ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

I am Christopher Caso, Credit Suisse semiconductor analyst. Our next presentation is ON Semiconductor. It's a home game for you guys, just down the block. With us from ON is Hassane El-Khoury, CEO; and Thad Trent, CFO. So gentlemen, thanks for attending.

Thank you.

Thank you for having us.

I thought maybe you'd start, and ON's been a company that's been in a bit of transition. And thus far, that transition seems like it's worked out pretty well with how the stock has reacted and how the fundamentals have reacted. So maybe a minute to kind of describe that transformation and how the strategy of the company has changed.

Yes. Look, at a high level, when I walked into the company, I could see from a technology with any semiconductor company, you have to have the value in the baseline technology before you can talk about what do you do for a strategy. So the company has very solid and broad-based technology components that we were able to take and through a portfolio rationalization reinvigorate the strategy of auto and industrial, but more importantly, focus on power and sensing as the go-to-market, and that allowed us to then tack on to a lot of the megatrends that are in automotive like electrification, allowed us to double down on silicon carbide. You've seen a lot of that investment, whether it's CapEx or R&D. We repositioned our R&D investments into those high-growth areas, meaning we reduced the distractions. That was part of the restructuring that we've done in the beginning of '21. And along the way, we had to fix the running business. So as we invest in the future and we over-index to the growing mega trends in auto and industrial, we had to fix the baseline through price to value discrepancies, where we've had components or we've had products that were priced below market, even below our own similar products between markets or customers. So there what we call value leakage that we were able to also bridge and close that gap over the last couple of years. And what you've seen is obviously an expansion in our gross margin, which was every quarter, record high. Our growth has been sustained with the growth that we are seeing in the electrification of the vehicle and the factory automation and renewable energy. And really a more sustainable financial model that is obviously starting to get exercised as we look into 2023. And that's, again, the sustainability of the model that we've been working on in 2 years is to make sure that we have a model that we can get through good or bad times.

Right. Right. and I mean, right now, and obviously in the mind of everyone right now is sort of the transition that's going on in the industry because of the macro, where there's some areas of weakness, but then some areas which seem like they're pretty rock solid in auto, industrial. And the previous leakers we've had and, of course, through earnings season, it seems to be a consensus view among the companies we cover as those sectors have held up very well. And I guess the pushback from investors we just rolled out coverage recently is that in every other cycle, it's just kind of a question of not if it slows, but when. And interested in your perspective on that. And both the if and when and then what ON does if that happens, if industrial should weaken because of macro concerns?

Yes. Look, I don't look at the market at that level because if you look at it at that level, you're going to miss a lot of the nuances that are driving the growth. So honestly, you can peg and put your finger on exactly what is causing the growth, you're going to get sucked into the cycle and the macro which is historical. I don't think -- I think everybody will agree, we're in no way shape or form going to be pegged on any industrial type or any historical type market because there are a lot of mega trends. There are a lot of force on the road that we have seen and we will see more. So if you take it at a market by market, consumer and compute great last few years, everybody wanted multiple PCs at home. You had a lot of that. The softness, we started -- we started taking down our wafer starts to get ahead of it from an inventory back in the second quarter before anybody was talking about the softness in those markets because it wasn't the backlog. People ask, "Oh, my god, did you see cancellations and push up?" Well, the answer is no. What we saw is a slowing down of layering of new backlog, so like almost a second derivative of the backlog. Because look, the backlog is what it is. Do you trust it, do you not? It's way above what we can supply, so it doesn't really matter. So you want to make sure you move your allocation to where you can make an impact on the end demand. So we started taking that and we start seeing -- we continue to see that softness, and we're managing through it through making sure we get ahead of it from this T-inventory is at an all-time low for us. So we're managing to stay ahead of that. And our utilization is already baked into our number, back to the resilience of our model. So we got ahead of all out of that. You go to the industrial. Industrial is holding up. And again, but you have to look within industrial, what is holding up. Energy storage, renewable energy, energy generation, energy distribution, chargers, solar, wind and energy storage. Factory automation remains strong because what drove that strength is the fact that people couldn't get labor and social distancing, which reduced the throughput. So company started investing and accelerating their CapEx for factory automation, that's not going to slow down. Once you get started, you have to go through the whole. Some areas in industrial, like, for example, the things that are closer to the consumer like power tools. And that is going to -- that's softening. That's expected because it's stacked to the consumer and the demand. So that, again, we're planning for that. And with that, you can see the resilience of our model through that. Automotive is actually remains strong, and it's going to grow into '23. Now what is that growth driven by? Again, you have to go -- you double-click on automotive, you have industrial or you have electrification and you have ICE engines. So if you look at the 3 numbers that people need to watch for, you have the 2022 number of vehicles, you have the 2023 demand from OEM and then you have the 2023 supply that semiconductor companies are going to ship. And those numbers are stacked that way. Our ability to supply is not going to match the demand. We're still going to be supply constrained in automotive in 2023. So even if the demand kind of fluctuates, it's not the supply boundary. And that supply that we're -- as the semiconductor industry, we're going to be shipping into 2023 is higher than what we shipped in '22. So it's going to be a growth year, even if you're going to see fluctuation on the SAAR. Why? Because no matter what happens, OEMs are going to ship and manufacture electric vehicles, even if it's at the expense of an ICE engine -- ICE car. If you get 1 part whether it's power semiconductor or mixed signal analog, and you only get one. I guarantee you the OEM is going to put it in -- to make 1 more EV versus 1 more ICE car. That is going to fuel the growth. Because for us, specifically at ON Semi, the content is about a $50 content in an ICE for powertrain, going to $750 in an EV when you do the silicon carbide and all the power semiconductors that go with it. That delta is going to fuel the growth net of whatever the SAAR and the units are going to do.

Right, right. So that in the constrained environment, the ability to produce EVs is so far below what they want to do, that EV growth will continue. That makes sense. That makes sense. What about the fungibility of supply, though? And for your own business, there is some fungibility between what's happening in some of the consumer-oriented sectors in industrial and automotive. Has that allowed you to reallocate your production to those areas? And of course, you're deemphasizing some of those noncore businesses.

Yes. So both of those obviously played a big role. So we've been on a trajectory for going to a fab lighter strategy, which is resizing our manufacturing footprint for where we want to be as far as mix. But over the last few years, obviously, with supply constrained, we did prioritize our auto and industrial customers ahead of our noncore markets. And that allowed us. We just reported a 68% auto and industrial as a percent of revenue. And that's been kind of increasing that of course helped with our mix that was accretive to margin as well. So it's a double positive where you are now exposed to high-growth macro markets and you have a better margin mix. So we're -- we've used the last 2 years of supply constrained to move more and more supply. Not every -- it's not 100% fungibility. That's why I mentioned earlier that we took down wafer starts. Those technologies, we can't just move to automotive. But to the extent we can, we have. But nevertheless, as a company, when I talk about we have to be -- have a sustainable financial results in good and bad times, you have to be willing to work with utilization versus just build inventory to keep the fab full. That used to be the old ON Semiconductor. We're not in a fab filler product. We don't do fab filler. As a matter of fact, we're walking away from a lot of that business. We have more to walk away from. But we'll take utilization down before we burn cash on the balance sheet with inventory.

Right. Maybe pivot a bit to 1 of the drivers. Obviously, a big part of the story is silicon carbide. And maybe you could speak to that. And I think you've already kind of spoken to why electric vehicles will continue to grow regardless of what happens with SAAR. What about ON's own competitive advantage in that space? So what's your secret sauce in silicon carbide allows you to win.

So a couple of things. Obviously, we've been in the power semiconductor for over 2 decades. We've been -- we're a big worldwide player with IGBT, which gave us a very strong pedigree in both technology and packaging. So when you talk about 300, 400, 500 kilowatts with a die size that is shrinking and shrinking, how are you going to get that heat out. So packaging is actually a very big competitive advantage we have. When combined with a very good technology from RDSon or whatever spec you want to talk about, you have to have both. Because if you have the best device and you put it in a less than optimal package, you're not going to get the efficiency. You end up paying for silicon carbide, which is more, and you're not going to get the benefit because the heat is trapped in. So packaging and device are very important. That's where we win business. Now what we've done also strategically is we've acquired GTAT in -- it's been a year now. October was the 1-year milestone in order to also extend the vertical integration from packaging device all the way to substrate in order to get that supply assurance, which is very critical. If you look at where the market is going and the ramp and the speed of adoption that we're seeing in automotive and in industrial, that market is going to be constrained for probably the next decade. And so having supply assurance that a customer can visit the site, touch and feel the capacity that's being installed and get that confidence that as they ramp, we're able to ramp with them. There's no dependency that I have to go and hope that somebody is able to ramp. That goes away. So that vertical integration translates into supply assurance is another angle of not just why we win, why we win a lot of the majority shares at the big accounts. And that translates into the $4 billion of committed revenue under long-term supply agreements that, that customer has signed up for.

Great. So as you look at silicon carbide, and I think you talked about $1 billion run rate by the end of next year, what's the constraint? Is it actually a substrate that is constraining your ability of supply? I mean, it doesn't sound like electric vehicle demand is a constraint.

Yes, it's definitely not demand. Yes. It's -- at this point, it's execution. It starts with the substrates. I talked about -- since the acquisition of GTAT, our -- we have -- we will have end of this year, which 4 weeks left, 5x the number of furnaces installed that will be online. So from our capacity and our ability to add capacity, we're on track. Like I said, we have 4 weeks left. I'm not worried about it anymore. I don't lose sleep on do we have enough furnaces installed. Right now, it's the ramp, which when you do a 5x ramp in the 12-month period, I can tell you if somebody is doing that and sleeps good at night, they're missing it. That's where a lot of my focus is and the teams focus day-to-day is making sure that every issue you have doesn't become systemic, the blocking and tackling, which again, yes, it's silicon carbide, it's very difficult. But the process of a ramp, we've been doing for decades on ramping fabs and ramping manufacturing even in power. So from that, we already expanded capacity on wafering and Epi. So once you get the pull you have to be able to make the wafers and put Epi on them. We expanded that capacity that we have. And then the fab, we've doubled the capacity in '22 from last year. We'll double that again next year and that the equipment. So I want to say it's all execution, it's all planned. The question is until everything is in-house and I put my finger on it, I'm going to keep managing it day to day. But it's execution, but it's not a demand, which is when we get to an execution focus, I feel much better because then you have control.

That's something in your control.

That's right.

So the silicon carbide for the next several quarters, 1 year, 1.5 years, 2 years, it's really in your control in terms of your execution and getting the product out.

That's right.

Yes. Let me pivot a little bit to pricing, which has been a good story for the industry and for yourselves as well for much of last year. Maybe you could speak to 2 things with that. One is what's happening within the noncore segment with pricing. And I suppose that's one of the reasons why you're exiting some of the businesses that you've been talking about exiting for a while. And then secondly, what about in the core businesses? And if there is price pressure in compute and consumer, does that leak its way into industrial and automotive at some point?

Yes. So let me take it in 2 sections. So the noncore business, a lot of the revenue or the business that we have said we're going to walk away from, we've walked away from about $270 million at an average of 25% margin in a good pricing environment. So I wouldn't call a good pricing environment. That is somewhere that I'm high-fiving about it. It's better than it was, but it's still dilutive. So we walked away from that. Where we are today? We talked about another $400 million to $450 million that we're going to exit in 2023. That's going to be market dependent. If a customer is able to get it somewhere else cheaper, you're welcome, go for it. We're not going to chase it down to kind of single-digit margin like it used to be. So from that perspective, we don't see -- we're not going to be under any pricing pressure. As a matter of fact, as we lose this business, it's going to be accretive to margin because it's below the corporate average, right? So from that perspective, it's actually going to be a good thing for us. So softness in that market may accelerate us losing that business. I'll be honest with you, by now, I thought we'd be done with it. But it's going slower because they can't get it anywhere else. We expect that to change. Now on the core business, specifically on auto and industrial, you heard me talk about LTSAs. Just to remind you, LTSAs are multiyear in nature. We have LTSAs that go 4, 5 years, some of them are 7, 8 years that have both volume and pricing in it. So in core businesses where we are investing, there's not a conversation about pricing. The conversation about pricing has already occurred. Right now it's just making sure that we ramp because we've been very consistent and transparent about the fact that we're only adding capacity where we have LTSAs. So for us now is making sure that the LTSAs are what the customers need and that we're getting ahead of it by adding capacity to support those LTSAs. But call it whatever the demand does, there's not going to be a pricing conversation. It may be a volume conversation, but it's not going to be a, hey, I can get it somewhere else cheaper. Well, at that point, the LTSA is a legally binding agreement.

Right. And how much of core industrial and auto business is covered by those LTSAs?

It depends on the technology. I can tell you, for example, silicon carbide, 100% is LTSAs. Businesses where we are adding capacity are 100% on LTSA. We're not going to add -- we're not going to invest in CapEx unless there's a signed LTSA. So to a first order between LTSAs and NCNR, 2023, I mean, we're sold out already. So that gives you kind of an idea of the -- it's a big percent in the shorter term and then -- and we keep layering on top of it.

And aside from silicon carbide, where else are you looking to add capacity for 2023?

So in 2023, our CapEx is going, obviously, silicon carbide, but also a 300-millimeter fab. So we will close the East Fishkill, 300-millimeter fab acquisition end of this year, called December 31, we'll be owners of the fab on January 1. So that -- we're converting that fab into a power discrete and logic fab. So the second CapEx, call it, intensity is going into that fab. Because if you track the story, we've been -- I referred to the fab lighter. We've divested or announced divestiture of 4 fabs. We closed 3 of them already. One, we've announced the agreement when we expect to close in the fourth quarter. So by the end of this year, we would have exited 4 fabs of our fab network. Our capacity, when we put East Fishkill online, will actually increase. So net-net, we are adding capacity, but we're adding capacity at scale, and we're divesting the sub scale fabs. So those are the 2, I would say, components that are -- where a majority of our CapEx is going.

Great. I'll pause for a second and see if there's any questions from the audience before I continue. Rob?

Thanks. Appreciate it. I think 1 point of differentiation on your SiC strategy has been your investment in the air wafer-level burn-in systems. I think your competitors have not come public with such investments. Can you explain how, if at all, this is aiding in your success in winning kind of SiC business?

You're talking about wafer level burn-in?

Yes.

Well, I don't know what they're doing. I look at it, I don't think wafer-level burn-in is anything new in the industry. Any new industry, any new technology as you get to maturity and you flush out a lot of the stuff, you have to get a wafer level burn-in and you have to get out of some burn-in. Some technologies will remain. So I can't comment at that level because that's not really where the competitive advantage comes. Wafer level burn-in is a quality in PPM. It is not -- we're not going to win because we have wafer-level burn-in. I'll go back to -- you have to have the best devices and you have to have the best packaging. Those are the 2 key items why a company -- why we are winning. I can talk specifically for us. How we manufacture, how you improve all of that stuff, the manufacturing cost and manufacturing baseline has a lot of things between test, the burn-in and all of that.

Maybe I'll direct the next 1 on you with regard to margins. And there are some puts and takes on the gross margins right now, where, again, the -- what's happening with Fishkill, with what's happening with silicon carbide. Maybe you can walk us through what the puts and takes are. And it's my perception that the headwinds that you face on gross margins sort of peak toward the end of the year. So maybe you can confirm that and kind of put the road map in front of us.

Yes. So let me start off by just calibrating where we are on margins. So we started with margins down in the 33% range, right? We pushed them up pretty close to 50%. Now what you're referring to as we're going forward is we have 2 major headwinds in margins. And we've been very clear about these for a long time. We've got about 100 to 200 basis points of headwind with the silicon carbide ramp. So we have all start-up costs with silicon carbide. That will basically be there for -- through '23. We think by the time we exit '23, we're about at the corporate average on silicon carbide once we ramp. So we've always said that's about scaling. We don't exclude the start-up cost out of our non-GAAP numbers. So that's 1 element. The other element is the East Fishkill. So as we take over the fab as this onset, in January, we're going to be providing foundry services to GLOBALFOUNDRIES for 3 years. So next year, we'll have about 40% of the capacity, and it steps up and as GLOBALFOUNDRIES steps down over 3 years. So for next year, it's about 40 to 70 basis points of a headwind, which that steps down again over in '24 and '25. So those are the 2 main elements and then obviously depending on what the market does, there may be some underutilization, but that's going to be macro driven. Now as we go forward to '24, we've got a couple of tailwinds coming as well. So Hassane talked about the 4 fabs that we're exiting. There's about $160 million of annualized fixed costs that come off the gross margin line as we exit. It takes 3 to 5 years to totally exit.

When does that start? And when....

We'll start to see the real impact in '24. We don't see much of it in '23, just because we just inked these deals here recently. But the big impact where we start to get the tailwind is in '24, and we get the headwind kind of coming off of the 2 other items as well. And then obviously, any new products that we have coming out are accretive to gross margins as well. So anything that we're working on in R&D is accretive. So longer term, we feel good about we've got a nice tailwind coming at us as well after we get through '23 as I look at it as more of a transitional year.

Right. And what sort of levers do you have to pull if industrial, for example, should weaken if some of the customers ask give us a little bit of time, what sort of levers do you pull to stabilize those gross margins if that were to in fact occur?

Well, I think it's back to the discussion earlier, what's fungible, right? So if we can move this. Now a lot of that business is on LTSAs, right? So we've got good visibility. We're going to have our customers coming to us well in advance if there is an issue in softness of demand, which might be below their LTSA revenue. If that's the case, and we can move that capacity somewhere else. We'll do that, right? There's got to be a win-win for us and the customer through that process. But that's the main thing is being able to reallocate those -- that capacity somewhere else.

Yes. Bottom line, from the LTSA, at a high level, the guiding principle is any flexibility offered cannot be dilutive to our financials. Meaning if we have underloading because of it, then customers are going to have to offset that if they don't want to take the product. So our goal is not to blindly shove inventory into the customer. Because you just pushed the problem down the road. So we're not focused on short-term kind of benefit, had a long-term impact. We want the long-term stability of our go-to-market and our revenue and our margin. But it can't be margin dilutive because historically, we're left holding the backlog and disappear in 30 days before. Back to that point, we're going to get a call 6 months in advance when there's a problem, which gives us a lot more time to manage it as a win-win with the customer.

Right. And it's very different than the way the industry has for quite some time. I guess 1 last one, just carrying this back to cash flow. And there's still some investment. You're still in investment mode right now, particularly in silicon carbide and other. At what point does some of that capacity addition, the spend start to moderate and you start to generate stronger cash flow on some of those investments?

Yes. So let's be clear on our CapEx. Our CapEx is all being put in place to support LTSA revenue. So this is not a build capacity and hope to fill it, right? So we're not going back to that fab filler strategy we'd be forced to go into. So on the last call, we said that our CapEx intensity in '23 would go up into the high teens, right, from about 12%. That's all to support the incremental LTSA revenue that we're locking up. Most of that, you can think about as being silicon carbide. Look, I think that starts to subside, if it doesn't, it's because we have incremental LTSA revenue over the long term. That's what would keep that at a higher level, but we do expect it to come back into a normal range.

Okay. Well, lots to talk about. Unfortunately, we're out of time. So we'll continue this conversation, but gentleman, thanks for your time.

Absolutely. Thank you.

Thanks, Chris.

Thanks, everyone.