NXP Semiconductors N.V. (NXPI) Earnings Call Transcript & Summary

Good morning, everyone. Thank you so much for joining. I'm Toshiya Hari. I cover the semiconductor space at Goldman Sachs. I am very excited, very pleased, very honored to have the team from NXP with us this morning. We have Kurt Sievers, President and CEO; and Bill Betz, Executive Vice President and CFO. I will kick off with a bunch of questions, but we'll definitely leave some time at the end for questions from the audience, so please be ready. Kurt, Bill, thank you for attending. Really appreciate the time.

I wanted to kick off with a relatively near-term question. You reported June quarter results that came in toward the high end of guidance. You guided September quarter to grow 3% at the midpoint. I don't expect you to give us an update here today. But perhaps, what were some of the puts and takes that you guys thought about, considered when providing guidance? And what are you seeing in the market today, I suppose?

Yes. So good morning, first, Toshiya. Thanks for having us. Good morning, everybody. Yes, the environment, clearly, is still in turmoil. But bigger picture, NXP is, first of all, seeing two things. So one is we have a relatively high exposure to the automotive and industrial markets, which have been much more resilient during the current semi down cycle. So our exposure to mobile, compute and consumer market is comparably little. But we've seen the down cycle there, and we had declared in the last earnings, which you just quoted, that we did see the trough there actually in Q1. So our Consumer, Mobile, Compute business did trough in Q1. Since then, we are seeing a gradual but slow sequential ramp-up. So Q2 was better than Q1 for those markets, and we guided Q3 again sequentially, slightly up. I say slightly because most of those businesses, in our case, are exposed to [ China ]. And clearly, the Chinese macro is weak, such that it's not the sharp rebound, which some people would have hoped for, but a relatively gradual improvement. And then on the other hand, the second point is clearly the automotive and core industrial business, which continues to be resilient for us. So we see solid demand trends, where, and I think it's a little bit behind your question, I know that some of our peers have been a little bit less positive in the more recent past relative to what they saw in automotive and industrial. We believe most of that has to do with how different companies have managed and continue to manage inventory. In NXP's case, managing inventory has two legs. One is the inventory being built or not being built in the channel, in distribution channel; and secondly, with direct customers. Now on the distribution side, and I think we've talked about this now for 6 or 8 quarters in a row, since we had burned our fingers on that 8 years ago, we really, Bill and I really wanted to do it differently this time. So we have put a lot of discipline on keeping our distribution inventory at 1.6 months, which is a full month lower than our long-term inventory target like $500 million. And that's one of the differentiators to, I think, what some of our competitors have done, which have been a bit more aggressive in shipping into the channel and face now more of a cliff in trying to adjust. Same thing on the direct customers. This year, for many of us, has been governed by the so-called NCNR orders, where we, I think, have simply shown a bit more flexibility with our customers than several of our peers, or put it differently, when customers started to come to us already early this calendar year and said, "Yes, we have this finding order signed with you for 100, but we really think we only need 85 or 90." We didn't send them away. We said, "Okay, let's find a commercial way that you compensate us in a different form, like give us future design wins, pay us a little bit more, so let's raise the price a little. Then we're going to be flexible, and we're going to adjust the 100 to, say, 90." We've done that in multiple cases because I've learned the hard way in the past that enforcing NCNRs doesn't create any demand, unless it is a commodity product, where it's about market share, but we hardly have that. Most of our products are application specific. So long story short, I think we have been probably a little bit more prudent and thoughtful in how we manage both inventory build with direct customers as well as in the channel, which lets us navigate this cycle in a softer way than what you've been hearing from some of our competitors.

That's a great overview. I guess as a follow-up to that, Kurt, how would you assess your ability in measuring customer inventory? I think a lot of the points you made very clear, makes a ton of sense. But I know it's not perfect visibility, but how do you go about...

On the distribution side that we can keep very short, we have total full transparency because we have a contractual situation that we get weekly reports of the sell-in and sellout of our product into the channel from each of our distribution partners. And Bill and I, and we've done this for a couple of years already, we review that every week. And by that, we can very, very precisely make sure that we don't ship in more than what they can sell out. So that's a contractual agreement, which facilitates this in a very precise manner. With -- and that's 50% of NXP, I should have said that. So 50% of NXP's revenue is going through the channel. So it's a pretty material part of the company. On the direct customer side, Toshiya, it isn't that easy because none of them has an obligation to show us their inventory. But admittedly, the NCNR concept, which especially in automotive and core industrial is a larger part of our market this year, forced customers to come to us. So we didn't have to know when they have more inventory. They would come to us and actually say, "We have a risk of growing too much inventory because we signed up for too higher orders." So that was kind of a self-regulating mechanism that they typically came to us and asked for a relief, which is why we could handle it more flexibly. But in the end, we have to trust in the goodness of the relationship in most cases because there is no absolute precision in them telling it to us.

Got it. I guess you touched on this a little bit. But over the past couple of years, you've talked about your customer engagements and the nature of those engagements evolving a little bit, right, more intimacy, tighter line of communication with your customers. Can you maybe spend a couple of minutes comparing and contrasting how you engage with your automotive OEM customers, industrial customers today vis-a-vis pre-pandemic?

Yes. I think it's a step function of how that has evolved. And it really came through the supply crisis. It's not because of the pandemic. It was really the supply crisis. There, of course, initially that was a very tactical enforcement of the desperate cry for help on shipments. But through that, I think, we've built much stronger ties through all levels of hierarchy. I would even tell you that especially in the industrial space, we learned about OEM customers we didn't even know about before because they got our product. And we only learned about them when they were bell stopped because we didn't serve them anymore. Now through that period of 2 to, I'd say, 2.5 years of supply crisis, that tactical supply relationship has [ moved ] into very deep innovation partnerships in most cases because most of these companies, just like we figured out that a lot of their future innovation projection is based on semiconductors, it's based on our road maps. In the past, we didn't have a lot of direct interactions, so they didn't even know what road maps we had, so they wouldn't know what is the next generation of processes, what is the software compatibility, which we can offer. So there was a big desire also from their end to actually intensify the relationship. Today, I dare to say that we have deep and very strong relationships all the way to the CEO level with all of the top 10 car companies worldwide, which did not exist in any case before. So it's a sharp contract. That in itself is super helpful, both for managing the logistics and inventory build question in a more measured way, but honestly, from my perspective, even much more importantly, to manage future business creation. And I guess we should speak a little bit later in this fireside about the role of the software-defined vehicle and the compute architecture in the car, which is a big deal for NXP, where we just announced now our first 5-nanometer tape-out for a very high-end [ vehicle ] computer. That would never happen without the relationship to the OEMs, because we need their insight, we need their requirements to understand what the compute architecture of the car needs to look like, going forward. We don't get that from the Tier 1s. That is really a direct result of that relationship. So with all the pain, with all the trouble of this supply period, I must say that piece is a fantastic outcome because it's a step function in how we can interface with them.

Got it. Just wanted to come back to NCNR for a moment. You shared a couple of anecdotes, which was really helpful. To level set, can you remind us if NCNRs were introduced over the past couple of years? Is it something that you've had for a very long time? And if I'm a customer, if you're willing to be flexible, as you stated earlier, why wouldn't I sign up for an NCNR? What's my downside as a customer?

So first part of the question, it was really only introduced through the supply situation. We didn't really have that before. There might have been very unique, exotic single, isolated cases. But in general, that didn't really exist before. I have to clearly state that it came into life at the request of the customers, it wasn't our idea. It was actually the customers because they felt if they offer us a binding order for a longer period of time, they would move up in the queue for supply. I mean that was the logic. Now over time, of course, we also felt that it's good because it creates more stability for us to see what future demand probably looks like. We also felt it should force customers to not lightly place double orders and triple orders but think a bit more strategically about what they really need and what not. So I think in itself, it is something which is also going to stay. Now I guess your question is coming from what I said earlier about the flexibility, which we allow. Then it's like, "Yes, okay, if anyway, it's flexible. So why wouldn't I sign up in the first place?" It isn't that easy Toshiya, because when I say we offer flexibility, then that flexibility comes against another ask. There is still a binding contract. And of course, we didn't let them just shift their cash flow problem to us. I mean my CFO would kill me if I just allow to absorb the customers' cash flow problem for nothing. So we didn't do it for nothing, but we said, "Okay, we understand, of course, it doesn't create demand, so give us design wins or raise the price or drop maybe another claim which you had on us." So it was always about finding a commercial bridge, which means you wouldn't sign it too [ lightly ] because you know that NXP would still ask for any commercial compensation. It's still a contract [ upon ] the commercial value. Now going forward, by the way, we gave the child a new name. So we do it again for next year. We call it now [ Early Visibility Program ] because the word NCNR is a little burned because of misuse by some people. There -- it is totally voluntary. I mean nobody has to do it, but we have a relatively high sign-up rate because people simply understand that while supply and demand is now, in many technologies, just in equilibrium, we are still tied on a number of technologies, tied means we don't have enough capacity. And they also know as soon as other parts of the industry come roaring back, which will happen, I mean this industry has always come [ under pressure ], we will be sold out again. So there is a lot of customers now we say "Nah, it's actually a good thing to put a stake in the ground and actually make sure that we at least get the minimum of what we need next year." And that's totally voluntarily after what I think was, for them, not a bad experience this year because we've been flexible in a way. And for us, it's also very helpful because that gives us a feel on where the demand is going in the world, which is not easy to forecast.

And just to clarify, this is 12 months volume?

Yes, I should have said it to say, yes, it is a calendar year-based program. So that's why we talk now about next calendar year and only in automotive and core industrial. It's not being done in compute, mobile or consumer, purely automotive and industrial and always calendar year-based.

You're calling it EVP instead of NCNR?

It has the same characteristics.

Right. Makes sense. You talked a little bit about channel inventory before. Again, to your point, you've managed it really tightly at 1.5, 1.6 months. I think the long-term target, if it hasn't changed, is 2.5 months. Has there been a change in your go-to-market strategy? Or is this just you being very prudent? What would you need to see for you to be aggressively selling into that distribution channel?

Yes. So no, absolutely no change in the go-to-market strategy. So I would forecast that NXP will continue to be 50-plus percent exposed to the channel. We need the channel. I know there is peers who walked away from the channel. In our case, with our product portfolio, with a strong focus on the industrial market, which is very fragmented, many, many small companies -- for us, it's a big part of our value proposition. So no change, we stick to the channel. The refill, so the $500 million gap to the long-term 2.5 months target, clearly depends on a more consistent pickup of the sell-through. Now I've had many, many questions. What is that? Where would it come from? How do you judge that? It is not a single metric. But I would tell you, it likely is strongly associated with China because a good part of our channel business is for China and in China. And outside of automotive, China is weak currently. So it is really waiting for a more solid -- I wouldn't necessarily call it a full rebound, but for a more solid uptick in the China demand, that would probably be the signal we need to go back towards 2.5 months. Now, A, it will be gradual. So I don't think anybody should expect that in one quarter, we suddenly ship another $500 million additionally in there. I don't think we want to do this. Why we could, I mean, on the balance sheet, we have the product more or less. But I think it wouldn't be prudent either. B, I dare to say from today's perspective, now we have early September, I don't think it's going to happen this year. And Bill and I have been so disciplined on this for such a long period, and I think we start to get credit for how we kind of navigate this cycle in a soft and prudent way. We don't want to spoil this on the last mile. So we will be careful in doing it. We will do it, but really only when the demand is solidly there. And I don't see this for the rest of the year. Now at the same time, the guidance we gave you for Q3, but also kind of the directional guidance we gave for Q4, which is, I think we said half 2 of this year is going to be bigger than half 1, we also said half 2 of this year is going to be bigger than half 2 last year, none of that needs the refill of the channel, just to be sure. So all of those metrics we gave you from a revenue performance projection perspective are not contemplating to go beyond 1.6, and they are also not contemplating a rebound in China. That would be really on top.

Got it. It's very clear. Let's talk about pricing for a moment. Pricing has been a tailwind for you all and for the broader industry for the past couple of years. I know you guys haven't been price gouging. I know that's been a clear statement from you guys. But as we think about the forward, particularly in light of the current demand environment, how should we think about pricing? Is it neutral? Is it positive? Could it revert? How should we think about that?

Yes. I mean first of all, it's been a tailwind to revenue. It has not been a tailwind to gross margin, as you say. So it's just something we needed to do to compensate he very much increased input cost, which we are suffering from, which, by the way, in itself is not a big surprise in an inflationary environment, but more importantly, in an environment where the industry which we are serving needed much more supply. And that supply comes at a cost. I mean, we and our foundry partners continue to have to invest into this. And that's the depreciation, which is sitting then on the -- on our input cost. Another factor there, which is also going to continue to impact pricing, is the, I'm not sure I should say, decoupling or derisking. I just learned from [ Gina ] that we talk about derisking globally. But the matter of the fact is the semiconductor industry is now building more factories in Europe and the U.S. than it used to. And that is good from a geopolitical resilience perspective. But the cost of those factories, the manufacturing cost of those factories is going to be higher than what it used to be when it was more concentrated in Taiwan. And also, that is going to sit obviously on the pricing of our chips. So for this year, pricing will be up. I think it's fair to say this now in September that when we exit the year, the conclusion is going to be that for NXP, the '23 pricing will be higher than the '22 pricing. And mind you, last year, we increased by 14%. The year before, we increased by 2%. This year, it's not going to be as high as the 14%, but it's going to be higher than the 2%. So it's going to be somewhere in that range. For next year, the biggest cost block NXP has is the foundry cost. Mind you, 60% of our revenue is made from [ foundry wafers ]. Any indication we have today -- it's not finalized. It's too early. But any indication we have today from our Tier 1 foundries for next year is they increase price again. And we will continue to follow our policy of passing on that input cost, as laid out earlier. So early indication from that perspective is that I don't see an environment where we would go down with price next year.

That's very helpful. Maybe shifting gears a little bit and talk about the drivers, particularly in automotive. It's been a while since you hosted your Analyst Day, but you talked about a 9% to 14% CAGR for the core business. You called out applications like radar, zonal processors and BMS within the context of broader electrification. I guess two-part question. One, how have those drivers paid out over the past couple of years relative to your expectations? And two, do you believe 9% to 14% is still the right target, if you will, given the outperformance we've seen from you over the past couple of years?

Well, so we don't do now the new Analyst Day. So I can't tell you what it's going to be beyond '24. But I'd say, in automotive, I don't think that I see a single reason why it should be less. We will have likely performed on or above those targets for the total segment but also for the specific growth drivers, which you mentioned. So I think radar was targeted to be $1.2 billion revenue next year. And I have every confidence in the world that we're going to hit the target, which simply has to do with our growing share. So we used to be on par with Infineon. And I think as [ Joel ] and other market research companies reported, we left them now quite sizably behind us and grow share over them and that based on the design wins we have that it's certainly meant to continue. So I think radar is much on track. I mentioned it as the first one because it's big. I mean $1.2 billion revenue on one single application is already a quite significant statement. The electrification one, which is battery management, inverter gate drivers and stuff, is running ahead of target. And that is certainly thanks to our good performance. But clearly, it is also rising with the tide of the penetration of electric vehicles. When we did the Analyst Day in -- I think it was November 21, we did not forecast that the penetration of EVs and hybrids would go as deep as it did. I think this year, I'm hearing 33% or 34% of the [ SAARs ] is electric. That's more than we would have forecasted. And of course, that lifts our growth here. The third one is the S32 platform for the compute architecture in the car, which is microprocessors and microcontrollers, is running at target or above target, where I would say, and I mentioned it earlier, the big breakthrough there is only going to happen in the second half of the decade, far beyond the success we have today, which has all to do with the software-defined vehicle compute architecture, which is coming. There is just nobody else which has a portfolio, which is fully software compatible all the way from a 5-nanometer, 4 billion transistors vehicle computer on the high end, all the way down to simple edge nodes 180-nanometer, high-voltage, fully integrated microcontrollers. And all of that with software compatibility. So most of our classic microcontroller peers cannot follow us because they cannot do microprocessors, so they are stuck on the microcontrollers. They can't move to the processor architecture. And the high-end entrants like NVIDIA or Qualcomm, which have the MPU, they cannot scale down. So that big OEM requirements for one compute architecture across the board, which offers full software compatibility and reuse, satisfying that has been our target a long time. I personally remember we started the strategy, I think, in 2018 also, which is now 5 years behind us, it starts to play out now with new design wins. And I think the second half of that decade will give us then also a significant breakthrough in revenue, mainly thanks to ASPs. It's just that this high-end 5-nanometer device has an ASP, which is 10x that of a normal microcontroller. So that actually in itself is going to be a significant boost to revenue. So all in all, automotive in a good place, while clearly, the past years were not easy because while price has been a tailwind, as you rightfully said, supply was a headwind. So we could have shipped more if only we had more supply. So it was kind of supply holding us back, price pushing us forward, the outcome of that is being above target.

That's great. And maybe double-clicking on the progress you made in radar and the share growth there relative to peers like Infineon, what sort of separate you from the competition? How are you guys different and better?

Yes, that's really -- we were faster in moving to 77-gigahertz systems from the 24. So the whole industry moved from 24 to 77 gigahertz for higher resolution of the radar system, and we were much faster in bringing that to CMOS. So competitors were stuck with silicon germanium, which is more [ bulky ], less flexible, more expensive. And we were the first to have mass-volume automotive CMOS, 77-gigahertz radar transceivers. And then we were also first to bring that into one chip, fully integrated products with microprocessors and the RF chip on one piece of silicon, which is today running in very high-volume mass product, announcing these products now for the future. So it's -- and I mean we are in mass production. And you know that in automotive to be now in mass production means you had to have it qualified already years ago. So it is innovation which drives the differentiation.

Got it. Makes sense. The geopolitical environment is a challenging one, and it will probably stay that way for some time. You remind me, talked a little about China being weak, outside of automotive. But can you remind us how big or how small China is as a percentage of your business today? Has your strategy, your approach to China, has that evolved at all? And how should we think about emerging competition in the country as well?

Yes. So last quarter, I think it was 31% of NXP's revenue. Now mind you, this is about half of that China for China, and half of that is just pulled through China and ends up in the West again. So I'd say the, if you want the risk exposure to China from a China-for-China perspective, it is 15% of NXP, 1-5. You very rightfully said outside of automotive because we think automotive China currently is in great shape. It's doing very, very well. If you look at the success of the automotive EV OEMs, less by BYD, but there's many more, it is smashing. I think they are very quickly displacing their [ rest of ] competitors. And their strategy is, I believe, also going to be to now expand beyond China, Southeast Asia, but also Europe. I think it's today that Europe -- the Munich Auto Show has opened, and it has more Chinese OEMs there than ever before. You really see them breaking into Europe now and stealing share. So from that perspective, it's a very attractive market for us in at least the automotive space. From a competitive perspective, it is amazing. I had you asked me 5 years ago, I would not have said what I'm saying now or I would not have anticipated what I'm saying now, at least today, we only face our Western competitors in China in automotive and Renesas from Japan. But other than that, no really strong local competition. However, we see local competition coming in, in the low-end microcontroller markets outside of automotive. That's not yet in automotive, but outside of automotive, which isn't a big for us because we are not really a low-end microcontroller player, but we see them. So it is happening. Secondly, and that's just -- we just witnessed this, it doesn't matter to us, clearly, China is investing big time in silicon carbide, both from a manufacturing perspective as well as from a design perspective, which doesn't matter to us since we have chosen not to play in that part of the business. But that shows how both capable and aggressive, clearly, local Chinese competitors are. So we look at this in the following way: I think it will be completely naive to think they don't come also in our fields, which would be analog, mixed signal and the higher-end processing business. They certainly will. We take it very seriously. But as long as this is just fair competition, it's our job to deal with it. I would say, we know how to do this from a customer intimacy perspective, from an innovation perspective, we know what to do. The trouble starts when it becomes something which is more government enforced. So if either the U.S. government would broaden the export control to hinder us from shipping the kind of products which we ship into China or if the Chinese government would start to enforce certain take rates out of local manufacturing or local competition, that's where it starts to be painful. If it's just fair competition. I'd say we know how to deal with it. Now on the manufacturing side, I'm also not that worried because our manufacturing strategy is more and more about working with third-party foundries. And we are clearly working to establish foundry services for our automotive portfolio in China, for China. So we can satisfy that requirement, going forward. It gets harder than it would be about being a Chinese company to be able to compete. And that -- but again, we don't have it yet. So what I'm offering is our paranoid view about the future. Today, it's no issue.

Got it. Bill, apologies. Question for you. Sorry to keep you waiting. So gross margins, you've managed to maintain above 58% over the past couple of quarters. And this is during which utilization rates have gone from above 80 exiting last year to now, I think, low to mid-70s. So obviously, you guys have executed really well. What are the key factors that have helped you maintain gross margins at these levels? How should we think about the key drivers behind that? And is 55% to 58% still irrelevant? Again, I don't expect an update here, but I'm going to ask it anyway. Is that still the relevant range?

It's the #1 question we get every time we talk to an owner. So fundamentally, there's two structural changes that are impacting our gross margin. The first one, Kurt talked about where we source about 60% externally from a wafer fab standpoint. And today, we source about 40% internally, think, 90-nanometer NXP IP proprietary technologies. And we'll continue to invest there. We're spending about 6% to 8% in CapEx. Now 10 years ago, that was swapped. We were probably doing 80% sourcing in-house and 20% externally. And as you know, the ramp of our products in the future are all driven by 90-nanometer and below technologies, which we source externally, so that has a major change on your fixed cost structure. So today, our fixed cost structure is 30%. 10 years ago, it was 70%. So two things happen. One, it reduces the variability of the gross margin impacts that we've seen in the past. And secondly, it allows us to navigate and control our inventory and where we want to put inventory through. And that's helping us in the short term. The second structural change, and again, it's nothing new other that the gross margins of NXP continue to improve and go higher. And so if you think about the revenue today, that was based on investment since we made 3 or 4 years ago that is ramping. And that hurdle rate, from a financial standpoint, was around, call it, the low 50s. Now today, everything that we invest has a hurdle rate of 58%. And therefore, in 3 or 4 years, you expect higher gross margin accretive coming on top of the corporate margins today. So it takes time, think about improved new product introductions over time adding to the portfolio, as well as the lower gross margins tailing off. And so we'll naturally grind higher from here. Short term, medium term, how are we doing against the lower utilizations? Again, the sweet spot to 85, we've been in the low 70s, and we have a favorable mix that's offsetting that headwind at the moment. If you think about going into 2024, I'd say you have two tailwinds that if the market improves, you have a utilization that can go from the 70s back to the sweet spot of 85%. And secondly, Kurt talked about replenishing the channel when the time is right. We don't think it's going to be this year. Most likely, it's next year over a gradual period. And think about $500 million focused on the long tail, which typically carries a higher margin mix for us. So that's another tailwind. But more longer term and to answer to your question, I think you'll probably see us update our gross margin range next year in November 2024. The current 58% is not our final destination.

And maybe important to add, we do not have a significant gross margin variation between internally manufacture business and foundry business and also not between the revenue segments. We have, within those, clearly a curve, but not between them, which is important. So we are not very susceptible to changes in the revenue structure between those segments, which is good, which is a big element of resilience. So we are confident indeed as well as that this year, which is clearly a tough year for the semi industry, we should see no decline from our gross margins and actually should march through this in a pretty steady way.

That's wonderful. I could go on, but we're out of time. Thank you so much for attending.

Thank you for all the support.