ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

Hello, and thank you for standing by. Welcome to onsemi Treo Platform Webcast and Conference Call. [Operator Instructions]. I would now like to hand the conference over to Parag Agarwal, Vice President of Investor Relations and Corporate Development. You may begin.

Thank you, Towanda. Good afternoon, and thank you for joining onsemi's Treo Platform Webcast and Conference Call. I'm joined today by Thad Trent our CFO, and Sudhir Gopalswamy, President of AMG and ISG. This call is being webcast on the Investor Relations section of our website at www.onsemi.com. A replay of this webcast, along with accompanying slides will be available on our website approximately 1 hour following this conference call and the recorded webcast will be available for our approximately 30 days following this conference call. Additional information is posted on the Investor Relations section of our website. During the course of this webcast, we will make projections or other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution that such statements are subject to risks and uncertainties that could cause actual events or results to differ materially from projections. Important factors that can affect our business including factors that could cause actual results to differ materially from our forward-looking statements are described in our most recent Form 10-K, Form 10-Qs and other filings with the Securities and Exchange Commission. Our estimates or other forward-looking statements might change, and the company assumes no obligation to update forward-looking statements to reflect actual changed assumptions or other events that may occur except as required by law. The purpose of this call is to introduce our newly launched to Treo Analog and Mixed Signal Platform. We request you to focus your questions on Treo, and we will not take questions on the current business conditions or other topics. Now let me turn it over to Thad. Thad?

Thank you, Parag. Good afternoon, and thanks, everyone, for joining us on the call. At our Analyst Day last year, we talked about building our leadership in power and sensing by expanding our portfolio of industry-leading Analog and Mixed-Signal products. After 2 years of investment, we announced the Treo Platform in Electronica last week. Treo is a highly differentiated platform, solving our customers' most challenging problems in a new and rapidly evolving markets. Treo also unlocks a $36 billion TAM in our focus end markets of automotive, industrial and AI data center. Consistent with our gross margin expansion initiatives, Treo products are expected to have gross margin of up to 70% clearly demonstrating our ongoing product portfolio shift towards highly differentiated products and technologies. Taking a page from our playbook, we will leverage our brownfield investments in our 300-millimeter East Fishkill fab in existing back-end facilities to manufacture Treo products driving best-in-class ROIC. Now let me turn the call over to Sudhir. Today's presentation is available on our Investor Relations website, so you can easily follow the call. Sudhir?

Thank you, Thad, and good afternoon, everybody. I'm super excited to be with you here today and to give you some insight into the Treo Analog and Mixed-Signal platform. Let's start by providing a bit of context, moving to Slide 3. So as Thad mentioned, at Analyst Day 18 months ago in New York, we laid the foundation for Treo by describing the company's strategy in 2 ways: first, winning at the core, and we define the core as the distinctive capability that onsemi is known for, silicon carbide and silicon power on the power side, and image sensing on the sensing side. And what we said as well is that we would expand from that core by providing ICs that complement these core areas of strength in delivering full system solutions, intelligent power and intelligent sensing solutions. Moving to Slide 4. Earlier this year, in March, we announced the formation of the Analog and Mixed-Signal Group, and our charter in that group is to execute on expanding from that core to deliver the ICs that really allow us to build out the intelligent power and intelligent sensing Solutions. And we do that by introducing products that target the main segments of the company overall targets: automotive, industrial and AI data center. And as we target that, we will deliver high-efficiency, devices that offer the highest level of integration, and as well best-in-class performance, which brings us to last week, moving to Slide 5. As Thad mentioned, we introduced Treo Platform publicly at Electronica last week. And while the customers we had known would appreciate the value proposition, we were pleased to realize that the reception was extended to all of the press. So we received a great deal of press attention. And the important thing for us was that the press not only understood the platform, but more importantly, the benefits to onsemi and to our customers. And so moving to Slide 6. If we talk about what the significance is of the Treo platform at the highest level, it is a disruptive platform. What the platform allows onsemi to do is to enter the high-performance analog and mixed signal market. And as we do, we will leapfrog the competition with market-leading technology and it's that technology that is the source of competitive advantage. The competitive advantage gets realized in the form of an unprecedented level of integration in these Analog and Mixed-Signal parts. We simply can integrate more than our competitors can. And we accomplished that through a combination of the proprietary BCD 65, Bipolar, CMOS, DMOS, 65-nanometer process technology, which, as Thad mentioned, is built on a 300-millimeter fab in East Fishkill, but we combine that with an SoC-like approach, meaning that we are able to give rise to a wide variety of parts by defining an architecture for the platform and creating IP blocks and then proliferating those IP blocks so that we realize a wide variety of products, and we'll walk through that a little bit later. And in doing this, we will expand into a $36 billion TAM, and our intent with Treo is to build a $1 billion revenue stream by 2030, and we start today with 0. And as Thad mentioned, because of the product categories that we will be able to address and because of the competitive advantage we offer, we'll be able to realize gross margins up to 70%. Moving to Slide 7. If we talk a little bit more about those markets, as I mentioned, the market segments that we target with Treo are absolutely consistent with those that we target at the corporate level. Automotive, industrial, and AI data center. And if you think about it in terms of product categories, the first among those product categories are power management ICs. So this is the range of IC products that combine with our silicon carbide and silicon power devices to create winning full power tree solutions. And so what I mean by this are gate drivers, DC to DC converters, power management devices and as well AC to DC converters over time. Sensor interfaces is the second category of products that Treo gives rise to. And in sensor interfaces, we span a full array of performance and power. At the high end, we offer high-performance sensor interfaces, like, for example, ultrasonic sensors that have applications in automotive and industrial. In automotive, those ultrasonic sensors are integral to park assist type of applications for vehicles. Inductive sensors, in sensing provides the ability to detect position. For example, the position of a steering wheel, the position of an accelerator pedal or going forward the brake pedal. And so these are instrumental in what's called x-by-wire or drive-by-wire type applications. At the other end of the sensor interface spectrum, we scale all the way down on the same platform to ultra low-power devices and here, the initial target application is medical segments, like, for example, continuous glucose monitors, where battery life means everything or hearing health devices, hearing aids were the same is true. It's all about battery life. The third category of products that we access as communication devices. Now we have a track record in communication devices in automotive and we're building on that track record by expanding into a very new category of communications device. And that's called Single Pair Ethernet, also known as 10BASE-T1. And the Single Pair Ethernet type of devices are a key component of zonal architecture. So the overall automotive segment is transitioning from domain control type of architectures to zonal architectures, in which the electronics and electrical systems are partitioned into zones. And what Single Pair Ethernet or 10BASE-T1S does is to provide transport within the zone and in that way, complement high-speed internet in the backbone -- Ethernet in the backbone to have an end-to-end Ethernet type of solution. And the last category is ASSPs. And these ASSPs range from simple, single-function devices like LDOs or level shifters, to very complex devices as well. And as I said, the aggregate of these product categories allows us to unlock a total $36 billion TAM at margins up to 70%. So let's talk a little bit more about what Treo actually means. So we move to the next slide, Slide 8. Treo is at the end of the day, the industry's most advanced Analog and Mixed-Signal platform for intelligent power and intelligent sensing. And if you think about the word Treo, it denotes 3 things. The first is the number 3, and what we mean by denoting the number 3, is that Treo offers the best of all worlds. Here, we're talking about the BCD technology process and BCD again, stands for Bipolar, CMOS, DMOS. Bipolar is used for analog functionality, CMOS for digital functionality and DMOS for power. And what we're doing is integrating all those 3 in the same process. This in and of itself is not unique. BCD technologies have been around for a while. What is significant, though, is the manner in which we implement BCD. First, we do it on an advanced technology node, 65 nanometers. That means that we can integrate a large amount of digital processing in our devices. The second key component is the voltage range. We scale from 1 all the way to 90 volts. And this means that we can integrate high-voltage and low-voltage functionality in a single device. And the third aspect is that we support up to 175 degrees in the full automotive grade 0 requirements. In automotive grade 0 is the most stringent of temperature and reliability requirements in the automotive field. So the number three. The next is the concept of a tree, and this is equally important. The concept of the tree, of course, is that you have a trunk. And here, we refer to the Treo Platform itself as being the trunk. And that gives rise to a set of branches and the branches here refer to different types of products or product categories. The branches in turn, give rise to specific product lines and in turn to ultimately products. So the tree expands and grows as we proliferate products. They all stem from this common platform, this trunk, and so Treo isn't just a product. It is a platform that gives rise to a broad range of products, and that's the tree. The third aspect, so it turns out the Treo in Spanish refers to it's a nautical term that refers to a square sail that's used to catch basically favorable wins. And so by this, we can note the concept of performance and agility. So the idea here behind Treo because we're building a platform comprised of reusable IP blocks is that our development cycle gets substantially reduced. So our time to market for introducing new and innovative products and of course, for our customers to do the same, is much shorter than it would otherwise be. So the best of 3 worlds, Bipolar, CMOS, and DMOS on a single process, branches stemming from a common trunk and a high performance and agility or time to market are what Treo is all about. So let's talk about it in the next level of detail, starting with the technology process and moving to Slide 9. Up until today, there has been a trade-off between voltage and advanced technology nodes or technology nodes. So if you look out at the competitive platforms that exist today, on the one hand, if you want high voltage, you typically are getting a capability that is implemented on a larger technology node. And because it's a larger technology node, there's lower performance and a fewer amount of features that can get integrated. So that represents the upper left of this chart. On the other side, if you want to take advantage of the advanced processing technology node, then you're typically making trade-off in voltage, meaning that you're limited to low-voltage applications. And of course, there are some kind of tweeners in between from the competitive platforms. Now these are actual competitive platforms that exist today. Now what Treo does is to essentially break this trade-off. There are no trade-offs, no compromises because we implement Treo on the 65-nanometer technology process, and we scale all the way to 90 volts. So in a single platform, you can have 65-nanometer and up to 90 volts and break the historical paradigm here where you have those trade-offs. And as well, of course, we support the high temperature operation. And so if we move to Slide 10, you'll get a sense of what this allows us to do for customers. At the end, we can integrate high-voltage power up to 90 volts, high-density digital leveraging the 65-nanometer node in the logic that we can integrate on it, and high-performance analog where we have optimized transistor devices that allow us to implement high-precision and high-performance analog. And we bridge all of those together on a single device, and we're the only ones that have the capability to do that. So it is the technology process that allows us to do this in a very distinctive way. And if we think about what this means to customers, and we'll move to Slide 11, because the platform is so broad and because it has so many attributes, the value proposition is equally broad. So you can see here the dimensions of features and what those features bring in terms of significance in customer systems as well as specific examples of how we implement those. And I won't go through all of these. I'll highlight a few of them to give you a flavor of what Treo Platform enables. So we'll start with the voltage. As we said, we scale from 1 to 90 volts. And what this allows us to do is to integrate what would otherwise take 2 separate devices, one high-voltage device and a separate low-voltage device, and we can integrate those in a single chip. And when we do that, we create BOM efficiency and we create area efficiency and value for the customers in doing that. An example of how we do that is in that automotive Ethernet application, where we literally take 2 devices, the high-voltage and low-voltage device, combine this into one on the Treo platform, so we can implement Single Pair Ethernet on a single device here. And the example that I'm talking about in automotive Ethernet extends to a broad array of devices that also require high voltage and at the same time, low-voltage capabilities. So you can think about the 48-volt transition that's happening in automotive and the AI data center as an example of how this capability could be realized in those types of systems as well. The second attribute then is the modular SoC-like architecture. So the idea of having a common architecture with predefined interfaces for the IP blocks to interconnect isn't unique. This exists in SoCs but it doesn't exist in analog and mixed-signal products as the type that we're talking about today. And what it allows us to do then is to deliver a very quick time to market. Now we will deliver from specification of a device to silicon samples in 6 to 9 months. We're already doing that. We have done that on several devices that we are sampling now in 9 months and what we will be able to do as we build the library of IP is realize this across the board. So the ability to proliferate products very quickly means that we can react quickly to customer needs as they evolve and the customers can react quickly to needs that evolve in their spaces. Now the other thing that the modular SoC-like architecture affords us is the ability to offer a high degree of specialization or customization of these devices. So we can target very fine slices of application functionality, and we can even go so far as to implementing customer-specific devices where the volumes justify. And this is something that our customers instantly appreciate and put value in it. So this is also something that's very distinctive between what onsemi can do and what our competitor approaches are, who pay more kind of standard product type of approaches. So again, this is something that offers value to onsemi and to our customers alike. The third capability that I would like to highlight is the power. And here, you can think about this in 2 ways. In higher power systems, it's all about efficiency. We know that. The most obvious example is in AI data centers, where it's all about extracting every little bit of efficiency. And here, the Treo platform will be used increasingly across the power trees. So for example, in the smart power stages. And the smart power stages are the things that are closest to the FPU, the GPU, CPU, TPU, that's actually the sorts of processing in these AI data centers. And the smart power stage is comprised of a driver device and FET device. And the driver will be implemented on the Treo platform and increasingly the overall product on the Treo platform. And when we do, we're able to deliver a higher level of efficiency than all of the current solutions. At the other end of the spectrum then, as we talked about earlier, our lower power devices. And here it's all about reducing power consumption and extending battery life. So for example, we introduced samples of a product that goes into continuous glucose monitors. And if you think about continuous glucose monitors, these are devices that fit on the patient's arm. And what you would like to do is to extend the life of those devices so that you don't have to replace them as frequently as you do today and by getting to lower power and getting to extended battery life, that's exactly what you can do. So these are several examples that illustrate how Treo provides value to our customers. And you can see that, that value proposition gets realized differently across customer types and across segments, but it's true really across that base. And so moving to Slide 12. And maybe taking a step back to give you an idea of how Treo is constructed. So at its core, I keep referring to this SoC-like approach. So think of Treo as a first order being implemented in these systems. So there are 4 subsystems. Typically, in the intelligent power and intelligent sensing products it's the power management subsystem or the sensing system that provides the core differentiation and value to the customer. That's most typical. The communication set system then at the top is what allows the chip to communicate to the outside world. Now in some instances, like Single Pair Ethernet, the communication subsystem is the primary source of value in the system as well. So you can see each of these subsystems play there own role. Now at the bottom is a very important capability that it's the compute subsystem, and this provides the ability to inject intelligence into the intelligent power and sensing type of solutions. So this is comprised of an MCU with optional hardware accelerators and DSP functionality and even going forward, AI -- optimized AI engines that really allow intelligence to be implemented across the range of products. And by virtue of integrating this type of compute, we not only make the devices smarter, but we can subsume functionalities that might be implemented otherwise in separate devices on the customer BOMs, super powerful capability. Now as I said, each of these subsystems is further comprised of IP blocks and the IP blocks then get reused across broad families of products. And so if we -- if we talk about what those products are, they range from very simple devices like voltage translators or LDOs or bus switches to ultra low-power AFEs like, for example, the continuous glucose monitoring products that we talked about, to ultrasonic sensing, which is a high-performance type of sensing application, Single Pair Ethernet they extend to compute DC to DC converter type of products, they used in compute systems. AI data center point of loads and going forward, automotive LED drivers, inductive sensing, LIN and on and on. So we're able to proliferate new product families from this common platform by leveraging common IP and reusing that IP intelligently across all of these devices. Now the products that are highlighted in bold here are actually sampling today, and some of them will start to generate revenue as soon as next year -- the first half of next year. And so let's talk a little bit more about the revenue, moving to Slide 13. We mentioned that our target is to generate $1 billion in revenue from Treo by 2030. And we do that by targeting the market segments that we've talked about in the way that we've talked about. We're going to gain share with our accelerated development cycles because we can move so quickly, we can respond to changes in requirements and as I said, so can our customers. And that predisposes them to adopting the devices built on the Treo platform. And of course, going back to the core concept, when we implement these devices, they combine with existing areas of strength. So think about gate drivers combining with the silicon carbide solution to deliver even higher efficiency and even more value to the customer. And in doing that, we build complementarity between our individual products and create more value and system-level solutions. And that's how we're going to realize this revenue. And if we talk about the financial benefit extends beyond revenue. So moving to Slide 14, it also drives gross margin expansion. We've been developing Treo for 2 years, which is actually a very short amount of time, given that we started without having this process technology and then develop the process technology, we're developing the IP and we're already sampling products in a span of 2 years. That's a very quick time schedule. And in that time, we're able to deliver high value and differentiated products, and we'll realize first revenue as soon as first half of next year. We have 10 products sampling now, and we're going to double by mid next year. And as I said, the Treo platform itself enables products that target higher value segments and offer differentiation within those segments and as such, help us realize gross margins up to 70%. Now there's 2 ways in which Treo benefits us from a margin perspective. One is it's a margin improvement as our legacy products and the customer systems in which those products reside end of life, we're able to replace those products at a higher level of functionality and therefore, a higher level of gross margin. The other way, of course, is that Treo targets new product categories, which we have not historically addressed, which inherently offer those margins. So it's that combination of capabilities that allows us to realize the margins. And then finally, as Thad mentioned earlier, we are leveraging existing brownfield investment, so the capacity that we use is at East Fishkill today the 300-millimeter fab and we built out the BCD 65-nanometer process technology in that fab. And because we're using that and using the existing back-end facilities, we're able to deliver best-in-class our return on invested capital. And so moving to Slide 15, and again, in summary, what Treo brings for onsemi is the potential to disrupt. We are entering the high-performance analog and mixed-signal market, and in doing so, leapfrog into competition. The source of competitive advantage stems in part from the technology process, but also from our design approach, our SoC-like design approach that we're now bringing to Analog Mixed-Signal products. And the result of this is going to be to enable us to deliver $1 billion of revenue in that $36 billion TAM with gross margins of up to 70%. And so with that, I'll wrap up and transfer back to Towanda to take questions.

[Operator Instructions] Our first question comes from the line of Ross Seymore with Deutsche Bank.

Congrats on launching the platform. Two questions. I guess the first one, probably for Sudhir, if I just look at this technology and think why haven't others done it before? Is it more the ability to have the BCD process on the 300-millimeter and 65-nanometer. So it's the core technology that's different? Or is it more that you're -- how you're putting it together, that's truly a differentiated approach?

It's more of the latter, Ross. It's this idea of not just doing BCD, but trying to take the risk of doing that the 65-nanometer node. And again, doing that with this SoC-like approach is that, that's distinctive. And I think that just takes someone that is coming in from the outside. I think that's the benefit of entering a new as opposed to having an incumbent position.

I guess for my second question, whether it be for you or Thad. On the gross margin side of things, the up to 70% is obviously very impressive. Is there a rough range these tend to be 50% to 70%, you have that 50% to 53% gross margin target overall for the company. Just, is there a range around that? And is it really scale based where you don't start off at these gross margins, you have to work your way up over time as you get more of the volume and cover the fixed cost, those sorts of things. How do I think about how this folds into the company from a gross margin perspective over time?

Right. So first, at the highest level, these are going to be accretive to the company targets even with the new margin targets that you're referring to where we will grow. These will be accretive. What we see right now is in aggregate across an array of products, the gross margin that we can achieve is at a 6 handle. It's got to 6-handle across the array. And you're right, it scales from the 50s all the way to the 70s to get to that average. And of course, this is going to depend on what actually happens in terms of adoption and the different rates of adoption of the -- and the mix across those products. And yes, there is an element of growing into this. So Thad, I don't know if you have more to say on that.

Yes, Ross, if you think about it, right, I mean, today, as you know, our fabs are roughly mid-60% utilized. This will help as this ramps to fill up that utilization over time. So that is also another tailwind. But when we talk about the margin range that Sudhir talks about, that is the -- think about it as the product margin and how we price it at scale, at utilization. So as we transition into these products, they're all accretive to gross margin.

Our next question comes from the line of Gary Mobley with Loop.

I apologize in advance for the background noise. I wanted to ask about the R&D intensity of the Analog and Mixed-Signal Group, I presume since this group was formed a while ago, it's already in the baseline R&D investment. But as most people on this line know you're going after some pretty good players that have pretty good R&D scale. So maybe if you can just talk about if, in fact, we're at a baseline R&D spend for this business already. And what you really have to do to compete against those guys, whether it be for more efficiency or just trying to outspend them on certain niches?

Yes, Gary, it's Thad, I'll chime in first here and then let Sudhir add if he wants to. So, yes, it's in our baseline. We've been developing this for 2 years. So the R&D effort to develop this platform has been in our P&L for really 2 years. As you go forward and you think about the efficiencies that Sudhir described, it is very efficient when it comes to an R&D standpoint because of this SoC-like design capability that we have. So from an intensity standpoint, doesn't -- it isn't going up. it's actually going to stay relatively in this -- and what we see the baseline today. As we continue to roll out new products, it's very efficient. Sudhir anything you want to add?

No, I think that's exactly right. I think as we scale the platform, the incremental cost to deliver a new chip from an R&D investment perspective, is lower and lower, again, across a range of complexity, but the incremental cost, all things equal, is lower and lower as we build up the platform and generate more IP. And to answer one part of your question, Gary, no, we're not going to outspend our competitor. We're trying to outsmart I guess, from a spending perspective given our relative budget and be able to create new products efficiently.

If I can ask a quick follow-up this modular approach to putting together different IP blocks, does this involve a partnership with a network of IP partners? Or is this all homegrown IP blocks within onsemi?

Yes. It's primarily homegrown IP blocks. Certainly, all of the differentiating IP, we will develop in-house. But as with all kind of SoCs, there is a component to working with external parties. So I alluded, for example, to the fact that we're putting in MCUs, there we would license from third parties, the MCU. And like that, there are examples of other things that are not necessarily core to our differentiation that we would source from third parties.

Our next question comes from the line of Vivek Arya with Bank of America Securities.

I wanted to go back to this competitive differentiation. So is the argument that none of your competitors are able to have 300-millimeter plus 65 BCD plus and SoC format? And if they aren't, and then why are you only targeting 2% or 3% market share, right? You mentioned a $36 billion TAM and only $1 billion. So in sales, if this thing is so kind of breakthrough, why not target much higher market share?

Yes. So it is true that the advantage stems from the combination of BCD-65 and up to 90 volts on a single device. So those are the 2 things from a technology process as well high temperature and from the SoC-like approach. So that's right. That's the basis for competitive advantage. In terms of the $1 billion, again, even though there is a $36 billion TAM, we're not pursuing every aspect of that TAM. So we're selecting the areas in which we can provide the highest value as part of system-level solutions for our customers. And of course, those that offer the higher value to our customers in a way that helps us realize higher gross margin. So we're not going after all $36 billion of that TAM.

Okay. And I think you mentioned we could expect to see some revenue contribution in '25, any kind of rough sense? Is it like low tens of millions, $50 million? Like any rough sense of how we should track sort of more intermediate milestones of what the revenue contribution might be in '25 or '26 for that matter?

Yes, Vivek, it's Thad. So look, we said it's going to start ramping in the first half of next year. We'll start to see some revenue. But as Sudhir said on the call, we're going from 0 to $1 billion in 2030, right? You think about some of these markets, they do take time to kind of layer in. So you will start to see some revenue. It's going to be small and it will be ramping. You can look at the CAGR that we've laid out there. So we think we should be outgrowing that as well.

Our next question comes from the line of Toshiya Hari with Goldman Sachs. Toshiya we are not able to hear you. Toshiya is having trouble - phone trouble. Our next question comes from the line of Joshua Buchalter with TD Cowen.

I wanted to ask a big picture about how you feel about your cost basis and competitiveness in this market. So totally here, you on a300-millimeter facilities being a competitive advantage from a cost basis. But I would imagine being on 65-nanometer and competing against primarily peers manufacturing on 130 or even 180 nanometers on BCD would be a more expensive process geometry. Could you maybe speak to how you feel you can merge into traffic in this market? And how you feel about your cost basis with those 2 vectors.

Sure. I think if you think about the type of devices that we're targeting, we are choosing devices that benefit from this 65-nanometer technology node in one way or another. And when you choose based on that criterion, you're in a position where you actually can achieve a lower cost basis. So in general terms, the products that we have will have a favorable cost basis with respect to competitors because we're able to do more on the 65-nanometer node.

Okay. And to follow up, I wanted to ask about your go-to-market. Is this primarily verticals do you expect to be more catalog-based or more design socket based? And maybe you could talk about your visibility into sort of the near-term or medium-term trajectory as you ramp the initial revenues? And how much of that is, I guess, a socket that you expect to compete with on a design win basis?

Sure. In the short term, especially, it's true that it's going to be predominantly socket based, right? We're targeting the automotive and industrial segments, and maybe there's a different mix across those. But within automotive, it's going to be fundamentally socket based. The products that I'm talking about are going after existing or emerging design sockets, that is true as well on the industrial side, especially true on the health side, where there are fixed sockets that we're going after. As we grow over time, this will -- the broad market capability will evolve as well, and you'll start to see more of a broad market phenomenon where these products just find their way into catalogs and on the shelves of the distributors.

Our next question comes from the line of Quinn Bolton with Needham & Company.

I just wanted to step back. I know BCD processes from foundries have been available for a while. But could you just kind of -- to level set us what -- typically, what nodes are the foundry processes available on? Are those like 180-nanometer or higher and similarly, what kind of voltage range could you get off of commercial BCD process just to try to give us some sense of how much further ahead you may be? And then I've got a follow-up.

Okay. Yes, there are quite a range of competitor BCD platforms out there. And in general, they could be on 130 or 180, in which case the voltage is typically -- you support up to high voltages there and you're more optimized for high voltage. On the other extreme, you do have more advanced technology nodes, but you don't have the full range of voltage. So they range for -- some of them are low voltage, some of them are low and medium voltage type of processes that exist. So there are a number of different technologies there, but none as broad as Treo, none -- and we do our survey. We're looking at everything that's out there, none of it as broad as Treo.

Got it. And then just you had mentioned sort of an example is the smart power devices. And wondering how integrated would your platform be, would you have the gate driver and high side and low side FETs all integrated on a monolithic BCD process, would you start with discrete FETs. How much integration can you bring to that kind of a device? And then I think one of the advantages of higher levels of integration in those types of devices would be switching speed. Could you compare what switching speed you might be able to achieve in that device versus, say, a discrete solution from some of the competitors on the market?

Yes. So in that particular product category, the short-term focus is on just doing the drivers on the Treo platform and having the FETs being discrete. It certainly is on our radar to look at integration of the FET as well. So that's something that is our radar. But I think the -- in the short term, think about it as being stand-alone drivers, discrete with the FETs in a package. And in terms of -- in terms of the switching speeds, because it's discrete, we're not limited. We can switch up to high speeds in that way.

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

And this might be a broader one for management. But you guys mentioned licensing MCU, IP and I was wondering if this might open the door to more microcontroller SKUs, either specialized or general purpose. Just as Hassane and Thad, you guys had a lot of experience here at your last company, a lot of success. Could this be another step in the evolution for ON moving more towards the processor side?

Our focus right now is the existing Analog and Mixed-Signal product categories where we're injecting intelligence into those product categories. we're not pursuing a general microcontroller strategy instead focusing where we have the most value today. And our value is derived from really that analog mixed signal capability and then complementing the other power discretes, for example, that we have in creating the full system solutions. So that's our focus.

Okay. And then just basically, as I look at all of these different IP blocks, there's a ton, I just think about all the permutations and combinations. And so I guess, are you going to have just a ton of different SKUs with all these different combinations? Or is there something configurable through the use of like memory and software or something like that where you can change these different blocks, so you don't have as many SKUs?

Yes, very perceptive question. Both are true. We're going to have to proliferate a lot of the IP blocks. And in some cases, you have to optimize for area and for performance. And so there is going to be a proliferation of IP but yes, you limit that proliferation by providing configurability where you can, where it makes more sense in a given product type, you will make it configurable. And there will always be that trade-off between configurability and device-specific optimization in the form of either area or power, and we'll make that trade off intelligently. But the answer is both.

Our next question comes from the line of Tore Svanberg with Stifel.

This is Jeremy on the line for Tore. I guess our first question would be the Treo platform. You talked about a 1 volt to 90 volt range. Is there plans to expand beyond that range to maybe address different markets? Or is this kind of with that AC to DC conversion application is going after?

Right now, all of our targets and the high volume that we see that we're going after right now is all contained within the 90-volt limits. There are definitely applications above 90 volts, and we have solutions for those outside of the Treo platform, but our focus for Treo is really up to 90 volts.

Got it. And I guess, if -- for those 10 products that you mentioned, you're sampling currently and that doubling of SKUs by 2025. Can you just give us a little bit more intent to maybe which product categories these are initially coming from? Is it all one product category? Is it all one application? And where are you seeing the most traction at this stage?

Yes. So it's -- no, it's not one single one. It's broad. So we have LDO's, single fusion devices like LDOs and voltage translators and bus switches and those are going to find their way to revenue. We have an ultra-power -- ultra-low-power AFEs that is sampling right now. That will make its way to revenue next year as well. And the ultrasonic sensor product, the compute DC to DC converter here, we're talking about a multiphase converter. Those are also sampling right now. So it's really a range of products that are already sampling and will convert to revenue starting next year.

Got it. And one final question, if I could. Regarding the East Fishkill fab that this is being installed into. Can you just give us a sense of how much capacity you currently have maybe what kind of CapEx you might need to get to that $1 billion target? And if it's all brownfield like or would you need to kind of expand the site? And finally, related to that, is there any benefit from the chipset that you can get for this project?

So Jeremy, it's Thad. We have made big investments in the East Fishkill over the last few years in terms of capital. And that capital has been put in place in anticipation of the launch of the Treo platform. So if you think about from a CapEx standpoint, most of the investment is behind us. And on our last call, we talked about lowering our CapEx intensity down to the mid-single digits from what's been running kind of in the 11%, 12%. So that's already reflected in our go-forward plan. So this is why we're saying this brownfield investment generates a very high ROIC. If you think about the gross margin favorability and operating margin favorability. This is a -- there's a high return on this product given that most of the investment is already behind us.

Our next question comes from the line of Kevin Cassidy with Rosenblatt Securities.

Yes. Thanks for taking the time to explain to us about this new platform, very interesting. And maybe 2 years ago when you started in on this, was there a pull coming from some of your customers saying that, "Geez, no one else is doing this? Can you help invent this for us?" Or maybe can you say how you started into this R&D?

Yes. I would say it's more of a push. The customers think in terms of what their requirements are at the system level and don't necessarily map those to, hey, we wish we had a broad platform and wish you had these attributes. What we did know is that customers want an ever-increasing support for a higher and higher voltages and higher efficiency in the power domain. We know that you have to get better and better battery life through lower power, for example, in CGMs. And so the general trends we were well aware of, but they didn't map those trends specifically to the Treo platform. It's really our conception of a platform that responds to those trends.

Okay. Great. And along with that, as of being a brand-new design, does your customers need to buy new test systems and just to be able to see if this device is working as expected? Or is there equipment out there already that would support it?

There's nothing really specific that the customers would do differently from a test perspective. So we're at the end of the day, introducing silicon that may offer a higher level of integration. So we do in one part in what previously they took 2 parts to do. There isn't sort of more from a testing perspective, but where we offer more integration at the platform level obviously, they would make some hardware changes on their board, but not specifically for testing.

Our next question comes from the line of William Stein with Truist Securities.

Typically, when we see companies enter sort of these new technology areas, they go after a very well-understood opportunity at least to start. So I wonder if you could clarify for us the stuff that you're sampling now where you say that's going to generate revenue in the first half, should we think about those parts as being sort of pin-for-pin compatible with existing parts that are being delivered by competitors, where they're just going to potentially swap out some portion and qualify you as a new vendor? Or are these brand-new opportunities that were never delivered before that are -- that your customers really taking a bigger risk on.

Yes. I would say they're definitely not pin-for-pin compatible type of devices, and that's not the intent. They're designed to introduce new functionality and or kind of new performance or power kind of specifications or dimensions. Now the building from a source of strength is that we have a leadership position in some of these segments today. So there is that. And while there's customer design effort, we know because we have intimate relationships with those customers, that they're willing to do that. And so when we deliver a sample, we know that the customer is taking that sample, putting it on a board and starting to exercise the part with the intent of getting it into production. So we have that confidence. But they're not -- we're not going after for pin-for-pin compatible type of parts.

That's helpful clarification. Maybe one other, if I can. I wonder the degree to which this would be complementary to the rest of your portfolio? Or is it possible that this would be this could cannibalize some of your existing portfolio, maybe still mix up in margins, but does it remove or consolidate revenue from other parts of the portfolio? Does it fit with some other thing that you do now that makes the sale easier?

Yes. So if you look at it at the company level, it's definitely complementary. So I talked about the fact that, for example, if you look at a power tree solution, you would find silicon carbide FETs or silicon power FETs. And the products that we would introduce in Treo are completely complementary to that. Now at the level of AMG, we do have some products that are earlier versions of these type of things. So it's not, for example, the first CGM, AFE device that we've done. It isn't cannibalization because the customer is doing a next generation, but it's replacement at a higher value for something that we have today. So you have that mix across the board. And then, of course, there are new product categories entirely.

Thank you. Ladies and gentlemen, I'm showing no further questions in the queue. I would now like to turn the call back to Thad Trent, the Executive Vice President and CFO for closing remarks.

Thanks, Sudhir for leading the discussion today, and thanks, everyone, for joining the call this afternoon. As you can see, we're excited about the Treo platform and the market opportunity ahead of us. I want to thank our teams that have contributed to the development of this truly innovative and marketing -- market leadership technology. Thanks again for joining.

Ladies and gentlemen, this concludes today's conference call. Thank you for your participation. You may now disconnect.