ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

Good morning, everybody. My name is John Vinh. I cover semis here at KeyBanc. We're pleased to have the management team of onsemi here. We have Hassane El-Khoury, CEO; and Thad Trent, CFO. Welcome, guys.

Thank you.

Thank you.

Obviously, front and center, everyone is talking about the cycle. It seems like you potentially finally found a bottom here. You're calling for a sequential growth into 3Q, and then you're still calling for an L-shaped recovery. What are the signs that you're looking for, for an L-shaped recovery? And what does an onsemi L-shaped recovery look like?

So when you talk -- look, when we look at quarter-on-quarter, we -- forget about the guide, which is up slightly. We look more for a demand trend that are more sustainable. We've talked about demand stabilizing. We've talked about demand actually recovering in certain pockets in industrial. We talked about auto and industrial being flat to slightly up. But net-net, what we're looking for is a sustainable demand signal that goes beyond a quarter. That's when we will start taking action and changing our kind of cautious approach that got us to sustain our financial predictability in our margin. That suited up very, very well. We're going to maintain that style of management until we see a more sustainable sign other than stabilization. But we'll take stabilization, obviously, from where we were. The next step is just monitoring across market, not just within a market green shoot but across market signs of recovery.

All right. Hassane, you've been pretty adamant about seeing an L-shaped recovery. What does that look like and why are you still so adamant that L-shaped is in the cards, right? Because obviously, the macro backdrop right now looks pretty weak out there.

Yes, it's -- there's some of it -- where we are today is really the inventory digestion that we've had. And people talk about, okay, well, how long is this L-shape? How long is inventory going to take? And my answer is always inventory time line or inventory digestion is directly proportional to what the demand does. If demand accelerates, all of a sudden, you have lean inventory. If demand doesn't, you need more time to digest the inventory. So for us, from an L-shape, it goes back to the uncertainty in the outlook. Nobody can call. People ask me, are we at the bottom or can you call the bottom? And my answer is very simple. I'll be calling the bottom when I'm standing on the top on the other side. That's the only predictable bottom that you can call is when you're out of it. So until we get to that sustainable growth that then you can look back and say that actually was the bottom a quarter or a few quarters ago, we're going to be managing to what we see and guiding to what we see, which is that L, which is a stabilization. In retrospect, it is a positive development from where the deterioration was coming up to this point.

Great. I assume you're spending some time with your customers to just try to understand where they sit on inventories at this point in time. Do you feel like you've got a handle on how much more inventory needs to be stocked at this point? And are you -- do you have a sense of just other parts of your business that you think you're further through this inventory destocking than others? I'm just thinking maybe across IGBT, SiC, CIS.

So the inventory destocking is not actually by market, which is why it makes it so difficult and why I always say it's actually tied to demand because it's not at a market or it's not at a technology perspective. Of course, there are some pockets that you can't put your finger on. But in general, the destocking is on a customer-by-customer basis. There are some customers that have a different approach to inventory management. Others are actually already at a lean stage. But we still talk about inventory digestion because it is not at an overall market segment. You can't talk about inventory dissection in automotive in general. We talk about it in aggregate at what it does to automotive demand, but we manage it on a customer by customer based on what we see and what we believe they have on their shelf. And some customers are more aggressive in drawdown of inventory because of their cash position. Other customers want to be there for their customers who are the OEMs. So they may hold a little bit more than their peers as a competitive advantage. So the inventory digestion is really on an account by account. We're very close to both OEMs and Tier 1s so we're able to navigate. So when we -- when Thad and I talk about, yes, we're under-shipping demand, that is a path to drawing down on the inventory. That accelerates if demand recovers. But net of that, we're just going to be very cautious about what we ship. The same with our channel inventory. We've managed a very lean channel inventory throughout not just the downturn but throughout even the peak. Today, even our distribution partners will take more if we ship it to them. But we're going to have to be very cautious about really matching what we believe the end demand is. So we're managing it very well. We're increasing in pockets in our distribution channel because we have a strategy of going broad at this point in the mass market that we have starved in the last 2 years during the shortages. So there are effectively strategic decisions we're making that impact inventory in the channel and at some customers, but those are very surgical and not at a market level.

Thad, some of your peers out there have said that they're starting to contemplate refilling the channel because they've been so lean and they want to be prepared for kind of an upswing. What do you need to see before you start refilling the channel? And also, where are we here right now in terms of channel inventories, and what's normal for you?

Yes. So we've been running about 7 weeks. We've slowly been taking that up, and we've been mixing the inventory in the channel to the mass market. So through the supply-constrained environment, we just couldn't support that mass market, right? So you think about a lot of customers, small volume, typically go through distribution, typically high margin. We starved that off because we just couldn't support it. So what we've been doing through this as the market got soft is we've been shifting that inventory. So we've been taking the inventory up intentionally, and as Hassane said, our distributors want to hold more inventory. We're actually limiting with the hold. So we've been shifting that inside the channel. We're just under -- we're at 8.9 weeks. I think we're going to run in this range for a little while here. Historically, the company would run 11 to 13 weeks. Now that was a lot of the businesses that we exited so I don't think we're going back there. But we'll slowly take inventory up as we see demand recovering. What we want to make sure we have is that when there is a recovery, you have inventory in the channel, you got inventory sitting in die bank that we can launch quickly and we can support an upswing as well. You don't want to get caught with too lean of inventory. So to answer your question, we're going to run about this range probably through the rest of this year, maybe taking it up slightly. But I think over time, we may go up to 10 weeks, somewhere in there. But for the foreseeable future, we're going to stay where we are.

Just last question on inventory before we move on. I think one of your peers, I think, surprised a lot of people when they said that some of their customers want to keep as little as 2 weeks of inventory on a normalized basis. When you talk to your customers about what the new norm is, what's your sense of what that new norm is? Are we looking at levels above or below pre-pandemic norms?

Yes. So I think I know the background of the 2 weeks. This is where I said there -- back to your other question on inventory, it's account by account. And in the automotive segment, I'll tell you, there is a kind of a friction between the OEMs and the Tier 1s. I think 2 weeks is too lean. Two weeks is a sneeze away from losing OEM volume if there's an uptick. At the end of the day, the OEMs are the ones that paid a very heavy price during the shortages. So I think getting down to the 2 weeks, but again, that's a customer choice. I think in an uptick where Tier 1s are going to be constrained and damaging an OEM ramp, I think that's where you're going to see companies that will no longer be around or companies that will get consolidated. I mean that's usually what happens. So I don't think that's a healthy level just because it's not even a sprint capacity for a week that will sustain it. Now is it back to pre pandemic? I think that overall, what you have to ask is what is the total inventory and what is the sprint capacity you can have? And that means it's not a question of what our customers have to hold or what the OEM has to hold or what we have to hold. It's what is the combined mix that we have to have. To give you an example, we're staging with customers that we have LTSAs, and we have a dialogue about how do we do BCP, which is business continuity planning, which is a managed inventory setup. What that means is we say we will hold X weeks in die bank. We will hold X weeks in finished goods. The disti, if there's a disti in between or the Tier 1 will hold this much and the OEM will hold this much. We all know how much we have in every pocket. So as the OEM starts ramping and sucking faster from that WIP, we can see it real time and replenish the die bank at that same rate or slightly higher if the ramp is accelerating. That's the difference between "BCP" and holding inventory that nobody sees, which everybody gets caught with. In Japan, BCP is a way of running a business. If you look at the pandemic, Japanese OEM were the last ones to cut production. Why? Because there was a managed BCP. So the dialogue between us and our strategic customers is more on BCP rather than how much do you have -- how much is -- do you think the Tier 1 is holding or how much the OEM or what do you think is the disti and so on. It is a dialogue about how do we react constructively to all these parties in order to ultimately support an OEM ramp and not damage a potential demand signal. That's the dialogue we're having, and that's the dialogue we're going to continue to have with our strategic customers. More and more of our customers are catching on. So it's not about the 2 weeks, it's what is it in context of everything else before it and after it.

Great. Questions? Right. Maybe we can switch over and talk about EVs. I think on your last call, you said that at a recent China automotive show, you guys counted that 60% of kind of the new EVs introduced by China OEMs were onsemi wins, so in silicon carbide. Obviously, the biggest, China, EV market in the world, but there's obviously a lot of angst and consternation about the China market, right? There's obviously a big push for domestic sourcing over there. They're obviously ramping a lot of silicon carbide capacity. Can you just help us think through and understand the China risk and how you're looking to manage through that China risk going forward?

Sure. Well, let me first talk about the China opportunity because the opportunity is going to mention how we manage the risk, which is no different than a risk in Europe or a risk in -- which is a competitive risk that always exists. The opportunity in China is the biggest penetration of battery electric vehicles is in China today. Therefore, you have to be present in China and you have to be winning in China. That's the comment I made in our last -- in my prepared remarks on our earnings call. And those are not -- it's not a funnel to be. It's not something we're designed in and so on. It's vehicles, platforms that were on the showroom floor that were introduced, that we are the silicon carbide for traction inverters on these vehicles. Now of course, the cars have to sell. That's where the demand comes in. But from a something we control and we execute, we're all in the right area and the right places with the right technology. Domestically in China, every one of these brands want to ship outside of China. Therefore, the competition is not on the available capacity, which today, we can't mix the available capacity of substrate with devices. Let's separate the 2. But from a competitive perspective for the OEMs, they want the best products out there because they have to compete on a global level. That's where we come in. So the local suppliers are not up to par. They're a few generations away from a device and package side that the OEMs have to measure because any lack of efficiency on the silicon carbide means a battery, a cost that the OEM have to put. If you lose range because of silicon carbide efficiency, you just have to add battery content. Battery content is way more expensive than whatever benefit you get from a local source. So today, the efficiency still matters and we're the leaders in efficiency. That's why we still win. That's why we, even up to the last few months of the show in China, we're still the -- a market share leader over there. On a competitive landscape, yes, there is availability of substrates. So when you talk about, oh, there's capacity and there's overcapacity, whatever you want to talk about it, we're talking about substrate. That's not a competitive disadvantage for us because we don't ship substrates in the open market. We have substrates internally for our own consumption. And we've always said we're going to be 70% to 80% internal and that 20% to 30% external. That external could be from China. And we're able to qualify in China. We can assume and really predictably assume that the quality and the consistency of the output from China is going to get better over time, so there's no reason to ignore or look the other way just because of a point in time. However, you still have to anchor on the geopolitical risk of it. Having the capacity and saying you're a global supplier like onsemi going for European design wins and going for U.S. design wins, you can't really underplay the geopolitical risk of having substrates solely from China. So having this flexibility but our ability to do our own substrate is a competitive advantage and remains from a supply assurance. And our announcement about an operation in Europe adds to that supply resilience that we are able to show and commit to our customers. So we believe we're in the best of both worlds where we can leverage that innovation that's coming out and the supply that potentially would come out from China while we maintain a supply resilience and supply assurance with the mix that we have.

That's great. Maybe just if we take a step back and think about the overall EV market, I know, Hassane, you've been in the auto industry for years even before you became CEO of onsemi. It seems like we have maybe gotten a little bit over our skis in terms of what we thought the slope of the EV market is. Now that obviously we're working through this correction here, can you help us understand what's the right trajectory that we should be thinking about going forward qualitatively in terms of the EV market going forward? And what are the challenges that we need to overcome as an industry to see greater adoption?

So I'll tell you it won't be linear but it is up and to the right. And the up and to the right is a penetration of EVs into the total SAAR. So even if you take SAAR as being flat, let's just take that to make it easier, if SAAR is flat, the percent of EVs year-over-year as a percent of total SAAR is going to increase. Now when I say it's not linear, you can't draw a line between now and the end of the decade and say, all right, this is what it is, which is where the industry was, it's going to be linear, it's going to be -- and there's not going to be any hiccups. Any new technology adoption is lumpy. Now what does that mean? With any adoption, there's always the positive and negative catalysts. And what we've had is a positive catalyst in the last few years, which is energy cost. Now we are in a potentially negative catalyst. When I say negative, it doesn't mean it goes backwards. It means the slope goes up. So you already had all the first adopters that drove the electrification. Now what we're working through is how do you have a charging network deployment? When you have more charging networks, you're going to have that secondary surge, not the early adopters, but the people or the users that actually want an EV, but it wouldn't fit their commute or it wouldn't fit the lifestyle that they have because they lived in an area that charging network was not there. Then you're going to have the charging time, which is another technological advance that we are participating in with our charging modules where we introduced at CES, we introduced a 350-kilowatt charging module that allows a charge to almost full charge, above 80%, in 15 minutes, 1-5. Okay, well, somebody who says, I can't deal with a 5-hour charge because I can't leave it at home or whatever it is, now 15 minutes, you're going to get the other surge. So that's where that lumpy adoption is going to happen. That's what we are investing in. But you have to invest on both sides. So we're investing on the EV side, and we're enabling those positive catalysts, the fast charging for batteries and the charger deployment. That's our energy infrastructure business. That's been very, very healthy. So we're investing on both because both of them have to ramp up at the same time in order to maintain that mega trend of electrification and penetration.

Right. Thad, maybe we can talk about silicon carbide gross margins. How much of a headwind is silicon carbide to your overall corporate margins today? And what's the path and timing for us to get to accretion there?

Yes. So when we look at the way we price the product, right, it's at favorable gross margins, at our corporate average or better. As we're ramping, we're growing into that capacity and we're bringing capacity on for '25, there's some underutilization charges. So right now, it's below the corporate average. But if you think about at scale, it will be at or above. Now when do we hit that? We'll start to see that happen probably late '25 and in '26, depending on the market penetration there. But we have line of sight to that. And because of our brownfield investments, right, it's capital-light when we think about the capacity that we bring on. So it's a high ROI on these investments. We talked about the Czech Republic and the investments there, the conversion to 8-inch as well. These are all capital-light investments, right? So we see a high return out of those investments and a short time to revenue if you think about it in each one of those. So it's just a matter of ramping into this capacity.

Great. And then you also had reiterated 53% as your long-term target model. I would imagine kind of near term, most of where you're sitting on your gross margins today is a function of underutilizations. But maybe can you just walk us through the bridge for how do we get from here to 53% longer term?

Yes. So our target is 53%. If you go back almost 2 years ago, we were bumping right up against 50%. We started taking utilization down at that time. Our utilization today is at about 65%. We peaked out on utilization around 83%, 84% so you can think about that as being fully utilized, right? So utilization is the bigger drive in the short term in terms of driving gross margin. So for every point of utilization, it's 15 to 20 basis points of gross margin improvement. So as the market recovers, we start taking utilization up. You can do the math there on going from 65% to 80%-ish. So that's the shortest term driver. Now the other thing that we have that we've talked a lot about is the dilutive impact of EFK, our GLOBALFOUNDRIES business that we do in there. It's about 100 basis points dilutive this year. This is just a low-margin business. We've done a great job of taking cost out of that fab, and we now have that at parity. So this is just low-margin business that we have to exit. We will be exiting that by the end of '25, so you can think about this 100 basis points rolling off from '25 kind of in a linear fashion as that business ramps down. And then the third piece of it is we divested 4 fabs in 2022. It's $160 million of fixed cost that we'll start to realize as we bring that production into our network. So we've built inventory to -- for the fab transitions. We need to burn through that inventory. And then every wafer that we start producing inside is accretive in terms of the margin profile in that business. So today, the way to think about that $160 million is if we have those fabs, it would be a headwind. So our margins would be much lower than what they are today. So we have a cost avoidance today that will turn into a cost tailwind or a margin tailwind going forward. And then the last element is just you talked about silicon carbide ramping. And we've got other products that we've talked about that are high gross margin, 60%-plus gross margins that are in R&D that as they ramp, that gives us a tailwind, too. So you start adding all those up, and you can get pretty close to the 53% right there.

Great. Speaking of higher gross margin products, I thought it was interesting that you recently mentioned -- you talked about kind of an analog/mixed-signal opportunity. Maybe, Hassane, you could just give us some perspective on just where did this come from? Was this part of the acquisitions that you've made of some of that ADI team are in the past? And then what target end markets are you targeting? And when should we think about just the timing and ramp of the analog/mixed-signal revenues?

So I'll give you a lot more detail later this year because I promised my team I won't disclose before they want me to. But what I will tell you is, first off, the investment that we're talking about in analog/mixed-signal, Sudhir did a very good job framing it and what the target market and all of that is at our last Analyst Day. So that's where it started. So it's an organic effort that we've done roots up from a technology platform, so it starts with technology. It's not just a new market and a new product line. It is an underlying technology that we actually have completed in East Fishkill, so a 12-inch East Fishkill-driven. Back to Thad's point on as those ramp, they will help with also utilization. So that's on the platform. The platform is a cross-market and cross-product platform. What I highlighted on my last earnings call is really kind of almost the boundary where we said it's an ultra low power for the medical and it's a high performance, high-power analog for automotive, for example. So it gives you kind of the boundary of the versatility of the underlying technology. The products and the segments and how we're going about it, that's why we're going to be announcing in the, call it, the end of the year, around the fourth quarter. But at least where we are today, we're sampling, which means what we talked about at Analyst Day, we focused in the last few quarters on really what we can control, which is our execution. That is part of it. We have proven the technology. We have sampled the customer and we're proliferating. The go-to-market for it, we're not going in general purpose, a mixed-signal/analog. This is not a TI or an ADI model. This is not the same market. What it is, is a cross-selling opportunity with our other core markets. So think about the silicon carbide or IGBT, we have the switch. The main event and the system is really the switch. The decision is made on the switch based on performance and efficiency. Well, if we have all the drivers and the mixed-signal/analog around that switch, the customer is more likely to use -- you can think of it as a chipset to use our mixed-signal around the switch, given that the switches is more strategic. And that's the cross-selling opportunity versus just a mass market or a general purpose analog. So that's the framework you can think about as far as how we're approaching the market, with highly accretive gross margins as well.

Great. I think we're out of time. Thank you, guys.

Thank you.

Thanks, John.