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

Great. Good morning, everybody. My name is John Vinh. I cover semis here at KeyBanc Capital Markets, and we're pleased to have NXP with us. We have Jeff Palmer, Vice President of Investor Relations. Welcome, Jeff.

Thanks, John. Nice to see you. Good to be here again.

Good to see you. So top of mind for everyone, Jeff, as I am sure you're having your conversations with investors is obviously the cycle. If you think about cycle, looks like you guys have done a better job of just managing through the volatility in the cycle than many of your peers. And you could argue that the automotive supply chain is much more complex than some of the other verticals. Maybe if you could just spend a few minutes discussing how you've managed to do this?

Sure, John. I agree with you it is topical. I think you have to kind of go back in time a little bit. And late in 2022, Q3 of '22, actually, is when we kind of realize this felt like the cyclical peak for our business. We started to see some kind of more weakening signals in the consumer side of the business in mobile and consumer IoT. And so at that point in time, we made the decision to limit how much inventory was in our channel -- distribution channels. About half of our revenue goes through global distribution, our normal target of inventory in the channel is 2.5 months. And with the supply crunch, we had gotten down to about 1.5 months. And so as we kind of felt we were at the cyclical peak, we said, look, let's keep the channel lean and tight, and we decided to keep it at about 1.6 months of which we've done since Q3 of '22. That's on the distribution side, right? On the direct side of the business, which is predominantly the auto. So automotive is 60% direct and 40% through the channel. On the direct side of things, last year in probably late spring, early summer, as you know, we had noncancelable, nonreturnable order programs. We started getting kind of the message from our direct automotive customers that say, "Hey, we know we have a contractual obligation with you to accept material we signed up for it. But if you ship it to us, we're just going to put it on the shelf." So we made the decision at that time to limit what we would force our customers to take. Now we didn't do it just because we're nice guys. The first thing we said to him is, okay, how much inventory do you have? And where do you want to get to? So kind of that was a stake in the ground that we want to have a conversation around. And secondarily, we said, okay, we can't just let you off a hook for a business or arrangement because that's kind of foolish. So we said, "Hey, give us design awards that maybe we were competing for. Let's talk about different price profiles, things like that." But basically, what we did is we said there's no benefit in jamming our direct customers. And so at that point, if you think about spring of 2023, we are limiting the channel, what we're shipping in, and we decided to limit what we would force our direct customers to take. Now some people might say, okay, that was foolish, you left the dollars on the table. But our view is there's no benefit. It doesn't create demand to shut the inventory into your customers. And that's really, I think, how you can profile how we navigate in the last couple of years in the cycle.

Got it. Maybe just a follow-up there. I think Kurt made the comment on the call that, hey, you are finding business combinations to help them work down the inventory you had referenced maybe agreeing to awardings and design wins. Is that what you're primarily referring to? Or are there other business accommodations that you've been able to construct with your customers in terms of helping them manage these [indiscernible]?

Do they tend to be around things like future business awards, pricing because if some -- if you're letting somebody off the hook on a volume commitment, that means they're taking less than they had made previously. And so ASPs are variable. It was more that type of the thing.

Okay. That makes sense. I don't know one question I have gotten from investors after your earnings call was that customer that wants to only hold 2 weeks of inventory. I know you're not going to tell me who it is, but I was wondering if you could just talk about how your customers are thinking about what the new norm is in terms of inventories in this post-pandemic environment? Is it they want to hold less or more than what they were holding pre-pandemic?

Yes. I think what's important to understand is that 60% of our automotive customers is direct. These are global Tier 1s. They tend to be primarily Western, European, North American Tier 1s. There's not a one size fits all. Each one has its own business conditions they want to try to achieve. We have some who want to hold inventory in the 10- to 12-week range. And we have 1 or 2, who think holding 2 weeks of inventory is the right idea. I'm not -- you're right, I'm not going to tell you who that Tier 1, who wants to hold 2 weeks is, but I will say it is a very large European Tier 1. We'll leave it at that.

So does it feel like the pendulum just shifted in their overacting too much to wanting to hold too little inventory on average? If you were just to take an average a sample across all your Tier 1 customers at this point?

Some of them are actually being pragmatic and are saying, look, 10 to 12 weeks makes sense, because NXP, our cycle time is about that. So that kind of makes sense. The ones who want to carry much less than that, they view as look, Mr. NXP, intellectually, we know it's too lean, but we are under very tight financial conditions here. Our working capital metrics are not great. And so while we know strategically, this might not be a smartest idea long term, tactically, that's where we want to land. And so it varies Tier 1 to Tier 1.

Approximately, how much was that customer holding pre-pandemic?

It's a fair question, John, and we're not going to drill into that.

Okay.

We do have -- I mean, I've reviewed your questions before we came in today, we do have some Tier 1s who want to revert to their battle pre-pandemic ways. Others are becoming little bit more pragmatic. What's interesting this time around is the OEMs where the actual customers of the Tier 1s are getting a little bit more concerned about it, because their view is if nothing changes, we're probably going to go into another supply crunch in '25 or '26. And so we have started to see OEMs either put pressure on the Tier 1s where they can with surprisingly, they don't have as much lever as we would have thought. And to some of them, some large ones coming to us directly and saying, look, we can't force the Tier 1s to take this material, but we know it's the wrong thing that they're doing. We understand why they're doing. So Mr. NXP, we'd like to place direct orders with you. We'll take delivery of that product and then we'll fork it off to different Tier 1s. Now it's not huge amounts of orders from these OEMs, but it's a change in trend and a change in behavior, and which is a positive coming out of the supply crunch in last couple of years.

We had heard a similar commentary from one of your other auto semi peers yesterday in terms of the -- maybe this tension that's kind of building up between OEMs and Tier 1s. I'm surprised to hear you say that they don't have as much leverage as you would have thought. Why do you think that is? Because ultimately, they're the ones that are making all the design decisions for the...

I think if you have to step back and look at it over a very long period of time, and the OEMs have not been easy customers to Tier 1s, and they've put some onerous tease and seize on them and they require top-shelf service. And it's -- in a lot of times, it gives is the working capital, right? If you hold too much inventory, your free cash flow goes to heck. So what I've heard is that a number of OEMs have said, look, we can't change the running business, but we can change the terms and conditions of new awards. So they see that they can modify behavior on future designs, which is good. And we also see some of those similar OEMs coming to us directly. So what we're seeing is a change for the old days.

Yes. So that's another interesting point that you're bringing up, Jeff, is it sounds like your relationship coming out of the pandemic is getting stronger with the OEMs that you're going more directly to them versus previously, maybe sometimes you'll be going through some of the Tier 1s. Is it fair to say?

Yes. So historically, all of our business transacted with Tier 1s. So that's our normal process. For a number of years, we've been working with the OEMs and design awards. But I would say it's more at the engineering R&D to R&D level. Through the supply crunch, there was a much heightened executives to executive type of relationship between Kurt and Bill, our CEO and CFO and the OEMs of different automotive companies. That's been helpful from a supply chain perspective, but it's also been much more impactful from some of the new design wins we're doing in the area of software-defined vehicle. I know we'll talk about that in a little bit. But a lot of the newer next-generation platforms are really being awarded by the OEMs directly, not through the Tier 1s that it normally used to be.

And going direct to the OEMs, has that helped also in negotiating kind of pricing on some of those new awards at all or...?

At the end of the day, the Tier 1s are still there converting ships into cars. And so no, it doesn't -- there's not what I would call a huge benefit. What does occur as we have better insights in the architecture of platforms, we understand the strategy of the OEMs, where they're looking to position these platforms. It gives us more of an insight. We're in the Tier 1 world, you basically only saw your part of the overall system.

Right. Maybe going back to channel inventories. A quarter ago, you had started to message that kind of complete the soft landing scenario you're going to start filling up the channel again and then you kind of came out on the most recent earnings call and said, "Hey, we're going to hold off here." What's the reason why you decided to kind of hold off here?

Two reasons. Actually 3, 2 of them external, 1 internal. So the 2 external reasons is, one, in the automotive Tier 1s. We were talking about how much inventory they have and how we've been trying to help them burn that off. We had a certain view of how that inventory burn off of a time perspective and where the landing point would be, that had shifted over the last 90 days. So the burn-off rate slowed a bit. And with the landing point of where they wanted to end up, went down even further in aggregate. And so we said, okay, we need to be a little bit more cautious here. Secondarily, in our industrial and IoT business, which is about 20% of the total company, we saw a weak demands in the core industrial side in North America and Europe. And since industrial and IoT is about 80% through the channel, we saw no reason to continue to refill the channels significantly. So it's really just a pushing off of the time line of when we will refill the channel. We do, at some point, want to get back to our 2.5 months. It won't be here this year in '24, sometime in '25 is all else being equal it seems reasonable at this point.

And is there a way to think about what do you need to see to kind of make that decision to start refilling the channel again? Are there certain metrics and inventory levels that you're watching for?

Yes. So over the last probably 5 or 6 years, we've built a pretty automated system to monitor the distribution channel on a global basis. We can see what we sell in on a daily basis, what gets sold out, how much inventory right part number, what geographies it is. And so I'd say we have a very good handle on what happens in the channel. So what we're going to want to see is a higher velocity of sell-out to the channel, right? As we see that go up and on hand inventories in the channel go down, that's a signal to us, okay, let's start to refill in more measured way. In either way, just as we measure the cycle down through the channel, we're going to be very measured and conservative, if you will, on the up cycle as well.

Great. Questions?

What is to quantify [indiscernible] but is that -- how do you think about that on your revenues?

Sure. Yes. So the question is, can I kind of give you a metric of what the value is of going from 1.7 months of inventory in the channel to 2.5. So the kind of the rule of thumb is for every 0.1 month of inventory, it's about $50 million. So I think when we were running it at about 1.5, 1.6 months, we use the aggregate target of about $500 million to refill. Now since that time, we've gone from 1.5 to 1.7. I'll let you do the chance on math on that.

Any other questions? Jeff, I think you also referenced kind of some company-specific drivers that should help things kind of recover for you in the second half. I think radar was one of -- I think you talked about ramping some new radar wins. I'm just wondering, are these the new 28-nanometer of CMOS solutions? Can you talk about what the advantages of these solutions are? And how do we think about content or the market share impact for you?

Yes. So radar for us is a very key part of our automotive business. We said that our radar business would grow from about $600 million in 2021 to about $1 billion, $1.1 billion in '24. I'd say that trajectory is still correct directionally. We don't have just a single part. We have a whole portfolio of different radar products we sell. The 28-nanometer RFCMOS is a 1 chip integrated 77-gigahertz radar product. What's the benefit of going to 28-nanometer CMOS, you can integrate the processor and the transceiver front end into the same die. And so you get a lower power, smaller die, good cost effectiveness, which enables the auto OEM to distribute those devices in multiple areas around the car. So that's one type of radar node we sell. We still continue to sell silicon germanium front-facing radar, which tends to still be a very good technology for distance, but that tends to be a 2-chip or multichip solution because there's no integration path for silicon germanium. So just the point to make, John, is it's not a single part portfolio, it's multiple parts. There is some 28-nanometer in the design awards, but some of it is some of the other products as well. And you have to also remember for automotive design awards, it's a 2- to 3-year design to revenue cycle. So while when we get close to when a product starts to ramp, is when our interaction with the OEM who awarded us that design goes up. But we've been baby sitting in these designs for some time. That didn't just pop up last quarter.

Are these multiple designs that are going in production? Or is this kind of a big platform?

It's multiple. So there's a smaller number of Tier 1s globally that actually are in the radar business, if you will, primarily in North America and in Europe, and it's just more expanding that footprint with those OEMs the Tier 1s.

Maybe we can also talk about software-defined vehicle. This sounds like a big opportunity for you. When do we start seeing cars roll in production with this new architecture? Can you talk about what the incremental content opportunity for you is in when you think about SDV, is this more of a TAM expansion opportunity? Or are there NXP specific advantages that's going to allow you to benefit more from this?

Yes. Excellent [indiscernible]. So what I do think it is a TAM expansion, but I do think that we, as a early movers in this area will have a disproportionate advantage from that TAM expansion. The first time you'll be able to probably buy a car that will have a 5-nanometer vehicle computer from NXP will probably be the model year '26 or '27, so late next year is when model '26 will start to go into production. We're already shipping domain and zonal processors into early kind of what I would call early versions of SDV. Our S32 family, which is the processor family that's behind software-defined vehicles. We started shipping revenue materially in '21. It was about $300 million of revenue in '21. We expect it to hit at least $600 million this year, and I'm pleased to say we're ahead of that target. So the adoption rate has been very good. But that's been primarily on our domain and zonal products, the 16-nanometer domain products and there are 20, 40-nanometers zonal products. The thing about the software-defined vehicle evolution is you're not going to go into a car dealership and they're going to say, "Here's a software-defined vehicle." What this is, is really the infrastructure of the car. It's the part of the car, you as a consumer really don't see and touch and feel day to day. But think of it this way, if we were to go buy a car today, the best year of that car is the day we drive it off the lot today. It's kind of a -- it stretched to waste from a functionality perspective. The whole idea of the software-defined vehicle is as you -- once you drive that car off the lot, over time, the OEM can update through the -- over-the-air updates to the software of different functions in the car that will keep that car feeling fresh and bring maybe new features to you. So the idea is when you actually drive that next-generation car off the lot, it will have quite a bit of processing headroom built into the vehicle.

Is there a way to think about kind of the range of incremental content?

What I can do is I can kind of give you a sense of where we're at today. So on average, a automotive microcontroller, which is kind of the run of the mill product in the industry. It can be in kind of mid-single digit, call it, $5 to $7. So a zonal product, which there would be multiple zonal processor, just think of a zonal processor is something that aggregates physical functions around the car. Those processors can kind of be in the mid-teens range. . Then you have domain processors. Think of a domain processes as functional aggregation, not physical aggregation or functions around the car. Those domain processors, they can be in the mid-20s range, $20, $25, and you would have multiples of those in the car. And then on top of that, what we call vehicle computer. This is our next-generation 5-nanometer product. It's actually a quite complex multicore product. That product would be probably about $50 and in multiple applications that we've seen, there's multiples of those vehicle computers for redundancy. I think what's important to realize, as the software-defined vehicle secular trend evolves, there's not one single architecture that every OEM is adhering to. They all have different ways kind of skinning the cat depending on what their pain point is. If you are an OEM as, let's say, the high volume and wiring harness weight is a key thing to focus on, you might first go towards a domain -- I mean zonal-type product. If you're a more complicated, say, a premium-brand product where you really are trying to get more logical control of the software, you might first go towards a domain type of product. Over the long haul, we think these different architectures sort of [indiscernible] and so there'll be different blends across the industry.

Right. Maybe we switch to EVs. Wondering if you could just remind us kind of your position on pure EVs versus hybrids? I think there's a view out there that given that you've got a strong position in hybrids, that's helped you kind of whether some of the volatility around EVs more recently?

Yes. So we're actually agnostic between pure electric and hybrid electric. Our content on a pure electric 800-volt vehicle might be something like $120. And that's just on the powertrain. So let's set that aside for a second. On a hybrid, it might be $60, because the battery pack is lower. Our battery management system, which is our focused product for the electric vehicle market is BMS, battery management systems. And our approach is a system approach, which combines precision analog front devices that sit on the battery packs, connected with in-vehicle networking technology to a back-end processor for doing load balancing.

Do you have a view on just how we should be thinking about kind of the ramp of the EV market? It seems like we kind of got over skis here. Do you think there's more staying power for hybrids? And then how do you think about kind of the ramp of EVs going forward from here?

Well, I think that this concern about the EV is the EV party being over is very much a U.S. Wall Street concern. I think, yes, some of the large North American OEMs are struggling a little bit to really put up competitive portfolios. But the Chinese OEMs are doing quite well and China continues to be where the EV game continues to be quite strong, both on the hybrid side as well as on the pure electric side. Now that growth in China will have to continue through exporting to Western markets. So we have to see how that plays out. I think looking at the most recent S&P data, which is not 3, 4 weeks old, they're still projecting China EV sales or unit production to be up at least 20% here, 23% to 24%. So that's still pretty good. And China is still about 40% of the total ex-EV market that's a combination of hybrids and electrics.

Maybe we can touch base on the Vanguard deal that you guys recently announced. I think you said when fully ramped, you're looking at 55,000 wafers per month. Can you give us a sense of just how much incremental capacity this Vanguard deal is going to give you?

Yes. So there's a bit of a history, maybe I should kind of go through it a little bit, John. So about 18 months ago, when customers were much more sensitive from the supply crunch about sourcing. We realize we had a fairly rich design-win pipe portfolio of mixed-signal products across different end markets and customers are wanting to know, "Hey, where is this capacity going to come from?" And as you know, TSMC is our long-term major foundry partner, and we engage with them and said, "Hey, look, we have about $4 billion of incremental annual revenue, which will start to become deliverable if you will, in the '27, '28, '29 time frame more in the latter part of that window." And TSMC said, that's great Mr. NXP, and we're really glad that you used our design kits to design these products, but we don't really have a lot of interest in building greenfield trailing edge mixed signal capacity, but they're a really good customer, a really good partner. I said this and look, what we think you should do is engage with Vanguard, which we own 30% of and build a joint venture and will help with the technology when the process gets -- you guys can build greenfield 300-millimeter facility. That facility's JV was announced in June. There's 2 phases. The first phase, as you said, John, is 55,000 wafers per month that will be -- we expect it to be fully loaded and up and running in 2029. There is a Phase 2 of that facility. It will be on the exact same campus, which will add another 45,000 wafers per month. Our ownership of the JV will be about 40%. In the early years, we'll have a little bit higher volume ownership because we're making some prepayments because of these design wins I talked about, but our fundamental equity ownership of the facility will be 40%. Now long term, there's some benefits from investors' perspective. So in the short term, it's cash out. Okay, you guys may not be real happy about that, but it's a good investment in the business long term. Beginning in '29, we'll see about a 200 basis point benefit to corporate gross margins and that comes from moving to 300-millimeter wafers, avoided margin stacking and things like that. And then long term, this is out in the 2030s time frame. This facility, this JV gives us a landing place to take some of the material that we run internally on our 200-millimeter fabs and move it over to 300-millimeter facility, because our current 200-millimeter fabs, we have one in the Netherlands and 2 in the U.S. that they're 40-plus years old. They're aging out. And so at some point, we are going to have to rationalize those facilities.

Great. Thanks. Looks like we're out of time.

Okay.

Thanks, Jeff.

Thank you, everyone. Appreciate it. Thanks, John.