ON Semiconductor Corporation (NASDAQ:ON) NASDAQ 47th Investor Conference Call December 6, 2022 3:30 AM ET
Company Participants
Hassane El-Khoury – Chief Executive Officer
Thad Trent – Chief Financial Officer
Conference Call Participants
Joe Moore – Morgan Stanley
Joe Moore
Sorry, we don’t have any transition time. So we are going to go ahead and get started. Excuse me, go ahead and get started now. Excuse me, everyone. I’d like to get started if we could. Thank you so much. So welcome back. I am Joe Moore. Very happy to have with us today the management team from ON Semiconductor, including Hassane El-Khoury CEO and Thad Trent, CFO and really excited to have you guys here. You really touch a lot of the key debates that we are dealing with right now across semis. So thanks for coming. Maybe we could start off with just a general commentary on the environment. You guys have talked about strength in autos, maybe pockets of weakness elsewhere. How durable do you see those trends looking forward?
Hassane El-Khoury
Yes. Look, as a general comment on the market, as you mentioned, the automotive strength is coming from a very sustainable mega trend, which is electrification of vehicles that we see as sustainable through, call it, the next decade as more and more OEMs introduce their electrification platforms. If you think about the announcements made anywhere between 2025, 2030 is where more and more of these vehicles will come in and that’s the underlying strength that we see in our automotive business with power. And of course, parallel to that, you have the autonomous driving, which is for us the cameras and the vehicles as that more and more gets into penetration on all vehicle platforms. We see that strength continuing again over a multiyear period.
On the industrial side, it’s a – call it a mixed bag. The industrial side, that’s closer to the consumer or more tied to consumer demand like power tools, for example, we see that softening, we saw it softening, and we took actions to get ahead of it. But the underlying strength in industrial factory automation, renewable energy, they are also tied with the automotive mega trend. We see strength remaining there and growth through this year and next. And then the consumer and compute, I think we see what a lot of our peers that are more exposed to those markets where we see softness and that softness seems to be maintained.
Question-and-Answer Session
Q – Joe Moore
Great. And maybe if you could talk a little bit to the general business strategy, you have talked about – a couple of years you’ve been here, you have talked about exiting revenue streams that have less enduring value. I think you talked $277 million of business that you have reduced in total, $39 million last quarter. Can you talk about the criteria for what’s a sustainable business, what businesses are you trying to get out of and how much more business is there left to reduce?
Hassane El-Khoury
Yes. Look, 2 years ago, when I joined onsemi, I refer to – we need to do a portfolio optimization. And what that meant is, looking at all our products we have over 20,000 SKUs in the company with a very broad manufacturing footprint. And the portfolio rationalization was purely based on value and value creation. Just like we look at value creation from the company side, we look at value creation for our customers. So, the criteria really is, do we see differentiated value enough that we can extract that value in the market for our products? And if the answer is not, we are not going to continue with that product. So, that’s really at a high level, the criteria, in order to maintain two things that are sustainable for our financials is the growth in our business and the margin expansion of that business. Both of those happen when you have great products that create value and differentiated products that you have a competitive advantage against a lot of your peers. Those are the high level criteria that we parsed our portfolio. And what we have done – deployed that strategy in August of ‘21, part of our Analyst Day, we put under that umbrella at the time about $700 million or $800 million of business that meet that criteria of, well, we are not adding value, and we decided to exit. To-date, you mentioned $277 million, average gross margin of 25%. So that proves the point of why we are exiting and that’s in a good pricing environment. So we have exited that. There is about $400 million to $450 million, call it, left. We expect that to exit that through next year as our customers have other options or other companies that are taken that business from us. That will actually be expansion of our margin from where we are today, which is part of our strategic intent.
Joe Moore
And can you talk about the remaining part of the business? You have talked about a relatively tight gross margin range over time despite a lot of moving pieces. So you really view that as kind of an enduring margin level that’s not cyclical on the remaining portion?
Hassane El-Khoury
That’s right. I mean we have – given our model of $48 million to $50 million, over the past few quarters we have actually achieved that model. Obviously, we have a few headwinds that are measurable coming up. Most of it are strategic decisions that we have decided to enter. One of them is a silicon carbide, which is our big growth. We have put a lot of CapEx. That’s about 100 to 200 basis point headwind. And as that business scale, which we expect it to get to scale end of ‘23 into ‘24, that headwind is going to go away. And then we have some foundry business part of our fab acquisition that we have done that will also work its way out over the next couple of years. As those headwinds go up, we are expecting just like we have in the last few quarters to maintain that 48% to 50% margin.
Joe Moore
Great. And I just want to get into some of those newer drivers in a sec. The – I think the interesting element of this is you guys really have executed to this plan. I think it’s clearer now when you have taken utilization down, you focused on reducing inventory at distribution and you remain – even though parts of your business have some softness, you remain committed to kind of taking out these other businesses. Anything that changes that progress, I think you have mentioned some. It’s just harder to get out of some of these businesses than you thought it was?
Hassane El-Khoury
So the answer is no. We stay the course. And if you look at the performance of the company, even in the last few quarters, we are kind of – some of these markets started seeing signs of weakness. We have kept performing and really outperforming the market, because we believe in our strategy, and more importantly, we are executing to that strategy. We saw the weakness coming. We took utilization down. We started reducing our fab utilization back in the second quarter before kind of we got to where we are today, which really gave us an advantage, have been able to see that coming and taking proactive action. People ask, you took utilization down in the fab. Did you see the backlog cancellation and push out? The answer was no. What we saw is new backlog layering on, that was call it the second derivative of backlog. But that was enough for us to take action because what we don’t want is get into a softer aspect of our – in some of our business and end up with inventory, whether on the balance sheet or in the channel. And what you have seen us do in the last two quarters is actually reducing both of these inventory, which sets us up very, very nicely for moving forward as a sustainable financial outlook for the company.
Joe Moore
Yes. I appreciate that. Thank you. Maybe if you talk a little bit about the longer term supply agreements and I feel like we are learning that every company has a slightly different philosophy about this, as we move into this elevated lead time environment, how do you view it? How long do the commitments go out? How much do you hold people to those commitments in different types of economic environments?
Hassane El-Khoury
Sure. Look, so the LTSA, as broad as it is today, our long-term supply agreements are new, call it, to a first order. But for us, both Thad and I, in our prior lives, we implemented and executed very, very well for – on LTSA with the same framework that we are in executing to at onsemi. It worked very well. So that learning, we carried over to onsemi in order to have that same infrastructure. So what does it mean and what do those commitments look like? They are essentially take-or-pay agreements, where the customer commits to taking products, except volume and pricing. So both of those are contractually – call it, contractual obligation for volume and price. And what we will do is we will invest in order to support those volumes at this price. So that’s at the high level structure. They extend anywhere. On average, it’s 4 to 5 years, to some extent, 7 or 8 years depending on the automotive and the new products that we are putting in. And on average, they have almost 200 parts in them. We identify it by part number. So it’s not this high level of are you going to pull this much revenue and this much margin. It’s a part number level list, volume and pricing.
Now the following question is, well, how enforceable are they, which is really the question that comes up now given some of the outlook in these markets. Let me just say it at a high level, they are legally binding agreements. So we will start with that. Now what happens when a customer says, hey, we have a – we see softness coming in, in 6-month or 9-month and we are not going to be able to pull. Well, benefit number one is I did get the call, where before, backlog will disappear 30 days before you ship you are stuck with inventory, then what? The LTSA, if they do one thing, they will guarantee a phone call. The bad phone is going to ring way in advance where the customer is going to be proactive and so will we. So what does that mean? It means we can manage inventory. I can move capacity somewhere else. So, it’s not dead inventory that ends up being reserved or scrapped. And the customer gets, we get a much longer signal of demand – of end demand than you do just by looking at the backlog, which is almost backwards looking signal. Now what – and we will negotiate, because look it’s not our benefit either nor is it the customers to shove inventory. You just push the problem down the road. So we want to manage inventory. We want to be proactive in managing inventory. What is nonnegotiable is the pricing? So we can discuss demand. We can discuss demand signal. But what is nonnegotiable and it’s extremely enforceable is the pricing conversation that we are not going to have. So that’s how we look at LTSA and the engagement we have with our end customers.
Thad Trent
And Joe, I would add that we have got a lot of cases where customers are co-investing with us. So they are actually putting capital upfront. So it tends to make that relationship very much like a long-term partnership, which is what our industry hasn’t had in the past and very sticky. So in a downturn, we expect to pickup market share well through that, right. So I think it’s an engagement that is totally different than what we’ve seen. It’s much deeper and it’s at a C level.
Joe Moore
Yes, that’s a great explanation. Thank you. Can you talk generally about the automotive market in this context of LTSAs? I have never seen anything like the last 2 years where the entire auto industry is bottlenecked by semiconductors, and in fact, a lot of industries. Can you talk to how that changes behavior going forward? Do you see customers trying to put safety stock into place? Are they – is that at the OEM level, Tier 1 level, what’s your customer behavior looking forward?
Hassane El-Khoury
So I would say it’s not a one size fits all. It depends on the customer strategy. But I’ll tell you the most proactive customers, both specifically at the OEM level, have engaged in the LTSA because that’s, from an engagement level both us and them, that is a sustainable model for us to move forward. We will invest CapEx. We will get ahead of it. They have to pull it, kind of like we had the conversation on the LTSAs. That engagement is a much higher level, meaning we have LTSAs directly with the OEM, although they may not directly purchase. They will do their directed purchase from a Tier 1 based on the LTSAs that they have committed with us. I would say that’s the model that the OEMs – when an OEM can build a $100,000 car because of $1 semiconductor part, it kind of makes you think something has to change. And that’s what changed, and the LTSAs have provided that. And we have seen in the last few quarters, even the number of escalations for LTSA customers have kind of died out because we committed, and the customer doesn’t have to call unless it’s a strategic conversation now, not a escalation calls. There are others where they are hoping to go back to the good old days. I don’t know what those days are after 2 years of being in this, where they look at it as potentially call it, a business continuity plan. So not inventory for the sake of inventory, but staging inventory between us and the end customer in order to absorb any supply chain disruptions. That is a much longer-term model because I will tell you today, there’s no inventory in between us and the OEM. Everything we ship is going into an end car and end module and end system that goes into the consumer – in the consumer hands. There are talks we’ve had with customers about how do we plan a business continuity as far as inventory staging, but that’s a conversation for, call it, beyond ‘23 because even in ‘23 for automotive, we have no intent nor do we have the capacity to put buffer because we’re still struggling to meet end demand.
Joe Moore
Yes, got it. And ideally, if there is conviction in the LTSAs, they don’t need.
Hassane El-Khoury
That’s exactly.
Joe Moore
Okay. Can you talk a little bit to some of the end markets, maybe starting with ADAS? I feel like EV has kind of jumped ahead of ADAS in the innovation pipeline. We were hoping for cars with no steering wheels by now, and we’re still looking at kind of single camera forward-facing ADAS. When you talk about the opportunities for image sensors, you’re starting to see proliferation, more cameras per car and where is that trend going?
Hassane El-Khoury
Yes. Look, ADAS, if I look back – and I’ve been on automotive my whole career. If I look back, call it, even 5 years ago, everybody is talking about Level 4, Level 5. Well, here we are. There is a new term, it’s called Level 2+, which is Level 2 and everything else. So whatever that is, it’s driving content. And if you look at those 5 years, the last 5 years, the number of cameras per car has doubled, and we expect that to double in the next 5 years as more and more vehicles put some level of autonomy, like I said, Level 2+. We don’t have to get to full autonomy, but the content is driving that adoption. What does that mean? We have vehicles today, not full autonomous vehicle, but call it Level 2+ vehicles that have nine cameras in them. So it’s not just the forward looking. We have seen that content proliferation. And for us, that content not only focuses on image sensing but our power and mixed signal analog content that goes as a cross-selling or a BOM coverage for – within our camera. So you get a PMIC and the camera. So the content for us expands as far as dollars from just the unit content that you see.
Now more importantly also, within that time frame, outside of ADAS, you have the content now in-cabin, passenger monitoring or driver monitoring where that’s driving a different adoption where the way you figure out if a driver is tired or not is how fast they blink. Any one of us who drives too late or too long, you start like trying to keep your eyes forced open because the faster you blink, the more tired. Well, that requires technology you can’t miss a blink of an eye. That’s putting us, again, as a competitive advantage back to the conversation about value where we bring value with our technology, and that camera is getting adoption. So it’s not just the ADAS outside the vehicle or the perimeter, but it’s also within the vehicle for safety as it relates to the driver. That’s what’s driving our image sensing business. So we don’t need to get to Level 4 and Level 5. The content is getting created today. And we have about an 80% market share in ADAS today, and that’s going to drive that proliferation as more and more penetration happens in ADAS for your day-to-day vehicles.
Joe Moore
That’s good stuff. Thanks. So maybe moving to EVs. Silicon carbide, very topical. You guys, a couple of quarters ago, sort of started talking about $1 billion number next year, which puts you among the market leaders. It’s kind of surprising how quickly you kind of got to that point. Can you talk to the progress there? What are the challenges that you have ramping that business and how much visibility you have to that $1 billion target?
Hassane El-Khoury
Sure. So let me start with the visibility and then I’ll give you kind of how we got there. From the visibility, we have very clear visibility. You heard me talked about the LTSAs. All of our silicon carbide 23 and beyond is under LTSA. So our strategy as a go-to-market strategy, we’re not going to put silicon carbide capacity unless it’s under LTSA. That protects our return of capital, which is very solidly tied to volume and pricing. So we maintain our financial model and the return on that capital. So that’s one. So it’s all under LTSA. Overall, over the next 3 years, we talked about a $4 billion of committed revenue, committed revenue under the structure of the take-or-pay with our LTSA. So from a visibility, clear visibility.
Now obviously, what we have to do to get there is ramp and ramp like hell because, like you said, we went from a few years ago to where we are headed, which is a market leadership. That doesn’t happen just by cruising into the growth. So what does that mean? If I take it a big high level components we have the substrate. So we’ve acquired a company called GTAT that provides access to that substrate, and we’ve been scaling that. Since the acquisition, maybe about 13, 14 months ago, exiting this year will be 5x the capability, capacity-wise, of that company. So we’ve invested heavily in that capital in order to get the growth to fuel our future intent and LTSAs. We’ve also, since last year, doubled our fab capacity to get to the silicon carbide out, and we will double that again next year. Now the conversation – then the following question is, well, how do you double fab capacity when equipment is lead times and so on? What you have to look for onsemi, this is not new for us. We’ve been doing power semiconductors for two decades. So we have fabs at scale that do power with silicon today. So moving from silicon to silicon carbide is what we call brownfield expansion, which is much faster and more predictable and lower risk. I don’t have to get a fab up and running from zero to high ramp and have to deal with yields and all of these things that you do with a hard ramp. We have the fab. It is producing. It is producing at scale or just moving silicon carbide into it. So these are the execution components that we have to do. So the beauty of it, it’s all execution. The market is there, the revenue is there. We have visibility. We have commitments. Now we’re executing towards that.
Joe Moore
And with the GTAT acquisition, I guess that seems like it’s pretty important to get that up and running. We sort of know what your wafer supply agreement is with Wolfspeed. To get to $1 billion number, you need a lot more than that. How do you – I went to an event in New Hampshire, which was great, where you kind of gave us an overview of what you’re doing on the substrate capacity. But how do you get customers comfortable with that? How do you – I think of this being something that requires multiple years of qualification. So you’re ramping GTAT very quickly. They need to qualify it fairly quickly. How do you get to that number that quickly?
Hassane El-Khoury
The great thing about GTAT is we’ve been – we worked with GTAT a few years before we acquired the company. So a lot of our sampling, a lot of design win and design-in that we’ve had was also based on GTAT substrates. So for us, it’s not a gap or it’s not a gate to get there. It’s always been in our baseline. But now we have the capability and supply assurance and more importantly, the vertical integration that requires. But ultimately, what does the customer – how does the customer get comfort? Just like you mentioned you visited GTAT at the event. We have customer visiting the facility and doing what they do, which is they audit, they review the data, they review the ramp, they review the capability, and that’s how they get comfortable. When a customer comes and signs billions of dollars of long-term supply agreement, they are due in diligence. They have done the diligence locally at New Hampshire and Hudson where we have that operation, and they have become very comfortable. So that’s the credibility that our vertical integration has brought to the industry and to our customers in order to – for them to sign up with the long-term supply agreement.
Joe Moore
Great. And so – and I have one last question on this, and then I’ll see if there is questions from the audience. You talked about the start-up expense of 100 to 200 basis points short-term of bringing up the silicon carbide capacity. Can you talk to the longer-term margin ramifications of this business? Does this fall into the 48% to 50% range at some point? And what has to happen for it to get there? What type of scale and how much internal substrate you have to have to get to that point?
Thad Trent
I can take that. Yes. So look, at scale, the silicon carbide margins are at or above the corporate average. So as you said, in ‘23, we expect 100 to 200 basis points headwind as we ramp because we don’t exclude all the start-up costs out of our non-GAAP numbers. We think by the time we exit ‘23, we will be back to parity and then it’s off to the races after that. So it’s favorable margin. We’re obviously pricing it that way at scale. It’s just a matter of getting through this ramp pace.
Joe Moore
And is that – whether – is that true whether you use other people’s wafers or your own? Or you need internal wafers to get to the above corporate?
Thad Trent
Well, the internal wafers obviously are more favorable, right, margin stacking. Yes.
Joe Moore
Okay. Great. Questions from the audience?
Hassane El-Khoury
We got one here.
Unidentified Analyst
Hi. From your angle, can you see changes in the EV leadership market, like changes in leadership in the EV market?
Hassane El-Khoury
From the OEM side?
Unidentified Analyst
Sure.
Hassane El-Khoury
Time will tell. I don’t – when people ask about our LTSAs for EVs and so on, I always make sure it’s – I mentioned it’s geographically distributed and OEM distributed even within the OEM multiple models distributed. We’re not here to bet on the right horse. We’re here to make sure our technology is available to anybody who’s got those aspirations. So that’s the reason we go abroad and we go wide in order – because even within an OEM, let’s not talk between OEMs. Within OEM, sometimes a vehicle will sell more than they expect versus the other. So you want to be able to have a very distributed model in order to sustain your growth and the margins. So that will be my expectation. But there is a lot of interesting things happening at the OEM level, not just the traditional, but you have a lot of disruptors that are pure EV players that are coming in. So the next few years will definitely be interesting to see.
Unidentified Analyst
So you want all your children to do well?
Hassane El-Khoury
Yes.
Unidentified Analyst
And you mentioned geographic diversity. Does that – your presence in China for EVs?
Hassane El-Khoury
Yes, but that’s the thing. Obviously, China is a big EV market, but we’ve talked about our engagement with NEO, but again, when you talk about it from an LTSA perspective, that $4 billion, it’s beyond just that one account in China. And that’s important for us to be able to get full access to our technology to the customers that see the value in adoption.
Unidentified Analyst
Thank you very much for the presentation. And you may have already covered this because I missed the first 5 minutes. Apologies, if you already have. So on the EV and ADAS, there is a phenomenal demand growth. We all know that. But could you explain about how your competitive landscape looks like and where your sort of value differentiation is?
Hassane El-Khoury
Sure. So from a value differentiation between the two technologies, not to get too much in EV, but from an imaging perspective or the sensing perspective that goes into ADAS, obviously, the differentiation is we’re the market leader today. We have 80% share, as I mentioned, for ADAS, specifically 60% in automotive. Now why is that? Because unless you design the imager for automotive, you’re not going to be able to solve the automotive use case. What does that mean? Anybody who tries to take a picture with their phone and the sun is in your face, you see dark. And what do you do? You tell the person let’s switch a different background in order to get the exposure. Well, you can’t do that when you’re driving a car. If you’re in a tunnel, the sun is in your face that could be a truck parked out there. And if the camera is blinded, it’s not really ADAS or you don’t want it to be ADAS. That’s called technically high dynamic range. Bright is bright and dark is dark. Our cameras are able to do that with, obviously, no shading, nothing, just normal same camera, same technology. That’s important. That’s why we win. Most signs today, speed signs are LED. If you take again a picture of an LED sign with your phone, it will look distorted because the LED flicker. Exposure breaks that picture. LED flicker mitigation designed into our products allows us to do that for autonomous driving, where a clear sign, no matter what they export. Those are technology enablement that our products do that allow us to service the use cases for automotive that are tailor-made for automotive.
On the silicon carbide side, it’s all about efficiency. We talked about GTAT, that’s more on supply assurance. We win for two reasons, efficiency driven by technology. What that means, technology and efficiency on the device, so the silicon carbide die with a package. When you do power semiconductor, package is as important as the device itself. So a lot of people failed to mention that or failed to invest in that. Silicon carbide is a high-power wideband device. What that means, if you can’t get the power off the chip, the chip is not going to give you the efficiency. You’re not going to get the range. You’re not going to save on battery. When we couple our device and our package, our end customers, the system designers, the OEM are able to get more range with the same battery volume or get the same range by removing battery volume. That’s cost and weight, which are key to a higher range. So, all of these are a competitive advantage that we combine both device and packaging in order to capture majority of that market we talked about.
Joe Moore
Great. Well, as on that, we’re out of time. We will wrap it up there. Thank you so much.
Hassane El-Khoury
Excellent. Thank you.
Be the first to comment