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

Good morning, everyone. This is Ross Seymore again here at DB, semi analyst. We're really happy to have management from NXP Semiconductor joining us next. We have both Bill Betz, the SVP of Finance; as well as Jeff Palmer, the VP of Investor Relations with us today. So gentlemen, thank you so much for joining us even if it's just virtually.

So before diving into some NXP-specific questions, I want to ask some market-wide ones that I'm asking of all companies. I think NXP is in a unique position, for better or worse, to be able to address some of these given how acute the supply issues are in the auto industry. So in the last couple of quarters, you've emphasized the importance of your relationship with your foundry customers, you've rebounded on supply. It appears like you're getting even more supply coming from those partners exiting the year. So talk a little bit about where NXP is in its ability to supply demand. And then when do you think the market as a whole will get back to some sort of equilibrium? And you can make it auto specific, if that's half of your sales as it is, or you can make it more general if you think that's a better answer.

Yes. So Ross, I think the challenge we're facing is the demand signal has not been static. So the demand segment continues to expand. We're seeing strong demand across nearly all of our end markets. At this stage, I think we probably are not going to be in any kind of equilibrium if there is such a thing until at least midyear '22 and possibly even later into '22. And that only gets you to a point where you're able to satisfy, let's call it, immediate demands and you're not on escalation calls every day of the week. Beyond that, you still have to rebuild channel inventory for NXP and our distributors as well as on our balance sheet. And we're also seeing more and more conversations from our auto OEMs wanting to carry strategic inventory out beyond the current kind of just-in-time model that they've historically adopted. And so that would be beyond that equilibrium point being achieved.

So if it's middle of next year, and you guys have more information than me, but these days, everybody's guess is as good as the next person's. Do you, in your conversations with customers, believe that there will be a structural change in the just-in-time model, probably most acute in the automotive side? Because I always look back and think it was only 2 years ago, in 2019, when those same customers that are clamoring for more inventory today wanted nothing to do with automotive semiconductors. And so I find it hard to believe something structurally changes necessarily given that just 2 years ago, the opposite conditions would have led to a different potential structural change. So are conversations that you're having with your customers actually indicating that something will change for them? Or is 2019 a point that they also appreciate?

I would say the conversations, and this is really at the executive level with the OEM -- auto OEMs and with our executives at NXP, that something will change. It's probably not one size fits all. It's probably different per auto OEM. But clearly, one of the kind of aha moments has come out of the last 18 months of nonstop supply shortages is the differences in the cycle times, manufacturing cycle times between the automotive OEMs, who tend to run in a just-in-time or historically just-in-time model, and the auto semiconductor or semiconductors in general, where the cycle time is 3 to 6 months. And I think, believe it or not, some of the auto OEMs did not have that realization. And they're starting to realize maybe their just-in-time model is not really suited for their business going forward. And so we are hearing conversations of different levels of, let's call it, strategic inventory being held throughout the supply chain. Some are airing at the end of let's try to put us on 3 months, some are saying let's go to 6 months. And that quantity is different per OEM, and it's not just kind of what the cycle time is for auto semiconductor development.

And do you -- the one thing I could see is a logical driver of more structural inventory, not for cyclical purposes, but longer term would be from 2019 to today, the importance of semiconductors in every vehicle has risen substantially. So in the automakers' defense, I can see the scenario where they just viewed semis as one more cog in a gigantic wheel few years ago, but now that's how they sell their vehicles. The electronics, the infotainment system, the safety systems, et cetera, are just so much more important now. Is that underpinning to these structural conversations as well, where you think there are appreciation shortages or otherwise of the importance of semiconductors has increased?

Yes. I think you're spot on there, Ross. I think if you look at what is differentiating the auto OEMs' products amongst themselves, it's electrification, it's safety, it's driver-car interface and features. And all those kind of those 3 main pillars are driven by semiconductor content. And I think if you think about the total BOM of a car in an internal combustion engine car, total semiconductor content, let's call it roughly $500. In an XEV, it's maybe $900. That's relatively small in the total scheme of the BOM. Why put at risk your launch of a new program and enticing customers to make moves over a small percentage of your BOM?

So let's dive specifically into the automotive side even more so. One question I get from investors all the time and I know, Jeff, you and I have talked about this, but Bill, feel free to chime in as well, is the content versus SAAR side of things. And there's been a relatively consistent relationship, semiconductor revenues, I think, Jeff, you and I have talked about the chart that we put in our quarterly automotive semi-tracker every quarter showing the relationship of auto semi revenues versus SAAR, unit SAAR. And it's maybe 6% to 8% delta every quarter. This year, you guys are probably growing somewhere 30%, 40%-ish and the SAAR keeps going down ironically because of semiconductor shortages, but let's say it's a round number of 10%, just so I can do the math. What explains that massive delta? And how do you not get concerned that you're out-shipping versus what's being built in the market today?

Ross, I think it's always the compounded math, the statistics of it all. I think if you use your point of reference as either 2020 or 2019, you're going to come up with, one of our good friends called, some crooked numbers. We would advise if you were to look, let's say, the first half of '21 versus the first half of 2018, the last kind of normalized auto environment, you'll see that global SAAR was -- I think at that point take a 3-year CAGR, global SAAR was down about 7% in that period and against that NXP automotive was up about 3%, so call it a 10-point delta, if you would. And I would think that, that feels a little bit more normal. Because if you remember back, 2019 was not a great year for auto and industrial exposed companies. We had the first full year where China auto sales were down after 10 years. You had a lot of companies, especially in Asia, who took a very cautious outlook on any extra inventory due to the China-U.S. trade wars. So 2019 was not a great year. We exited 2019 thinking, okay, maybe we're going to get back to more of a normalized environment just to find ourselves globally walking into the buzz saw of the pandemic. So '19 and '20 really are not what we would consider appropriate periods to measure against. Now to your question, does it concern us? Sure. We are in constant conversation with our customers. But if we use it as a kind of the litmus the conversations with the auto OEMs, everything that we're shipping to the OEMs is going out the door or going into production. We are not seeing anyone telling us to take a foot off the brake in terms of queuing up material for them for an extended period. And supporting that also is the auto OEMs and their Tier 1 partners wanting to place longer, noncancelable, nonreturnable order backlog on us, on NXP, and us in turn placing longer-dated wafer demands on our foundry partners. So it's a little bit of a balanced situation. But I think you have to look at it relative to the last one normal period in 2018, Ross. Sorry for the long answer.

No, no, no. It's -- where you start the math is incredibly instructive and important in that. So when you look about these relationships, the other dynamic that people point to is you have hundreds of thousands of units of nearly finished vehicles sitting on parking lots next to automotive factories all around the world, waiting for some magic component. I'm sure it's a wide array of components. Are you concerned that all of a sudden that one magic component gets shipped and the demand can drop very fast because you don't even have the cycle times of the OEMs building the vehicles as a buffer any longer?

I don't think it's that instantaneous for us, and I don't think it's any just one magic component. And I think it's different components with different OEMs. And it may not -- it may be beyond just semiconductor. I mean we've heard cases where just generally, there's a lot of tightness and a lot of input materials for cars. So I hear you. Are we concerned? Are we watching it? Yes. But all the indicators we can see, conversations looking at dealer inventories, commentaries that the auto OEMs are making themselves indicate to us that we are still probably in the early stages of a longer-term upcycle in the auto and in the industrial end markets.

So let's talk about some of those longer-term agreements that you put into place. I know you guys have run inventory very consistently over the years. And at times, people will yell at you for having too much inventory and sometimes they'll yell at you for having too little inventory, but you've kind of stick to your knitting in that. To the extent you get the visibility with these long-term supply agreements, who holds the inventory?

So we're talking about strategic inventory, Ross, specifically. That's how we are alluding to. I think we've been pretty clear, NXP is not going to hold that inventory. So it may be held somewhere in the supply chain, whether it's at the OEMs, the Tier 1s, even some Tier 2s. There's a possibility that maybe some of the Western auto OEMs, North American and European OEMs adopt more of the Japanese model where you might have this kind of kidding distributors that hold inventory ahead of production for the Tier 1s. I think it's still very fluid exactly how that's going to play out. We're at a point while there's conversations going on, there is no material to even build strategic inventory. So that event is still at some point in the future.

I think the last aspect of these relationships becoming more structural and more of a partnership with your customers is when you have these long-term supply agreements, you can think about duration, so extend the visibility of your relationship. And then you can also exact some rents by -- via the pricing lever. And I know you're always very hesitant to talk about this. So we don't have to go into any details on it. But relatively speaking, how do you balance those 2? Do you try to mainly get the visibility longer term? Or is there a locking in of price regardless of if you're passing along wafer price increases from your foundry in the near term? I'm talking more structural and longer term.

Yes. I think, Ross, it's the former. It's -- we see this opportunity as building closer and deeper ties with the auto OEMs. If you remember, if you think about how we go to market in the auto industry, while we engage with the auto OEMs and we win designs with the OEMs, we transact the business with the Tier 1s. Now what this supply chain challenge has created is a elevation of the engagements between our executives, our general managers as well as Kurt Sievers, our CEO, and the CEOs of the different OEMs. And so we are seeing a deeper, closer engagement and understanding of what both parties can bring to each other. While we may share with the engineering community at different OEMs some of our new road maps and technologies, you would think that probably doesn't bubble up to the C-suite at some of these large global OEMs. Now they are getting a better sense of what NXP can bring and help them solve and differentiate their products. So that's what we see as the positive aspect of this tough environment. In terms of pricing, we value our customers. You and I talked about this many times. Our business specifically in auto is a very long-cycle business. It's usually 2 to 3 years design-to-revenue cycles, 5- to 7-year product cycles. It is a high value-add application-specific and in many cases, sole-source business. The last thing we want to do is take that business and turn it into a commodity business like the DRAM business, where in tight times the supplier extracts pain on the semi vendor, and in loose times, it's the opposite way around. We're not looking to extract pain or rent, as you may call it, in this tight environment. Now we are unfortunately having to pass on increased input costs that are put on us by our foundry suppliers to our customers, but we're being very transparent about the nature of those inputs.

So 2 last kind of high-level questions on automotive, then I want to get into some of your secular drivers within this segment. The first one is, in adding supply to meet the demand, are you doing it internally, via foundries or both?

So Ross, if you think about front end and back end, so on the front end, about 57%, 58% of our total wafer consumption is sourced externally from wafer foundries. This tends to be processes anything below 90 nanometers that requires 300-millimeter wafers. Now on the back end, we actually have about 75% of our production internally at a variety of locations around the world. So as we raise CapEx to improve capacity, it's going to be focused primarily on the back end testers, packaged systems to assemble equipment, but some increases in equipment to break any bottlenecks or improved throughput on our front ends. Now what we're not going to do is we are not going to build a new foundry. We're not going to build a 12-inch wafer, that will be $10 billion a clip. And so we're not going to do that. I think over time, it would not be surprising if you see us become more and more fabless in nature as more and more of our newer products depend on advanced process nodes or this is what we consider advanced process nodes.

Got it. And the last question on this is going back to that content versus SAAR dynamic. You mentioned the math of -- I think it was first half '18 or second half '18 going to the current and those growth rates, et cetera. But it generally pointed to about a 10% increase in content on average over that period of time. That's a bit higher than the 5% to 8% we've talked about and seen in the past. Do you believe that content per vehicle dynamic is accelerating faster than it historically -- well, do you have a bigger content per vehicle today than you would a couple of years ago? And is that growth rate accelerating as you go forward?

I think the -- for NXP-specific, the content per vehicle opportunity is improving. What you and I have talked about in the past is kind of a rule of thumb that says pick global productions, global SAAR, and add 300 to 500 basis points, and that would equate to automotive semiconductor growth rates, in aggregate, not just for NXP, but in aggregate. And we said that we can usually aim to grow 50% faster than that. And so that's how we would underpin that 10-point delta between global production and what we've seen looking at first half '18 versus first half '21. We do think as the market moves more towards XEVs as things like safety become more sellable features, these will add content that we can address from our portfolio. Yes, we do believe that.

So the final topic in automotive is just talking about the secular growth drivers. And I assume at your analyst meeting later this year, you're going to refresh these numbers. But last I remember, about 1/4 of your automotive business was in things like radar and e-cockpit and BMS, those sorts of things that are growing 20% plus. One, is that still the right framework to think of, 25% of the segments and growing over that rate? And then two, update us on the progress in those 3 areas: radar, eCockpit and BMS.

Yes. So I think as automotive business has rebounded and the pie is larger, yes, that roughly 25% is still accurate, but it's a larger number today. So we kind of decompose that kind of growth bucket, if you will. So as you highlighted, radar is roughly 10% of automotive business today. We think it can grow at kind of a 25% to 30% clip. And that's a 2020 to 2023 window. So as the starting point gets larger, that rate may slow a little bit. But that is the current public statement we've made. We feel very good about our radar business. I think we use a term called true leadership, where our relative market share is at least 1.5 times the #2 player. And we believe, over the last year or so, we've emerged into that true leadership position in radar. We'll go into more detail in November, but I think all things are full steam ahead in the radar market. And as you know, in radar, the part that I think investors have to get their mind around, it's kind of a 3-layer multiplicative factor. There's more cars with radar every quarter, every year, there's more radar nodes per car, and each one of those nodes in the car is becoming more and more complex. And so that's all adding to a nice tailwind of growth to that business, very well underpinned with design wins that are still into the future to monetize. When we turn to our eCockpit business, which is the secular trend from electromechanical gauges to these flat screen panels and cars that you've probably seen on modern vehicles, we sell a application processor family into those opportunities. Currently, today, it's about 10% of automotive as well. We said that, that should grow kind of in the mid-teens in that 2020, '23 horizon. Good business, a little slower growth than radar because there's always some existing footprint, if you will. It's not a pure greenfield opportunity like radar may be perceived to be. So that's kind of 20% of that growth bucket. Now the last 5% is made up of multiple things, some of which are not yet large enough for us to articulate or go into a lot of detail. But we have talked about our battery management systems. That in 2020, that was roughly, call it, a $50 million, $60 million run rated business. We think we can grow that at about a 60% CAGR, 6-0, in the 2020, '23 horizon. That's about 2x the SAM that we're addressing. We've been designed in at 16 of the top 20 auto OEMs in one shape or form. So these aren't exclusive across the board on every platform, but we've got a footprint that's fairly material and continuing to expand. Our solution is fundamentally different than our large competitor. Our solution is a system-level solution where we take the precision analog components to sit on the battery pack, we connect them to a back-end processor, and we run load balancing software on the processor for the efficient charging and discharging the battery pack and to monitor the health of each and every cell in the battery pack. And that's a unique solution that our competitor is primarily delivering only a precision analog solution to their customers.

So 2 questions on BMS. Does your competitors plural as of last week becoming -- well, one of your competitors now -- with ADI and Maxim being a single company, does that change the competitive landscape? Second question would be, what happens when some of this becomes wireless with how they're enabling like the LTM platform at GM? What's the competitive response from NXP to that technology?

Yes. So on your first question, yes, with the completed merger of Maxim and ADI, they do emerge as the largest BMS solution provider. Yes, the math points to that. In terms of wireless BMS, it's an interesting idea. We're doing some research with customers on that. The one thing I would highlight that if you think about an automobile moving down the highway at 50, 60 miles an hour, it's a noisy from EMI noise and bad environmental conditions. We are a little suspect that wireless can be as efficient in that environment. But look, more power to ADI and Maxim for developing that solution. We'll see how the market goes. We think the -- if you think about the BMS market as a pie, the pie is expanding robustly. We think there is enough for at least 2 players to have very, very good positions. As you know, dealing with us, Ross, over all the years, we're never happy being #2. And so it's a great challenge to our General Manager, who runs this division, to go and be as competitive as he can in that marketplace.

And then last thing quickly, the size and scope now of the combined ADI and Maxim entity, has that changed anything from a pricing perspective? And I know they just merged last week. But given their scale, is there any sort of increased competitive intensity? Or again, as the pie is just growing so fast, that's not the primary concern.

I think the pie is growing fast. I think we've -- in our analysis, we feel that both ADI and Maxim's individual solutions are similar. So it's not 1 plus 1 equals 3. We think it's 1 plus 1 is 2. It is more of the same. They'll probably have to decide which road map they're going to follow moving forward. And you always have to wonder whether or not having that much concentration even on one part of the total system takes some choice away from the customers who buy in that marketplace. So we'll have to see how it plays out. We have a lot of respect for ADI and Maxim. They are both great companies. They both know what they're doing. But as I said earlier, I think it's a good market, it's growing rapidly, and we enjoy competing with them every day.

So I'm going to go a little more lightning round here for the other end markets because automotive has taken up so much time, and I think it's worthy of that. And then I also want to get Bill to chime in on some of the financial questions. So in industrial and IoT, you guys seem like you're going to grow 25%, 30% this year. We've talked a little bit about the structural increases there. If you take the cyclical out of the equation, what do you think are the unique drivers for NXP, whether it's crossover processors, WiFi, et cetera, that make that growth more sustainable and not just another example of industry-wide cyclicality that's great now, but could turn nasty in the next year or so?

Right. So as you remember, back in early 2020, we completed the merger or the acquisition of the Marvell WiFi connectivity assets. From our vantage point, that was the missing piece in our industrial and IoT strategy. Our strategy of going to market in industrial and IoT is to present and provide connected, secure processing solutions. Nice marketing words, but if you unpack it, it starts with engaging with the long-tail customers first on the processing needs, where they're going to need to write the software. You then marry to that and complement that with analog attach and you complement that with WiFi, Bluetooth, Zigbee, NFC connectivity and you wrap that with a common kind of firmware layer that our customers can write to so they can extract the level at which they add their value. Having those pieces now altogether with the broadest portfolio of industrial application processors, crossover processors and broad 32-bit ARM microcontrollers we think give us a really unique solution to go to market and address the requirements of a very long-tail business. Remember, industrial and IoT is tens of thousands of customers primarily serviced through distribution and primarily in Asia Pacific, China markets. And so it's a very unique market, very different than the automotive market. Very complementary to automotive's highly concentrated marketplace.

Is the WiFi asset doing what you expected it to do? Or are there some puts and takes due to COVID, et cetera? I think when you bought it, it was about $300 million in revenues, and you thought it could double in the next year or 2, if I recall.

Yes. We probably lost a year, Ross, because of COVID in that road map. We still believe the trend is to double the business fundamentally. Remember, we don't report all of the connectivity business in industrial and IoT. It's mostly in there, but it's spread across in our comm in front and a little bit in our auto business as well. Directionally, we're very pleased with the integration of the team. The team has come aboard. The product road map looks very rich. But COVID did impact us on that road map to doubling the business. So maybe lost a year.

If we go over to the communications side of things quickly. Give us an update on your view on your HPRF. I know you have the LDMOS side and the GaN side of things. How are you guys positioned for that? Obviously, there were some changes injected by the U.S. government courtesy of the Huawei side of things. But generally speaking, how are you feeling about the ramp in that business?

I always have to say this at the front end of talking about that business, it is a great business, but it's very lumpy, very choppy. You'll have a couple of good quarters and you'll have pauses and a couple of good quarters. It's just the nature of how the build-outs occur. But how we currently view it is, as you remember from our prior conversations, we were looking at the second half of '21 as being a positive period and uptake for our GaN module business. So these are hybrid GaN products that come from our GaN factory that we opened up in the fourth quarter of last year. That's progressing as we would have thought. And these are for build-outs outside of China. So the C-band build-out in North America, some other 5G build-outs in Europe and some in other parts of Asia. So that's moving along as we plan. Layer on top of that has been what they call the CP3 build-out in China, which is the rural macro base station build-outs in the 700 megahertz, 2.1 gigahertz frequency band. That is the sweet spot of LDMOS. So while we have this one curve of GaN ramping nicely as we purchased -- anticipated, we also can layer on top of that some of the LDMOS business in China, which will be nice. But let's be real upfront of you, Ross, it is a lumpy, choppy business, right? So it's not always up into the right forever.

Yes. And for everybody on the phone, it's also not the entirety of that communications and other segments.

That's correct. That's correct, Ross. That's correct. Inside of comment where there are also 2 other portfolios of business that are completely all unrelated. There's our historical security card business, bank cards, passports, transient access and RFID type of products, and then there's also our legacy communication processor business. Both of those businesses are nice businesses. They're not really our focus, and we would not really have investors spend a lot of time thinking about unique growth in that area.

So the last end market, and I only have one question on it, is the mobile side of things, about 13% of your sales. I'll make this one really simple. Why are you guys guiding it, I think, down sequentially in the third quarter at a time when most people believe seasonally, that would be up? And seasonality does not apply right now? Or are there some shortage issues to deal with?

Shortages, pure and simple, wafer shortages.

Got you.

We think we should be able to address those issues as we move into the fourth quarter. Not a good situation, but one that we cannot avoid right at the moment.

Got it. All right, Bill, you've been waiting patiently. So let's hit some questions for you on the margin side of things. You guys have done a great job getting the margins back up within your guided long-term guidance range of, I believe, 53% to 57% and even at the upper end of that range. Not to front-run your analyst meeting this fall, but if I think about the puts and takes as you go forward, I think we all know what could make the gross margin go lower if utilization fell something like that again. But if the gross margin were going higher, what would be the drivers to make that happen over the next couple of years?

Yes. Thank you, Ross, for your question. First, I'd like to say we're very proud of the team achieving towards this high-end gross margin range target like you mentioned. There's been a lot of hard work over the years since the Freescale merger has occurred. Now I think Jeff talked about how to think about our manufacturing footprint, where he mentioned about 58% of our front-end fab wafer fabrication is basically external and 42% is internal, and that will shrink over time as we move to our leading-edge technologies externally. And then the assembly and test is the opposite effect, and we'll get some leverage there in the future as we add more capacity put in place. And it just makes a lot of strategic sense and financial sense to expand our internal footprint for the back end. And again, as Jeff mentioned, that's about 75% in-house today. So at these current revenue levels, we feel very good in the short term, medium term, even though we're supply constrained. We believe 56% is the right range. As we think about how do we get to that 57% and maintain it, we believe this is a bit longer term, and I shared before by our new production introductions that will ramp and become more of a meaningful portion of our future part of our business. And then with your lever question, again, in any given quarter, we're going to move around based on the many elements impacting gross margin. Remember, we're talking only 10 basis points is $3 million on a quarterly basis. And in Q2 alone, we saw 50, 60 bps movement just purely on mix. So overall, I'd say we're very confident of sustaining that 56% level, and we'll look to bring this 57% level on the longer term through our NPIs.

Then a similar question on the operating margin side of things and maybe OpEx intensity will be a more accurate way to say it. I think you guys have a long-term model of roughly 23% -- or 22%, if I remember right, and I think you guided to a 23% number in this last quarter. If you're not able to hit it now, kind of when would you be able to hit it, given just the strength in revenue growth snapping back, probably more difficulty in hiring people than growing revenues, shortages aside? Is the goal just to continue hiring as fast as you can to feed the beast going forward? Or can people expect you to get some incremental operating margin leverage going forward?

Yes. Thanks, Ross. First, our range is between 20% and 24%. Like you mentioned, we're roughly around 23% today. If we look at R&D, the 16% really geared towards the R&D investments. And the way we look at this is we look at this at a multiyear perspective on our strategic plans and our projects. Throughout the year, we're constantly measuring our progress in our portfolio and our projects. Jeff has shared lots of those what we're working on. I mean we're managing over 200 projects at any given point in time. On a quarterly SAM basis point from a tactical standpoint, again, we know what's happening in the next quarter out with our projects, what's going on with mass, what's going on with headcount. I think we'll have some leverage as we go forward on the G&A side of SG&A as we move forward. But I would -- we feel very comfortable with the 23% model that we have in place, and we believe 16% is the right number to be competitive and outgrow the market from an R&D perspective.

Great. Well, guys, we are exactly at the end of our allotted time. So Bill and Jeff, thank you so much for attending the conference. I appreciate all the detailed answers, and we look forward to monitoring the progress of the company and the analyst meeting later this year in New York, attending that as well. So thanks for your time, guys.

Great. Thanks, Ross. Appreciate it. Thank you for your support.

Thank you, Bill. Bye now.