ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

Great. Thanks, everybody. I'm Joe Moore, again, Morgan Stanley semiconductor research. Very happy to have the executive team from ON Semiconductor, Hassane El-Khoury, CEO; and Thad Trent, CFO.

So I wonder if you guys could just start out. You've overseen a pretty big transformation. This company has much higher profitability than we had seen historically. You're going even through a period of low utilization now and you're maintaining those margins. Can you talk about what you did to get us to this point, and then we'll go into some of the details?

Sure. The last couple of years, we've been dealing with a lot of -- obviously, we've had the height of the pandemic, which saw a lot of shortages that we were able to support customers, so we took the opportunity to accelerate the transformation. So some of the components that we've done is, one, we walked away from a lot of the business that you typically would see as commodities, where our focus right now on our portfolio is being able to extract the value and therefore, the customers see value. If customers don't see value in the product, that's a product that we're not going to continue, number one, to support; and two, to invest R&D dollars in. So we went through a very big, large portfolio focus, portfolio rationalization. Following that, once you have the right portfolio, then you go and say, okay, what is the manufacturing footprint that we need to support that growth? And due to that, we ended up divesting 4 subscale fabs and acquiring a 12-inch fab. So again, a mix shift, a more healthy 12-inch. I guess I got the microphone.

Yes, I think it's coming in and out.

All right. So divesting the 4 subscale fabs put us in a position today that with the utilization being in the mid-60s, it doesn't have the same fixed cost absorption that we need to have, and that lifted the margin to, what you said, the mid-40. So a lot of it is execution, a lot of it is portfolio and a lot of it is product focus. To put it all in perspective, the mid-60s utilization that we are seeing today, the last time we saw that was in 2020. And at that time, the margin was at 30%. So we're already at the same utilization rate. However, our margin is holding up in the mid-40, which is the floor that we've said a while back.

Great. And it has been impressive definitely. Maybe we can walk through some of the drivers of the product transformation that you talked about, starting with silicon carbide. You've had a really good progression there, I think, up about 4x last year to $800 million. You talked about growing 2x the market rate this year. I know that you didn't end quite the way you wanted to. But can you just talk generally to your progress in silicon carbide?

Yes. When we started on the silicon carbide journey, we've had a few guiding principles and a few things that we've been very consistent about. And one is you're only going to win in silicon carbide when you have the best technology. And we define technology as being the devices and the packaging because both together, when you're talking about high-powered devices, you have to be very competitive on the efficiency side, that's the device, and you have to be able to get the heat off the device, that's the package. That's what allowed us to win most, if not all, of the designs at OEMs. And because of that superior performance both on device and package and, therefore, system-level performance translating to the OEM, we're able to capitalize on that, and you've seen the growth at 4x. And to put it in perspective, the growth, the 4x from $200 million to over $800 million that we saw from '22 to '23 is the highest growth percent in the industry and the highest dollar growth of all our peers. So it is not just because the market is a hot market or a growth market. We capitalized on the technology. We've increased share, and we went from #6 to #2. That's the momentum that's going to carry us forward as the mega trend of electrification continues in the future.

Yes. Maybe you could talk to the mega trend a little bit. As we've sort of seen sentiment fade a little bit on EVs in the United States, our auto team is pretty negative on that. But we're also seeing a very significant investment in China in EV. There continues to be a lot of momentum. Can you talk about those different forces? And is this an inevitability to down the road that internal combustion engines kind of fade away as part of the mix?

Yes. And that's the -- I guess, to define a mega trend, you can -- you're not going to draw a straight line to get from where we are to where I would say the market is going to take us. So let's talk about where the market is and why this is a mega trend that we're investing in and why it's not really a short term. And in my view, short term, '24, '25 is kind of short term. The investment thesis is a much longer view here. Why do we feel that? The penetration rate, think about it at the end of the decade is going to be maybe 50%, okay? We can argue maybe it's 2032, not 2030, regardless. It's only going to be halfway there by the end of this decade, so you have a runway of a decade-plus. So that's the investment, that's the mega trend that you're going into. Now is it inevitable? Yes. Is it going to be lumpy on the way there? Yes. What we're dealing with today is what I call a negative catalyst, meaning, okay, there's a price point disadvantage of EVs today. A lot of people are not going to write a check for $100,000 for an EV. And with the rates where they are, there's an affordability thing. That doesn't mean EVs are not going to remain the wave of the future of mobility. What needs to happen is the positive catalyst, which is a different price point or optionality where if you have a $30,000 to $40,000, maybe $50,000 vehicle, you're more likely to make the jump to EV. If you have better battery technology that allows fast charging, that is a positive catalyst. The last few years, we've had positive catalysts in electrification where it was ahead of what we thought EVs are going to be. Now it's slightly below what we thought. The positive catalyst is going to push it above what we thought. That's that lumpiness but the trajectory is up and to the right still.

Okay, great. And then your commentary of growing 2x the market growth this year, can you talk about the gating factors there? Because it seems like you're supply constrained, so you're constrained more by supply rather than demand growth, but you're also talking in terms of demand growth, like 2x demand growth would be great, don't get me wrong, but just trying to understand how we should think about the upside and downside to that.

Yes. Our supply is, I guess, intentionally tied to the demand, meaning we're not installing supply or capacity unless we see credible demand. So by definition, we're always supply constrained or we're always sold out if I look at it the other way because we're doing brownfield investment. And brownfield investment, in our view is, one, the most reactive to market. We don't have to build a $5 billion shell and then work for the next 5, 6 years dealing with the ups and downs in order to fill it while you're dragging margins down. What we do is we have a brownfield footprint. And within that same footprint, we will increase capacity as we need it. So therefore, you don't see that large impact for us on utilization from outlook in the market in, call it, '24 and '25. '24 is kind of -- the designs are qualified, customers are ramping. The primary question is, is the end demand going to materialize the way we expected? But there is no designs that we have to win now to make 2024. It's all done. We know exactly what everybody needs, and we're already starting to work on that. Right now, RFQs or engagements and design wins that we're doing are primarily, we're talking '27, '28. That's how far visibility we have.

Okay, great. And one of the dynamics in silicon carbide is ramping your own internal substrates through the GTAT acquisition. You got your milestones there of majority output by the end of last year. Where are you going forward? It seems like you're always going to have some that's outsourced. How do you think about that balance?

Yes, our view has always been, we're not going to be 100% in-sourced. We're going to be, think about 70% to 80% internal. That leaves about 20% to 30% outside. And we think that's a good mix in order to get a quick ramp or flex capacity externally rather than having to install capacity internally for a peak or a ramp. We can fluctuate and we can modulate externally for what we need while we maintain our internal capacity, our internal substrate manufacturing fully utilized, that's a good mix for us historically in other technologies, but also it gives us a market read of what's going on out there. What is the quality of the substrate in the merchant market? Is that going to be different? So it gives us a read versus heads down only internal. So we're not blind to kind of the world around us, but we still believe vertical integration is a competitive advantage. One, from a supply assurance, from a customer side, we are able to give the supply assurance that customers do need. And two, unfortunately in the short term, you have the geopolitical risk, where you can talk about substrate sourcing in China. But if that's your only source or a primary source, you're not adding a lot of comfort to an OEM in Europe or an OEM in North America because of geopolitical uncertainty. So you have to overcome these 2 hurdles, and we believe the mix of internal, external services that.

Great. And my last silicon carbide question, I mean, you mentioned China. A lot of the EV momentum has shifted to the Chinese OEMs. You've talked about having LTAs with 4 of the top 5, I believe, is the right number. How do you see that opportunity for ON? And do you worry about, over time, losing market share to Chinese device suppliers?

So I look at -- so you're right, a big opportunity in China. If you look at percent penetration and units, the biggest market is in China. We talked about silicon carbide, 50% is Asia. That's China, Korea, that's for us. So it is a big focus because that's where the market is today. Over time, I don't worry about the share loss or design loss because it's China. That's not -- we look at China and we look at design wins, whether it's China or Europe or North America exactly the same way. You're only going to win if you have the better technology, whether it's against a European peer, whether it's against an American or whether it's against a domestic Chinese. Because people think about, "oh, well, if it's going to be available in China, it's going to be cheap, it's going to be commoditized." Okay, well, if it's commoditized, then it doesn't really matter because that's -- we're not going to play there. And therefore, if you tell me it could be commoditized, then you're not adding value. Then it's not a business that's going to drive the margin regardless of what your cost structure is. Therefore, our focus is how do we add value? So when we talk about competitive advantage on technology, what does that translate to for everybody here? We talk about percent efficiency. You say, okay, well, great, that's a data sheet or a number on a PowerPoint. Percent efficiency in a car, if I give a customer 5% to 10% efficiency gain on a silicon carbide module, that same OEM can get 5% to 10% less battery volume in the vehicle. Battery is the biggest dollar content for a car. 5% to 10% benefit on a price of a battery because of lower volume while maintaining the range is a -- it's a given. We're not talking about the price of silicon carbide at that point. We're talking about how can we get more efficiency so they can save it on the battery. As long as that's the formula that we use to win in end market and a design, it doesn't matter where the competition is coming from. That's, at the end of the day, how we win in Europe and how we win in North America.

And you're going to get wins where the market is evolving?

That's right.

Okay. Maybe if we could talk about the non-silicon carbide portion of your business on the auto side. You guys had signaled some weakness in the automotive business about a quarter ago. I think that's becoming increasingly a consensus so you were kind of early to see that. Can you talk about that? What's your visibility on how long that might last?

Yes. We've always said our LTSAs provide us a phone call. As soon as the customer is going to see some softness or some changes in the demand environment, we're going to get the phone call because it's legally binding. They don't want to be stuck with inventory. We don't want them stuck with inventory so we're going to get the call. That happened, just to rewind further than last Q3, Q4, we saw the industrial softness because of LTSAs in end of '22. And then it took about 6 months for people, middle of '23, for people to start talking about how industrial is getting soft. So fast forward to Q3, I talked about we're starting to see softness in, I called the European Tier 1 at the beginning, because they -- from a market share perspective, they're a big indicator for me of a worldwide auto demand picture. And of course, at the time, if you remember, everybody is like, no way, automotive is still going to grow. It's going to grow next year. And then fast forward to CES, all the pre-announcements started and all the guide-downs. So we did see it earlier. What does that mean for us is the LTSAs work and customers see those as strategic because we are working constructively with all of them. Where do we see that moving forward? That really depends on demand. At this point, yes, there's some inventory digestion, call it, in the current quarter, maybe next quarter. But it's more of a demand picture at this point because inventory, whether it's elevated or not, is purely a denominator of what the demand is going to do.

Yes. And there was a notion, it never really came from you guys but from others that the auto OEMs would direct Tier 1s to build large buffer stocks, large safety stocks of inventory. You've been pretty consistent that the LTAs are designed so that people don't need to do that. But that idea seems to be fading a little bit anyway. So do you think that people will run with -- we had 2 years of very severe shortages that were very impactful. It seems logical to run with a little bit more inventory, but do you think that's just not happening?

So it's -- I would say it's bimodal bifurcation here. That is still what the OEMs want. However, if you look at the financial situation of a lot of the Tier 1s, it's thin margin and therefore, you can't carry a lot of inventory, right? So when things are tough and demand is slow, you're not going to strap your cash in inventory so they're reducing inventory. The issue with that is that's exactly what happened when COVID hit. COVID hit, demand went down. Everybody drained the inventory to a subscale level, and they struggled with the way up. However, OEMs learned a very big lesson. Our LTSAs today are most of them directly with the OEM, which means there is no judging in between. We do the judging. It's not a judge upon a judge. So we get the numbers from the OEMs. We will challenge the numbers with the OEM, and then the OEM will do either direct or directed by. Directed by meaning, "Hey, Mr. OEM, you tell me who do you want me -- which Tier 1 you want me to ship to." But the OEMs are seeing inventory levels getting drained below critical levels at the Tier 1s, and they're starting that dialogue either directly with the Tier 1 or directly with us to not let it happen because they understand when the demand signals happen, we're 6 months away. And congratulations, we're back to where we were in '21.

So you think -- so in Q1, you're relatively -- you're shipping kind of to end demand. And so the question from here is a function of end demand, is that right?

So no, Q1 is not end demand. We're still draining in Q1. What we talked about on the last call, Thad mentioned, we talked about normalization. So we're not planning on a recovery in the second half of the year but more of a normalization of demand. So it's not in the first -- it's not in Q1.

Yes, okay. Okay, great. And then can you talk about some of the other growth markets in autos? Image sensors is a big one that you guys have focused on. Can you talk about your progress there?

Yes. So for ADAS, obviously, we have a leadership position with our image sensor. That here, the growth is stemming from a couple of things. One is penetration of ADAS in general vehicles. So more and more vehicles have we call it, Level 2+ autonomy. Therefore, you need a camera or image sensor content. But within that, you have an ASP uplift. As we migrate from, call it, an average of 1-, 2-, 3-megapixel cameras and you go more into the 8-megapixel camera, that is an ASP uplift. That's over a 2x ASP uplift just by doing a 1:1 swap with higher resolution. So that's already. So our revenue in 8-megapixel in Q4 doubled year-on-year. So you're starting to see that uplift. So the growth is coming from these 2 areas. We'll continue to innovate, we'll continue to introduce competitive products specifically tailored to automotive to capture the content and capture the technology uplift as those cameras get adopted in more vehicles.

Great. So can you talk about pricing in the automotive market? Is this an area where you talked about getting out of lower-margin commodity businesses? Was some of that in autos? And what do you see going forward? Do you see in markets like sensors or power more pricing aggression?

Yes. Some of it was in automotive but most of it was in non-auto and -- because it's not a -- it wasn't a market-driven or product-driven necessarily. Some of it was market-driven back to the value. If a consumer company doesn't really care about the value we provide from a power device, they're not going to put the value nor pay for it, but the value in automotive can be extracted and we're able to do that. Moving forward, however, most of our revenue is tied in LTSA, and that's both volume and pricing. And look, the conversation is not about pricing. We're having a lot of dialogue about LTSA, about relief on price, on volume because of end demand. And those are dialogues. Of course, we're very open to because it's not to our advantage to kind of blindly ship to -- because the bills will come due. So you want to be as close to demand as possible, but it's not a pricing discussion because we made the commitment to support the customer and the customer made the commitment to pull. The other thing is we can talk about pricing, but pricing in general has not come down. And talk about energy, talk about wages, talk about all the things that went up during COVID has not really come down.

I would also point to, when we talked about exiting that business, we planned on exiting $800 million and $900 million over a 3-year period. To date, it's been $475 million. So we haven't exited what we thought. We priced ourselves out of that market, but the fact is pricing isn't sensitive and those customers aren't leaving. Today, that business is all at favorable margins. So we're going to stop talking about exits. It's good business, we're keeping it at this point. But if the pricing was soft, those customers would have found another supplier.

Yes. So maybe talk to that dynamic a little bit. I mean, this has been a big element of the margin improvement that you guys have seen. And now you're kind of signaling, hey, we're not walking away from anymore because it's just good margin business. Any prospect that, that could change? I mean, you thought it was a business that would fluctuate a year ago. Why wouldn't it kind of get -- if we get through a protracted downturn, start to see that?

Well, if it's a -- we would have seen it accelerate, right? Even in the fourth quarter, we didn't meet or even come close to kind of what we thought. So in an environment where we talk about a lot of companies that could support it have underutilization. So there very likely that they could have already dropped the price to get the revenue because we've been very public about our position. Yet that didn't happen. I can posture what the reason is but there's value. Maybe it's not a value where the product sees value, but there is a value for onsemi shipping that. Customers get it when they need to. They don't have to worry about it. It's high-quality product. So when -- and some of these products is, call it, $0.05. Okay, let's say we doubled it to $0.10 in a $300, it's going to cost them $1 million to change. Why change and go to potentially another risky supplier that has bad quality and jeopardize the business? So again, it goes back to customers perceive the value and therefore, they pay for it.

You kind of priced it to value and then maybe you're surprised how much of it is sticky in the limit?

Yes, that's right. As long as we don't need that capacity for something else, it's good business at favorable margin.

Yes. But you're going to stay pretty disciplined about the margins of the business?

Absolutely. We will not chase that down if it does get soft somewhere, absolutely.

We have -- back to the first question, we have rationalized our footprint, manufacturing footprint, in order when we are where we are today, we are very disciplined in no more fab fillers. That's how margin gets degraded and there's no recovery. Because if you degrade it on the way down with fab filler, you already set the price and the value of your products. So we changed the footprint of the company from a manufacturing side so we don't have to do that.

Great. That makes sense. So I'll come back to the gross margins. But maybe you could talk to industrial a little bit. You mentioned you guys were earlier to see weakness there. What are you seeing now? And are you seeing -- there's silicon carbide in that market, too. Can you talk about what you're seeing in power in the industrial market as well?

Yes. So in industrial market, we did see the softness. We took utilization down early in '23. And really, we haven't talked about industrial kind of changing because it's been in our guide, it's been in our outlook. So when a lot of our peers are talking about kind of peak to trough within a quarter and we don't see it, it's not because we're blind to it. It's because we saw it before and it's already in our baseline. However, if you look at overall industrial for onsemi, the results are more muted because within industrial, we do have a lot of strength and growth. So in industrial, we have -- we talked about the electrification mega trend in automotive. There's a same mega trend in industrial for energy storage system or renewable energy. And the categories are energy storage systems. You have a charging infrastructure and you have renewable itself. And I'm talking not the residential, which impacted a lot of my peers, but more on the commercial and utility scale renewable. So these 2 categories have lifted our industrial result. I mean, that business for the last 2 years have grown 60% to 70% year-on-year for 2 years. So the softness in industrial is still there but it's kind of muted by stellar growth in that business. We have the medical business that's also seen a lot of growth within the industrial. So net-net, if you look at our industrial result at the top line, it's more muted than some of our peers because we do have a unique position of growth within the mega trends.

I mean, if you aggregate industrial across all of my coverage, we've seen, I would argue, the worst inventory correction that we've ever seen. I mean, you had worse numbers in 2001 and 2009, but there was demand destruction on a much bigger scale, whereas we're almost as weak now and it's all inventory. So it feels like it should be due for a decent snapback when we get to the other end of it?

That's -- you hear a lot of some of my peers when they talk about what their views on 2024. Let me give you my view before I venture and talk if it's due for a snapback. The view on what we're managing the company is there's no recovery in '24. And let me -- because I'd rather be wrong that way than plan for recovery and be wrong the other way. Because planning for kind of a no recovery in '24, we manage, we're very disciplined and very cautious about managing inventory, managing utilization and managing the expectation. If we're wrong and there's a recovery, utilization goes up, revenue goes up, margin goes up, profit goes up. It's all tailwind in this case. If we sit here and are planning on a recovery because it's due or the model says so and it doesn't happen, now I have elevated inventory and I still have to take utilization down. That's not a position we want to be in. So our approach is we're going to be more disciplined in our view of the market. We will react much faster to an uptick in the market rather than have to snap back if we didn't get it right.

Yes. I mean, that makes sense to me. I mean, it is also true, though, that your visibility is always worse at the bottom, like it could get better quickly. But you're ready to adapt to that if it does?

That's right.

Okay, that makes a lot of sense. Maybe if you could talk to the financials a little bit, Thad, in terms of gross margins, where you are today with the utilizations that you have. I would assume that gives you a lot of confidence in hitting your longer-term targets when that utilization comes back up. But can you talk to that generally?

Yes. So '24 is all about utilization. So our utilization is mid-60s. As Hassane said earlier, I think it's worth repeating. The last time we were at mid-60% utilization was in 2020 and the gross margin was 30%. We're currently there and we're holding a mid-40% gross margin. That's our floor on margin, we believe we can hold it. So if you think about where we go from here, our target is 53% long term. In 2022, we were at 49%. So you got utilization, which will be a nice tailwind when the market starts to recover or normalize. So you think about every point of utilization, it's 15 basis points of gross margin improvement. We have the East Fishkill fab that we acquired where we got the surprise on the cost structure there. We'll have that under control by the time we exit 2024. That's about 200 basis points currently dilutive to the company. And then we've exited the -- or divested the 4 fabs that we'll start to monetize those starting in '25. It's about $160 million on an annualized basis that starts to hit the P&L as we move that production in-house. So we leverage that fixed cost. So once we start producing those in-house, you start to see that $160 million start to kick in. We'll start to see that in '25 and beyond. And then anything that we have in R&D today, in development is at or above the corporate target. So longer term, we feel really good. You start adding those things up, you start getting pretty close to that 53% target.

So everything you're developing now is about 53%?

That's the goal.

Yes. And then from -- I know you -- over the course of last year, you had start-up expense in silicon carbide, which I appreciate you never tried to strip out your numbers. So the -- where are you now with that? I mean, you're kind of growing at more of a rate that's...

Yes, we're at the corporate average at this point with silicon carbide, right? So we had that headwind in '23. Now we've mitigated that. And we'll be adding some capacity in '24 for the '25 growth. So we'll probably be kind of running at that level for most of '24.

Okay. And is there still some benefit to migrating to the GTAT substrate within that silicon carbide?

Sure. Any time we're going to internal substrates is favorable for margin.

Yes, absolutely. Okay, okay. And then last question before I turn to the audience. Maybe if you could talk to CapEx, where you guys sit. I mean, obviously, utilization is pretty low but you're also talking about a bunch of factory consolidation, things like that, where do you stand on that?

Yes. So we laid out in '23 a plan where we would be high teens for a couple of years in terms of capital intensity. We're at 19% last year. We said that because we're performing so well in silicon carbide, we actually take that down in '24. So we're in the low teens. That's the target. Our long-term model is 11%. So we've narrowed in on that a couple of years earlier than planned. The reason for that is, obviously, silicon carbide is performing well. But when we think about bringing capacity on for silicon carbide, it's a very capital-light structure because our fabs are already 6-inch and 8-inch qualified. So we've got those up and running. And if you think about on the substrate side, it's a retrofit to the current furnaces. So it's not like we've got to go break ground on new facilities and bring equipment in. It's actually a retrofit. And so we can modulate that very carefully depending on how the market develops as well, so we can derisk the business on that side as well. So we feel really good about where we are on the capital and the investment on that side of the house. These are all brownfield investments so it's about 40% cheaper than doing a greenfield investment.

Yes, yes. That's great. All right. Good. Do we have any questions from the audience? Maybe just last question for me. Is there anything from a growth standpoint that you're excited about that you'd want to talk about that I didn't bring up?

Yes. So I briefly -- I'm going to give a little hint on the mixed signal analog. I briefly talked about it on the last call. Sudhir did a great job presenting it at Analyst Day, which is a market expansion that we're doing. It is a new technology platform that we're introducing. And to put it in focus, it's more of a cross-selling opportunity of the mixed signal. So don't think about going after kind of mixed signal analog as a whole. But next to every silicon carbide, MOSFET or a IGBT, there's a driver, there's a controller. Having these mixed-signal analog products jointly sold or jointly paired with our -- the silicon carbide is a high attach rate. That's the opportunity that we're going after. And the benefit for a customer because you say, okay, great. Well, if you have it, why would the customer change? Having our driver, for example, paired with our silicon carbide, you're able to drive the silicon carbide switch much differently and extract even more efficiency back to the savings on the battery. So it is a system-level benefit. That's the cross-sell. I'm very excited about that. We announced it at Analyst Day and where we sit today, we've already sampled products. So it's a very, I guess, benchmark-setting development time line. So I'm very excited and very proud of the team for having done it so quickly to be able to sample customers now. And that's going to proliferate and you'll hear more about it in the -- through 2024.

Yes, okay. Exciting stuff. All right, good. Well, we'll wrap it up there. Hassane, thanks.

Absolutely.

Thanks, Joe.