Thanks for coming, everyone. Thanks for being here on this afternoon of the annual Citi TMT Conference. I'm Chris Danely, your friendly neighborhood Semiconductor Analyst. It's our pleasure next to have onsemi, the dream team. We have Hassane El-Khoury, the CEO and resident gearhead at onsemi, and then Thad Trent, CFO and resident math wizard at onsemi. So it really is my pleasure to have both of you guys up here. Thanks again for coming. You've supported the conference every year. So we had a couple of your competitors this morning talking about, you know, what's going on, both reiterating guidance.
But, you know, I, I'd ask you guys, it seems like, you know, we've had, I think, five or six, of your peers here today in the analog space and asked them about the various end markets. So let's start with, with auto. I would say it's, you know, I guess, slightly mixed. Some companies are saying that auto's a little better. Some companies are saying, "You know, we don't know," talking about the quarter will depend on September. Visibility is a little bit murky. Any, you know, of your expertise on what are... I guess start with the underlying demand trends, in the automotive market, and then we'll move on from there, 'cause it seems like that's, like, the most controversial area, of semis right now.
Yeah. Look, I think we'll, I'll start at a high level what we talked about last earnings call. You know, we saw stabilization, overall. I don't think you can call, recovery yet or, anything beyond what I talk about L-shape, which means first positive is a stabilization.
Mm-hmm.
However, we did acknowledge some green shoots in some of our end markets. You know, we talked about automotive being call it flat to slightly up, but that's not up because of a recovery, but it's the lumpiness of you know some customers-
Yeah
... decided to take inventory down lower than they typically would.
And that's for Q-
Um, sequentially-
- sequentially, just, uh-
I'm talking for Q3.
Yeah.
Because, look, beyond Q3, then you need to talk about an industry recovery or industry market.
Mm-hmm.
You know, I know some of my peers you mentioned talk about, you know, it depends on September. Look, at this point, you are in September. You either have the order or you don't.
Yep.
At this point, you're almost packaging and shipping. So I don't think that the visibility is as grim as, you know, we don't know what's gonna happen in September.
Mm-hmm.
We as a company don't have that, I would like to say we have better visibility. Look, we had better visibility on the way down, better than most of our peers, where we called both the industrial market, when the softness or the inventory correction started.
Mm-hmm
... call it, you know, probably six to nine month-
Oh, yeah
... ahead of everybody-
Yep
... and so did we in automotive. So I'd like to say we have better visibility because of how we run the business.
Mm-hmm.
But I wouldn't say, you know, I would say at this point, I'm comfortable to mention we have stabilization. As far as calling a recovery, I think that needs more time, and it has to be a change in the market.
Yeah.
Because to me, a recovery, I'm not gonna peg it on, "Oh, we got more backlog, therefore, here's the recovery." I have to see recovery over at least a two quarter basis, because you may get the backlog, you may ship it, you may get a good quarter, but if the end demand doesn't catch up-
Yeah
... then it's gonna be inventory, and the hamster wheel starts again.
Yep.
So you have to see a sell-through of our end customers before you start talking about the backlog being sustainable, and i.e., a recovery.
You know, one thing you talked about that I did want to ask you about is, you know, you guys lowered utilization rates. You were at least in the first two or three companies, if not first. You seem like you're, you know, a little more aware of what was coming and, you know, infinitely better prepared for the correction. What did you see? Like, you know, what caused you to be so proactive?
Yeah. So we look at... You know, both Thad and I look at, you know, many of the KPIs and the data actually, on a weekly basis with the working team. You know, it's not like we sit somewhere, and we get a report that's synthesized with a conclusion. We look at all, and we use our combined experience with the executive team's experience of, what does this tell us, and how do we react? But I would say at a high level, you know, you've heard us talk about the LTSAs.
Mm-hmm.
You know, I know there are mixed views. When we started talking about the LTSAs during the, all the shortages, even when kind of the market started to loosen up a little bit, mixed feelings about the LTSAs. But one thing I've consistently said is, if nothing else, the LTSAs are legally binding, therefore, we're gonna get a phone call.
Yeah.
That's the phone call that I want because in every other cycle or every other market dislocation, you don't get a call. You don't get a courtesy call of, "Oh, we're gonna wipe out your backlog.
Mm-hmm.
It is, the backlog just goes away, and you start wondering, "What the heck happens?" Therefore, you react too late.
Yeah.
By the time the backlog is gone, you're already over-shipped.
Mm-hmm. Yeah.
But because the LTSAs are multi-year, the customers can't ignore them. You know, you can ignore it for three month, four, five month, maybe a year. Some of our peers actually did it for a year, and it didn't work out too well. When you have three- to five-year LTSA, you're gonna get the call, and the call goes like: "We know we signed up. We know we're on the hook. The demand is not there. What can we do?
Mm-hmm.
That's the start of the conversation that we wanna have, which is it's not to our benefit to push inventory to the customer. 'Cause congratulations, we made the quarter. They got it on the shelf. You just pushed the problem down, down, you know, down in time. So when you get that phone call, we do engage with the customer and say: "Okay, you know, we see the demand also softening.
Mm.
That is a true problem. What can we do? What share can we gain? How can we offset it with new designs?" Et cetera, et cetera. But fundamentally, you get a very clear signal of what the real demand is from the customer.
Mm-hmm.
... So when you start getting those calls in industrial, in Q4 of 2022-
Yeah
We started taking utilization, 'cause the first thing you do is you stop building. You don't wanna compound the problem by having inventory now on your balance sheet that you can't ship. So if you know you're gonna have a win-win discussion with the customers, well, stop making the problem worse for ourselves and them, and then you start discussing. So we did that. Then when the, we started getting calls on the automotive ahead of everybody.
Yeah
... if you recall, Q3 of 2023, and everybody was in denial until the negative pre-announcement started around CES.
Mm-hmm.
Started a quarter late. By the end of two quarters, you know, end of the month in the March quarter, everybody's talking about automotive, so we're getting the calls.
Yeah.
So that's, I would say, a big portion why we were able to see it, and we expect on an upturn, we're gonna get the calls to make sure we're ready. I would like to say we will get the calls ahead of probably everybody, because for everybody, let's say in automotive, OEMs increases their numbers, puts it on the Tier 1. Tier 1 may or may not believe it. Then they order the... We're gonna get a call from the OEMs-
Yeah
... saying, "This is coming, get ready." That would be a signal that I would actually feel good about.
So we're definitely keeping the LTSAs. I got that loud and clear. Would you say that, you know, most of your top 10, top 20 customers have them? Have you seen a decrease in the overall number of LTSAs, or have you kept them, but at a lower level? How has that sort of evolved as the year goes on?
Yeah, so, you know, what is it? What, what could a win-win? So the number, dollar number, and really number of LTSAs have kind of relatively stayed flat, which means that customers actually do also still see value.
Yep.
So, but what is a win-win? A win-win could be, okay, we have this part, we don't have it on LTSA. Well, we can add it and make up the volume, so a different mix within the LTSA. Share shift is another one. New design is another one.
Mm-hmm.
When you have a new design, you put the volume kind of in the outer years, 'cause you gotta design it in and qualify it.
Yeah.
But hey, that's a win-win for us, 'cause that's a long-term share.
Mm-hmm.
So in the win-win, it's not about optimizing for the quarter, it's optimizing for a structural benefit for the company, where the customer wins in the short term, if that's the problem they're trying to solve, and we win, whether it's short or long term, is what we will optimize for.
Got it. What would you say your typical LTSA backlog coverage is for the quarters these days?
So if you look out over the next 12 months, and you look at the street numbers, we've got about 60% coverage for that, right? So if you think about it, and that's pretty evenly spread throughout the four quarters, so roughly about 60%. And as Hassane said, if you look over the lifetime of the LTSAs, it's roughly flat.
Okay. So a year ago, it was still 60%?
Correctly. Yeah.
Yeah, and do you anticipate that to remain at that level going forward?
Yeah, yeah, I think so. Now, obviously, you know, volumes have changed a little bit in here.
Yep.
But, you know, the top line's come down too, just with the market softness.
Yep, yep. Okay. So just going back to the auto supply chain, 'cause I get a little off track. Whenever you say LTSAs, I get excited. How would you characterize inventory in the automotive supply chain? I think, you know, you're a car guy, so I kind of, you know, rely on you more than almost every other exec as to what is really going on there. And you came from the automotive industry, so-
Yeah
... what's your take there?
So the difference is, this time around, it's not an industry-specific inventory conversation, it is a customer-specific inventory.
You mentioned that-
There is-
... in the conference call. Yeah.
Yeah. So there's not a trend. So if you look at and if you pull on, let's say, 10 Tier 1s, let's just do the Tier 1. If you pull on 10 Tier 1s, you're gonna get 10 different answers. Some of them are holding, in my view, way too little inventory.
Mm-hmm
... which is, I think, a bigger problem, and the OEMs are taking notice of it. What I mean by that, think about it, you have two weeks of inventory. That's all great when demand doesn't move. As soon as you have an inflection-
A suicide, yeah.
As soon as you have an inflection, the two weeks overnight become two days, then your go line's down. So the OEMs are worried about that level. Some Tier 1s have learned a very expensive lesson by taking, you know, the big OEMs down, costing them billions of dollars. They go, "We are gonna use this time to show we're able to have supply resilience and supply-
Mm
... assurance, so we're gonna take a more comfortable revenue." And those are the Tier 1s that are actually gonna gain share. The other ones may not know, but over time, they're not getting new RFQs.
Mm.
We get the benefit of we see a much different view of it, because we work directly with the OEMs. So the conversation with the OEM is, "Here's the volume that's coming. We will tell you later who to ship it to.
Yep.
Right? So that's gonna be the the discussion. And look, the level of inventory of the, you know, hypothetical ten Tier 1s, depends on how strong their balance sheet is. It's, you know, Tier 1 balance sheets, I mean, it's all public information. You can take a look at it. Some balance sheet cannot support more than that in inventory.
Yeah.
So it's not malicious or anything-
Mm
... it's just survival.
Yep.
But that's not good in the long run.
If you took an amalgam of, you know, that constellation of auto OEMs, would you say that there is a general, propensity to take inventory a little lower at this point or higher? Or do they feel like they have what they need, if you just took, like, an average, 'cause I know different folks are doing different things.
OEM?
Yes.
Higher.
Higher. And what else is there? Like the supply chain or distributors, or-
So the distributors, we manage.
Yeah.
Right? So you got OEMs, Tier 1s, you know, distributor, and us. We've done a very good job on the distribution, side of it.
Mm-hmm.
We've kept the weeks of inventory very tight.... even on a lower revenue base-
Yep
Dollars actually drained out of the channel.
Yeah.
We manage that to our comfort. You know, if we're able to stage some, we'll stage some. You've seen us stage that up for the mass market, so we manage our distribution inventory very, very well. Tier 1, they manage the inventory.
Mm-hmm.
OEMs, they want more than what's being managed, too. And today, we have LTSAs directly with the OEMs-
Yeah
... so first order, where we may not ship to them directly, but they're holding a portion of our capacity for their needs, part of their-- and we ship it through the Tier 1. But if they wanna share shift, the OEMs, if they wanna share shift, then they own the allocation-
Okay
... and they will share shift it.
A couple more questions for the automotive Yoda. So EV versus ICE cars, what would you say your, you know, relative exposure is if you had to split out your automotive business between EV and ICE?
I'd say, obviously, our-- Let me talk about the content from a drivetrain perspective.
Mm
... 'cause it's hard to differentiate. You know, interior, doesn't matter what the drivetrain is.
Yeah.
You know, the content in the cabin is both the same for EV, hybrid, or internal combustion.
Yeah.
The drivetrain is really what the difference is.
Mm.
For onsemi, the content in an internal combustion drivetrain is $50 worth of content. That content goes to about $350-
Yeah
... for hybrid and $750 for
EV
... EV. We have EVs, obviously, that are more than that and EVs that are less than that, but that you can think about it as, as an average.
Mm-hmm.
And look, I'll use the opportunity, given that I mentioned the word hybrid. A lot of people go, "Oh, my God, you know, hybrid is in the future. What are you doing on hybrid?" Well, it's $350 worth of content, so it's still 7x from the internal combustion. But let me make a point on EV. There are two types of EVs or two types of hybrid. One is what you hear about range extender hybrid. Okay, from a content perspective for us, and China, by the way, primarily is going to range extender hybrid.
Mm-hmm.
From a drivetrain perspective, that's a pure battery electric vehicle with a small internal combustion engine that just charges the car. It doesn't drive the drivetrain. It's a generator to charge the batteries.
Yep.
So for us, that's a full EV content. That's the $750. Then you have the hybrid, plug-in hybrid, that's the $350. That's more on the silicon rather than the silicon carbide . We do both silicon, IGBT, and silicon c arbide , so regardless of where the market or what the customer mix is, we can support both of them.
And so it seems like, the EV growth prospects have slowed, but perhaps that's being taken up by hybrid getting a little stronger. Do you agree with that? And then do you think that there's, like, an ICE comeback out there, given the, you know, potential slowdown in EVs? Or how do you sort of assess the overall market?
I would say I have a longer-term view of the market. Longer-term view is EV is going to be the drivetrain.
Mm-hmm.
That's a fundamental belief I have based on, one, investments at the OEM levels and consumer demand. Now, affordability is all of that you hear about in the short term-
Yep
... that's all gonna work its way through the system.
Mm-hmm.
We're gonna have more affordable EVs. You're gonna have charging. You're gonna have fast charging. All of those are roadmaps that we've introduced and I know customers are working on. So long term, EV, as a percent of total vehicle, is going to increase.
Yep.
That's, again, where hybrid, is it gonna be 100% by X year? The answer is no. It's not realistic to expect that. So first, you have to expect the growth through penetration, which is, I think, is a safe assumption based on all the RFQs, based on the introduction. Other than, you know, one handful of OEM, maybe one and a half, nobody canceled or pushed out EV plans.
Mm.
What they've done is they reduced the volume that they would be. But most introductions for EVs coming out of Europe have been introduced.
Yep.
Most introductions of EVs happening in China have been introduced. The sell-through is not what we expected two years ago, but the penetration is happening-
Mm-hmm
... as platforms. So that's why you have to but have to see that long-term view, which is the view that we're investing in. Short term, there is a problem, and it's multifaceted. You have high interest rates, not a lot of cheaper alternatives, and those two together are not conducive to growth in EV to the level we did. There's growth, but not to the level that we thought a few years ago. $30,000 car is very different than $80,000 car when you have 9% interest rate.
Mm-hmm. Couple of, couple of questions on that. Do you think that, EV, like, price/content needs to come down to drive more penetration?
Not yet. It's what, what I call optionality. You're not gonna get an $80,000 car, and you say, "I'm gonna buy this car when it gets to forty thousand." You need a different car at $40,000, which is the same thing in internal combustion engine.
Yeah.
Right? So I think affordability to get to mainstream is a necessary need, which is the same in ICE. Otherwise, everybody will buy premium, or everybody would make, you know, lower-end vehicles.
Yeah.
Ranges and affordability is, of course, a requirement.
Mm-hmm.
... People talk about range is another one. Range is an adoption, it's not a requirement. You know, today, anybody can buy a 300-mile range car. How many of us drive 300 miles in a week without having access to one evening that you can charge it overnight?
Yeah.
You see what I mean? So it's not that. Charging speed is a prohibitive. Like, if I have to wait three hours to charge a car, well, guess what? That's more than my lunch break.
Mm.
Right? So a 15-minute, one, five-
Mm-hmm
... That's a much different EV profile.
Yeah.
So those adoption curves are gonna accelerate as you solve these structural problems-
Mm-hmm
... while you provide affordability for the vehicles. And affordability, people ask me, you know, "Is IGBT or silicon carbide ?" That is not the affordability. The affordability is how much battery you got.
Yeah.
Silicon c arbide, we provide silicon c arbide, and the reason we win is efficiency. 10% efficiency on silicon c arbide versus IGBT is hundreds of dollars on battery volume that you can reduce to keep the same range. There's no way you get a hundred dollars benefit by going from SiC to IGBT.
Mm-hmm.
You see what I'm saying?
Yeah.
So it's not about the technology of the drivetrain or the traction, it's what that technology, although more expensive at a piece price, it provides a more affordable vehicle if you can reduce the battery...
Yeah
... between IGBT and SiC. That's the math that the OEMs are doing. Not a, you know, I need $0.10 cheaper than the-- $0.10 is not gonna make a difference when you're talking about hundreds of dollars per liter of EV-
Yeah
... or, you know, battery.
On that point, it seems like most of the more affordable EVs are coming out of China. It seems like every other, you know, week I see some new Chinese EV company who's launching a car for $20,000 or something like that. I mean, do you think that most of the growth in EVs is going to come out of China? And then, obviously, what's the impact to ON? Is that good? Is it bad? Is it indifferent?
Today, adoption in China is the highest from the rest of the world. That's a fact. Now, like you said, every week, every month, every quarter, there's not just a new EV coming out for $20,000, it's a new brand of EVs coming out.
Yeah.
So there's going to be, at some point there's gonna be consolidation, just like any industry, including in China. You get the broad innovation and push-
Mm-hmm
... and then you got consolidation for financial, performance.
Yeah.
So that, you know, keep that aside, that's a timing thing. From our side, you know, the commentary I gave at our last earnings call, I just attended in April the Beijing Auto Show, which is where all the new China vehicles come and they present. Those are vehicles that are to be launched, not concept, and so on. I would say most of them were silicon c arbide , most of them were 800-volt batteries, and over 50%-60% were onsemi silicon c arbide inside.
Six, six, zero?
Six, zero. So that's kind of the, you can talk about the market share from our perspective in China. Now, volume remains the same topic we talked about.
Mm
... as far as penetration, but whatever you believe the market is this year or next year, you have to be in the socket to grow when the market recovers. So our focus has been, even in the lumpiness, short-term lumpiness of, EVs-
Yeah
... is we need to make sure we're designed into these sockets, so when actually volume does recover through optionality of price points or interest rates or you're gonna grow with that volume.
One more is, the big Volkswagen deal you guys signed up. When can we expect this to be a material impact to ON?
So you gotta think through your typical automotive design cycle. So it's not a, you know, it's not a 2025 revenue.
Mm-hmm.
I'm not given the timeline. I think VW has done a good job on their SSP talking about timeline. But you can think about it from an automotive design cycle as far as when you're gonna start seeing the ramp, and then material within, you know, one year or so after that, that start-
Yeah
... just to get to that, maturity of, peak. Important thing is, you know, a lot of people ask: "What's So what is it? What's a power box?" You know, we talked about in the press release. So everybody talks about silicon c arbide as, as the die or the device. I've been talking about silicon c arbide as a technology play, which is device and packaging. You're not gonna win with the device if you don't have competitive packaging to get the heat out and the efficiency, and so on. If you think about the power box, it is every module, think about it this big, has eight to 16 multiple dies in it.
Yeah.
Eight to 16 of them to get the performance. The module, you need three of them to run the traction inverter.
Mm.
Then all three of them are sitting on a cooling jacket that you pump coolant in to water-cool it.
Mm.
We provide that box, that power box, and we provide it, in this case, directly to the OEM. So we're able to vertically integrate, not just down to the substrate as we've historically done, but vertically integrate also upwards, towards the OEM, by creating more value through system-level integration and system-level performance. Because at the end of the day, if they can get the efficiency we can provide through that, they save on the battery, and they save on the system, and that's a benefit for the OEM. It's something we've been working on. You know, in our last Analyst Day, Simon, our, the President of the Power Group-
Mm-hmm.
-held one of them in his hand-
Yeah
... before the win with VW and said, "This is the future. This is a new platform." And the VW Group win, not just the VW brand, but all the VW brands, landed on it as a scalable platform.
Great. Shifting gears, I guess, but to stay in automotive, let's touch on the silicon c arbide market. So one of your dastardly competitors talked about 30%-40% growth in silicon c arbide in 2025. Does that make any sense? Are we in the right solar system? I won't even ask ballpark, but are we in the right solar system? And what do you think are going to be the key drivers up or down of the silicon c arbide market in 2025?
Okay, look, at the end of the day, it doesn't matter what. Let me ask you this: What's the assumption of the demand? I can give you any number, and I'll have the disclaimer at the bottom, "And here's what I think the demand's gonna be." What's the number of EVs are gonna happen? What's your start, and what's your market recovery? Are they expecting a boomer year like we had two years ago? Or, my point is, it has no credibility, as far as I'm concerned, that far in advance to be talking about 2025, when I will. Let me ask you this: Is it the same one that said they don't know about September? Or is it that different?
Perhaps.
Do you understand my point?
Who is the accent? I couldn't tell.
Do you understand my point? If you don't know what's happening in September, and you can't even guide for the Q4-
You should be on the sell side.
... then it doesn't really matter what you tell me about 2025 at this point in time. I'm sorry, but there's a credibility that I don't want to engage in at that level. So-
You see-
I'll just leave it, I'll just leave it at that.
Yeah.
But let me give you a data point that might help as well. So, you know, there's the trajectory of the EV ramp-
Mm-hmm.
And then there's penetration that sits on top of that, of silicon c arbide So today, it's 22% penetrated.
Yep
... of silicon c arbide . What's in production today?
Yeah.
If you exclude the market leader, the number one guy in North America, it's 6% penetrated.
Yep.
If you look at our design wins, and we sell silicon and silicon c arbide products, you look at them, at what we've stacked up over the years, it's like 85% silicon c arbide .
Mm-hmm.
We sell either one. We're happy to sell either one to those customers. So the fact is, the penetration rate is going up. It's just a matter of time until those get into production. The question is, what's the sell-through gonna be?
Yeah, so that-
That's totally unpredictable. What we can tell you is, from a design perspective, penetration's going up for sure.
Yeah, and by the way, on the, let's just take the EV market. Can we pretty much, roughly map the silicon c arbide content to the ASPs of these EVs? Like, if there's an $80,000 EV, would it have 3X the content of a $25,000 EV? That makes any sense?
Yeah, no, it's... Yeah, I understand. It's actually more tied to the battery.
Yeah.
800-volt battery, primarily over 90%, if not more, is silicon c arbide .
Yeah.
So as the batteries get to the 800-volt, which most of them are-
Yep
... regardless of high end and low end, then you're getting more into the 1200 volt silicon c arbide , which is much more efficient than 1200 volt IGBT. So it's more on the battery, and the reason I say it's more on the battery, as battery prices also normalize-
Mm
... in the market, and you're able to get an 800-volt at a different price point than you did two years ago-
Yep
... you're gonna see those getting into more entry-level vehicles-
Yep
... than just the 400. So that's it. It goes back to really the battery pricing.
Yeah.
That's the biggest driver for price points of EVs. Because the content, you know, you still need four seats.
Yeah.
Right? Okay, leather seats or cloth, all that is optionality that you have from internal combustion, so it doesn't change the ASP. What does change it is the drivetrain.
Mm-hmm. Okay, one more, since we'll stop the forecasting there. How about your gross margin assumptions of silicon c arbide ? What's it going to take to get to accretion? I mean, granted, your whatever, non- silicon c arbide margins move around, but you know, give us some sort of milestones or timing on when we can expect silicon c arbide to be accretive, or let's just say parity, and then accretive to the overall model.
Yeah, look, and it goes back to today. I look at the margin performance of silicon c arbide in product margin and gross margin, and the reason for that is you got the underutilization.
Mm-hmm.
Meaning, are we, are we pricing on a value, and are we providing a value that is at or above? That's the product margin.
Yep.
The answer is yes. Now, you have the underutilization, because we've added capacity in order to be ahead of the ramps.
Yeah.
That's what is bringing the margin down. So that's a timing thing.
Mm-hmm.
At scale, we already know, and we expect our gross, the gross margin for silicon c arbide to be at or ahead.
Can you define scale? Give us some sort of revenue number.
I'm not gonna say a revenue number, because a year before we ramp, we add more. So you're gonna see kind of up, then flat, then up, then flat, because as you ramp, you're gonna add more capacity as well-
Yep
... until you get to capacity maturity. So we're not giving a revenue number. And look, a few quarters ago, we did say it was already at the corporate. That was a few quarters ago.
Mm-hmm.
So I'll just leave it at that, and I'll leave it at that purely for competitive reasons, because when I disclosed it, it didn't work out too well. A lot of pe-
True.
I'll just tell you, we got there. There's no reason we can't stay and grow from there. It's just we've proven it, so I don't need to keep proving it. I'll just keep that for competitive reasons.
Mm-hmm. Have your forward assumptions for silicon c arbide in terms of pricing changed at all over the last three to six months? I'm not asking what they were, but have they changed at all?
No. The value didn't change.
Yep. That is, just seemed like ASPs. And then, maybe talk a little bit about customer concentration. It seems like more and more applications, say, on the industrial side-
Mm-hmm
Are starting to use silicon c arbide . Can you give us any sense of your customer concentration, and then maybe auto versus industrial, and how you expect that to change over the next three to five years?
Yeah, I mean, overall, the split between auto and industrial really has not changed for us. You know, if we looked at the LTSAs, even from a few years ago, we talked about eighty/twenty. You can talk about 15/85, in that range.
Mm-hmm.
So call it 85/15, give or take. That's auto and industrial-
Yep
... and it really matches if you look at the market data. The market data is split about that 80/20 ballpark.
Yep.
We're no different than the market, which means we're pretty equally distributed between where the actual market demand is.
Mm-hmm.
On the industrial side, well, that's been a trend for a few years. It's not new, where silicon c arbide is making its way. Our energy infrastructure, specifically energy storage systems, where we provide fully integrated silicon c arbide , IGBT, and hybrid, where in the same module you get silicon c arbide and IGBT, we provide both, so we can make those innovative products available. That business has grown 70% year on year, 2021 to 2022, 2022 to 2023.
Mm-hmm.
Now, that took a pause, part of the inventory digestion, but last quarter we talked about green shoots in alternative energy.
Yep.
That's one of them. Another data point we have given is where we have LTSAs with eight of the top 10 energy infrastructure companies-
Mm-hmm
... that command 80% market share. So we already have that market share. We're already designed in, and we've been growing that revenue very aggressively. Like I said, 2021 through 2023, and are starting again.
Yep.
That puts us in a very good position to also capture silicon c arbide value into industrial applications.
Great. Before I switch gears and start to pester Thad over here, would anybody in the audience have a question? Going once, going twice, I'm fine. Laugh at my gums. Do you want to use the mic?
Thanks, guys. Hassane, when you talked in the Volkswagen press release about, like, the customization of the efficiency, like, what do you kind of mean by that in terms of, like, achieving some of these unlocks and efficiency unlocks for your customers?
Sure. So it's always the... So I'm gonna put it this way, the solution is not an off-the-shelf solution, which means it's what, you know, if you wanna put it in an umbrella, it's a proprietary solution umbrella that is, however, based on a scalable platform for us and the customer. But what are these customization? Look, I've been very consistent of saying our technology and our differentiation is both the die and the package. We co-design and co-develop the die and the module, the package and the module together. Because if you make trade-offs only at the die, I'll give you a perfect example of semiconductor. Everybody goes for smaller and smaller semiconductors, right? Die area, smaller is cheaper, you win margin.
Okay, if you take a square inch, it's a big die, I know, but just for the math, if you take a square inch of power, it's very easy to get the heat off of it. If you cut it in half and you have a 25 square millimeter, more than half, 25 square millimeter die, if you don't have the right package, you're gonna have a very expensive package to get that heat off a much smaller surface. So degrees Celsius per surface is way different and way more expensive to get cold, cooler than it is when it's bigger. Counterintuitive. So if you just make a die that's smaller, you may be high-fiving that you have an innovative device, but you just moved the cost to a very expensive package.
So when we say you have to co-develop the two in order to achieve the best performance of efficiency and cost, that's what I mean by that. Otherwise, you just squeeze the balloon somewhere else.
Great. Is there another one over here?
Thank you very much for this opportunity. So just want to ask a bit about the silicon c arbide growth of 2x of the market comment you have earlier. Could you comment a bit about where you see that coming from? So I think I've heard from previous calls that I think largely it's about the confidence about the performance that you have versus your peers in this product. Would you say you are more thinking about that penetrating into, let's say, the biggest North American player in EV, or is that in China, or is that maybe in the upcoming European players that are doing more EV, or is it maybe more broad-based? I wonder if you have any comment there you can share. Thank you.
Yes. The answer is all of the above. So at this point, you know, when you talk about 2024, those designs are already won. So at this point, it's ramp. You know, earlier in the year or end of last year, we said the second half of this year is gonna see a European OEM ramp, in addition to the wins that we've had in China already, that will ramp in the second half. And obviously, the proof point for that is the market share that I've talked about from the Beijing Auto Show. So designs, everything is in play today as far as where we need to be. As far as North America, specifically, we are the market share gainer over there. We've been very consistent since last year, and that is consistent.
Now, you're a market share gainer, but the volume has to catch up. The units have to catch up, and that is more public than I can comment on. So at this point in time, it's purely what the end demand is gonna be. That's why we peg it to market rather than a certain number, because that's gonna be what the denominator is gonna be, but we know where we're designed in and the sockets. Over a longer period of time, Europe is gonna ramp for us. You know, we just talked about VW. North America will continue to ramp for us. China will continue to ramp for us. Those are the primary regions, really, where EV penetration is high. Right now, we're focusing on design-in and market share.
What happens to the denominator, which is units, will just lift us along with everybody, but we expect it to lift us faster because of the share of design-ins we have.
Great. So Thad, you're looking a little lonely over here. Maybe just run through the gross margin drivers going forward, top two or three, and then we'll dig into it a little.
Yeah. So, look, I think we're in a really good spot, given where the market is, first of all, just to calibrate. We're at 65% utilized. We're holding a mid-40% gross margin floor.
Yeah.
The last time, or historically, when we're at 65%, margins are in the low 30%. So structurally, we're a different company, right? We're positioned really well for a recovery. When it does recover, we've got lean inventory in the channel. We've got lean inventory on our balance sheet. We can crank up utilization, and that starts to flow through to gross margin, operating margin, free cash flow. So let's walk through the pieces. So in the short term, the biggest driver is utilization.
Yeah.
So for every point of utilization, it's 15-20 basis points of gross margin improvement. So we maxed out around 84% utilized. That's you can think about it, is because we run so many technologies, that's pretty much at full capacity for us. So you can do the math there, going from 65%-80% + -
Yeah
... and calculate there. That's what I'm saying. The shortest term, that's the benefit of the gross margin. The second piece is, today, we have a 100 basis point headwind from the GlobalFoundries business that we run-
EFK
... EFK. That's just a matter of time. That's gonna roll off in 2025. So that 100 basis points starts stepping down early in 2025, and by the time we exit 2025, will be completely gone. Just, just a matter of time for that business to roll off. Nothing else other than that. The third element is we divested four subscale fabs in 2022.
Mm-hmm.
That's $160 million of fixed costs, and when we move that production into our footprint, we start to see the benefit of that. So today, if we own those four fabs, our gross margins would be lower. We'd have more under-utilization charges.
Yeah.
So great move that we did in, when the market was blowing and going at the time as well. But $160 million, we'll start to see some of that in 2025, probably most of it in 2026. If the market comes back faster, we'll see it faster. Just a matter of moving that production in-house. And then the last element is accretive new products that are accretive to gross margins.
Yeah.
So analog, mixed signal, everything that Hassane's talking about, silicon c arbide coming up to the corporate average, all of that is a tailwind that gets us there. So if you start doing that math, you can get to our 53%-
Yeah
... gross margin target pretty quickly.
Pretty easily, yeah.
You know, it's gonna be the number one driver is demand. It's, you know, when this market recovers.
W-
There's not a long latency for us, given the way we position the company.
Would you say your overall visibility now, versus, say, three months ago, is the same? Is it better? Is it worse? How would you characterize it?
I would say visibility is about the same. I think, you know, as Hassane said, we've seen some stabilization, so you know, we monitor push-outs, reschedules, you know, LTSA changes, and that's starting to slow-
Mm-hmm
... which is a good sign. So that's where we see the same, but I would say with visibility is about the same.
Yeah. And in terms of that, gross margin utilization rate calculation, what sort of pricing does that assume? Does that assume normal pricing, flat pricing, up pricing, down pricing?
That's, that's-
Price agnostic?
... consistent pricing. I mean, we are not seeing pricing pressure, right? I mean, pricing pressure is only on everything that we do in LTSAs, as well as non-LTSA business.
Yeah, I'll give you, for example, a lot of the designs that we're winning now, whether they ramp next year or three years from now, pricing is we know the pricing, right? Whatever that pricing land, there's no uncertainty in the price, even if it's three years out. Yeah.