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

Great. Well, good afternoon, everyone. My name is Craig Hettenbach, one of the semiconductor analysts here at Morgan Stanley. Very pleased to have with us NXP and Jeff Palmer in Investor Relations. So welcome, Jeff.

Thank you, Craig. Nice to be here.

I do have to just read this before we kick off. So please note that all important disclosures, including personal holding disclosures and Morgan Stanley disclosures, appear on the Morgan Stanley public website at www.morganstanley.com/researchdisclosures or at the registration desk.

So with that, Jeff, I know you guys had a call this morning, which was helpful just to kind of update the current environment. And then you can maybe just Lead with that as we drive into the discussion.

Sure. So if you go back a couple of weeks, February 3, to be exact, when we gave our fourth quarter update, we provided guidance for Q1, at that time, we were pretty explicit that it was near impossible for us to size any impact from the virus at that time in Q1. We just -- we weren't able to get our arms around a true dollar amount. Now since that time, we've been working pretty diligently to really understand what's been going on. And so what we've seen is about a $50 million reschedule or pushout of orders, no cancellations, out of the quarter. Now it's kind of not intuitive to think about how it played out. The first 2 weeks after Chinese New Year, we actually saw these pushouts, right, as people were trying to come back to work, factories were getting -- coming up slower. The last 2 weeks of February were actually more normal like we had seen or we would anticipate before Chinese New Year. And so knock on wood, we would hope the rest of the quarter continues along this more normalized rate. And so -- but we wanted to offer up what would it look like if that was not how the rest of the quarter played out. And that's how we kind of size the worst case downside, which is the $150 million if we went back to kind of that pushout environment we saw right after Chinese New Year. That makes sense?

Got it. It does. And I think you, along with other companies, have talked about most of this impact is within China and maybe Asia a little bit broader. But just North America and Europe has been kind of in line. Is that the case?

At this stage, relative to the virus, that's correct.

Got it. Well, maybe we can just discuss kind of your thoughts kind of leading into this. From a cyclical perspective, things were bottoming prior to this. But I would say, more importantly, your approach to inventory this cycle. And I think you guys did a good job in terms of keeping inventory low within the channel. But just give a backdrop on that in terms of that approach and how you think through managing the channel as we go into...

Yes. So if you've been following NXP for a couple of years, if you go back probably 5-plus years ago, our channel management strategy was not completely hands off, but if our distributor partners placed orders, we would ship it at the end of the quarter, we reconcile how much inventory was there, and we report that to you at the end of the month. And for most of the time, it was fine. Now every once in a while, we kind of -- we'd have excursions outside of our model. Now around 2016, we made the very concerted effort that we were going to manage the channel much differently. And so what we started doing is working with all of our channel partners, and we're able to get sell-in and sell-out data on every product line on every distributor on a weekly basis. Some of them we can actually get daily, but it's almost too much information. So what we've also done is that we've now maintained -- in the channel, our target is 2.5 months of supply, plus or minus 0.5 month. That's our range. And if you've been following our calls, we've been at about 2.4 months for the last 3 to 4 years. Last year and last quarter, I think we had 1 quarter where we were at 2.3. In Q4, we were at 2.3, so just a little below that. Given the current environment that we just talked about with the pushouts, we are going to maintain that steady hand on the distribution channel. We're not going to let the channel inventory grow. We're going to keep it at that kind of 2.5-month range because we think that's the right thing to do, at least from an investor perspective.

Got it. Okay. Well, let's shift gears to some of the longer-term growth drivers, particularly within the automotive business. I know there's a focus on ADAS. You have EVs with BMS. At CES, I think you guys were demo-ing kind of digital cockpit. So a lot of exciting growth drivers in autos. Pick your choice in terms of where you want to start with and then we can take it from there.

Yes. So let's maybe take it at the highest level and kind of drill down. So as you know, automotive is just under half of our total revenue. So a little over $4 billion. What we have said is the total auto business we anticipate to kind of grow at a 7% to 10% CAGR over a 3- year window, 2018 to 2021. And that makes the assumption of a 1% to 2% global SAAR or global auto production. Now if you have a different number you'd like to plug in there, you can kind of get a sense of how that would change the long-term growth curve. We've said that automotive semiconductors grow 300 to 400 basis points above global SAAR, and then the rest is NXP-specific. Now if you drill down a little further, we said that about 75% of our auto business is what we would call our core business. These are businesses where we have high market share, high barriers to entry, and it's unlikely that we would outgrow the market given the size of our franchises in those areas. So we've said that that 75% would grow at call it, SAAR plus 300 to 400 basis points. All right? Now the other 25% is what we term our growth businesses, and it's really -- can be broken into 3 big buckets. The first is our radar business, which is about 10% of automotive, so just north of $400 million. We -- when we use the term radar, we talk about it as a radar system. So this is a radio RF device in the front end, a back-end processor, in-vehicle networking connection between those devices and then a power management for the processor, all tied together with our software. The trend that's going on in radar, though, is kind of a multiplicative effect. First, you have more cars with radar and more radar nodes per car. And so if you look at some of the standards like in Europe, like the NCAP 5-star rating, if you want a 5-star rating for your car, you've got to be able to support automatic emergency braking, automatic lane departure warning and a few other requirements. Those features are supported or implemented by a combination of radar and cameras. So we have pretty good mandate-driven adoption of radar. We think that our radar business can grow 25% to 30% CAGR over the next several years. And as you know, Craig, the design to revenue cycle in automotive is 2 to 3 years, life cycles are 5 to 7 years. So for us to be able to stand up and say, "Hey, we think this business can grow 25% to 30%," we have a very high confidence based on design awards we've already been granted, and we have to ramp into production. Why are we doing so well in radar, you might say. So the industry, up until about maybe a year or 2 ago, had primarily been implemented using 24 gigahertz radar. What we did as a company is we did not have a footprint in the front-end 24 gigahertz radar devices, though we are the market leader in the back-end processing. What we did is we doubled down on the engineering to intersect the market transition to 77 gigahertz. So we have both a SiGe, silicon germanium, part, off of the high-volume production today; and we just started to ramp into volume production in Q4 an RFCMOS part, which, over a long period of time, will become the workhorse part and allow us to really integrate the processor and the RF front end into a kind of a monolithic solution. Okay?

Got it. Great. Help with just kind of the path along towards autonomous. I know, typically, there's a lot of excitement and then things pause. I think right now we're talking about kind of 2.5. Kind of where are we from a progression?

Yes. So we think of true autonomous vehicles, level 4, level 5, 8 to 10 years out. Now technologically, you might be able to build a car that is fully autonomous, but you have to understand that putting them out into the real world is -- faces some other challenges. So we think what that means is the life cycle for L2 and L3 is probably going to be longer and probably more profitable.

Got it. Well, maybe we can segue into just digital cockpit. I think going back to the -- what I -- the comment I made about CES, I think you guys were demo-ing a quad core product there. Where are you in terms of adoption? And what's kind of the growth rate for the digital cockpit?

Yes. So digital cockpit, or some people call it digital clusters, is the secular trend auto OEMs are implementing to move from electromechanical gauges to flat screens. And you can see it ranging from 1 to 3 screens, even some cars I saw at CES having almost a panoramic screen across the whole dashboard. So for us, that business is driven by our application processors, i.MX, things like our QuadMax or i.MX 6. Currently, that business is about 10% of automotive revenue today, and we think it will grow at about a mid-teens growth rate over the next several years. Again, based on design wins we've already been awarded.

Great. Then switching over to BMS. It's a smaller piece today, but there's a lot of excitement in terms of the potential opportunity in that market. And so if you can touch on just kind of from a design perspective today where you are. And then also just relative to some incumbents in the market, be it ADI and Maxim, what is NXP doing differently? And how is that...

So BMS products is actually kind of an interesting outcrop of the merger between Freescale and NXP. Freescale, as you know, we merged with them in 2016, brought to the table a very deep and knowledgeable base of automotive processing. NXP brought to the table a good, deep base of advanced analog. And so our BMS solution, again, is back to a system-level solution. It's made up of 3 components. The first is the precision analog front end. These are the devices that will sit out on the battery pack in the car. We connect those analog precision devices with a back-end processor, where we run ASIL-D functional safety algorithms. This is -- what we do is system load balancing, both for charging and discharging the battery pack, providing a safety capability. How that's different from, let's say, an ADI or Maxim, both great companies but we believe ADI and Maxim both just provide a precision analog front end. We're not really sure how maybe their partners implement the total solution, but we know what we're providing is that complete analog, plus interconnect, plus processing and functional safety algorithms running on the processor. The go-to-market on that business is kind of bimodal. One mode is we engage with the big battery OEMs in Asia. So the CATLs, Samsung, SEI, LG Chem, which are basically chemical companies who don't have active electronics capability, and they're looking for a partner to provide them with a battery management solution that they can then incorporate for the full solution to take to their customers like the Tier 1s and other OEMS. So that's where the business today is primarily sourced from. The other mode is actually a direct sell model into certain OEMs who basically are looking at, hey, they have multi-brands, different brands across their -- under their umbrella, and they want to roll out a platform that can be scalable from -- all the way from 800-volt high-performance sports cars, all the way down to 48-volt mild hybrid. And they don't want to redesign each one differently. They want you to be able to have a scalable BMS solution. And the one OEM who's talked a lot about this is Volkswagen with their MEB platform. And I think you can probably see the battery management system in those platforms beginning in the '21 model year.

Great. All right. Maybe we can segue into another exciting technology, ultra-wideband. And really interested in your perspective kind of as a base technology, talking about kind of your position in mobile wallet secure element, what that means as ultra-wideband comes to market for kind of automotive and mobile applications.

Yes. So I think the way to think about ultra-wideband is think about how, over the last several years, you watch things like your credit card or your building office card morph into your cellphone. Think about ultra-wide enabling you to morph your car key or your house key into your cellphone. So the 2 big applications that we've talked about have been secure car access, where you'd use your phone as a mechanism to get in and out of your car. And the other opportunity is in secure building, both home and commercial access. If you think about those type of markets, you would have to have interoperability both between the devices that are in the car and what are in the phone. And so this is an area we've been working on for several years. We only started talking about it in Q2. We anticipate that, that business will begin to ramp for us in the second half of this year, 2020. It's negligible revenue today. There's a little bit in automotive and some early key fob implementations. We've said that it should be about a $900 million market by 2024, and that's just the RF component that sits on top of the mobile wallet. So what makes the ultra-wideband unique and differentiated is the ability to manage digital credentials in the secure element in the mobile wallet. So let's say, Craig, you had a rental house, and you want to loan it to me. You could electronically send me the key and control when I would be able to have access to the property. And you might say, "Well, Jeff, you need to be out of there by noon on Sunday." You could deactivate that key. Or you could loan -- somebody could loan you a car or you can imagine you're going to take your car into a car dealership and you could say, "Hey, Mr. Car Dealership, you can use this car for this period of time, but then the key is no longer working." It's very conditional access.

And how about your capabilities in this market? Qorvo recently did an acquisition of a private company that's been around for a long time, but like how do you kind of stack up, if you will, versus that solution?

So the ultra-wideband by itself is just a 5.6 gigahertz RF frequency technology. What really adds the value is mating that RF capability with the embedded secure element, so the digital credential management. I don't know a lot about Decawave, but I believe that they only provide the RF component. I do know that they have been helpful with the IEEEs as long as -- as well as we have been with getting the standard set. But I'm not really sure how Qorvo's going to take them to market.

Okay. Great. Maybe we can switch over to just the connectivity market, the Marvell deal that was recently closed towards end of last year. I know you guys have talked about that was one piece that you really felt can really enhance the portfolio. And so really, how it fits in with the core business and how the integration is going so far.

Yes. So if you go back again to 2016 with Freescale and NXP merge, what you saw was Freescale brought a great portfolio of processor and processor know-how, NXP brought great analog capability. But what's missing in both portfolios is connectivity solution for the ability to provide secure connected devices. As we went into the aborted merger with -- or acquisition by Qualcomm, one of the selfish things we wanted was we wanted access to Qualcomm's connectivity assets. That deal broke, can't cry over spilled milk. But when that -- after that deal broke, we started to look very strongly at what other assets were in the marketplace. And we looked around and we felt that the Marvell team had a number of really attractive things. One, they have a long history of being in the market; a cohesive team of engineers who have been very involved with a lot of the standards bodies; very, very strong in the certification of software. And we thought it was the best asset that was available to us. Now when we look at that market and say, "Okay, where are we going to deploy it first?" It's really in our industrial and IoT business. And what we did is we looked at our design win pipe of our i.MX application processors, and we saw that about 2/3 of all those designs in industrial and IoT had a connectivity design in the overall system. Now we didn't provide that device. It was provided by other vendors. So our ability to take the Marvell asset, combine it with i.MX processors and our channel to market, we believe, will enable us to grow the business quite nicely.

Got it. And then if you can talk about just from a margin perspective how you scale that business, if you will, right? I think from a margin percentage, I think it starts a little bit lower from an operating perspective. But just as you grow that, what type of operating leverage do you see?

So on a gross margin basis, yes, it's a little below our target, but we believe it's a manageable thing or a manageable task to improve that to kind of our target rates. In terms of the people, it's roughly 500 people. So in the grand scheme of the overall company, we don't think that that's going to be a big disruptor to the overall OpEx model or the leverage in the model.

Okay. Great. And then how about on the crossover MCU side? I think that's another unique aspect if I think about both NXP and Freescale from microcontrollers, the processing side of Freescale and just your ability to kind of intersect that market versus maybe some stand-alone MCU companies.

Yes. So if you look at our processor portfolio, it ranges from very, very broad embedded 32-bit ARM microcontroller family all the way from very low-end LPC family, which is a lower performance, very low-power product, all the way up to the higher performance Kinetis 32-bit ARM products. We also then have a very wide portfolio of i.MX application processors. And what we heard from customers is they wanted something that sat between both. They wanted a product that looked like a high-performance embedded microcontroller, but they wanted certain types of capabilities to make these industrial devices have either audio control so we were able to embed audio codecs or video interface, embed video codecs into that microcontroller chassis. And that's what we call the crossover business. Crossovers today, after just a little over a year of being in the market, the family is running at about $50 million, $60 million a year run rate. We anticipate that, that will probably double in the not-too-distant future, and we'll have a very nice runway going forward.

Got it. And as part of that, in the growth trajectory, I think you guys have talked about kind of edge computing. Kind of where do you fit in there?

Yes. So edge computing, so I think that's a -- it all depends on your perspective when you use the term edge. When we say edge computing, we're not talking about edge in the sense of like in a data center. We're talking about edge in terms of everyday devices we may have in our lives. Your car is an edge device, if you will. A lot of IoT devices are edge devices. Now what we've heard is a lot of customers, they -- currently, today, they move a lot of data to the cloud, they process it in the cloud, and they bring it back. What they're facing is, is that the latency of doing that is impacting their business efficiency. So they want to process more data at their edge, and they want that to be secure. And so a lot of the things they could be processing at their edge, and you might think about in a building with -- cameras around a building, right? You want to be able to process and make sure all that data is safe and may take actions at the edge. And so it's really -- think of it almost as a kind of a processing engine that could go in all sorts of things around your life.

Okay. Great. I want to switch gears and talk about 2 kind of themes, if you will, one in distribution and the other kind of China in-sourcing as a longer-term kind of thing to think about in semis. On the distribution front, TI's been very vocal about kind of pulling back from distribution, taking more business direct. What type of opportunities, if it does, does that create for something like an NXP?

So we're believers in distribution. For us, the distribution channel provides a great force multiplier for our sales teams. If you think about our industrial and IoT business, which is, effectively, tens of thousands of smaller customers who buy their products through distribution, there's no way we could touch all those customers ourselves. So distribution provides an invaluable asset to touch that wide, long tail of customers. So to us, it's very valuable. We also have a number of customers, strategic customers who, by their own choice, want to be serviced by global distribution for kind of kitting and things like that. So that provides a real benefit. Given that certain distributors are now going to have maybe some holes in their line card from TI's actions, I think they will be highly motivated to maybe focus a little bit more on the NXP portfolio. And I don't have anything more to share with you than that at this time.

Got it. And then the China front, you're also somewhat unique in terms of having IP and technology here as well as in Europe, right? And so I think one of the investor debates is around a secondary impact of the trade war in terms of China saying we can't overly rely on U.S. suppliers. And so understanding your markets, auto is an industrial long cycle and it takes time, but what type of behaviors are you seeing from customers in China around this theme?

So first and foremost, it's important to realize we are a Dutch company. And when we are in China, we make sure that, that's the face we wear, if you will. Now there's not hard rules, if you will. What we've heard is that if you can design in a Chinese-based company, do so. If you can't do that, design in a product from a non-U.S. entity. And if that's not possible, design in a product from a U.S. company. I think these are more nonofficial instructions, if you will. We have heard it from customers. Our view is to be, I think, a valuable partner to our customers in China. I don't think you can stick your head in the sand and say -- and pooh-pooh their aspirations to be independent long term, but we're going to be there and engage with them and try to help them with the products we have in our portfolio.

Okay. All right. Before we open it up to the audience, I did want to address the target model kind of margins, what you guys would like to expand margins to as well as capital allocation. So just maybe starting with gross margins. You've been very vocal about kind of the 55% threshold at a $2.4 billion in revenue. Just kind of what are some of the things that kind of get you there and then the longer-term target of 57%?

Yes. So our official model is 53% to 57%. We've said, intermediate term, we'd like to hit 55%. We believe we need about $2.4 billion in revenue to achieve that. After that, the ability to string together a number of quarters together at 50 -- at least 55% such that we can deliver a full year of 55% is kind of the second phase, if you will. And then from 55% to 57% is really going to be achieved by new product introductions. And I think it's really important to understand this is not going to happen in a short amount of time. This is going to be measured in years. Because if you think about it, new products that are just going to be introduced that are very margin-accretive need to ramp to enough of a material size in the overall portfolio and certain maybe older products need to roll off a little bit. So it's really probably important to realize this isn't going to happen in 1 quarter overnight. It's going to be multiple years.

Understood.

From an OpEx perspective, our target is to run R&D at 16% and SG&A at 7%. And those are targets you should probably see when we get back to more of a normalized revenue level, which is probably north of $2.4 billion, right? $2.4 billion is really the sweet spot for the gross margin, but it's not necessarily the normalized target for hitting those OpEx targets.

Got it. And then on capital allocation, you guys initiated a dividend over a year ago and an increase of 50%. You bought back a lot of stock, and you have a $2 billion authorization. So how does the company and Board approach returning cash to investors?

So our standard capital return policy is pretty simple. We're going to return all excess free cash flow to our owners. There are 2 mechanisms to do that, either just buying the shares back or providing a dividend. Our target on the dividend is to pay out 25% of cash flow from operations. In 2019, we were at about 17%. So we still have some runway to be able to increase the dividend in a measured, meticulous way over the next several years. And then the rest of the capital will go back to the shareholders via buyback. So last year, in 2019, we returned 95% of all free cash flow generated to shareholders.

Got it. And I think you've said, from an M&A perspective, connectivity was the one area where you really felt you can enhance the portfolio. But beyond that, how do you think about longer-term M&A?

Yes. So right now we feel the connectivity asset really filled the one obvious hole in the portfolio. And given that it's now filled, we don't feel that there's anything that we need to absolutely go out and do. Now as a standard matter of our approach to business, we'll always do small tuck-ins. We'll also do small carve-outs. We're always looking at the portfolio in a very active managed way in real time. But at this stage, I don't think you should -- there's not an appetite to do another Freescale-size deal. One, even if you could get it through the regulatory environment, there's not an appetite for it because it is disruptive to the organization and there's no real obvious hole in the portfolio that we need -- we feel we have to go fill right at the moment.

Understood. All right. I have a few more I could get through, but we have a few minutes for the audience as well. If you have a question, you can raise your hand, and we'll bring a mic around. All right. As we're waiting for that, maybe just going back to gross margin. Really just from the philosophy, right, sometimes, companies can kind of talk about gross margin percentage, gross margin dollars. I know NXP is focused on growing gross margin dollars. And really what drives that philosophy?

Well, we see there's a balance between revenue growth and profitability. I mean we -- it's not that we don't like high gross margins or anything like that, but we feel if we want a business that will continue to grow above the overall market on the top line, we think the 55% on a blended basis is about the right target.

Got it. Okay. And then maybe we can talk about com infrastructure. Maybe a good thing it's taken this long to get there. Always interested in 5G. Clearly, there's been some delays and pause. But just talk about kind of that market, how you see it playing out over the next 1 or 2 years and what it means from a content perspective.

Yes. So I think it's important to understand what we do in 5G or in base stations overall. So we do very specialized RF power amplifiers. These are the products that go up on the cell towers that you may see around town when you're driving around. So if you see those big rectangular boxes up on cell towers, that's where our products go into. We've said that the market for RF power amplifiers is about $1.4 billion in 2019. We've got about 40% market share. We think the overall SAM, if you will, will grow at about a 13% CAGR through 2025. Given how large we are in the market and given how customers tend to want to not let anybody get too powerful, if you will, we think that that's probably our target market share. 4G versus 5G, we see it as basically about a 3x uplift on a kind of like-for-like basis.

Got it. And I know from a technology perspective, you have LDMOS, which it feels like that's been kind of extended, if you will. You also have GaN, silicon germanium. So just how do you think about your position in the market?

So I think if you look at 5G as a standard, it can kind of be broken into 2 big areas. There's the sub-6 gigahertz band and there's the above 24-gigahertz for millimeter-wave. In the sub-6 gigahertz band, we have 2 products we can offer: LDMOS, which you mentioned, which is really the industry workhorse, which we've improved the efficiency to about 3.5 gig. Now from about 3.5 gig to 6 gig, GaN, gallium nitride, is the right technology. And then above 2.4 gig -- 24 gig, we would use silicon germanium from massive MIMO. I think what's important to remember in the 5G generation, almost all of these power amplifiers will be deployed as array-based, so MIMO-based products. Whereas in the 4G, it was mostly 2 and 4 T transistor implementations.

Got it. Okay. Well, I think we are coming down for time now. So really appreciate the discussion today, Jeff.

Great. Thanks, Craig. I appreciate everyone, thank you.