of the power supply are taken away from being discrete components, which they're shown here, into integrated into a single package. As a consequence, the overall power supply becomes a lot simpler and a lot smaller, a lot more compact.
As Balu mentioned, it also removes the responsibility from our customers of dealing with all of those components, whether it's procuring them, or whether it's sorting out how they work in the power supply. That is now all integrated in InnoSwitch. If I tabulate a comparison, it'll perhaps give you a better understanding of the level of integration that we bring. Both the competition power supply that I showed a moment ago and the InnoSwitch power supply that I'm showing here, they both integrate the GaN transistor and the driver and the current sensing goes with it. If you step through all of those other boxes that were highlighted on the original schematic, those are all integrated in InnoSwitch, they're all external with the discrete solution. What that means is the eventual component count drops by more than 50%.
Now, in our world, that's a revolution. That completely changes the way that our customers have to behave in terms of procurement, sourcing, which of course, in today's environment is very, very critical. Anytime you can save components with the supply chain issues that we have, it's very critical to customers. And all of this we integrate into a single package, and I show a photo there, of the package in the bottom right-hand corner. You'll get familiar with that package by the time I finish this, 'cause you'll see that strategically, we plan for that package to be used across many different products. That's part of our manufacturing efficiency drive in our strategy. Now, InnoSwitch was so significant as a breakthrough technology that I'm gonna spend a couple of slides talking in a little bit more detail about how we did that.
What InnoSwitch does is it integrates something called the isolated interface or data communication channel, which communicates what is happening on the output side of the power supply, the low voltage side, to the high voltage side of the power supply. There's a picture here. The arrow there is pointing to what we show as the FluxLink on the circuit diagrams. There's a dotted line. What makes InnoSwitch unique is that the photo of the power supply circuit board you see in the bottom right-hand corner there, it has a light green part on the PCB. That is the isolation barrier. It is a physical barrier between the high voltage and the low voltage side to ensure that the output voltages is safe for the user.
What is unique about our component, the InnoSwitch, as you can see, is it bridges across that barrier. Within the InnoSwitch, we have a controller that is situated on the low voltage side, a controller on the high voltage side, and in between, we have this FluxLink communication. What does that do for us? Well, once you have that communication within the chip, it opens up a whole world of possibilities in terms of control. One of them, and Balu already touched on this, is that we can now integrate in that controller in the InnoSwitch package, either a silicon high voltage switch, gallium nitride high voltage switch, or even a silicon carbide high voltage switch. We don't have any reason to push one of those technologies over the other. We use which is the best for the application.
It so happens at the moment, silicon is optimum at very low power levels. Gallium nitride comes into its own at much higher power levels. Silicon carbide is superb at very high voltages, and that's important in electric vehicles, for example. Some of the new electric vehicles being under development at the moment are using 800 V batteries instead of 400 V, which is currently used. That 800 V battery requires, because of the way the power supply works, requires almost double that voltage on the high voltage switch. The silicon carbide that uses 1,700 V. Complete flexibility, that's one of the advantages of InnoSwitch. Complete flexibility in the high voltage switch technology that we use.
The other thing that FluxLink allows us to do is some very proprietary control techniques which enhance the energy efficiency, and I'll talk about that in just a moment. Before I do, I want to illustrate what I mentioned to you a moment ago is that a great deal of our IP has nothing to do with the semiconductors within our devices. Some of them are to do with the package itself. FluxLink is one great example of that. This is a picture that you can see here at the bottom left of the insides of that IC. It's a zoom in of the IC. The FluxLink that's being pointed to with that arrow shows that there's no silicon there. It looks, to all intents and purposes, like a hole in the package.
All it is is copper from the lead frame and the molding compound that goes around it. That is the FluxLink. That is the high voltage communication, isolated communication link between secondary and primary. Nothing to do with semiconductors, all to do with package. This is why the system knowledge that we have is so critical to developing and giving system solutions to customers. Now, I also mentioned a moment ago how the InnoSwitch helps us in terms of safety and in terms of energy efficiency. Here's a graph that shows that benefit of energy efficiency. This is a comparison, again, between a discrete GaN, gallium nitride power supply, and a Power Integrations PowiGaN, gallium nitride power supply. Along the x-axis is the load.
To the right is 100% load, and then as you move more to the left, that's light load condition. Now, if you first of all look at the 100% load position, we are more efficient. It's not surprising because our PowiGaN is developed specifically for this application as opposed to the discrete GaNs which have to cover many different applications. What's really noticeable is that if you go down that curve from right to left, our efficiency actually goes up as the output power goes down. That is less to do with GaN. It's far more to do with the control schemes that we use within the InnoSwitch controllers.
Again, it's a combination of understanding the high voltage semiconductor with the control schemes that are needed for efficient operation in particular topologies and for things like FluxLink, the fact that the package itself is used to create the isolated feedback network. If you think about this, the standby condition is very, very critical. Most of these applications are the things that are powered in this room or in your home. They spend most of their life sitting in standby. They're not doing the full function. Look at your TV, your coffee machine, your dishwasher, your washing machine. Much of its life is sat there, not doing its main function, waiting for us to tell it to do so. That's what's called a standby condition, and a huge amount of the energy that's wasted in the industry is appliances operating in that standby condition with poor efficiency.
The things that we bring with our control technology is to make sure that that efficiency doesn't go down. It actually goes up a little bit at standby, and that is why there's such a huge and dramatic difference between the discrete circuit and ours in terms of energy efficiency in standby. Now, so far I've talked primarily about AC/DC power supplies with a single output. Here's one of those again. But in typical and many applications, whether it's television, whether it's air conditioner or a refrigerator and so on, they require multiple outputs from the power supply. The way that's normally realized is that our front end AC/DC power supply chip, InnoSwitch, is followed by a bunch of controlled power management DC/DC converters. Sometimes they're referred to as DC/DC converters and sometimes as power management ICs, but they're here shown as green on this schematic.
Because of our FluxLink, again, coming back to the FluxLink, we've derived a technology now which entirely eliminates those converters, those downstream converters. They waste energy, they take a large number of components, they use space on the board. What we've done is absorb all of those converters into a version of the InnoSwitch. We do this by very high frequency multiplexing of energy to each of the outputs that's necessary. We can do that all from one magnetic component and eliminate the need for the DC/DC power management stages altogether. Again, that's a revolution. That's something entirely different. No one else can do this. It's based on our packaging technology combined with our IC technology and our high voltage process technology. Same thing goes for motor control.
Oliver already mentioned that many of our appliances now are switching from the old AC motors, which had basically no electronics inside, but lower efficiency to much higher efficiency brushless DC motors. Those require electronic control to switch the currents in the windings in a way to generate the rotation of the motor. For many years, we've supplied AC/DC components. Again, I showed InnoSwitch here as an example. But they have a bunch of other high voltage semiconductors in them as well, and those are the semiconductors that are shown here on the right. We haven't previously been involved with those, but they are high voltage switches. They switch the currents into the three windings of the motor. I've ghosted the three there so it doesn't get too complicated.
There are basically three of those combinations of high and low side switches needed. There's a bunch of external circuitry around that. Again, what we've done, we looked at our technology, we've now integrated those into three much simpler packages. If you notice, it's the same package, same package we use with InnoSwitch, same package we use for the DC/DC eliminating chip we saw on the previous slide. We reuse the packaging technology as much as possible. That's part of our strategic planning to lower manufacturing costs. Not only that, we've established some clever techniques, and one I've highlighted here, for example, we call a fault bus. The fault bus transmits information from the motor itself to the microcontroller at the center, the heart of the control of the motor.
It allows manufacturers to implement diagnostics and the preventative maintenance, which is becoming popular in certain applications. For example, if this motor was used in a dishwasher, if the motor was becoming noisy, using higher current than normal, going to overtemperature conditions, our BridgeSwitch product can communicate that to the microcontroller. The microcontroller can send that through Wi-Fi to the manufacturer. The technician can come to the home, replace the motor before it even breaks down. That's one example of a feature which we have introduced here. Now, the microcontroller, you may say, "Well, why don't we make the microcontroller?" Well, it's not really a high voltage semiconductor. It also has multiple other functions and duties within an application, like a dishwasher or a washing machine because it may be controlling wash cycles and timing and so on.
It's not a business for us. What we have done now is introduce some software that we provide as a free service to customers using our BridgeSwitch and our InnoSwitch products that allows the motor to operate more efficiently and also simplifies the external circuitry even further. Although we don't sell the software, it's monetized in the sale of our chips that go with it. Talking of software, we've developed over the years a number of expert systems. The one on the left here we call PI Expert. That's a software suite that we can deliver online. The customers can use it online, and it automates the process of developing an AC/DC converter. On the right-hand side is a more recent introduction called MotorXpert.
does the same job for the BridgeSwitch that I just showed on the previous slide. These are expert systems because the customer can introduce their specific requirements. Let's take, for example, a power supply. If they introduce the fact that they want the highest efficiency and the smallest size, this software suite will make expert decisions based around that to optimize the design for those particular purposes. If they want the lowest cost, they'll do other things. It is really applying some of our expertise that we've built up over the last 30 years or so in a form that customers can access online. Again, free service, but we monetize it through the sale of our ICs that go with it.
Lastly, I'm just gonna touch for a moment on our high voltage gate driver, isolated gate driver business, which as I mentioned at the beginning is a board business. We do manufacture ICs that go into the boards, but we sell the whole board because that's the best way of delivering the performance and the system performance to the customers. Incidentally, these two photos that I show here on the screen are on display in the demonstration room next door. I'd encourage you at lunchtime to take a look at these. It's the same story as the AC/DC power conversion in a way. That on the left-hand side is a competitor gate driver IC. It's mounted on a, what's called an EconoDUAL module. That's an IGBT module, very high power.
Same as on the right-hand side, same module. The board on top is the gate driver board, the isolated gate driver board. You can see just from the photograph, the complexity of what's on the left compared to our board on the right, the SCALE EV. Again, we're bringing the same level of system knowhow and integration that we did in the AC/DC conversion market as well. By the way, the competitor board, if you turn it upside down, there's more components on the back of that as well. A significant reduction in component count, and very much a simplification for the customers.
Again, you'll note just I've pointed it out there on the board on the right, that again, it's the same package that we use for InnoSwitch is used for this high voltage gate drive market as well. Again, manufacturing efficiencies is a strong part of our strategy and our long-term thinking. Let me just wrap up by summarizing, if you will. What I said at the beginning was multidimensional. What I meant by that was that we don't just innovate in semiconductors. We innovate the package and external to the power supply and external circuit in the power supply. I hope this has demonstrated some good examples of that, some real world examples. The IP portfolio we've developed along the way has proven to be highly defensible.
We have had to defend it on a few occasions, 100% success rate. In building all of that, we have a great deal of high voltage expertise. It's not easy. You know, you could say, "Well, why don't competitors just integrate all that stuff?" This is not integration in the way we might normally think about it. We're not just integrating transistors and resistors into silicon. Many of the components in these power converters are not integratable in that way. They're high power, high voltage, very bulky components. They don't lend themselves to conversion into silicon. The way we integrate them is we eliminate them. We eliminate the need for those components. That's why our integration is a little bit different from what you would classically think of as integration. It all comes from this high voltage expertise and know-how.
The culmination of this is there's many, many barriers to the competition in entering this market. As Balu already mentioned, you know, not only have you got to know about the way the power supply works, you have to have safety certifications and knowledge about high voltage and the packaging and so on. So a number of barriers. What this means is, and we've proven this over the last 30 years, is a highly defensible, sustainable market model. Sandeep will talk later on how we've sustained the operating margins and gross margins of our products within that model. In culmination, this is what we believe is innovation in power conversion. Now, I'm gonna hand over to Radu Barsan, our VP of Technology, in a moment. He's gonna talk about high voltage semiconductor processes and our developments in that area.
Before that, I'm gonna show you a video from a company called Diehl Controls. Diehl Controls is, if you haven't heard of them, you may not have done, they are one of the world's largest ODM suppliers to the appliance industry. They provide subsystems, electronic and electrical subsystems to people like Electrolux, Whirlpool, Miele, Bosch, Siemens, and so on. I personally know this customer very well. I mentioned at the beginning, I started off life as a technical support guy. I was the technical support for Europe for many years when we were a tiny company. Diehl was one of our first customers for TOPSwitch, and I personally built and populated the first demonstration board that I provided to them, which they subsequently put in production in 1996.
They have been a customer for us for over 26 years, and I think you'll see in this video is that it's clear we wouldn't sustain that level of business and growth of that business over so many years unless we had a very strong partnership with this type of customer. Thank you.
Diehl is a worldwide operating international company with a turnover of EUR 3.2 billion. Diehl is producing and developing controls for home appliance, heat pumps, and air conditioners. Energy efficiency is a major aspect for our controls. Having the energy efficiency in the front line is giving us a competitive advantage, and this is supporting our customers, and by that, the consumers.
We use Power Integrations because of the high quality and because of the efficiency of the part.
Designing in Power Integrations is really easy. Our R&D teams are also in close contact to Power Integrations, so we have really great technical support which make us really capable to design in Power Integrations in a really wide range of applications.
We always get the great support from Power Integrations in all categories like quality, technology, supply. Power Integrations won the Supplier Award in 2021 from Diehl Controls.
From engineering point of view, it's really pleasant to work with Power Integrations. We get samples in very reasonable time. We have very well prepared app notes, data sheets in exactly the required level of details, which make it really easy to design in Power Integrations into new applications. Diehl use Power Integrations products mainly in switching mode power supplies for our power boards, which are used in home appliance like dishwashers, washing machine, refrigerators.
Diehl Controls is working with Power Integrations since decades, so this cooperation we are having since that long time is bringing us forward to the future.
Good morning. My name is Radu Barsan. I'm the VP of Technology at Power Integrations, and I've been with the company for the last 10 years. I'm very excited to present to you the technology story of Power Integrations, which is in fact the story of the integration of proprietary power devices for high efficiency power conversion. Now we don't usually talk about technology, but today I'll share with you how technology has enabled our growth for the last 30 years, how it supports the company's strategy to help reduce energy waste through improved power conversion efficiency at affordable prices in applications that span a very broad range of power levels, and of course, the unique business opportunity and prospect it afford us going forward.
With very few exceptions, silicon carbide being one as of today, our products use proprietary devices and technologies tailored for our chosen applications, products and markets. Why not use a foundry process like every fabless company? This attribute of our approach is not fortuitous. It's a cornerstone of our business model, and here are a few reasons why. First, it offers the best performance for our devices and our integrated system level products. The second is that for power devices, the control of process technology and device design is absolutely critical. From Infineon's CoolMOS product line to STMicroelectronics, silicon carbide, and so on, you can hardly find a power device company that has been successful other than vertically integrated. The reason is the power device technology and the power device design are very closely intertwined.
Unlike integrated circuits, ICs, where the designers don't need to change the process in any way, they don't even sometimes need to know how the millions of transistors they are hooking together are being made, in the power transistor, it's extremely important to be able to change the process and the design at the same time in order to obtain the best performance and the lowest cost. This model affords us the ability to have a very unique business model. I'll talk about it in a moment, but it's what we call fabless IDM, and I'll explain the reason. It gives us a win-win ability to partner for long periods of time with our manufacturing partners.
Our products are very difficult to reverse engineer, and that's because we own not just the design, but also the manufacturing process. It's very difficult to copy both the process technology and the design. It gives us end-to-end control over manufacturing parameters, quality, reliability, performance, and so on. It gives our product development teams full flexibility to choose the best approach to engineer our system-level products at the highest quality and reliability and at the lowest cost. I mentioned this fabless IDM model. What does this mean? This contradiction of terms is meant to illustrate how our model uses the best of both worlds. Fabless being asset light, and IDM being having full control of our internally developed processes and technologies. We develop these technologies in order to provide our customers the best-in-class assurance of supply, quality, and cost.
We own specific equipment if needed when it's specific to our process technology, and we replicate the process technologies to multiple factories, anywhere between two and five different fabs that run our proprietary process. That ensures multiple sources for our products as well as operational flexibility. The factories are geographically dispersed for seamless business continuity in the event of local disruptions of any kind. We develop long-term strategic alliances with our partners. 10 years, 20 years. We have manufacturing partners that we've been working on for 25 years. These agreements bake in upside capacity for assurance of supply and also resilience in terms of disruptions and business continuity. We also share the pain and the gain across multiple semiconductor business cycles, which is also a feature baked in our agreements.
The integration of power switches such as MOSFETs with advanced controller, as you heard already, is at the core of our business model. In fact, the company was founded on a novel high-voltage CMOS process technology that allowed the monolithic, meaning on the same chip, of high-power, high-voltage transistors with controllers. Since then, we introduced multiple generations of this technology approximately every five years, each generation enabling higher product complexity and value. To fully utilize the best attributes of advanced high-voltage processes, which require more advanced lithography, and we did develop two generations of such controller, high-voltage controller technologies. In order to utilize those attributes in conjunction with power devices, power transistors, which require different technologies, as different as a separate material, as you know already, we diversified from monolithic to co-packaged technology integration.
That allows us to mix and match the optimal controller technologies with power device technologies. As far as power device technologies, we've introduced a couple of high-voltage lateral silicon MOSFETs. Again, these are silicon transistors that are being co-packaged with controllers made in more advanced lithography processes. We developed a couple of high-voltage vertical silicon transistors, which are, in this case, optimized for motor control, and you will hear about our BridgeSwitch product, which uses these vertical FETs. PowiGaN, our GaN flagship technology, which I will describe in a little more detail soon. This ability to integrate the best technologies into a single package solution has enabled our relentless development of ever-increasing value products. This chart shows the functionality of our product families released at various times.
From the monolithic era to the co-packaged multi-technology products you see in more colorful colors. Blue meaning silicon-powered FETs married again with silicon controllers, yellow meaning GaN, and the two-tone meaning a portion of the family being supported with silicon-powered FETs, and another portion of the same family being supported by gallium nitride in a completely seamless manner. This provides higher levels of integration, higher power levels to enable us to address multiple new markets, and industry-leading functionality, performance and cost. You're already familiar with this picture, but in order to illustrate the breadth of our integrated solutions, I'm gonna use the same example of our InnoSwitch3-PD product. This is a fast-charging multi-protocol solution that is compliant with USB power delivery 3.0, and with the USB Type-C 1.3 standards.
It incorporates two controllers, one on the high voltage side, on the main side, the other one on the client side or the secondary, as we call it, a digital controller that interfaces with the client, our proprietary FluxLink communication Mike described earlier, and the power switch. The power switch can be anything from a lateral silicon power FET, a vertical silicon FET, a gallium nitride switch, or a silicon carbide switch. The same ability to pair the best fit for function power switches with controllers is pervasive and spans across all our product lines. Now, you've been surely hearing a lot recently about the competition between silicon and these new so-called third generation semiconductors or wide bandgap materials for power markets. Today, most power applications still use silicon. The majority are insulated gate bipolar transistors, IGBTs.
IGBTs can go to very high power levels and very high voltages, but they're not very efficient. For high efficiency, which is what this slide is focused on, you use silicon MOSFETs today, and I'm sure you're familiar with Infineon's CoolMOS and other vertically integrated companies that dominate the silicon MOSFET space. However, silicon MOSFETs are not the most efficient solution, and actually, in terms of power and voltage, silicon carbide continues from where silicon stops. Really shining at very high voltages and power levels. The only problem is that the cost of silicon carbide is limited by the manufacturing, the fundamental way silicon carbide devices have to be made, which involve temperatures in excess of 2,500 degrees Celsius, so a lot of costs in the energy required to make silicon carbide.
Both the starting material, which also has to be silicon carbide, and the wafers themselves, silicon carbide, conspire to keep the cost of silicon carbide relatively high. Gallium nitride overlaps with both. It's more efficient, higher performance than silicon and lower cost than silicon carbide. The reason is, gallium nitride can be made on relatively lower cost substrates with processes and equipment that are not very different from the standard CMOS toolset you find in silicon factories. To recap, silicon carbide can go to very high power levels, very high voltage levels, very high temperatures as well. Whereas in the space that overlaps the silicon and silicon carbide areas in this two-dimensional chart, GaN wins on both performance and cost. Let me share a few details of our PowiGaN roadmap. It has basically two branches. One is device scaling for size and cost reduction.
Another branch is for extending the voltage at higher ratings. From the first PowiGaN generation we shipped commercially in volume in 2018, PowiGaN is now in volume production with on the second generation, in qualification on its third, and in development on its fourth. This entire roadmap is rated 750 V. While this mainstream PowiGaN roadmap is rated 750 V, newer generations are being qualified to support much higher voltage applications, matching silicon carbide up to a certain voltage and power level, if you recall the previous chart. There's a new version coming up that can support 400V bus automotive, and a lot of industrial applications that require more than 750 V rating.
Another generation that can support 800 V bus automotive, and also on the roadmap is the generation that can support 1,000 V rating. A 1,000 V bus automotive, which means a lot more than 1,000 V for the device rating. How is PowiGaN different than other GaN? What are the unique advantages? I break them into three categories, technology, device design, and product level. Unlike competitors who use generic foundry GaN processes, both the GaN epitaxial process, which is a fundamental contributor in both the cost of the device as well as the performance. Both the epitaxial process and the fab process are internally developed, just like any other technologies we use in our products. Like silicon, their GaN is no different. It's all developed internally, 100% ownership of our IP, very strong protection, as I mentioned earlier.
Second, the GaN power device design, reliability qualification, quality management, they're all 100% proprietary with themselves well-protected IP. Finally, unlike competitors who sell discrete devices, PowiGaN is incorporated in our system in a package product and tested all the way through applications, real application system-level validation. Very unique, very different. To summarize all this, PowiGaN is the fruit of continuous collaboration of all in-house engineering teams. We don't have to go ask customers what they need. The customer is inside. From epitaxial growth to foundry to fab process to testing, probing, system-level validation, product engineering, they're all in-house. This allows us to produce the best possible technology since we own all the pieces of the technology. Now, a word on competition. Yole is a market research company that has been following GaN for a long time.
This data I'm showing here is from their latest report they released in July. The advantage of GaN have been recognized long ago, and a lot of companies have been started, whether large companies got into GaN or small startups were started over the years. They all have cumulatively invested billions of dollars in making GaN work. The majority, I have to say, have failed, but new companies have started over the years to replace them, especially after Power Integrations have proven commercially that GaN is viable in volume in 2018. As Balu mentioned, we never made the news about this. People figured over time that we're, what we're shipping is not no longer silicon, we're shipping GaN. Along with product companies, a lot of epitaxial foundries have been started and purchased and sold.
Generic GaN process is also available from some foundries. TSMC is one of them that gives startups access to a black box GaN process and allows them to make GaN transistors, albeit at foundry prices. As of today, this is the list of our competitors. You recognize some large companies, some small companies. What I'm very proud to claim still is that we were first to ship high voltage GaN in volume. To this day, we are still number one in high voltage GaN market share. To recap a few of our key business model characteristics as they relate to manufacturing excellence, we discussed our unique manufacturing model, fabless IDM, that affords win-win supplier partnerships with mutual support over semiconductor business cycles.
Very important, especially as you can see today, the world transitioning from super demand bubble to signs of softness. We develop our own technologies. We install them in partner factories, specifically designed for cost and performance, geographically dispersed, a multi-source strategy for all process and package technologies, which form a solid foundation for a resilient supply chain, as demonstrated over the recent pandemic. Last but not least, we're very proud to have a best-in-class delivery performance of over 95% success to customer request date. I want to conclude by sharing the outstanding level of quality of our technologies and products. This chart shows the number of parts returned from customer production lines or from the field that failed for any reason whatsoever, per million units shipped by date code.
To give you a sense of a benchmark, a best-in-class quality level for a low voltage standard semiconductor company is 1 PPM. We achieve that and go below that level with our process technologies which operate at 700, 800, 900 and more volts. How can we achieve that? Well, first we invest significantly in quality and reliability. Second, we design in quality through our rigorous technology and product development and qualification processes. That allows us to continually drive the failure rate towards zero. As a result, the Powi brand stands not only for exceptional performance, but also value and support, and last but not least, for outstanding quality.
We're very excited and proud to continue to lead the power conversion industry, as we move forward and really reinforce our value proposition, and I hope you will better understand from today's presentations how we do that, and you'll enjoy the ride as well. Thank you. Next, we're gonna have a Q&A session. Balu and Mike are gonna join me here on the stage, and in the meantime, you'll see a next video from one of our leading home and business automation customers, Leviton.
Leviton is an electrical wiring device manufacturer that designs and builds and supplies low voltage, line voltage, commercial, industrial, and residential products. Our brand strategy is engineering possibilities, which leads us to why we use Power Integrations, because we are an innovative company through engineering. When we look to become the most innovative devices in the market, we look for partners like Power Integrations who will help us achieve that goal.
There's great support we get up front in terms of, getting the designs off the ground and through the process of development. Power Integrations team has been very helpful in helping us get the product qualification done and get us all the way through launch.
Without the Power Integrations components and the efficiency, we would not be able to achieve the levels of power that we have achieved on our USB devices. It's critical to have energy efficiency and heat management and all of these aspects to make these devices work.
Ross, I'm sorry, yeah.
Hi, guys. Thanks for the great day so far. Just had a key question on, as you expand into these new markets, it seems like the R&D is very leverageable across it, as Mike talked about. That makes a ton of sense to me. What about the scalability of the go-to-market, the sales process? Can you stay as kind of 75% via disti and handle this expansion of markets that way? Or do you have to, Balu, you know, from long ago, you would be more of an evangelist. I used to see you on flights to Asia all the time trying to go tell people about the benefits of integration. Do you have to put those hats on again and, you know, hire a bigger sales force, go direct to customers, those sorts of things?
Yeah. Excellent question. This is something we've been working on for several years now. In fact, we have very-
Yeah, yeah.
Yang Chiah Yee, I'll let him answer the question, please.
If you look back history like 10 years ago, Power Division is about like 10,000 customers.
The microphone.
Today, we are looking at, like, more like 15,000 customers, and it's growing as we speak. Our top 10 customers is like about 40% of the revenue. You know, we continue to have a lot of customers that because of the design tools that we have, like PI Expert, we talk about, you know, the BridgeSwitch, you know, the MotorXpert software that we have, we're able to bring in a lot of smaller customers, you know, to use our Power Integrations. You know, we recently launched an online application engineering support that allows small customers to, you know, basically, send in their requests. We are in the process of making this like able to support through live video as well.
We are basically making sure that we can support a wider and broader customer, because the design effort for a small customer and big customer are almost the same, but we need to make it easier for those customers to adopt it. Yes, distribution is very important. We have a global distribution partner like Avnet, Future, and then we have a couple of more focused and, you know, specialized distributor in China. Those all, you know, help us in the expanding our customer base. Also, we have DigiKey, Mouser, Premier Farnell, all those online distributors allow customer who want to buy just a few pieces to test out our solution with our demo board and so on and so forth. That whole strategy seems to work really well so far.
I think it's fair to say that there are markets that require scaling the go-to-market forces such as automotive or motor control, and we are investing in those areas.
Yeah. Just to add to the automotive market, we have dedicated sales and marketing today focusing on that market.
Tore.
Yes, let me echo the thank you for putting together this first analyst day. So Balu, you talked a lot about the SAM going from $4 billion to $8 billion. Obviously, you know, we've seen the new products, the new technologies you've introduced. As we start to think about, you know, units versus ASPs, and you know, I think it's a big discussion right now in the investor community, you know, pricing has obviously been very firm, very strong for the industry over the last few years. As we think about Power Integrations moving from that $4 billion SAM to $8 billion, how much of that is coming from higher ASPs?
Not just pricing, but higher ASPs, 'cause you're clearly trying to capture, you know, a much, much higher value with your customers, especially in some of these new markets like industrial auto and so on.
Yeah. Well, actually, Sandeep will talk about it more as to how our ASP is growing over time, and that will continue. The part of that will come from higher level our higher value products, and therefore, the ASP higher. I think that'll continue for the foreseeable future. Of course, it's also coming from us going into new markets with new products and expanding content and unit volume in a number of markets like electric vehicles. Again, Doug will talk about that as well. Did you have a question? Yeah.
Yeah.
Yeah.
Thanks. David Williams from The Benchmark . Just wanna see if you could give maybe a little more color on the automotive side. I know you've made some nice progress there, and it looks like you're in a couple of in-production designs. Maybe talk about that and then perhaps your next generation, kind of the engagement cycles, or maybe engagements with customers, as we kinda look at that next generation product.
Yeah. Actually, I'm very surprised how receptive this market is to a newcomer like us. We have not been in the automotive business, like many of the big analog companies are. But I think what is really changing the view is the compelling nature of our products. The how many components we replace. The best example is our InnoSwitch AQ, which we offer both as a 900 V and a 1,700 V product. The 1,700 V being silicon carbide. Every customer we have shown it to wants to use it. We already have four large OEM customers going into production this year. Some of them are already in production.
In fact, when you go at lunchtime, take a look at one of the inverter boards that has an emergency power supply using our chip. They removed the name because they don't like us talking about the name, but to give you an idea of how well-suited our product is. We replaced probably about 90 components in that system, and that used a discrete solution, which you know you had to put a lot of components to make it do what they required it to do, which is work from 30 V all the way up to 1000 V. InnoSwitch has no problem doing that. Similarly, we have LinkSwitch TN2, which also has a very wide voltage range. It's used in non-isolated applications like your seat heaters that requires a power supply.
We have, of course, drivers, gate drivers. Again, there are samples of that, and some of them are in production. Currently, we have four large OEMs using our product, and then we have another three design win but not gone into production yet. Some of them are in Asia, and some of them are in Europe.
Thanks. Hey, guys. It's Matt Ramsay from Cowen. I wanted to ask a couple questions about the GaN business. It looks like from one of the slides, revenue sort of tripling this year from $10-odd million to $30+ million, and $100 million design win pipeline. I was particularly encouraged by how diverse across end markets the design pipeline is for GaN. Is the GaN-based product you view it as the driver of this diversification of end markets that you're talking about, for the business over the long term, or are the traditional products gonna be more diverse as well? I'm just trying to figure out the momentum of GaN, and I was surprised that it was as diverse as it was across end markets.
Yeah, just to clarify, the 3x the revenue is last year, the revenue. We expect it to grow this year as well. In terms of diversification, all of our products are diversified. As I mentioned, every one of our products go into all these applications. It's just that some markets have shorter design cycles, like the cell phone market has a shorter design cycle, whereas appliance market, industrial markets have longer design cycles. We have a lot of design activity going on across industrial and appliance markets, but the revenue takes a little bit longer for it to develop. We are not entirely surprised with the diversification, although I must say, I would've thought because GaN is more expensive today, in cases like appliances, like washing machines and dishwashers, nobody would care about size, right? They don't.
What they do care about is efficiency, and they care about lack of heat sink. We can deliver 250 W without a heat sink. Nobody else can do that. That's a apparently a huge advantage because they say the biggest reliability link, weakest link, I should say, is that these heat sinks fall off when, you know, when somebody jerks the machine as they're unloading it or doing transportation, and then it's dead on arrival. You know, it goes into thermal shutdown, and it's gone. They love the fact that they don't have to put a heat sink, which either breaks off or it breaks the board off. A huge problem for them. And that's true even in the motor control.
If you look at the motor control board for a washing machine, the heat sink's about this big. It's really unbelievable. With our BridgeSwitch, we can get rid of it, and that's all because of the efficiency of our products. We are very pleased that it's actually going into appliances and industrial in parallel.
If I can add a little bit to the GaN part of that. One of the things that Mike and his team has been able to do is make the InnoSwitch look very seamless from very low power all the way through to high power. And it transitions from silicon to GaN at depends on the application, but somewhere between 40 W and 75 W. Many of those customers on that list that on the graphic that Balu put up showing the diversity of GaN, it's really a lot of them are looking just more power, but they're looking for more power in the same form factor, using the same algorithms, using the heat sinkless sort of piece of hardware.
They don't wanna have to add a heat sink just 'cause they increased power. As they add wireless capability into a refrigerator, or they wanna add some other communication technology or LED lights or more lights or a screen or something like that, they just need more power. They wanna have a consistent design philosophy, a consistent product line that they're all familiar with, and it just means the GaN is the next highest power unit.
Yes, Tore.
Thank you. I had a follow-up question for Radu. I think you said, you know, the only technology that's not proprietary, and I think you said for now, is silicon carbide. You know, you clearly have some very high voltage products on your roadmap. I was just hoping you could elaborate a little bit on what you meant by that. You know, should we expect, you know, Powi to eventually have its own proprietary silicon carbide technology? Thank you.
Well, as you probably know, we did flirt with silicon carbide ten years ago with a startup called SemiSouth. Then we realized that for our at least next decade or more period, our focus is very well served by gallium nitride. What we're also discovering is that as we learn more, as we develop GaN at higher voltages, as we ship GaN in higher and higher volumes, is that if you recall my overlap chart, there is this boundary in voltage between gallium nitride and silicon carbide is not 100% fixed. Philosophically, we're gonna look at whatever technology makes most sense.
We will most likely want to always own the technology and have total exclusivity, IP, etc. For now, silicon carbide is not developed internally, but you know, there's nothing that would stop us if it will make sense. Okay, one more question before we break. Oh, there's one up there.
Hi. Thanks again for hosting the analyst day. It's been great so far. Quick question. On the $100 million pipeline, can you just speak for a second around how you came up with that number, and what's the sort of timeframe to see that realized as revenue? Thanks.
Great question again. The number we come up with what's called flat rate. In other words, if we are getting designed into a product, and once it goes into full volume production, what would be the yearly flat rate of revenue from that? That's how we calculate that. In terms of, you know, how quickly it'll develop, in some markets it'll be faster. Like notebooks will be a lot faster because design cycles are shorter, and appliances will take longer, and industrial will take a little bit longer, I should say. What is surprising to me is how quickly we got designed into automotive. I'd have thought automotive would take a long time, but the compelling nature of the product has really accelerated that, you know, designing.
It also so happens that the emergency power supply is not a safety component, so it doesn't have to go through all the safety approvals. Whereas when you go into the drive, the gate driver, it takes much longer. It takes three to four years to get designed in. Even there, we are seeing a significant activity in terms of design. Okay, we are gonna take a 10-minute break, and then we'll continue.
Okay. Is everybody reseated comfortably? Okay. I'm Doug Bailey. I'm the Vice President of Marketing. I'm gonna talk to you today about the Served Available Market. I've carved it up in a way that is a little bit non-traditional. Generally, we look at it in terms of consumer, computer, industrial and so on. What I've done here is I've put it into a form how, you know, we think about it and in markets that I believe have some special significance. Either we're growing very fast into those markets or they're markets where we have a very high share and it's worth digging in and discussing them a little bit. You'll notice that SAM is growing. It's quite a spectacular growth. It's about a 16% CAGR, SAM growth.
That's driven partly by the natural growth in the markets, macroeconomics and the general move towards electrification, and it's also being driven by new product innovations. There's some step functions in here that I'll go through, and those steps are when we have a new product that we're introducing. What I'll do is first talk about power everywhere. There's a market. Actually, I should have pointed it out, broad-based power applications at the bottom here. That's, that market talks about just everything that you can plug into the wall. There are multiple small markets, tools, bikes, drone chargers, printers and TVs. I've lumped all of these into this sort of general market because there's a lot of them.
Anything that can plug into the wall, anything that's the other side of the wall providing power to a wall socket could have a Power Integrations product in it. Most of these kinds of applications require high efficiency. They require a high reliability, particularly things like metering systems and industrial systems. The cell phone tower there, once again, very high reliability requirement. Many of these smaller markets tend to be favoring Power Integrations type high reliability products. This market is growing at around 7% CAGR. It's roughly 1/3 of our business right now. It's going to grow at the 7% level, but become less significant over time, and that's because we're adding in some very large markets, automotive.
We're adding in some appliance motor drive business, and that grows at a faster rate than the base 7%. That 7%, just for reference, is about 2x the predicted global GDP growth over time. That's about just a little more than 3%. You know, even 7% has a very, very significant compounding effect. Here's an example. I don't know how many of you guys are lawn mowing, leaf blowing kind of guys, but in California, they're banning the gasoline version of these things. It's not for the classic reason, which would be energy efficiency or carbon. This is purely a smog abatement activity, and it's gonna happen in New York also. There's a similar program happening in New York.
All of these kinds of programs worldwide, where they ban one kind of gasoline or fossil fuel-powered technology or product, brings customers to Power Integrations. The alternative to a gas-powered leaf blower is gonna be a electric-powered leaf blower, and most of those are now battery-powered, and we provide high-quality ICs that charge those batteries highly efficiently. This is a great market for Power Integrations, and it's illustrative of how the world is coming towards us as electrification of everything continues. Mobile phones. Smart mobile phones, tablets, and notebooks are a market that we're predicting will not grow in a substantial way. There are applications within this market where we can still gain share, and the ASPs are going to continue to grow. What I've done here is I've broken that market out.
You'll see the top line number is somewhat flat, but this part of it is right in line with our gallium nitride products. This is something where we can grow revenues within a market that is relatively flat. There are two segments here. There's the yellow segment, and that's the largest segment, and it's got something called power factor correction there. That is an additional chip in every power adapter that requires, you know, in our case, gallium nitride. There's two chips in the 65-W and over power adapters from Power Integrations. One would be the power converter, and then the other one is the power factor corrector. Power factor correction is a mandatory requirement in many jurisdictions to manage power quality on the network, on the grid.
Then there's the green segment. Some jurisdictions don't need it. That's one chip from Power Integrations, typically also gallium nitride. There's the gray segment there is a 15-W to 65-W market. That's the bulk of the market. It's where most mobile phones, particularly the inbox products, reside. Some of the 65 and overs are inbox, but many of them are aftermarket. Serving the people who want a little extra or who have to buy a charger because the vendor doesn't put it in box for them. That segment that I was talking about is growing at 22% compounded, and so that's a great market for us, and we can grow into that very nicely.
The next market I wanna talk about in some detail is the industrial gate driver business. This includes the gate drivers for wind turbines, for grid-scale solar. It also includes mining equipment and big pumps and material handling applications. It's hundreds of kilowatts, and they drive the big white modules, big black modules that contain a lot of IGBTs and more and more increasingly silicon carbide, and we have some of those in the demo room, so when you go have a look at those, you'll be able to pick them up, but those are heavy pieces of equipment. That's growing also at about a 7% rate compounded. The 7% is gonna double our SAM over the next five or six years.
This is how that SAM breaks down. What I've done here is I've grayed out a section that is what we're calling the captive market, and that is our assessment of OEMs who are intractable makers, not buyers. Part of our goal is to introduce some new products that will convert those guys from makers to buyers. That expands our SAM by a step function. The first small step next year, we're introducing a new board-level product for gate drivers, and that'll expand the SAM a little bit. In 2025, we have a whole new platform technology coming.
The reason that that's important is that as time passes, there's more requirement for telemetry, and that means information coming back from the module to the control system, and there's more a requirement for using silicon carbide, which is very tricky to drive. This new platform will, we hope, entice these the makers to become buyers, and so we're adding them into our SAM at that point. When the IoT buzzword first hit in about 10 years ago, we took a look at what does IoT mean for Power Integrations? What we decided was there's two kinds of IoT. There's sensors only, and then there are sensors plus actuators. We put our focus on anything that does actuation, heat, light, motion.
That's because there's additional power required beyond just the radio, because we felt that a lot of the radio technologies were such low power technologies that they really weren't gonna be very interesting for PI products. We focused on actuation. These are some examples of that. You've got speakers, obviously that's mechanical motion. Then you've got something with a screen on it that's providing that requires power over and above the communications technology. The wall controls, that's the things like vacancy occupancy sensors, anything that switches the lights on and off in a home. There's some motorized products here. There's one that somebody would be very familiar with, which is, you know, opening and closing blinds and shutters.
There's another one which is a recloser, and that's a strange application, but what it is, in an apartment complex, if you manage to trip a fuse, you don't want the apartment owner fumbling around in the dark trying to turn the fuse box on in the basement. This is a mechanism that automatically can reset the fuses in an apartment. Motorized products are IoT. There are some smart sockets, and I'll talk a little bit more about those in my next slide. Finally, protection security. These arc fault detection, ground fault detection, but also things like smoke alarms. The traditional way to build a smoke or CO2 alarm was to put a battery powered unit.
Increasingly people are, you know, it starts to bleep 'cause the battery runs down, you throw the battery out, you stick the thing back on the wall and forget it without a new battery in it. The various jurisdictions don't like that, so they're asking people to wire those in now, and that's all an opportunity for Power Integrations. As standby is particularly important there because it's plug in all the time. It's sucking power under all circumstances, so having a very good and high efficiency standby power supply is very, very important. This is a market that's a growing market. If you look around your benches, there are power sockets there are power sockets on the walls. The building code in California says every 10 feet in a home you gotta have a power socket.
If you can move from it being a replacement for a charger at 60 W to being something that's 240 W, which is very possible now using the USB PD Extended Power Range. When you make that transition, almost everything in your home that isn't a major appliance can be powered from one of these wall sockets. Safe, you know, you don't have to worry about, you know, your two-year-old with a fork hurting themselves because this is all low power and starts out at 5 V before it negotiates up to 48 V. This application is going to grow. You heard from Leviton on one of our videos. We believe that that's a great application. If you think about adapters, you won't need adapters anymore.
You'll just be able to plug your cable into the wall. Every one of these wall sockets has to have a high efficiency, low temperature gallium nitride device just to be able to produce 240 W. They've gotta be everywhere. It's a huge market that requires a very high level of technology, which we bring to the market right now. The market itself is actually relatively small right now. Having spoken about it is a bit skinny, but it's growing at about 14%. It's double the growth rate of the other markets that we're looking at. It's very useful market for us. The designers happen.
For many of the designers happen to be close by with the usual suspects in Silicon Valley. Data centers, that's another expanding market. We already have a substantial share in data center. We have gallium nitride products in data center right now, but it's in the standby power. The standby power for data centers is a relatively small market. What we're anticipating is that the deployment of gallium nitride into data center with some new products that we're working on. This is what happens to the SAM in this market. The standby power is relatively small right now in terms of the overall market size. We have a good share, but of a small market.
As the GaN products become available in 2024, we gain a very large SAM, something around $500 million SAM, opportunity in the data center, telecom, datacom market space. Okay, refrigerators and appliances. This is an example of how government regulation has helped people save money. If you look at this, back in the 1990s, your standard refrigerator was allowed to consume over 1,000 kWh a year. That's 1 MWh, right? That's a lot of energy being used just to keep food cold. Government regulation has driven the allowed energy usage down and also market forces. From having something, you know, an AC motor and maybe electromechanical systems in it, no interest for Power Integrations.
The products moved in the mid-2000s, and this includes things like, you know, dishwashers and air conditioning systems. There's a whole bunch of appliances. I'm illustrating it with a refrigerator, but many other appliances have similar energy usage constraints. Now we could put a TOPSwitch in there to provide power for the electronics that were doing the controls for the refrigerator. In a more modern fridge, you get a lot more power requirement. You get things like wireless, you're going to get screens, you've got a lot of additional power requirement, plus you're using a brushless DC motor, a high efficiency brushless DC motor to power the compressors and the fans in the fridge.
This is a major step forward, and it provides us with an opportunity of about $5 per refrigerator. One of the pieces of learning out of this is, I don't know how many of you remodeled your kitchen and put your fridge in your garage and filled it full of beer. Crush that beer fridge. It's costing you around $400 a year. If you buy a fridge for $700, you can, you know, year and a half later, you got free beer for the rest of your life. Plus, we get $5. Okay, $5 for me, free beer for the rest of your life for you. Okay. Brushless DC motors, they are proliferating everywhere.
For pumps, fans, compressors in all kinds of appliances, up to around about 1.5 kW. We're going to be able to support with some products coming soon. The current BridgeSwitch supports up to about 400 W. So we can do everything except the drum motors in washing machines at the moment, and things like pool pumps, which require a little bit more power. Compressor motors for air conditioning systems is something that's also gonna add on as time passes. Air conditioning systems are a fantastic opportunity for us. You require a high power brushless DC motor for the compressor, and then you got a fan on the inside unit and a fan on the outside unit.
That's a total of three potential brushless DC motors for just one air conditioning system. This is what happens to the SAM in the appliance space. We have the power supply portion, which is the darker purple, that's growing at around about 7% CAGR. That's our favorite 7% CAGR again. The brushless DC motor segment is growing very fast, and that's as we add some new products, but also as the sector grows. There's a new product coming out next year that increases the power level and that adds some SAM. Then in 2025, we have some gallium nitride motor drivers coming, and that's going to provide another step up.
Overall, that market grows around about 33% compounded. Let's talk about cars. EVs, you know, everybody's, you know, aware of how well that market's been growing lately. We have around about a $66 opportunity per vehicle right now. It's not only in the inverter. You know, you might think, you know, you've got a battery, you've got the motor. Battery, motor, that's what makes the EV go. There are many, many, many other applications that wanna be attached to the high voltage rail. Once you've got high voltage available to you, your air conditioning can run off high voltage. You need a charger. You've got to have bus converters that convert between the low voltage and the high voltage. You've got potentially active chassis components or power steering components.
There are many components there that need high voltage, and all of those have not only the power processing part, but the command and control and telemetry also requires a power supply because you've got to You know, the user of the vehicle has to be able to turn these things on and off and control them and communicate over CAN bus and what have you. That means there's a lot of little high voltage power supplies scattered around inside every vehicle. At the moment, around $66 a vehicle, using the LinkSwitch, InnoSwitch, some diodes, the Qspeed diodes and the SCALE-iDriver parts. As we add some GaN products over the next couple of years, we're going to almost triple that opportunity per vehicle.
Those products being used in the actual power conversion stages in the car. We're not limited to passenger vehicles. We've put a lot of work into the heavy vehicle market. It turns out buses, trucks, construction, agricultural equipment, airport tugs, scissor lifts. There are a whole variety of applications, even aircraft moving to being electric or having some electrical components related to fuel cells. In those applications, we can get up to about $365 a vehicle right now. That's partly the hotel loads that are in there. Those are the loads that are there to keep the passengers comfortable, but also the actual main drive inverter.
The SCALE EV board, which we launched a little earlier this year, is a major contributor to the high value per vehicle in that market. As we add the GaN products, we're anticipating that's gonna go up to almost $500 a vehicle. It's a much lower quantity market, but highly diverse and very much in line with requiring high efficiency and also high reliability products. It's very in line with Power Integrations' product philosophy. The overall market, as it develops over time, you'll see a couple of step functions in there, is growing at about a 70% CAGR for us. That's obviously a major focus.
As Yang Chiah Yee mentioned, we have dedicated sales and support activity into the major automotive manufacturers and tier ones develop products their systems using our products. Overall, getting back to the original SAM slide, we believe we have no practical limit on our SAM growth. There's no cap here. The world is coming our way with the decarbonization, electrification impetus as society moves away from using fossil fuels both for the generation and then also for the use of energy on the consumption side. No practical limits. I'm gonna play a video now from one of our customers who is in the high power space. They make locomotives or locomotive power systems.
They make wind turbines. They do grid-scale solar. Let's hear it from WEG.
WEG is a global supplier of electrical machines and automation for industry, infrastructure, and energy systems. WEG initially start producing electric motors in the year of 1961. WEG nowadays is involved with energy. It means energy generation, transmission, distribution, also conversion and industrial automation, traction, renewable energy like wind converters, solar converters. Power Integrations can help us a lot in the development of proper converters for this industry. The company is doing all efforts to improve processes that reduce environmental impacts, such as the conservation of natural resource and the management of waste and emission generation. During the design phase, the engineers of Power Integrations are always available to help us for design review, so this is very helpful. WEG is also involved in battery energy storage, and certainly Power Integrations can help us incorporate this field as well.
We have been working with Power Integrations for the last decade with great support, great products, and we have a lot of challenge for the future, and we are confident that we will still continue this cooperation.
Good morning. I'm Sandeep Nayyar, CFO of Power Integrations. Before I share with you the financial perspective, I want to summarize what you have heard today, and hopefully it reinforces what you've been hearing me say for the last 12 years: focus, execution, and long-term thinking. Balu talked about the unique requirements of high voltage and the requirement for focus. As Doug showed, we really are not limited by this focus because of the SAM expansion that we will continue to see in the future. Execution. Focus makes for better execution, but you need talent, and you need a culture of accountability. I think today you have heard and met some of my peers, and that should give you an idea of the team we have built and the culture that we have developed at Powi.
I think the third attribute of long-term thinking is really what differentiates us from all the others. Radu talked about our manufacturing strategy, and you can see as a small company, you know, we could have chosen the lowest cost, but what we chose was strategic partners where we could get capacity and our IP was protected. It is not limited to just manufacturing strategy. You heard a lot about GaN. It took us over a decade to reach here. For a small company, spending that kind of money impacted operating margin, but today it is the cornerstone of our technology. Another example Balu gave you was the development of InnoSwitch, five years in the making, FluxLink technology, very innovative. The chart here shows you the semi cycle that I've talked about, and even here, it reflects our long-term thinking.
In every downturn, we have actually added to our inventory levels. In fact, the last pandemic, we went up as high as 178 days versus a normal model of 120-1 25 days. This enables us to utilize the capacity, keep our fab partners economically viable, and when demand swings, which it did, we were able to actually satisfy the needs of real demand by our customers, which enabled us to get significant market share. Now, why are we able to do this? We have very unique products, but these products, as Balu showed you in the chart, last very long time. There is no obsolescence, and also the same products can be used in different end applications as well as different markets. Another example of our long-term thinking is our capital allocation. This chart reflects the share count timing.
Many companies keep buying shares on a regular basis in up cycles, higher valuations. What this chart reflects for you is the opportunistic way we use for capital allocations. The perfect example is in 2008, during the financial crisis, we bought back nearly 15% of our company. Nearly five to six months ago, when we got promoted to the midcap index, we had 4 million shares that were dislocated, and again, we used that opportunity. This chart really shows you that every time there's a downturn, we saw the opportunity to buy back shares and return value to our shareholders. Now, why are we able to do this? It's, again, because we believe in our long-term model, and I'll show you later why we are able to walk our talk with our model. Back to the keys of success. What does it all add up to?
It gives you a greater appreciation of the defensibility of our franchise. We have IP in process technology. We have IP in packaging, in design, and Balu talked about the knowledge and the system-level knowledge that we have, which has been put together for the last 25 years. You've heard a lot from Doug about the size of the opportunity and the ability to grow the SAM in the high voltage in a very focused way. The other thing is the stability of the workforce. Take the leadership team that you're meeting here today. Majority of us have been here for over a decade. People are everything. High voltage requires special expertise, and it's these people who have enabled us to generate this moat. Now, it's the long-term thinking what differentiates us even here. Let's take the pandemic.
When it struck, most of the companies started laying off people, cutting down their production, cutting salaries. Powi did just the opposite. We did selected hiring, kept our inventory levels high, gave normal raises, and in fact, in the U.S., increased the contribution to the 401(k). It's just how we think differently. What I believe that these characteristics add up to help us with the financial performance that we have delivered over the years. Balu had talked about and showed, we have grown at 2x the analog semiconductor industry for a very long time. The margins that we have gotten in this business are the highest, as you can see, the discretes, which is 50%+, and the discretes make far lower margins.
What this has enabled us to do is to have consistent cash generation, which has enabled us to return value to our shareholders. This chart clearly shows you the return that we have given from 2008 to now on free cash flows of 15%. If you look at 2021, we gave about 26%. If I look ahead, I feel very comfortable that we can give the 25%-26% free cash flow generation in the years to come. I had earlier talked to you about the buybacks, but I don't know how many of you know, since 2008, we have bought back more than $800 million of our own shares, and the average price is $27, which probably gives you, looking where the share price is today, the return we have garnered through this capital allocation.
The other thing that you can see that we have a dividend that we have instituted, and obviously you can see, I'll show you in the chart right now how that we have grown that over a period of time. We instituted the dividend again during the financial crisis of 2008, and have consistently increased it over a period of time. Because we believe it is an important way of returning value. In fact, it is an important part of the four-prong approach that we have, where we invest internally, we do selective buybacks, and selective M&A, and dividend being the fourth.
Even though we have not a formulaic approach to dividend, if you look at the past history, we've been giving about 25%, 20%, 24%-25% of our free cash flow to our investors as dividend. What are the drivers of our financial model? As you can see, we have made substantial gains in 2021 because we capitalized on the product availability and capacity to accelerate these gains. We were probably one of the few companies who was able to provide to our customers product that they required when they need it for the real demand. I think this also gave the customers an idea of our integrated solution, that they needed lesser components. The share gains that we got here are gonna really serve us well in the future when this whole cycle turns around.
Additionally, Doug and Balu showed we will deliver or double our SAM by 2027. I think many of you have heard me say that 2025 to 2035 will be the decade for power in EVs. This chart in the middle shows you our ability to gain share. If you look at the long-served markets, because you know our markets, all of them don't move that quickly. Over a period of time, if you look the long-served markets, whether it be the appliance or the consumer, we've been able to get 20%-25% market share. You saw from Doug and Balu the new areas of SAM. What that gives you at least an idea of the potential we can get in these new markets to grow our business in the future.
You've seen here on the right-hand side is another chart which shows the ASP trend. Our ASP has gone up from about the lowf 20s to the 40s. We just got a question earlier about this from Tore, and I'll address this question. The reason we have seen this ASP growth is integration, high power, GaN, and content increase. I think Doug showed you a really good example of the refrigerator, where a content increase has gone up 4x-5 x. To answer Tore's question, I'll take another example. In the communication segment, you know, our share used to be teens about four to five years ago, and it has grown to over 30%. If you look at the average ASP in the communication segment, it has grown in the same direction as this chart.
That should give you an idea of share and ASP that have gone in the direction of how we have grown by content and by. In the communication segment, what has happened is basically it's the power levels have gone up. Now we are also getting, you know, multi-port devices. All that has increased our content. For the further drive, talking about the drivers, I think Balu and Doug talked about how our revolutionary products are providing opportunities in new verticals. Doug talked about IoT, EVs, and the USB plugs as an example. I think Radu gave a really good example of the fabless IDM model of how that is benefiting us. Because in this case, we control the process technology as well as the high voltage switch technology.
What that is enabling us to get is some of the margin benefits of vertical integration without the CapEx burden and also the margin volatility that comes with owning the fab. I think Balu gave you another great example of how the other analog companies require hundreds of products, whereas we require few products that go over multiple applications and markets. What that does is give us tremendous amount of leverage in R&D. Another, again, supporting the model that we have. The other thing that impacts us quite a bit is the mix. Our margins are the highest in the industrial segment, followed by consumer because of appliances, followed by computer, followed by communication. If you see how we have it in this chart that I'm showing here, from 2002 to 2027, how we have changed our mix.
Believe me, this was intentional. We really wanted to diversify our business. In 2002, 43% of our revenue came from communication, which was primarily cell phone. If you see where we are projecting 2027 to be, it'll be more than 70% of that will go in industrial auto and consumer. Consumer is now roughly 90% appliances, which is more industrial-like. To summarize the model, in the three to five years, we believe we will continue to grow low double-digit, and we have confidence of that with all the GaN expansion, the share gains that we have gotten. The non-GAAP gross margin model is 50%-55%, even though we are running above that. The reason we are running above that is mainly because of the pricing environment that we have.
You know, we do value pricing, and when the discretes prices increase, we capture the same value. Also, the cost increases have not yet kicked in, which will kick in, but we are also starting to get some yen benefit. I'll show you in the next few charts. The non-GAAP operating margin will be about 25%-30%, and our capital expenditures will be about 5%-7% of sales. I've been asked this question many times. "Well, you're running your operating margin above this, so why don't you change your model?" I just wanted to remind everyone that not long ago, our model was 20% with an aspiration to get to 25%. Not too long ago, we changed our model to 25%-30%.
While we believe, and I'll show you in the charts a little later, that because of the mix and because of the share gains, we will be at the higher end of the model. Being Powi, we like to walk our talk. We will reassess our model down the road. I'm pretty confident that in the near future we still will be on the higher end of this model. Now, this is one of the reasons why I feel comfortable about the model. If you go back over a period of time, you will see that our margin is between 50%-55%. Yes, we have some aberrations. In 2011 and 2017, yen went in a different direction. It was a massive headwind. In 2023, it'll become a tailwind.
I talked a little about how mix can also impact us. The same thing happened to us when in 2011, communication went down to, like, 10% of our revenue. You can see what happens to it when it goes to 35%, which happened in Q1 last year. All these things you know, have what I call gyrations. If you take over a period of time, our model holds very well in the 50%-55%. Now, I believe in the near future it'll be on the higher end because of the mix, because of the pricing environment benefit, but also yen. I think I have provided the guidance for next year, or, you know, to be slightly above the higher end of the guide, of our range.
This is another chart which reflects the revenue and spend to give you an idea of the consistency of our model. I've always talked about, we talk in terms not in quarters but in years. I've always talked about that our model is to be viewed in a three-to-five-year timeframe. Because we don't try to manage years, we try to manage our business, which we run on a multi-year dimension. Here you can see from 2007 why I feel comfortable about the low double-digit revenue growth rate. If you see, if you can grow low double, we can see that our expense rate of growth is only 6%, which is very much the 60% of the revenue growth rate. You can see here that there is leverage in the model. To conclude, this is truly a unique company in many, many ways.
Being in high voltage, we address different challenges than our industry peers. Our markets are different. We have a small number of products that last an exceptionally long time. These differences require a different approach and a different way of thinking. I hope the discussion today has given you a better idea of how we have addressed those differences and have been successful. As we compete with the much larger companies, which Balu talked about, with very large R&D budgets. I hope the discussion of the market opportunities and the new products and technologies we are using to address them has given you a sense. As you've heard me many times say, I am more excited about our future today than I have ever been in my 12 years at Powi.
Now I'll turn it back over to Balu, who's gonna conclude with comments about contribution to low carbon future, a very important element of our story. Thank you very much.
Okay. Well, you heard a lot about energy saving aspects of our products. Saving energy has become a mission of the company ever since I first read a paper from a scientist in Lawrence Livermore National Laboratory. That paper basically said that a huge amount of energy is wasted in standby, billions of dollars of energy every year. We were working on a new product called TinySwitch. I immediately knew that we can solve this problem. In fact, I actually drove up all the way to Berkeley to meet with this professor, and I showed him what we can do. The problem with power supplies is they only focused on efficiency at full load, or when the product is fully operational.
Nobody cared about what happens when there is no operation or it's idle. As Mike showed you, the efficiency in the standby in classical power supplies is very low. It just falls off the cliff. All it took was a smart way to control the switch. That's how this EcoSmart came along. In 1998, we introduced TinySwitch, which had the EcoSmart technology, and since then, every new product development, we focus on energy efficiency. Not just at full load. You saw the full load efficiency is very high, which is important, but it's even more important that you have high efficiency at light loads. In fact, our latest product, which is InnoSwitch, has a far more sophisticated EcoSmart technology that actually increases the efficiency in standby.
We don't realize this, a lot of the products, even like your computers, they're not operating at full power unless you're doing some gaming or some video. Most of the time, they're operating at 20%, 30% power. It's really important that you're efficient at those power levels. That's also true for automotive, automobiles, by the way. Now it has really become part of our culture, and it's a great tool for attracting talent. When we talk to people during interviews, we emphasize this aspect of it. Our employees are very proud of it, being part of this mission. Because energy saving and reducing carbon emissions is integral to our culture, we look at the entire spectrum of ways we can contribute to low carbon future, including not just our products, our supply chain, and our internal operations.
Now let's take a look at our supply chain. One of the byproducts of our fabless IDM model is that we source our wafers from Japanese and German partners. These partners already have very strict environmental standards. On top of that, they have a culture of environmental stewardship, so it kind of gives us an advantage right away. As you can see, all of our wafer partners or fab partners are Sony Green certified. I mean, certified, I should say. They have meaningful or actually robust plans to reduce carbon emissions. Same is true with our backend suppliers. They're also Sony Green certified. I won't go into a lot of details. There are a lot of different certifications they pass, and they also have plans to reduce greenhouse gas emissions.
We as a company are committed to using solar at all of our major locations where we own the buildings. In 2010, we installed a 600 kW solar array. At that time, that was one of the largest solar arrays in Silicon Valley. Now we are just about to break ground to triple that capacity. The first installation took one of our buildings almost entirely off the grid, and now we're gonna triple that. We have three buildings, by the way. Similarly, our Switzerland facility, which we recently opened, and German facility, which we opened about 2 years ago, they get significant amount of their power from solar, from the rooftop solar.
Now, reducing internal carbon footprint is important, but in our case, it pales in comparison to amount of impact our products have on the planet. This graph shows an estimated savings from EcoSmart over the last 20 years. If you just take 2021, we saved through our products 15 TWh. Now, what does that mean? It's equivalent to roughly 2 million households, you know, the power consumption of 2 million households through the entire year. If you translate it to CO₂ emissions saved, it's greater than 6.5 million tons of CO₂ emissions that we prevent. You know, we talked a lot about EcoSmart technology. We didn't stop there. We introduced a product called LinkSwitch after TinySwitch. That was targeted to cost effectively replace these iron and copper transformers, which are unbelievably inefficient. I mean, they're so inefficient it's actually shocking.
If you touch those transformers, even when they are not loaded, they'll be warm. That's because there is a copper winding that's permanently connected to the grid, and it's burning power all the time. Later, we came up with a CAPZero product, which eliminates the bleed resistor losses. What are these bleed resistors? These are resistors that are required by safety, so when you unplug something and you accidentally touch the terminals, you don't get a shock. You have to discharge all the capacitors with a resistor. Unfortunately, it always draws power. With CAPZero, we are able to do that with no power loss. It only discharges when you unplug it, otherwise it's open circuit. Similarly, we have a product called SENZero that eliminates the losses associated with bias resistors, and I won't go into a lot of detail about that, and so on.
We have LinkZero and InnoSwitch, and now of course, we have GaN technology. We are the first ones to bring GaN technology to the power supply market, and that will drive a massive energy savings for decades to come. Now I'm gonna show this again. I know you saw this with Mike Matthews' presentation. You can see here that our PowiGaN is not only very efficient at full load, more efficient than any of our competitors' GaN because it's very specifically designed for what we do. When the load is reduced, like when you're not watching a TV or not washing you know, using your washing machine, the efficiency actually increases, and that's actually nothing to do with GaN. Actually, it doesn't matter what switch you use. The switch makes a big difference, you know, at full load.
It is all about how you control the switch that makes a difference in standby. That's possible because we have very advanced control system in InnoSwitch, and that control system is made possible by our FluxLink, because it requires extremely high speed link between the low voltage side and the high voltage side. It's our system level approach that allows us to do that. Look at the discrete solution. You can see that it's just dropping off. Yes, great to have GaN at full load. Doesn't help you at standby. There's another way we contribute, which is not at all obvious, and that is the influence we have on the rest of the market. Now, we can tell you all you want, but you really have to talk to the experts in this field, and they will tell you what our technology does.
They come to us constantly to figure out what the technology can do to reduce energy consumption. I can give you a couple of examples. You know, when California decided that they wanna eliminate these transformers because they were burning so much power, their first concern was, oh, they have to actually give some money, some incentive to the manufacturers because they thought, you know, technology will be too expensive to replace them. When they came and visited us, and we showed them our TinySwitch and LinkSwitch product, they were convinced all the industry required was a nudge because we already had cost-effective solutions to replace these transformers. Another example, when we first introduced EcoSmart, I actually got an invitation by the President of the United States. I went to Washington, D.C., gave him a demonstration of the technology.
The reason he was interested is because there was an energy crisis at that time, and all it took was 5% increase in capacity would solve the problem. The standby waste was more than 10% across the United States. We could completely eliminate the standby waste, and the whole crisis will go away. He was so impressed, he literally immediately signed an executive order to limit the amount of standby that is allowed for products purchased by the government. Unfortunately, he can't do it for the whole country. He has executive orders for products developed by. Of course, it followed on with CEC regulations and EU regulations on the same thing. Essentially it became a standard worldwide. There is yet another way we contribute to reduction of carbon emissions, and this is also indirect.
Now, Mike talked about the number of components we replace. Each one of our products replaces dozens of discrete components. These components obviously are, you know, consume energy. They take energy to build. They are e-waste, so we reduce e-waste, we reduce raw materials, and so we have an indirect benefit to the environment because of reduction of components. Finally, we not only save energy, we make energy-saving technologies much better. They work better. Take, for example, our LED lighting technology. The weakest link in LED lighting is not the LED, it's actually the driver. The driver fails before the LED does, and we make those drivers very reliable because of the level of integration we bring. Take for example, renewables and traction and so on.
We provide gate drivers that not only make them efficient, but they dramatically increase the reliability of these products. Take brushless DC motors, very efficient motor. Why would you connect it with a driver that's not as efficient? Our drivers are the most efficient drivers for brushless DC motors. More than 98% efficiency. Basically we amplify the benefit of brushless DC motors with our bridge switch solution. Of course, electric vehicles are by far the most exciting transformation, and high voltage semiconductors have a huge role to play to our strengths of efficiency and reliability. Okay, we have finally come to the end of our presentations. I hope you gained greater appreciation of the keys to success, of what differentiates us from other companies, including competitors, and also other semiconductor investment opportunities.
We are possibly the most focused company in the industry, and we have chosen to do so because high voltage is very unique, very different from low voltage technology. Revolutionary innovation is not possible without focus. Our products take a long time, and they last a long time. The system-level approach has allowed us to compete successfully against big analog companies and small companies. We also have successfully competed against discrete components, which have much lower margins. Now, we talked about secular trends driving high voltage semiconductors like IoT, electrification, renewables, advanced charging, automotive, and so on. We are perfectly positioned with our technology like FluxLink, like GaN, to take advantage of these, opportunities. You know, Mike showed that the gate driver, it not only uses our package, it actually uses FluxLink to provide isolated communication between the switch, the module, and the microcontroller.
We showed you a strong record of growth. We talked about stability of our business model, our commitment to returning cash in both consistent ways and opportunistic ways. Consistent being the dividends and opportunistic being buybacks. I hope you got an appreciation for how our products make a difference to the environment as well. Energy savings is in our DNA, but in our case, it also meets the market needs. For many companies, there is a friction between the two, and they're going hand in hand for us. With that, I'll stop and take questions now. By the way, you people who are online, you're welcome to ask questions and here we, you know, we will answer those as well. Actually one more thing. There is one more video and after that we'll ask questions.
There is a video from one of our customers, actually early adopters of GaN, one of our largest GaN customers. Right after that we will take questions. Sorry.
PI is one of the most trusted partners of Anker charger products. Thanks to PI's leading AC/DC technology and first-class delivery capability, the shipment of Anker Nano and Nano products has exceeded 10 million units, achieving a new milestone of any single product sales in the accessories market. In our exclusive Gen 2 cooperation, we together impress the world again. I hope that both sides will continue to maintain in-depth cooperation and move forward hand in hand towards the next more brilliant decade. Thank you.
Okay.
Hi, guys. Christopher Rolland, Susquehanna. Thanks for doing this day. Really appreciate. So Sandeep, all my questions are gonna be to you. So that's great. Revenue unchanged. You know, profitability on the gross margin side and the op margin side, I think people were expecting a little bit more here. On the gross margin side, you talked about yen, you also talked about pricing. Maybe you can break that apart for us. Then lastly, on the OpEx side of things, I think you reiterated a 60% flow-through of OpEx expenses. If we still have that revenue growth, why don't we have some operating leverage and an increased op margin range as well?
Yeah. As I mentioned, right now, the op margin is higher because of the gross margin. The pricing environment has been quite favorable, as you know. You know, normally you have price declines and to offset them in manufacturing efficiencies. We have had a few years where the discretes and the availability has caused the prices to be at a different. We do value pricing, so we have benefited from that. Our cost increases have not yet flowed in, and I've talked about that. That'll flow in next year and will have an impact on our margin. Mix is another thing. In the short term, I do expect next year a rebound in cell phone and in the notebooks compared to where it is.
When I talk about the margin, the model is in the three to five years. now, as I also mentioned, we have just recently upped that model, and we like, as I said, to see this model through a little bit. Are there opportunities for this model to go up? The answer is yes. I think when you look at 2025-2035, with the opportunities in EVs, can the growth rate accelerate? Very possible. But I think the way we'd like to do this is we look at our business in the three-to-five-year horizon and model it that way. At its best, at this time, the best I can estimate in the near term will be on the higher end of the model.
Let me add to that. You asked the question about leverage. If we can maintain our expenses at 60% of the revenue growth, there is leverage. The reason the expenses will grow is because as you get to be a bigger company, you have to feed the beast with more products. In growing, you know, 20% or 10% at $1 billion is different from growing the same amount at a much lower revenue. I think it's a very reasonable, you know, expense, given, you know, how much innovation that we have to bring to the market to stay far away from competition, competitors.
Yes. Tore Svanberg. Thank you again for this very informative analyst day. Maybe a question for all of you. I mean, feel free to answer, but you know, the two to three years, the supply chain has been highly disruptive, and I know you talked a little bit about it, but you know, the benefits of integration, of course, is that you know, your customers are sourcing from much less suppliers. And I assume that you know, your customers going forward are gonna be thinking increasingly in that regard. So you know, why source from 100 companies, you know, when you can source from maybe 50? So I don't know who's gonna answer the question.
You know, maybe you have anecdotes, or you know, anything like that, but theoretically, that should allow you to gain more share in the future than what you have even done historically.
Let me take that question. You know, it's really surprised me, especially through the two to three years, how much share a customer is willing to give us, even though we are a single source product. Our product can't be replaced with somebody else's. Nobody else does what we do. I think they've gotten very comfortable. I mean, whether it's an appliance customer who has been with us for 25 years and had no issues with the supply, and they're willing to give us very high share of their business. I think the whole idea of multiple sourcing is getting less and less as long as you're able to supply. That's not true with the other companies or other competitors, because they've had serious issues with supply. People are taking a serious look at that.
The other thing that is done is the OEMs are now much more directly involved with us. Usually, we used to sell products to power supply manufacturers, and we were kind of completely shielded from the OEMs. Thanks to the supply issue, they wanna actually understand what our business model is, how they can get comfortable with having that higher share with one company. I think it's brought us closer to the OEMs, and I think this is gonna be a very significant benefit going forward.
You're trying to say, "Why aren't you growing more than the lower levels?" Is it? I think that's where you are. That's where your question is. I'll address it straightforward. If you look at the SAM, the SAM is expanding, but in communication, it's not gonna expand, and it's still a big piece of our business. There is a portion of that where the growth rate is not as the same, as high as others. We are also realistic that the macro conditions are changing and the normalization of the pricing environment is gonna happen at some point of time. I think in the near term, I think that should answer why. I think a little longer out, I think your point is very valid with the opportunity in EVs, could that accelerate? Absolutely.
I think as you know in our business, it takes a little while to penetrate, you know, things like cellphone and notebooks, it accelerates. The other things, I mean, take GaN, for instance. It's penetrating into other areas, but it's taking a little longer. I think the reason why you are hearing the answer you're hearing is, yes, there is potential, but in a little longer run.
Let me just add, in the same spirit of us taking on responsibility away from the customer and assuming the pain, so to speak, goes a long way. That may, you know, give you some favorable food for thought in terms of the model. We never disappointed, ever, supplying to a customer. Even though we're, as Balu pointed out, they're all single source with us, we don't have any package that's pin-to-pin compatible. We never disappointed, and that's because of our fabless IDM business model. We diversify our supply chain so that the customer doesn't have to.
Hi, guys. Again, thanks for the great presentation. Ross Seymore from Deutsche Bank. Sandeep, I wanna go back to the gross margin side of things again. I realize where we are today and why that can be somewhat cyclically elevated across the entire industry. If I take that three- to five-year view you're talking about and the 50%-55% you've had for a long time, I'm still a little puzzled why it wouldn't be higher. I appreciate you guys are a conservative bunch, and so maybe that's the answer.
If the beginning of this period, next year, you're gonna be near the high end, maybe even a little bit above it, you've said, and say auto and industrial is gonna go from 28%-38% of sales, I think you have in your slides, and you've given us the mix, is there something changing within the segments where GaN penetration is at a lower gross margin, some assumption you make on competitive intensity rising even though it seems like you're pulling further away from your competition? Is there some other assumption other than conservatism that would stop the gross margin from being higher over the next five years than it's been in the last five?
Firstly, I think, you know, the cost increases that we talked about are clearly not yet reflected. Also we are benefiting what we have guided to next year from yen. You know, yen, if you go back historically, has been around 115. If you look at where it is and when the inventory turns, the benefit we get, yes, it does offset our cost, but there's also pricing normalization that I've talked about. In the short term, I believe you're gonna have growth back, not only in communication, but in notebooks, which is another area, if you look in the hierarchy of gross margin, is on the lower end. Longer term, as you're talking about, with the mix shifting to that, you know, we believe we will be on the higher end of the model.
If you look historically, we used to average about 52.5%-53%. Now what I'm talking is as the mix shifts over a period of time, we'll be on the higher end of the model.
Hey, guys. Thanks. I guess it's third time the charm on gross margin. My team was pinging me some stuff about the business, and like, the last three years have been unusual, right? You've had the post-COVID rebound, you had 2020 was COVID, 2019 was kind of an inventory correction. I mean, your gross margin is sort of right in the middle of the range you're presenting us today back to 2018 with $400-odd million in revenue. If you go out to 2027, I'm just, I guess to Ross's question and Chris's as well, just trying to figure out with 2.5 x the revenue if we go out to 2027, like how we're still in the same range we were at in 2018 and the mix is probably getting better.
I don't know, Sandeep, maybe you could try to quantify for us, like what margin impact you're assuming from yen, what you're assuming from cost increases, 'cause that's the. We're all trying to turn a-
Yes.
Round peg into a square hole here. Thanks.
Yeah, that's a pretty fair question. If you look at it, every 10% change in yen impacts 120-130 basis points. The yen is in the 130s. I mean, obviously it just went up higher, but let's assume for average it was in the 130s, and it reverts back to 115. You got a couple of points of margin erosion right there from yen. As we talked about, there are inflationary pressures still. Yes, fab capacity, I know Ross asked me this question that, you know, what'll happen. Yes, fab capacity may be freeing up, but the question is what happens with the inflation and what that has impact on the cost. Now, on the opposite side, that if the inflation aside also helps on the pricing because you're not gonna give price reduction.
I think the yen impact is not less if you normalize it back. Yes, we have factored in when I'm talking to you that yen will get back to the normal of 115.
Let me also add that in the last two years have been unusual in the sense that we haven't had to give the yearly price reductions. Once the supply chain issues get better, there will be pricing pressure, there is no question, across semiconductors. We have to account for that. There will be, you know, price pressure on in most areas. Hasn't happened yet, but it's coming, I think.
The inflationary pressure is real.
Right.
We start seeing it across the food chain.
Hi, Kevin Cassidy from Rosenblatt Securities. Just a question on your manufacturing for GaN. If you're at 150 mm wafers now and going to 200, is that a cost reduction? You know, maybe this is another way of asking about gross margin, but
Persistence pays. We don't think so. The reason is, the majority of the cost in the GaN technology is in the epitaxial growth, which does not benefit from going to larger diameter wafers. The cost per square inch is about the same, regardless of whether you do it on 6-inch or 8-inch or whatever. The one benefit that going to larger diameter wafers will bring is the ability to use 8-inch fabs. It's primarily a question of accessing capacity as opposed to cost reduction.
Sure.
The answer is, it has no impact on gross margin.
Just to be clear.
I'm getting a lot of help from my friends here.
I'm usually the, you know, the optimistic one.
Hi. Thanks. Jeff Walkenhorst from Copeland Capital Management. I'm not gonna ask about gross margin. Something that I think I'm surprised hasn't come up yet, but the stock was hit the other week significantly with NVIDIA and this idea that they need a license and certain IP for China. You've talked a lot about your IDM model and your approach and how you build a product, but your assembly test package does happen in China, Malaysia, Philippines, and maybe one other location. The first question is, how much of those things happen in those countries? If you could kind of give a size of that for us. How much risk is there to Powi ? And also, how do you square that with trying to keep your IP protected, and not have any leakage in those other markets, especially China? Thank you.
You have the right person to answer that question. Yes, Sunny.
Thanks for the question. In terms of breakdown, I think Radu shared, we don't have anything single sourced. In terms of being able to move stuff around, we've demonstrated that over the last two years. You've seen almost zero impact to delivery. Frankly speaking, we can manufacture in all those locations and move things around seamlessly.
To your question about you know government action or any other barriers to us exercising our supply chains, we have a small amount in China. That was your specific point. The majority is in Malaysia and Thailand.
Malaysia and Thailand. Even what's in China, we have all dual sourced.
Yeah. The important thing is that what we have in China are legacy packages primarily. We also have some of the newer packages, but whatever has a lot of IP, we try to keep it out of China. So those can be manufactured elsewhere if there is a problem. As Sunny said, in the last two years, it's not like we didn't have disruptions. I mean, obviously.
Right.
There were issues in China, but we were able to manage through them, thanks to our excellent operations, where we were able to use different logistics, go through different backend suppliers. We are multiple source in so many ways. You know, we you know, that didn't affect us. You know, COVID didn't affect us. Earthquake in Japan didn't affect us. It affected some fabs, but it didn't affect us overall.
Thank you. Just to try and summarize, the idea that Powi would be impacted by government regulations and/or trade war type situation with China, that really should not, as yet or going forward, be something that we as investors should be concerned about. Where do you put that in the risk factor spectrum? Thank you.
Well, if they do something very suddenly, it'll impact all companies, not just us. Everybody. They just shut off the borders. I don't think that's ever gonna happen. If it happens over time, all we have to do is shift all of our legacy packages outside of China to some other place, and all of our existing backend suppliers have the capacity in those areas. We distribute it so that we have, again, a geographical benefit in terms of risk. But I don't think that's a problem for us. So that's on the manufacturing side. On the other side of the coin, of course, the markets get disrupted, that will have some impact. But you know, we had the Huawei situation that did impact everybody.
There was over-enthusiasm in taking share away from Huawei from some of their competitors, which caused some unnatural exuberance in demand, so to speak. This kind of disruptions could.
Yeah.
Could happen.
Well, one other thing that could happen is more internal sourcing. Roughly about 60% of our products get sold in China, but not all of them stay in China. Actually, most of them come out of China. To the extent the OEMs are shifting their manufacturing outside of China, over time, it'll become a less and less of a problem. However, there's a portion of it, roughly about 20% of our revenue.
Which is actually consumed in China. That is a risk over a long period of time, and we have actually taken that into account in our projections. 'Cause we believe over the long term, we will have less and less access to the local Chinese market. Although we haven't seen it yet, I mean, except on the low end, which is a commodity, and we don't even play in that right now. That's a risk, and we have to account for that. That's the only risk I see in the long term.
The other part is that's on the lower end. One of the things that's really differentiated us is the innovation that you heard about. Even in a cell phone space, you know, which is commoditized, what innovations we have done, and my personal belief is that economics play in the end. If you have a great product, it will get demand. I think, yes, in the low end, we may have the challenges that Balu refer, but I think because of our innovation, a need for our product will always be there.
We have a question from online.
We do have an online question. In fact, Balu, your last response there regarding China might be relevant to this one. It says you laid out a compelling long-term growth story underpinned by an increasing and diversifying TAM, which paints a picture for potential growth rates in excess of low double digits. What are the primary governors on growth that are constraining your long-term guidance relative to the opportunity?
Okay, one of them is a potential decrease in internal usage in China, and that we have taken that into account. Maybe that won't happen, in which case we could potentially do better. The other one is some of the markets are not growing very much. You saw that cell phone is not growing. Notebooks are not growing very much. Yes, we will get share of the notebooks, but the market TAM itself is not growing as much as other markets. If you take the overall growth and other markets like automotive and all the new markets we talked about will take some time for us to get significant share. It doesn't happen overnight like in some other products.
Our power supplies always take longer time, but they also last a very long time, as you saw. Okay, where we start here? Go with Ross first.
I have the mic. Just wanted to ask about the supply side of the equation. You talked earlier in response to a question about diversifying outside of China on the back end, et cetera. The front end is incredibly concentrated in Japan. Can you talk a little bit about why that is? In the future, is there a benefit to diversifying that manufacturing footprint?
So far, this model has proven very resilient, even though, as you say, the majority of the fabs were in Japan. The fabs in Japan meet multiple criteria that are very critical to our business. First and foremost, the IP protection. Second, the collaborative business models that they practice and that we have developed over the years with them. We use many fabs in Japan that are scattered across the island from north to south. That has proven on multiple occasions resilient to natural disasters or whether it's a nuclear incident in the north or an earthquake in the south. We never had any disruption as a result of that overall, even though individual facilities were impacted. We're also using fabs that are outside Japan.
We use one in Europe, we use one in the U.S. You are right, our focus remains Japan, for the reasons I described, and I really don't see a risk going forward for capacities that will support the higher end of our revenue expectations. Okay, who gets the mic?
You said you're the leader in GaN, so I'm sure you know a little bit about GaN. What are some of the bottlenecks right now for GaN deployment? You know, whether it's horizontally, you know, end market deployment applications or vertically, meaning really moving up to higher voltages and eventually, you know, going after silicon carbide. What are some of the specific bottlenecks, you know, for you as a company and perhaps also the industry today?
Well, you use horizontally to describe markets.
Yeah.
I think of them as verticals.
Okay.
I get your question. I don't think there are barriers across market segments, and I think our presentations clearly articulated that. Not only we can use the same product across multiple markets and applications, but customers don't even know what's inside, whether it's GaN or silicon or silicon carbide. In terms of voltage and power voltage boundaries, I had a slide with the profiles, and you saw there is like a cutoff in GaN when you go to both high voltages and high powers. You may have noticed there were no units on either of those axes. That was intentional in order to give you a directional sense of where these technologies sit, but these boundaries are not fixed.
Even though fundamentally there are some differences that I articulated, benefits, plus and minuses, we think that our GaN technology, because of the specific choices we have made, and I articulated that in, if you recall our list of advantages that our GaN technology has the potential to move those boundaries beyond what they are today.
Perhaps I can add a little bit. That is, you'll notice in my presentation there were some step functions where we're introducing new products. It turns out that some of those markets have some levels of sophistication of the control systems and the peripherals around the part that our business model is to make it easy for customers. No heat sink, minimum number of external components, design quality, essentially, so that the entire experience of the customer is of a high quality. In order to address markets like the server main power, it's a higher power level. Plus, there are some requirements of that market that our current products don't address in a neat way. That's one of the. I guess your question was, you know, what's the constraint?
Well, one of the constraints is simply the system-level selling part that we do is, you know, we haven't completed that level of systemization with it around the GaN device. There's no fundamental change to the GaN required. It's simply the system that goes around it.
Hi, guys. Chris Rolland, Susquehanna again. I know it's been a month since your earnings release, and there was some great stuff in there. Maybe you guys, and Sandeep, this is probably another one for you. You are the macro maven out there. But if you could talk about, I think you mentioned a slight increase in order cancellations. There was a channel inventory build that you guys were seeing as well. You know, this discrepancy between sell and sell-through, bookings trends. And I think it was across all the end markets. I guess talk about how things have progressed since the Analyst Day, and then can you kind of help us look into next year and how you see these things playing out?
As we mentioned, nothing materially has changed. I think I'll just take a step back to say what we said so that I can give a better perspective. If you look at it, we were among the earlier companies in Q2 when we gave a guidance for Q2. We said we had orders, but we really wanted to see what the sell-through would be. I think our call was right, because all through the quarter, the sell-through deteriorated. It was not that it was even June was lower than May. It was abrupt, and it was mainly the Chinese OEMs where the cell phone sell-through really declined into nearly half. You know, we have a very large share of the Chinese OEMs because of the products we have and their adoption of higher power levels.
When we guided the quarter, we told you that, you know, our range was $165+ ± at that time. We said again with sell-through, see how the sell-through will be. That's where I think we are waiting to see where the sell-through will be. With all the macro conditions, and I'll tell you the challenges as the war is still an issue. You've still got the Chinese lockdowns, which we thought would go, and you know, keep having them because you've got the political situation in China, which can, depending on what happens, change. Just the whole inflation things where things become more expensive and people can defer demand. You know, if you wanted a dishwasher and you know, you can't afford because of your gas, you're gonna defer it.
The best I can tell you right now for the near term is, I don't know what Q4 will be. It's a little early, but directionally, because of the macro and where I'm seeing, it'll be directionally lower than Q3. What I have typically seen from a macro and seeing what that happens, and again, I'm not a, no expert on this, but it typically takes about three to four quarters for this to turn. Even in cell phones, I think you're gonna start seeing the OEMs clear out their inventory and start taking from distie and start seeing some demand come back in Q1. If you follow the three to four quarter rule, then probably Q2 things start. Q1 should be directionally better, is what I'm hoping for, but it's a little early to tell.
Unfortunately, what happens when you have this macro thing hit you in the back end, it messes up two years of growth. You have a lot of moving things this time, you know, between inflation and the war, the China lockdowns, and just the whole macro, where the demand can be impacted. I can only guide based on the prior experiences, what we think will happen. One thing which is really important is we tend to see this earlier because we go into power supplies, and nobody waits for a main product for a power supply. If you look what has happened, every time we come out of it earlier and stronger. With all the market share gains that we have had because of the ability to supply, I think it's gonna benefit us very well when this thing turns.
Hey guys, Charlie Neal from Port Capital. I thank you for all the great detail today. I had a two-part question on the long-term growth forecast that you guys provided. First is on GaN. What is the risk of cannibalization among your existing product portfolio if you really see increased adoption there in the out years? Secondly, on the automotive end market, great to see the traction, the early traction that you are getting there with some leading OEMs. Historically, that has been a market where OEMs have consistently gotten a lot of price out of their suppliers. The past year or so has been relatively anomalous with the supply chain issues. What is the risk of more material ASP degradation with the large OEMs and some of those projections for the automotive growth? Thank you.
Let me answer the first.
First one.
The first question. I think the overlap between our GaN products and the rest of our products is almost zero. They're positioned to be seamlessly continuous. Even within a product family, there are power levels supported by one, like one material. The next power level up starts with GaN and continues with GaN. Within the GaN families, we define the transistors in such a way that they themselves form a continuous power level as we express it in on resistance. There's no overlap even at the device level. That's a very simple answer to your question.
The answer to the automotive pricing question is really straightforward because it's just a function of what the value we bring, right? If you're replacing, you know, 50 or 90 components, it's the value of those components that determine our price. Yes, it could come down, but I still believe we can make very good margins in that market. We have such significant levels of integration that we are very confident that it'll be the margins will be consistent with our company margins.
One more.
One more. Okay. One more, and then we have to go have lunch. Any online? No.
No.
No.
Oh, we got one there.
Thank you. This is Trevor Janoskie from Needham. With the data center and industrial opportunity expanding meaningfully in 2024 and 2025, I'm wondering how big of a role GaN reaching price parity with silicon plays into this timing. To add, when GaN does reach price parity, what do you expect to change within the market?
The GaN, I think we already said in the next couple of years, we think that it'll, you know, become in parity with silicon. GaN also has the advantage that we can go to much higher power levels without causing heat problems because, you know, we don't need a heat sink for very high power levels. It's not just cost parity that gets us into these sockets, it's the integration value we bring. We're not just replacing the MOSFETs, we're replacing the whole system. We are designing the whole system. We bring a significant level of integration. Otherwise we won't even play in this market because we are not a discrete supplier.
Now we compete with discretes all the time, we compete with the cheapest MOSFETs made in China all the time and still make, you know, 50%-55% margin. That's no different than GaN, but the fact that GaN is getting close to silicon is a huge advantage long term, which means that above certain power level, there is no reason to use silicon. Now, there are some applications where GaN is trickier to use, like motor control. One of the advantages we have is because our technology group is right next to our engineering group there, we are able to fine-tune the GaN technology to work for any given application. It is actually not easy to do. I mean, it's easy to say, okay, if you have a discrete component, oh, I can go into motor control, I can go into
The problem is now on the customer to solve. I mean, if you notice there are several companies who are working on automotive GaN solutions who supply discrete components. You still don't see that in a system level. Yes, it takes us longer to solve the system problems, but it takes the same amount of time for our customers to solve it as well. It's just that we are doing it for the customer rather than they doing it. I think that, you know, I hope I answered your question in terms of, you know, value pricing we do, and so we don't see any issues at all. I think we will be very competitive and still make good margins.
To echo what Balu just said, as you know, data centers, the biggest cost in a data center is energy consumption. To the extent that you provide a solution that has better energy efficiency, such as GaN, that adds to the compelling value proposition of a GaN solution, even at cost parity or maybe even slightly different.
Thank you.
Okay. Well, thank you very much.
Thank you.
I think we have demonstrations next door. You can see a lot of products that use our technology, and then we also have lunch in the pavilion.
Thank you very much.
We'll be around if you have any questions.
Thank you. Thank you very much.