Okay, good morning everybody. I'm John Vinh. I cover semis here at KeyBanc Capital Markets. We're pleased to have the onsemi team with us today. We have Hassane El-Khoury, who is the CEO, and Thad Trent, the CFO. Welcome, guys.
Thanks, John.
Hassane, maybe we just start with kind of the big picture, kind of cycle question. I think, in previous quarters, your basically message is, "Hey, I'm not going to call it till I see it." You clearly are signaling that you're seeing clear levels of stabilization. Obviously, the next part of the cycle would be the recovery stage. Maybe just walk us through what you're seeing that's giving you confidence to call stabilization in the business. Then how are you seeing kind of the recovery play out for you at this point in time?
Yeah, look, there's not one KPI that we look at that will say, "Okay, this is why I call a stabilization," or "This is when I would call or start talking about a recovery." It is a multitude of signs. You can think about it from the fill rate for the current quarter versus the same fill rate at the same time last quarter, turns needed. You look at changes in ordering patterns. There's a lot of data we look at. Both Thad and I look at it along with our operations team to make sure that we're not, one, getting kind of a false sense. With the disruptions and the uncertainty outside, any small sign could not mean anything or could mean everything.
That's why we have to look at all of these patterns, compare them to historical as far as we can, because even historical is not a clean comp anymore. Look at it from what is the customer doing. Forget about what the supply chain, what inventory, and so on. What does the customer want to do? That's the stabilization that we talk about. In order to get from that to me talking about a recovery, it has to be more sustainable signs. What I mean by that, it all lands on the end demand. At the end of the day, end demand drives everything from the supply chain. If you think about it, you think with the inventory depletion that customers have done, we would be already talking about the fulfillment cycle where you have nothing in the supply chain. You start getting the replenishment.
That will start getting the order patterns more healthy, and then a recovery will drive that. We're not even seeing that yet. It's not degrading, that's the stabilization I talk about. I have to start seeing demand driving replenishment, replenishment driving order pattern. That's what I would call a sustainable outlook and sustainable demand.
Got it. Thank you. Can you talk about just, you know, obviously, you know, automotive is a really important end market for you. It's been historically a very important part of your growth story. Can you, and obviously your background is deeply rooted in the automotive market, just give us your perspective of what you're seeing within the auto market, maybe talk about geographic trends, and then what you're seeing within EV versus non-EV at this point.
Yeah, so overall, the automotive market is more of a geographical discussion at this point. I'll take it by geography. What we've talked about is, overall, as a market for us, automotive was down in Q2. We called that being the bottom, and we've already talked about Q3 being up in automotive. Now, within that environment, even in the second quarter where automotive was down for us, China automotive has grown. China overall was up 23%, driven primarily by the automotive market. We expected that. If you recall in the Q1 call, I talked about the design wins and really the platforms that were at the Shanghai Auto Show that we are in. We expected China to be good from automotive, specifically driving the EV. Automotive in Europe and the U.S. is still weak. That's from a geography.
Now, if I go on submarket, specifically on EV, EV is still growing. Demand, and I would call it xEV. You got the battery electric vehicle, which is kind of what's the primary push for us specifically when we talk about EVs with silicon carbide. Right now, you hear also from our side, but also in the market, talking about plug-in hybrids. We've announced a big platform win that we are partnering with the Schaeffler Group for a plug-in hybrid for a large global OEM. Even in China, we have a new platform for plug-in hybrids in China. That's why I call it xEV. It's not just a BEV, but it's also a plug-in hybrid. That's still growing year-on-year. It's not growing at the expected rate that we all thought, you know, three years ago, but it's still growing.
From an EV, you have the secular growth still happening, but it's distributed differently on the geographical. It's more China. China's got the highest penetration. On top of all of that, from a technology specifically, if you think about silicon carbide, the silicon carbide penetration in the existing EV is still very low. All of these together, we still believe EV and silicon carbide, or EV, and even in some cases, IGBT, is a secular driver for us. We'll continue to see that.
Great. One of the things I'm wondering what your thoughts are. It seems like obviously the Chinese OEMs are doing extremely well in this market. It seems like a lot of it is at the expense of some of the European luxury automakers. Do you think we've gotten to the point within the auto market where it's turning into a zero-sum game? When you think about this strategically internally, are you starting to kind of think about, should we be focusing more on the China market at this point, just given the trends that we're seeing in terms of share shifts going on?
I think share shifts always happen in a technology dislocation, and EV is a technology dislocation. We know China started the EV conversion way ahead of everybody else. There are some share shifts in China. We know that. The European OEMs talk about it. Even the North America OEMs talk about it. Japanese OEMs talk about it. That's not what's driving the growth. That's not a zero-sum game. When you have to look at the Chinese OEMs and China EVs in general, a lot of people look at China EVs as domestic and China. I look at China EV as a global footprint. If you, you know, Australia is all China EVs. That's not a domestic market, but that's still volume out of the China OEM. Africa, South America, you go anywhere outside, even Europe, but you go outside the U.S., China EVs are an almost dominant EV.
I look at it, it's not a zero-sum game because all we hear about is what's happening between Western OEMs and Chinese OEMs in China. To me, yeah, that may be a little bit of a zero-sum game, but that's not a market driver. What's the market driver on the upside or market driver on the positive trend is China outside of China. We still have upside in China because it's not 100% penetrated, right? From an EV perspective, there's still a long way to go in China, but also China's doing the globalization of their manufacturing. There's a lot of volume outside of China as well. That's how you need to think about the China automotive market and why it's important.
Great. I had a follow-up question on the cycle for Thad. I know you spent a lot of time looking at your turns-based business and your booking trends and your book-to-bill. I think one of your peers had recently commented that the full second half recovery picture is a little tough to call right now because there's a lot of just short-term turns-based businesses because lead times are so short right now. There's a little bit less activity going into backlog at this point in time. How are you seeing things on your end? Would you kind of agree with that statement? Is that also making things a little bit more difficult to call on in terms of just the second half and the recovery?
Yeah. First of all, I think Hassane mentioned this. We entered the quarter in a better position than we did 90 days ago. That gives you confidence, right? When we think about the turns business we need this quarter, it's less than what we needed last quarter to meet our guide. That makes you feel good. We look at cancellations, we look at order bookings patterns, things like that. If you go back to early Q2 when the tariffs were initially announced, you saw a pause, right? Then customers kind of got on with it later in the quarter and kind of moved forward. I think as we sit here today and we look at it and we listen to what our customers are telling us, look at that order pattern, we feel like the second half is still going to outgrow the first half.
There's a lot of uncertainty out there. Our customers are telling us that. I think they're being very cautious. Our lead times are around 14 weeks. That's short for us, right? I do see the backlog layering in probably at a faster pace than it was 90 days ago, and especially 120 days ago. That gives us confidence that, as Hassane said, we're not calling it a recovery. We're seeing stabilization, but we think the second half outgrows the first half just based on the order patterns that we can see.
Okay. Thank you. Hassane, just on Silicon Carbide, you talked about the penetration rate still being relatively low. Can you talk about what is the silicon carbide penetration rate that you're seeing at this point in time? I know people are always really nervous about the China competitive threat. I think on the traction inverter side, you clearly still have a lead, but I think there's a lot of angst about the progress that the Chinese are making on the substrate side. I know you keep track of that. I think you sample some other volumes. Maybe just talk about that.
Yeah, so the penetration, and you have to think about penetration two ways. The penetration I referenced being very low is on the production today. You know, cars in production or platforms in production. Now, if you look at design-in activity or RFQs that are being out, we're talking about a 90% penetration rate. We know that the growth of silicon carbide, that secular growth that I talk about, is happening just because the RFQs are primarily silicon carbide today. That's going to go naturally. Even if EVs are flat, there's more EVs with silicon carbide than there were in the prior RFQ cycle. From a competitive thing, you're absolutely right. We have a lead. If you talk about the competitive environment, are there companies in China that can make silicon carbide today? Devices. Yes, of course. It's not, you know, it's a transistor.
To make the best transistor, that's where the IP and that's where the know-how comes in. We today still make the best transistor. Our new generation trench that I talked about on the call is ahead of not just the Chinese device maker. I'm competing with European and North American device maker, and we're winning against them. That gives you kind of where we are on the scale of technology. That's really what customers buy into is the device and the performance of the device. Therefore, a lot of that concern on the substrate side, it doesn't really impact us, because we don't sell substrates in the open market. We have sampled, we have qualified, Chinese substrate vendor along with non-Chinese substrate vendor along with our internal substrates. We believe we have kind of the best flexibility as far as what substrates to use.
As costs evolve because of the supply, we're just going to go with the best quality and best costs that we can get. Even if the mix changes internal or external, we have the best flexibility to position us. One thing I always say is having the internal and external is good risk mitigation or continuity because look at the uncertainty we're talking about on tariffs, and it's all fun and games to be able to get all the substrate you want from China until you can't. The question is, what do you do then? We have to be able to manage all of that uncertainty, and we're managing it with supply resilience and supply assurance strategy, which is where a mix of inside and outside. Quality is improving, supply is improving.
Like I said, we have qualified some, so we're in the best spot where we can take advantage of that availability now. It's not really a threat because we don't compete on substrates. We don't sell our own substrates in the open market. For us, it's the best of both worlds.
Great. Other questions? Okay. A couple of follow-ups, Hassane. You mentioned the move to Trench FET from Planar. Many of your competitors also have Trench FETs as well. What is unique and differentiated about your trench versus your competitors?
It's IP. I've always said at the end of the day, even before some of our competitors and some of our peers had trench before we introduced ours, yet we were winning. Why is that? Our Planar technology was better than the trench. People get hung up too much on, oh, trench versus Planar. I hear trench is better, so if you're on Planar, you're kind of falling behind. The answer is a customer doesn't really care. The conversation with the customer is not like, is it, I want your Planar versus your trench. The customer is, I want this efficiency. I want this resistance. How you get it, it's your business. It's our know-how. Our Planar resistance was better than our competitors' trench already. Having introduced trench for us is not the breakthrough of trench. Our trench is now a leapfrog from what's available in the market.
Does that make sense? Planar or trench were equivalent. Our trench is that much more superior than what's available in the market. That's the difference. It's not the trench itself. Our fourth-generation is on trench, and it's much better. We're already winning. Like the plug-in hybrid. You think, oh, plug-in hybrid is more IGBT or not. Plug-in hybrid was on our trench because it gets you even more efficiency when you don't have a lot of battery in a plug-in hybrid. You want whatever battery you have to take you the distance. That's where our efficiency comes in. That's why we get selected. Those are what matters to the customer. Range or efficiency that drives range. That's the only thing our customers can monetize. That's why they pick us.
Great. Speaking of range, you talked about the transition to 800 Volt. From a financial perspective, is there a benefit to you in terms of just higher ASPs, better gross margins? What's the implication there of 800 Volt?
No, if you think about it, the ASP, you have to look at an ASP per, what we, the way we normalize it, an ASP per kilowatt. The ASP is really driven by, from our side. Of course, you know, 1,200 volt device is more expensive than, or higher ASP than its equivalent 750 volt that we use, or 650 volt that we use with a 400 volt battery. There is an incremental, but the primary advantage is really dollars per kilowatt. The more power in the car, the more ASP we can drive because there's more content to get that power, especially if we go into modules, then that ASP gives you an uplift. There are other indicators that get us the ASPs that we want. 1,200 versus lower voltage is one of them, one of those components, but it's not the only one.
Got it. Oh yeah, one other follow-up on silicon carbide before we move on is, as you know, one of your U.S. silicon carbide tiers is really going through some financial difficulty at this point. Is this helping drive additional engagements and potentially could lead to future market share gains for you in silicon carbide?
The struggle, I guess I want to say the short-term struggle, which is, you know, the bankruptcy and all of that, that's not really what's driving the increase. If you recall, it started two years ago when the first kind of fumble happened, where the yields crashed. They had some, there were unfortunately some, execution issues in their scale and their go-to-market. That's when the conversion started. The financial struggles were just the last kind of thing that happened. Customers started diversifying away from them. We've won some designs. We've won a lot in China. We've won in North America. That's more on, you know, from two years ago when the cage was rattled and customer confidence was shaken because they couldn't deliver, they couldn't scale, they couldn't ramp. That's when customers started doing the work. It's already embedded in the outlook. The financial struggle was not a trigger.
It was just the last thing.
Okay, moving on to data center, you said the AI data center doubled for you guys on a year-to-year basis, which is pretty impressive. Can you just give us a little bit more color on kind of the scale of this business? Where in the power chain are you participating right now? You know, how big could this business be for you going forward?
Yeah, we haven't broken out the business yet. What I've talked about is where the growth is coming from and how we are coming at it. If you recall, when we started the AI journey, I said we're coming at it from the high power side, from the PSU, from the battery back, and so on. That's where we came in, especially as the racks get into the over 100 kilowatt, it falls right in our wheelhouse. It's the start. People talk about now 800 volts as this is the breakthrough. We've been doing 800 volts in automotive with our devices for a few years now. That fits. We're coming at it from here. Over the last few years, we've been introducing more controllers, more POL and SPSs, and so on. I refer to the 5x5 SPS, for example. We're getting closer to the core.
Those are kind of what we call the Power Tree. On our IR page, we have a nice description of the power tree as we see it. Of course, it evolves with every new generation of compute. If you think about it as a power tree, where and what products we give each. We started with the high power and we're making our way in. The growth is really across all of it. We also have parts anywhere in between. As you get to the 48 volts, we do 48 volts for automotive. High reliability, always on. We're into those products, are also available in the AI data center and they're ramping. We're very happy with our portfolio and how to map onto the power tree. We're coming at it from high power and already introducing with Trail, kind of the core level PMIC.
Okay. On the 5x5 side, do you already have design wins in place there that you expect to ramp?
I won't specify the 5x5 because it's very specific to platform. I'll just say we are in production and we're sampling the dual.
Okay. Maybe another way to come at it is you also put out a press release about maybe some work you're doing with NVIDIA. Can you just give us a little bit of background of just what drove that press release and how you guys have been engaging with each other?
Look, what drove the press release, of course, is the engagement. I think that the details of the engagement are contained in the press release. I'll just keep the comments of what we talked in the press release. In general, what drives engagements like this one is technology. As customers in general, as customers, as same in automotive, as customers try to work on their next generation, what they're architecting, they always have to get from technology companies like us, "Hey, what's the latest?" If you go on the website, you're not going to get the latest. The website is already in production, it's already designed in. We're innovating on things that are not even out yet. If the customer comes in and says, "Hey, we're thinking about this for next generation," you may end up in a case you go, "Hold on a second.
We have a part that we haven't really announced yet that fits right in." Now you have an alpha customer. We always work with alpha customers on our new platforms because you always want a teaching customer. That's how it works. By the time the customer announces the architecture, there are products that actually support it versus here's a great architecture, wait two years to get the products out. That's the engagement that we're looking for in architecturally heavy discussions. Traction Inverter is the same. Nobody goes off the shelf, whatever you got, give me the silicon carbide you have because every car is different.
Thad, just can you give us an update on kind of the low margin business you're planning to exit, and maybe just talk about as part of that bucket of business you're planning to exit, how much of that is ISG, which you talked about on the recent call?
Yeah, on our last call, we talked about 5% of our 2025 revenue wouldn't repeat in 2026. That included Intelligent Sensing Group, it included our planned exits, and then just EOL of some legacy products. We wanted to make sure we scoped that for everybody. We've been talking about these initiatives for a number of times. This is really putting kind of a box around it to quantify it for everybody. For the Image Sensing Group, that's about $50 million- $100 million that we'd walk away from next year. There's also about $100 million of the exits that we had planned. If you remember, we walked into this year saying we thought there'd be $300 million of exits. Year to date, we're about $100 million of exits so far. It's taking longer. We've undercalled this for three years now. It's taken longer.
Maybe that's a good thing because customers are sticking with this. We still do plan on exiting that. We think this year it'll be $200 million rather than $300 million, which pushes about $100 million over into next year as well, just on the planned exits. The other piece is the legacy kind of EOL products that we're going to have, which makes up the delta to that roughly $300 million exit. I would tell you that we think those are unplanned, but again, we've kind of undercalled that. I think we've kind of hedged that. Now, the Image Sensing is just the repositioning of what we talked about in that business for a long time. It's just a headwind that we've talked about and really just trying to make sure that the investor base and the street knew what the magnitude of that was.
Right.
Yeah, if you think about it, just to add to Thad's point, two years ago in our analyst day, I talked about repositioning image sensors to machine vision. That's where the value is, the quality of the products and so on. That's putting a number around it. If we take a step back, John, the important thing is, people ask, what, why now? Why are we talking about it with that, not just determinism, but with the scope of it? If you look at the rest of our business, we talked about AI doubling year-on-year, for the second quarter in a row. We talked about Trail, first revenue coming in in Q1 instead of the second half of this year. We talked about 5 million units shipped or over 5 million units.
Everything we've been investing in for the last three years is starting to drive the growth in the right markets that we want. However, you're not seeing it at the top line because there's that headwind that's happening. We said, we almost have two different businesses within one. One is the headwind and one is growth in the right areas with the right margin. Think about Trail as 60% -70% margin. For us to be able to really position the company as a value company, which is what we've been investing in, you have to put this wrapper around this box, like Thad called it, around that headwind and say, yeah, that's going to happen. Don't be surprised. I don't want to get hung up next year on, oh my God, your image sensing business is going down $50 million. What's going on? What's broken?
No, we expected that, it's scoped. The rest of the business is going to be growing above market. That's the clarity on the value and the R&D dollars we've been putting in for the last three years. By the way, while we do all of this, even in 2026, we'll be generating enough for cash-free cash flow. We've committed and we've upped our share buyback to 100%. We've done 107% year to date. We're going to continue that buyback. The transition is really positioning out to the company we've been investing into to be in. Now we're going to get that, we want a clean kind of separation of the legacy stuff that's the headwind versus the growth value stuff.
Great. Looks like we're out of time. Thank you, guys.
Thanks, John. Sure, thank you.