2023-08-07 17:12:09 ET
ON Semiconductor Corporation (ON)
KeyBanc Technology Leadership Forum Conference Call
August 07, 2023 01:30 PM ET
Company Participants
Thad Trent - Executive Vice President and Chief Financial Officer
Simon Keeton - Executive Vice President and General Manager, Power Solutions Group
Conference Call Participants
John Vinh - KeyBanc Capital Markets Inc.
Presentation
John Vinh
Good morning, everybody. My name is John Vinh. I cover semis here at Keybanc Capital Markets. We are pleased to have Thad Trent, CFO; and Simon Keeton, EVP, GM of the Power Solutions Group from ON Semi. Welcome, guys.
Thad Trent
Thanks for just having us.
John Vinh
Thad, maybe just talking kind of big picture from a cycle perspective, it looks like you've successfully navigated to kind of a soft landing scenario here. I mean, however, there's still a lot of concern investors out there that maybe auto and industrials just haven't seen it and potentially it's the next shoe to drop. I'm just wondering if you could just comment on that. And then the other thing I think everyone is looking at is if you look at aggregate inventories on balance sheets and distributor balance sheets, we're all kind of running at all-time highs. So maybe help us square that?
Thad Trent
Yes. So I think what you really got to do is you got to break down our auto and industrial business further, right? Because that broad category doesn't really fit well. I think in our auto, we're over-indexed to EVs, ADAS, everything that's got a rapid growth rate on it. In industrial, we're being driven by alternative energy, some medical, things like that. Last year, mid last year, we saw consumer facing industrial gets soft, right? So I think it's similar to what others have seen. We've seen that go basically flat from that point forward. And I think it's been fairly stable. But the alternative energy grew 70% year-over-year.
Our auto business is doing fabulous. And that's because if you think about an internal combustion car, we've got $50 of content. In an EV car, we've got $750 of content. So there's 14x content story here, which gives us that growth. And I think as best we can tell, we don't think there's large pockets of inventory out there. Our medium and high-voltage products are still highly constrained. Customers are still expediting from us. I think things are stable. If we look at our LTSA signing long-term agreements, customers are still signing them, which is a good sign. We added $4 billion of long-term agreements just last quarter alone.
So these are customers that are extending their agreements or expanding, adding additional products onto those agreements or new customers that are coming in and engaging with us. So I think given our silicon carbide exposure, which is obviously fast growth in EVs, I think that's why we're a little bit unique compared to everybody else's. We have secular drivers. I think the other part of it, even if you exclude silicon carbide, out of our revenue numbers from last quarter, our automotive business still grew without silicon carbide. So we've got a lot of products that are over-indexed into that EV that is really creating that content story for us.
John Vinh
Great. Thanks. Maybe we just focus a little bit more on autos. I think auto demand for you seems like it's been pretty strong. Several of your peers recently have come out and talked about seeing weakness in autos and some have talked about seeing weakness more specifically in China ICE. And we've even heard some companies talk about weakness amongst certain EV OEMs. I'm just wondering if you could just talk us through more broadly about what you're seeing within the auto spaces. Are you seeing pockets of weakness? How is EV compared to ICE? Is ICE weak, but it's just the content story that's offsetting that? And then maybe just talk about geographically, what are you seeing in North America, Europe and Asia?
Thad Trent
Yes. Well, the auto industry in terms of units is forecast to grow about 5% this year. I think next year, it's forecasted to grow about 1.5%. Our exposure in EVs is obviously going to outpace that. If you look at the EV growth rate, it's projected to be a 35% CAGR over the next several years. So that remains kind of as I was saying, that that content story there. When we think about it by geography, you mentioned China, we don't have a lot of exposure into ICE engines there in China. But we're a net share gainer in China. So it's all incremental for us, right? It's not like we had a large embedded revenue base there.
So for us, our China revenue last quarter was up slightly. And I think that's – but I would also say, generally, in China, we haven't seen a big snapback in the overall market. We are seeing growth in automotive, but I think the other stuff is similar to what you're hearing from many of our peers. It's just – I think there was an expectation that there's going to be a snapback and we are not seeing that broadly.
North America and Europe auto very stable. Again, I think order patterns are very stable. Again, we don't think there's a lot of inventory out there. We've been managing our channel very tight. We've been managing our internal inventory very tight. We took down our utilization starting Q2 of last year. So really, for a year, we've been keeping under – utilization kind of low in that 70% range exiting last quarter at 70%.
So I think we've been managing very cautiously just given all the uncertainty in the overall market as well. But auto generally across the globe from an EV perspective is stable. We've also been supply constrained on our sensors for three years now and just having been able to keep up with the demand there. So I think as we've been able to get more supply externally on sensors, that's been helping us grow as well.
John Vinh
Great. Maybe Simon, on silicon carbide, $11 billion in LTSAs. Just curious, what's the average duration of these LTSAs and with the incremental $4 billion in LTSAs is getting into a cumulative number, $11 billion. How far do you have visibility out on that business?
Simon Keeton
Yes. I think you see a variation across the LTSAs in terms of duration, I'd say, on average, it's about five years, but some go out over a decade. And it gives us very good visibility into the order patterns and what people are expecting over that duration. And that helps us with the development of next-generation products within that time period.
John Vinh
Great. And obviously, with the progress that you're making there, it does feel that you're gaining market share within silicon carbide. Can you just walk us through what are the top kind of two or three factors that are driving these share gains for you? And why do you win?
Simon Keeton
Sure. I think one of the pneumonic I used before at Analyst Day were the four Ss of silicon carbide, and it still rings true. It starts with superior performance, right? We have a value proposition and the performance of the device. And when I say device, it's silicon carbide in a module really gives you the most enhanced performance to extract all that value. So it's having that capability. So superior performance is still critical to every single application, then it's supply chain, having the substrate, the device itself and the package all in-house, not only can we guarantee a supply, but we can tune the performance because we have substrate device and package to give an optimized solution to the end customer. Then we can scale as a customer wants to scale. We were talking about $1 billion of silicon carbide this year, and it gets even bigger. We've been able to scale up very rapidly and very effectively as evidenced by the performance in gross margin last quarter, which was really quite good.
And then you look at the scope. It's not only the silicon carbide, but the surrounding products that go with it. So our silicon carbide traction inverter will have a gate driver. That gate driver is tuned to that silicon carbide to provide optimal performance. And so we have all these other non-silicon carbide silicon devices that go along with it into the EV. So those are the four Ss.
And the last one I would add, and I'm going to add a fifth S to it is systems. So because we're a first mover into EV, it gives us a better picture of what the end customer wants from a solution space now and over the next five and 10 years. So we can adopt that into what we call Horizon three developments and really create optimized solutions, not only for now, but for 2025, 2027, 2030, 2032, so we have that visibility because we've got first mover with that OEM in that particular space.
John Vinh
Great. One thing I wanted to kind of touch base on Simon. I think you touched based upon this topic at your analyst event is just China risk. I think there's some concerns out there that the Chinese are investing a lot into silicon carbide substrates. At some point in time, does it make sense for you to completely outsource it? And maybe can you talk about kind of the value that you see with the integration of the substrate, the drivers and the modules? And then what are you currently seeing right now from the Chinese and their ability to ramp substrates?
Simon Keeton
Sure. So as you know, we have internal substrates and then we also buy from the external market. We expect to have north of 50% of our substrates coming from internal by the end of the year. Part of that value proposition is we have a very good cost structure on those internal substrates, which you see evidenced in the gross margins that we're getting on silicon carbide. What we're seeing from China, and we source from China as well as part of that external play is we get to see the quality of what's coming in and that quality is quite – not quite there. And we also see – it's not indexed completely, meaning if I have a ball, I can see slice one through slice, whatever slice 30, index one after another. What I'm seeing from China is slice one, slice seven, slice 15.
So it's not contiguous. So maybe we're getting kind of the best of the best. And the question is, is it scalable? So we have to see it scale. And at some point, could it be there, yes. And that could be a tailwind for us. We could actually buy more external substrates if it made sense, but there's also a geopolitical factor and particular customers will not want to take that risk of a supply chain, having a substrate coming from China. Some are okay, some are not. So it's a bit of a mixed bag, and we'll be watching it very closely. So I think we're in a very good position. It won't necessarily – substrate doesn't necessarily drive the performance of the device itself. But with supply chain, it becomes incredibly important.
John Vinh
Great. Are there any questions? Great. Thad, I wanted to follow-up on a comment you made. You had recently talked about seeing a pretty meaningful uptick in your image sensor business. What's driving that uptick?
Thad Trent
Well, so clearly, there's more sensors in the car every year, right? So you've got advanced safety. Where we focus is on machine vision, not human vision. So if you think about a backup camera, that's just not something that we can differentiate on. But all the sensors around the car even inside the cabin, which have some type of machine vision is where we play, very advanced, sophisticated things that many of our competitors can't do.
So we've got north of 70% market share when it comes to ADAS across the globe. That's because we can do this that other competitors can't do. So the growth there is really the adoption rate of more sensors in the car. As I said, we've been constrained for three years. We've just been limited because all of that is manufactured externally, which supply was very limited externally over the last couple of years. As I said, as supply is coming online, we're actually able to catch up with some of this as well. But we pick our spots in automotive.
The other piece that I should mention is factory automation. So that's another big area that we play in when it comes to sensors. So both those areas are expected to grow at really nice clips over the next several years. And like I said, we've kind of got a great portfolio that supports all of our customers' needs, whether it's auto or industrial.
We've completely restructured that business from what it was three-plus years ago, three-plus years ago, it used to be a lot of webcams, a lot of consumer type devices. Today, it's 90% auto and 10% industrial. High margin, sticky business, just great business for us versus a commoditized business that turns quickly. So when we look at our projections going forward, we look at our long-term agreements, we have a nice view on what our customers are doing.
John Vinh
Great. Speaking of exiting business, I wanted to talk about your non-core business. I think you've exited $100 million this year so far. What's been the gross margin profile of the businesses you've been exiting these days? And I think you said you have $350 million to $400 million for the rest of the year? And then how much is remaining in total that potentially could be exited beyond this year?
Thad Trent
Yes. So that's right. So let me just calibrate everybody. So we're exiting the highly volatile price sensitive business that historically has been there. And the way you exit is you price yourself out of the market, you keep raising pricing until your customers go someplace else. We announced our plan to do this in '21, 2021, and it's actually taking a lot longer for us to exit this business than we expected. So our gross margin of this business that's still remaining is in the mid to high 40% – I'm sorry, low to mid-40% gross margin, right? So it's not bad gross margin today as we priced it.
We're just not losing it as fast because supply is not coming online. I think our customers are also valuing assurance of supply because we can support them. But – or the switching costs might be too high for them as well. So we think there's about, I think I said $350 million to $400 million for this year. We've exited 100 thus far through the first half, and we'll exit the rest of it here coming in Q3 and Q4. At the end of this year, this business is still here, it's going to be deemed good business, right? Because it's a decent gross margin.
As long as we don't need that capacity for something else, we'll continue to support it. Now also remind you, we divested four fabs last year in anticipation of exiting some of this. So some of it is going to happen and has been happening in others because we'll just run out of supply. Others may stick around in its sticky business, we'll keep it if it's attracted gross margins. But it is much slower than we anticipated just because supply hasn't come out. And I would say the pricing environment hasn't dropped like we thought it would. So if that's the case, and the pricing holds, we'll continue to support it. But after this year, I don't expect there's more exits. This is it.
John Vinh
Okay. I wanted to also ask you about medical. I don't think I remember you guys mentioned medical in your prepared remarks ever. It sounds like this is kind of an emerging opportunity here. What is the medical growth that you're seeing all about?
Thad Trent
Yes. So there's a couple of key areas. The continuous glucose monitoring, the CGMs a big, big play for us is that's taking off. We've also do a lot in hearing aids. So if you think about hearing aids going over the counter now, that reaches an audience that didn't have access to hearing aids in the past across the globe. So that's driving really nice growth.
If we think about our medical, it's still a small piece of our business, but we think that's got a 20% per year growth rate on it when you think about those applications. And all of those are kind of in their infancy just starting to take off right now.
John Vinh
That's great. Maybe we can talk about margins. On the silicon carbide business, I think you talked about being at 15% margins in the current quarter. As that continues to ramp by the end of the year and as internal becomes greater than 50%, are the margins on your silicon carbide business going to be at or above corporate at that point?
Thad Trent
Yes. So we've always said long-term, our gross margins would be at or above the gross margin – the corporate average, right. So as we price the products without all the start-up costs, we would achieve those margins. Last quarter, our silicon carbide gross margins doubled quarter-over-quarter. Simon's business for silicon carbide was profitable for the first time. High-teen operating margin.
So it gives you an indication of kind of where those margins are. As we continue to ramp by the time we exit the year, we believe the silicon carbide gross margins will be at the corporate average. Q3 is the high point for the dilutive impact because it is still below the average and you just have a higher revenue number. But by the time we exit the year, we think we're back to the corporate average. And then longer term, we're back off to the races again in terms of margin expansion.
John Vinh
Okay. Maybe related to that, can you talk about just the progress you're making on 8-inch right now? And is there a way to think about potential cost benefit to you once you start ramping 8-inch?
Thad Trent
Do you want to talk about that?
Simon Keeton
Yes. So obviously, we've got 8-inch engineering going right now. We've got the furnaces, and we've got the flow through Bucheon through the factory. So we're still on target. We're hitting all the milestones that we need to hit – so we expect engineering through 2024 and then production in 2025. And what you see as you go from 6 to 8, the real savings comes from the processing after the substrate. Because at that point in time, that manufacturing flow that tool set is 8-inch capable. So it doesn't know if it's doing 6 or 8. So you're going to get 6 to 8, 1.8x more [indiscernible] wafer on that tool set. And that's where you can get some of those savings run.
Thad Trent
Yes, the big benefit that we get out of that is it's – the capital required to scale the 8-inch is already there, right? So all of our furnaces are retrofittable, so it's not like this next wave, we've got to put a lot of capital into it. And then in the fab, we are 6 and 8-inch compatible capable already. So it's not like there's another wave of investment we need to make that move.
John Vinh
Great. Any other questions?
Question-and-Answer Session
Q - Unidentified Analyst
Just curious on fab strategy. Has anything changed over, what everybody's gone through supply chain in terms of how you think about fab and…
Thad Trent
No. No change to the fab strategy. We divested four fabs last year for subscale fabs. We just acquired the East Fishkill fab from GlobalFoundries, which gives us a lot of capacity so no change in our strategy there. We manufacture internally about 65% of our own product as we fast forward with our expansion. It will still be about 65%. So we think we have the right footprint now. Now it's about what I call fab right, getting the right mix in the right locations to optimize the cost. So step one was fab lighter, just downsized it. Step 2 is Fab right optimize it. That's what we'll be doing over the next few years.
Unidentified Analyst
Just a quick follow-up on 6 to 8-inch [indiscernible] targets for internal production with the substrate versus intention – are those targets the same for inch?
Thad Trent
They will be over time, right? So there's a ramp time. We're going to be very deliberate on 8-inch. We're not going to rush it. It will be at our pace and very deliberate we're getting extremely good results on 6-inch making sure all those learnings go into 8. And when we do it, it's going to be at or above moving forward. And so we're not going to rush it. And there's a lot of learnings from 6 that we're going to apply but we're doing very well on 6. So I don't feel like I've got a lot of pressure like, oh, I've got to do it immediately. We have to be very thorough and deliberate with this and very engineering-focused. So we'll get there for sure, but it will be on our timing.
John Vinh
Is there another question there? Okay. Thad, I wanted to ask you about East Fishkill. I think some of the incremental margin headwinds there were somewhat unexpected. Can you give us just some perspective on what's causing these headwinds and just the time frame and your level of confidence of getting these headwinds resolved?
Thad Trent
Yes. So we closed the transaction December 31 last year. The deal was announced almost three years ago, right. And over those three years, we were slowly ramping up production while GlobalFoundries was running the fab. And over the next three years, we'll continue to ramp up while they ramp down. The surprise that we got after we closed the transaction, when we got full access to the fab was just the cost structure was much higher than we anticipated.
I characterize this as like blocking and tackling. We run lots of fabs. We know exactly how to run fabs. We can benchmark to our internal fabs know what the cost structure should be. It's a matter of being able to make those changes and that takes time. Now the fab is fully loaded today. So it's really hard to tinker with a fully loaded fab. You got to be really careful that you don't break the recipes while you're adjusting for this. So it takes time. We're really confident.
It's about a 250 basis points headwind was there last quarter, and we'll be there for several quarters. My projections are by the time we exit 2024, we will have that cost structure back into place. So that headwind is going to be with us for a while just as we continue to take cost out. But it's everything from supplies, chemicals, cycle times, it's blocking and tackling, and it's what our teams know how to do. We've got a long history of running fabs. We just need to go tune that one up. I would really say this was probably from being neglected for three years, right? It's just nobody was really focused on it, and it's just a matter of focusing and getting it tuned up.
John Vinh
Great. Simon, I wanted to maybe follow-up with you on a comment that maybe Hassane made on the call that you're starting to sign these LTSAs with upfront capital commitments from your customers. And I think you guys made it clear that you don't really need customers to commit capital for you to build out this capacity. Maybe can you talk about how these agreements and these capital commitments came about? Are they kind of suggesting in? Or are you kind of suggesting it? And how do you guys figure out what amount of capital commitment needs to be put out there to secure these agreements?
Simon Keeton
Yes, I think a little bit of both. We do have customers approaching us that they want to be a part of, right? They want a tangible part of the silicon carbide growth and to know that supply is going to be there. And so in some cases, it's everything from consigned equipment to capacity reservation, there's different models that we use with different customers. So we try and be flexible with the customers and have it done in a way that makes sense for the customer and does at the same time, financially. So I think moving forward, we'll still see that. People want to be a part of this, and I think we have an effective way to do it, but also gives us a good financial return at the end.
Thad Trent
Yes. And I would also kind of to clarify Hassane's point, right? Last year, our free cash flow margin was 20%. We create a lot of free cash flow. Our long-term model is 25% free cash flow. I'm confident we're going to get there. Now Hassane's point is we're not hostage to customers to fund our investments in silicon carbide expansion. Now as customers approach us and they want dedicated supply, we're going to say we want you to co-invest with us, right?
We want skin in the game. We want you to have a piece in there with us. And that's required to achieve our financial model longer term, right? So if we can get them to pay upfront, and then that helps with the cash flow. But Hassane's point, it was really like we're not relying on customers for the expansion, but we will require they co-invest when they want dedicated supply.
John Vinh
Just a follow-up on that. That is when you've got a customer that's putting up capital to build out capacity on behalf of that customer, is that capacity that you build out for that customer dedicated solely for that customer? Or are you also able to utilize that capacity for other customers?
Thad Trent
Yes, good question. I get asked that a lot. It depends a little bit on the model, right? I mean we've got cash from customers, we've got consignment from customers where they're actually going to consign us the equipment. In most cases, we structure it to where we have the ultimate flexibility. So it may be dedicated for that customer, but is – they're not using the capacity we can use it for other customers. So we take that incremental capacity and be able to use it.
So that's a win for us. It's a win for them, obviously, as well because they're getting that dedicated capacity. But we structure it so that we can scale with it and use that excess capacity. We don't want a line sitting in a building not being used, right. I mean, that doesn't make sense even from our footprint, that doesn't make sense. But yes, that's the intent is that we can use that capacity for our own needs.
John Vinh
Great. Last question for me. I need to ask the obligatory AI question. Are there opportunities for you to benefit directly or indirectly from AI at this point?
Thad Trent
Do you want to take it?
Simon Keeton
Certainly, if you look at the power development that we have, I think it's complementary to what we have in like servers, AI servers, absolutely because it's about power efficiency. So I think, yes, as this adjacent player that we have in power that we can participate in the AI market certainly.
John Vinh
So most of these AI servers are running at 48 volts are you – do you have share in that area?
Simon Keeton
We do already. So I can't give you the specific names, but you can kind of guess who's doing what in AI and servers. And so we already have a position – and I think there is the ability to scale that position as well.
John Vinh
Is there an opportunity for silicon carbide in servers?
Simon Keeton
Potentially. I think the verdict still out between silicon and silicon carbide and GaN and servers, right? It's pretty crowded when it comes to a technological solution. It remains to be seen who's going to win. And it may be all three, depending on what the server is looking for. So I think we're in a pretty good position. We'll see where the market changes with it, but we want to drive that.
John Vinh
Okay. Sounds good. Thanks, guys.
Thad Trent
Thanks, John.
For further details see:
ON Semiconductor Corporation (ON) Presents at KeyBanc Technology Leadership Forum Conference (Transcript)