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

Okay. We'd like to get started. Good morning, everyone. Thank you so much for coming. Our next guest is Jeff Palmer from NXPI, Senior Vice President of Investor Relations. Jeff, thank you so much for coming.

Thanks, Toshiya.

I will go through a bunch of questions, but we'll definitely open it up to the audience. So to the extent you do have questions, please be prepared. So kicking off, Jeff, you guys reported really solid Q2 results. You guided Q3 revenue up 4% sequentially. At the same time, you did modestly take down your full year revenue outlook from, I think, flattish plus or minus year-over-year back in April to down slightly. Again, very modest change. But what were the major industry or company-specific developments that drove the update?

Yes. Good question. So the way to think about what was different between April and after we gave our Q2 guidance was; first, in automotive, we did see the rate of inventory digestion at Tier 1s be a little slower than we had anticipated in April, still coming down, but coming down at a lower rate than we had originally anticipated. And a number of the Tier 1s also lowered their landing point. So they want to bring inventory down to a certain number of weeks on hand. Some of them took that kind of landing place down. That's one. Secondarily, in our industrial and IoT business, we did see weakness in our North American and European industrial -- core industrial business. And I will say that, that weakness persists. We're hoping it gets better, but we're watching. It's not an inventory situation. It's more end demand by our customers' customers, if you will. So keeping a very close eye on that. But I would say that, that has not gotten better. And then thirdly is more an action that we took internally, which is, as you know, through the peak to trough cycle over the last 2 years, we've managed our distribution channel very -- I'd say, very aggressively, if you will. Because of this kind of weakness in automotive and in industrial, we made the decision not to refill the channel at a rate we had originally contemplated back in April. So we originally probably thought we would have gotten to probably closer to 2 months if the cycle has kind of really come back better into the second half. But instead, our Q3 guidance only calls for the channel to go back up to about 1.8 months. And even that is -- it's always difficult to actually understand where the channel will end in advance, right, is it mix related and things like that. So I'd say those 3 big things were the biggest difference between the post Q1 and post Q2 results and guidance.

Okay. Great. Jeff, you touched on this a little bit, and I realize you don't have perfect visibility into customer inventory levels. But curious, how would you characterize current customer inventory at the OEMs and the Tier 1s relative to 3 months ago and relative to what you all would consider healthy? And do you see any stark differences between the respective end markets or the device types?

So I think that question really is around our automotive businesses what you're talking about. So first thing to remember is each one of our 4 segments has a certain amount of distribution exposure and a certain amount of direct exposure. Automotive being just over half the total revenue of the company, about 40% of that business goes through the channel, and that's primarily for customers in Asia and China. The other 60% is North America and Europe-based Tier 1s. So we do not transact business with OEMs directly, a little bit, but not significantly. So the majority of our business in automotive is to Tier 1s. As I said in your first question, inventory started to build kind of probably coming into '23 at our Tier 1s. We've been working with them to kind of help them moderate that amount of inventory. It has come down. It continues to come down, but at a rate less than we would have ideally thought. In terms of the amount of inventory there, we're not going to disclose it because it's not auditable. It's kind of -- best case, it's art to figure out how much inventory is actually there. But I would say that ideally for us, we'd like to see our Tier 1 partners carry about 10 to 12 weeks. And that makes sense when you think about our cycle times of 3 to 6 months. So we do have some Tier 1s who, for their own decisions, would like to get down to as low as 2 weeks. We do have a few who are already at 10 to 12, and these Tier 1s are already back reordering and maintaining a certain amount of safety stock on hand. So it's really -- we're watching what the Tier 1s do right away at this moment.

And obviously, your customers and your customers' customers during the pandemic went through a really challenging time from a shortage perspective. Are you concerned at all that some of your customers are running on too little inventory?

Absolutely. As maybe some of you know, we did send a letter out to our customers, both our auto OEMs and our Tier 1s, basically highlighting just that fact, Toshiya. If our cycle times are 3 to 6 months, and if you're a customer and you want to drive your on-hand inventory down to 2 weeks, you're effectively setting up the potential for another supply crunch sometime in the future. Because if we do snap back or we just get an incrementally better environment, that 2 weeks is going to burn through very rapidly. As you know, the auto market is very much driven by sentiment. It's a very deep supply chain, starting with OEMs at one end and multiple Tier 1 transformation sites in between and companies like us at the other end. As soon as sentiment changes at the OEM level, it ripples back through that supply chain and gets amplified at each stage. Unfortunately, semiconductors do tend to be the dog of the tail, a bit of a bullet. It goes with the neighborhood. And so what we're trying to do with these conversations with Tier 1s and the letters to the OEMs is to say, guys, look, we're not trying to say it's not a cyclical industry. So let's be prepared and let's try to learn something from this last cycle. We'll see how it plays out.

Curious, when did that letter go out?

It was sometime after, I think, Q1 and sometime in that time.

Q1. Thank you.

We've been working with -- I'm sorry to interrupt. We've been working with our Tier 1s since about spring of '23 to help them and understand how much inventory they have and what their burn-off plans were. And as I said earlier, it's not scientific and auditable. So really, we learn about how much progress they're making by our ongoing interactions with them. We kind of -- just like you would do, build a mosaic out of how much inventory each one has.

Got it. Okay. Very helpful. Similar question for your distribution business, which, again, at the company level, I think, is 50%, 55% of revenue. To your point earlier, at Q2, it was 1.7 months. Your Q3 guide assumes 1.8 months at the end of September. Remind us what your philosophy is as it pertains to managing the channel? And what would it take for you guys to say, okay, we can take it up to 2.5.

So our long-term model, all else being equal, is 2.5 months, plus or minus half a month. Pre-pandemic, as you remember, we were running the channel at 2.4 months for over 5 years, consistently solid. So we do have a very auditable systematic methodology for knowing what we ship into the channel, what sits in the channel and what gets shipped out on a part number, geography, customer base. So I'd say we probably have one of the best channel management strategies in our industry. Just like we've managed the channel very tightly from peak to trough over the last 2 years, 18 months, we're going to manage the channel very conservatively on the way back up. We don't want to get over our skis, right? It'd be very easy to read a signal in 1 month and say, okay, things are back to normal, ship inventory and find out you just got head faked. And then you're stuck. You got to digest it. And I think that's unfortunate situation some of our peers found themselves in, right? They thought there would either be a longer cycle or a faster snapback and they positioned inventory into the channel. And now before they are able to see resumption of growth, they've got to digest that because we all recognize revenue on a ship to into the channel.

Got it. A question on NCNRs. I think you guys did a really, really good job in sort of managing the upturn and managing the downturn, other companies a little less so. Hypothetically, if a very sharp upturn were to hit you guys in the industry over the next couple of quarters, and you had nonstrategic customers come to you asking for long-term supply, would you sort of be open to that? And when you sort of think about the NCNR in the most recent cycle, what aspects of it would you inherit and sort of keep -- and what aspects of it would you tweak or change?

So NCNRs as a mechanism served its purpose during the time. And if we found ourselves in a similar supply crunch again, we would reuse them again. They did serve the purpose. I'd say the learnings for us from the past cycle is, we would more than likely only extend NCNRs to our largest direct customers. We're not going to -- we don't -- we didn't do NCNRs with our distributors. It's just there's no need to do it. There's probably less need to do it with our less strategic businesses. So that kind of narrows you down into that automotive and kind of core industrial side of the business, which is still about 70% of the total revenue of the company. Our model, which was different than others, which was our NCNR program was calendar year based, and we think that, that served us very well. So while maybe you had a customer who participated in '21, as we went into '22, maybe they decided they did not want to participate. So that turn of the year was where that decision was made. We didn't have a kind of open-ended book, if you will. Lesson we did learn is, it probably does not make as much sense to build NCNRs with some of our mobile customers or nonstrategic customers because I think there was a feeling that NCNR meant free pass to the front of the line for supply, and that's not how we managed it. It was still a bit of an orderly first come, first serve. It was really a way for us to understand how much capacity we needed to put on our suppliers, how we had to plan our own back ends and front ends.

Got it. Okay. Shifting gears a little bit, let's dive into some of the businesses. In automotive, your current long-term growth target is 9% to 14% on an annual basis. Level set us by breaking down the automotive business into parts of the business that are tied really to ICE or SAAR, if you will, and those that are more secular in nature. And I know you've got an Analyst Day coming up. But if you could remind us how radar, zonal processors, electrification are doing relative to the expectations you had thrown out a couple of years ago.

Okay. A lot of questions. Yes. So let me first say, you're correct, Toshiya. We do have an Analyst Day. It will be on November 7 in Boston. We hope you'll either attend in person or online. Given that it's so close, we're not going to update our new model. But I will say this, this upcoming Analyst Day is probably more evolutionary than revolutionary, and many of our businesses have very long cycle times, both from a design to revenue cycle as well as life cycle. So I don't see a revolutionary change in some of our growth outlooks. So as it pertains to automotive specifically, you articulated very clearly, Toshiya, we -- at our '21 Analyst Day, we parsed the business kind of into 2 large parts, one being a core part of our business where we have very large market shares in the automotive marketplace. And where -- when your market share is of a certain size, it's unlikely you're going to outgrow the market. So that bucket, the core automotive business, that was about 60% of our revenue in '21. It was expected to kind of grow mid-single digits kind of similarly to auto semiconductors. The remaining was built up of, as you rightfully called out, through what we -- 3 accelerated growth drivers: our radar business, 77 gigahertz radar. That business was about $600 million in 2021. It's expected to hit about $1 billion, $1.1 billion here in '24. I think we're on track, a little behind, but not significant. And that will continue to grow into future periods. So it's not like end of '24, it all comes to an end. It's just -- it continues to grow and penetrate. The 3 ways to think about the radar business is, every year more cars with radar, every year more radar nodes per car. And every year, our customers asking us for more functionality and capability in each node. And so each one of those stages creates the flywheel to drive continued growth. And that thinking is still in place. Second most important growth driver for automotive is what we call our S32 CoreRide platform. This is a family of both -- of MPUs, both for domain processors, zonal processors and vehicle computers. Vehicle computer is a 5-nanometer product. It actually does not start volume production until the end of next year. The domain processors are in production today. Zonals are in production today. The whole idea is to move the modern car from a flat point-to-point architecture to a hierarchical architecture of a hierarchical processing fabric. So the customers -- our OEM customers can update software in different functions around the car after you drive the car off lot, so to speak. That business was about $300 million in '21. We had expected it to grow at about $600 million in '24. I'm pleased to say we are ahead of that. Our design win rates for that business have been the strongest in the history of the company. Every year surpassed the prior year. So we have a very rich pipeline of new opportunities, most of which go into volume production in forward periods. We'll talk about that at our Analyst Day. So we feel very good about that. And then the third area is our electrification business. It's the only part of our auto business that is exclusively tied to either EVs or hybrid vehicles. This is battery management systems and inverter control systems. That business was about $200 million in '21. We said it would hit about $500 million here in '24, and we're on track there. I know you had a question in your list that you sent to me about the difference between pure EVs and hybrids. We're agnostic, right? Our business grows and scales -- our EV business grows and scales depending on the battery density. So when you look at a fully electric 800-volt car, that maybe $120 of content for NXP. A 400-volt electric or even hybrid vehicle, about half of that.

Okay. Great. So the recent change in mix sort of in favor of hybrids and EVs coming down for you guys, it's?

It's neutral.

Neutral. Okay. Got it. So on the industrial side, as you spoke to earlier, you've got core industrial, you've got consumer IoT. Again, if you can level set us by giving us a rough split of where that business stands today. And I think that business has been coming back, but you also spoke to some persistent weakness on the core industrial side. So if you can talk about the 2 subsegments in that, that would be helpful.

So it's roughly split about 60% for core industrial. What we mean by core industrial is factory automation, building automation and health care. Those are the primary major parts of that core industrial business. We do focus on other things, but those are the primary. On the consumer IoT makes up the other 40%, and that ranges traditional consumer businesses from wearables all the way to home automation. That's kind of where we draw the demarcation line. Our go-to-market or our product set there, it's really driven around a broad portfolio of industrial processors. So you start with the processor footprint, if you will. You pull along analog attach, analog connectivity, then you have wireless connectivity or wired connectivity and then security as well. And so we try to sell subsystems or reference designs to our customers. One of the things that I think investors are challenges with this segment is, it is a super long tail business. It's tens of thousands of customers. 80% of industrial and IoT goes through the channel, right? And there's no one customer or one customer product type that makes up more than a few points of revenue. So I can't say to you, Toshiya, track our white goods in China and you have your perfect read on our industrial IoT business. It's too diffuse. We even struggle to understand what drives it in a quarter-to-quarter basis. You have to look at it over kind of multiple years.

Got it. Got it. Okay. And as you think about the business over the long run, similar to automotive, again, you're not going to preview the Analyst Day, but how we should be thinking about growth relative to the market relative to GDP, that shouldn't change all that much.

No, not at all. I think it's a business that is very correlated to global PMI. I would say the effort and the -- our go-to-market and why we build our solutions, the way we do resonates with customers still. I'd say our channel strategy over the last 3 years did take some of the wind out of its sales, given how exposed to distribution it was and how tightly we control distribution. We did limit a bit of the growth there. But the fundamentals of where we're focused, what we're focused on, the product set has not changed.

In your mobile business, you've had 2 consecutive quarters of double-digit growth, which is great. Curious what's driving your content growth today? And what are your expectations as it pertains to both smartphone units and content for you as AI starts to become a thing, if you will?

So I'll just state the obvious. We are a niche player in mobile, right? 10%, 12%, 13% of total revenue. We do kind of one thing primarily. We do secure mobile wallets. So our whole go-to-market in our mobile business is around virtualizing physical things like your identification, your credit card, eventually your key for your car or your home into a secure element in the phone. That's our primary business. We have a very good exposure to both iOS and Android in equal shares. We do some custom business and some custom analog interface business for one of our customers, which we don't talk a lot about. I would say in the last couple of quarters, it's really been a rebounding from the trough in Q1 of '23 when the industry in general saw a bit of inventory digestion at primarily Android customers. That's behind us. But there's not been, I would say, a new socket or something to talk about. It's continued focus on developing and delivering increased functionality to our customers.

Okay. I appreciate it's a relatively niche sort of business for you. But as sort of the phone becomes more intelligent, if you will, over the next 2, 3 years, any sort of apps that could drive above-average growth for your mobile business? Or should we just sort of assume...

I'd say the 2 functions stay the same, security and analog custom interface. But in terms of that security business, the killer app that we've been working on over the last couple of years has been the virtualization of your physical car key into your phone. And that's facilitated with a technology called ultra-wideband technology. We are unfortunately only focused on the Android market because our customer in the iOS market has their own ultra-wideband solution. On the phone side, it's gone a little slower than we probably would have thought back in '21, more because Android players tend to be very pragmatic. They'll design you in, but as they get ready for production, they make a decision, can I monetize the socket or can I not? And they either populate it or they don't. Whereas in iOS, it tends to be a single SKU type of business, right? I would say on the car side of ultra-wideband, it's gone actually very, very well, probably a little bit better than we thought. We'll update you a little bit at the November Analyst Day. But I would say that part of that application is going well.

Got it. I tend to get a bunch of questions on pricing across your business. You guys have been pretty transparent on pricing over the past couple of years. I believe blended ASPs were up in the mid-teens. In '22, I think, it was 14%, up about 8% last year. And I think you've been staying sort of flattish for this year. As you look ahead, I realize negotiations haven't really started yet, perhaps for next year. But what are your preliminary expectations? What should we be assuming specifically, particularly given sort of the inflationary environment we've been in and the capacity that's been added, both on the IDM side as well as the foundry side?

Yes. So our approach to pricing has been pretty conservative, I think. Our view is if we faced inflationary input costs from our suppliers, we would pass that on to our customers just grossing it up to maintain our gross margin. We did not pad our pricing. We wrote that into all of our contracts. So that's there today. It's never a comfortable conversation with your customers to talk about pricing. This year, as you rightfully said, '24, it is likely to be flat, all else being equal. Into '25, we do have the ability, if at our discretion, to offer low single-digit pricing to our largest direct customers. It's not across the board. It's at our discussion depending on the business relationship.

Okay. So again, we should assume sort of the big automotive Tier 1s, those kind of customers.

That's the right kind of customers, but you also have to understand that we've taken a much more business approach to pricing. It's not just you're our customer, we're going to give you price reductions because it's [ Monday ], right? It's what's our business relationship? Where do we see it going together? We want a symbiotic relationship with our customers -- and that's -- and pricing plays one aspect of that.

Okay. And you would still have sort of -- you're not jeopardizing margins. Your costs are coming down, and therefore, you have the ability to.

That's a great point, Toshiya. So our -- and this goes from before the pandemic. Our whole view about annual price concessions was really around sharing with our customers. Operational efficiencies that we can drive out of our business to a certain extent. Pre-pandemic, we were down to about, I'd say, 1%, 2% annual costs, if you will. A decade plus ago, we were probably 3% to 5%. So we brought that down over time.

Okay. So the benefits that you're getting from your foundry partners are sort of reverting to pre-pandemic times, if you will, plus or minus?

There's also you have to remember, 40% of our front end is internal, right? And 80-plus percent of our back end is internal. So we have a certain amount of levers on our own to try to just improve functional and operational efficiencies.

Right, right. Okay. That makes sense. The other topic we get questions, and I'm sure, Jeff, you get as well is competition in China, given the geopolitical environment. You're in a little bit of a unique spot, if you will, you're European/American depending on how you sort of look at the setup and where you're located. But can you remind us what percentage of your business today is China -- for China? And what sort of portion of that would you consider to be at risk on a 3-, 5-, 10-year view? And what percentage of that is sort of safe, if you will, if there's...

Sure. So about 35% of our business is shipped to China. All semi companies recognize revenue on a shipped-to basis. Our best estimate is about half of that 35% or 17% is for Chinese-headquartered companies. The rest is for multinational companies, primarily for export. I'd also say as a footnote, those Chinese-headquartered companies, a lot of that is for export as well, but it's near impossible to figure that out. In terms of competition, we take a very pragmatic view. We think over the next decade, there will be a larger indigenous Chinese semiconductor market. There's no reason why it shouldn't be. We've been a little surprised it has not become more material to date. But we continue to believe the trains left the station. They're going to continue to try to innovate. They're going to continue to drive valuable products into the marketplace. Our snapshot today is as follows. In the power discrete area where we don't participate, there are Chinese start-ups of material size and Chinese foundries in that business. We see it today, primarily for the automotive business. There are some start-ups doing, what I would call, low-end catalog analog products, probably at the low end of our portfolio as well as low end of ADI and TI's portfolio, but small. And then there is consumer microcontroller players. They tend to be limited in scope and size, but they are there. We don't see any indigenous Chinese semiconductor companies being materially in our SAM, if you will, in the automotive marketplace. Could there be in the next decade? Sure. But the onus is on NXP to constantly innovate. If we don't innovate, we get lazy or we just fall back or we miss a cycle, we deserve to get our lunch eaten. But so that's no different than any other market we participated in. Today, in China, in the automotive market, we really compete with our Western peers, Renesas, Infineon, ST, to a certain extent, TI and ADI. There's not a new "bogeyman" under the bed in China who's been popping up.

Okay. Helpful. In terms of your manufacturing footprint, you've been focused on diversifying through your foundry partners, particularly in China. You also recently announced a JV with Vanguard. Maybe speak to your overall manufacturing strategy and more specifically, what drove the decision to team up with VIS?

Sure. So our strategy, we term it as a hybrid manufacturing model. 60% of all of our wafers we buy from third-party foundries, TSMC, Global, UMC, Samsung [ usual suspects ]. The other 40%, we build internally at 3 200-millimeter mixed-signal factories. That's on the front end. On the back end, as I said earlier, we do about 80% of our business internally, 20% we outsource. Our -- what drove the focus or the JV with Vanguard was as follows: About 18 months ago, we realized we've been very successful with a family of multiple products, mixed-signal products. And these products have been designed using TSMC's design kits, so they were very tied to the TSMC flow. So Kurt, our CFO -- CEO, went and spoke to TSMC and said, "Hey, we have this very large amount of business that we're going to have to start to supply in the coming years, and we want to know that we can count on you to provide it." And to TSMC's credit, they were very transparent and said, "look, we're not really that interested in building trailing edge mixed-signal capacity. We appreciate you using and being a dedicated customer to us, but it's not our cup of tea going forward." But this is the thing that I really give TSMC a lot of credit. They're a very honorable vendor to work with. And so they realized that our commitment to them might have been misplaced, right? So C.C. Wei, the Chairman of TSMC, and Kurt, worked on a deal that said, look, why don't you go and work with Vanguard, which TSMC owns 30% of and since they want to expand into the 300-millimeter mixed signal marketplace. See if you can build a JV that will work for both of you. If you can find that business arrangement, we TSMC will agree to sublicense to the JV our technology process flows, which is a very unusual step for TSMC. Fast forward, June of this year, we announced that JV. Total cost is $7.8 billion. We are buying or investing in a 40% share of that JV. We don't want to consolidate into our financials. That 40% share will be $1.6 billion for the equity portion, and we're also making a $1.2 billion capacity prepayment, if you will. The nice thing about this JV is, first off, it won't be up and fully running until probably 2029, fully loaded. But at the time at which it's fully loaded, fully running, that portion of JV for us will generate about a 200 basis point benefit to corporate gross margins. So whatever your models are for 2028 gross margins, think about over the period of '29, we'll grind another 200 basis points on top of that. So it's very positive. That's kind of the first step. The second step is, if you think long term, we have these 200-millimeter fabs, which are starting to age out. And at a certain point, we are going to have to mothball them. And so the JV gives us a landing point to start to rationalize our own internal fabs over longer term. That will also lower our fixed costs to improve margins and also make us more of a variable cost business. But that's over a decade kind of a window time.

Right. I'd like to say we have a 2029 estimate, but we're not quite there yet. No, you're not. Okay. Gross margins, again, you guys have done a really good job, 58% plus, I believe, over the past 2 years. During a time when utilization rates went from 80%, 80% plus, to now the low 70s. So there were offsets to that headwind. I guess, what were those offsets? And how should we think about the puts and takes going forward?

I think the biggest offset to think about it a longer term over the last decade. If you invested in NXP a decade ago, our cost structure was probably 70% fixed, 30% variable. At that time, our external wafer sourcing was probably 25%, 30%. Fast forward to today, that's almost completely flipped. Our fixed costs are about 30%. Our variable costs are 70%, goes with how we source our products, 60% on the front end, external, 20% external on the back end. And we've just also become much more operationally efficient since the years of the Freescale integration. And I think the way to think about gross margins going forward is 55% to 58% is the current model. I would say 58% is not the destination. It is just a stopping point to drive gross margins higher. The way to think about what are those levers for gross margin expansion over the kind of intermediate term, first, we'll refill the channel. The channel is effectively a margin-accretive go-to-market channel, all else being equal. So that -- as we start to go from that 1.8 months back up to 2.5 months, that will be a tailwind to gross margins. And a material or a reasonable up cycle, we will start to raise our internal utilization from the low 70% range back up to that mid-80% range. That's kind of the sweet spot for our fabs. That will also be a margin accretive tailwind. So those are the 2 things to kind of look at tactically over the next, call it, 12 to 18 months as this new cycle kind of materializes. We also believe that all of our new products -- new product introductions do carry accretive margins because it's just how we allocate capital internally. We're always looking for the best opportunities to invest R&D in.

That makes sense. I'm going to pause here and see if we have any questions in the audience. Can we get a mic in the front, please?

Large auto OEM this morning preannounced -- out of Europe preannounced some pretty negative news. And just trying to get your reaction when you see statements like that, how you sort of process it real time. You talked earlier in the presentation, which is very helpful about seeing some weakness. It sounds like industrial side, certainly not getting many better. But do you see these type of announcements as sort of like, well, maybe there's actually a step function down that's going on here. And curious to get your view, is this cyclical, if it is more structural and structural being because of China and so forth. So I'm just trying to square that circle. However you want to answer it would be helpful.

Yes. I think I'm going to try to paraphrase the question. I think the question, tell me if I've got it wrong, is the current environment with this kind of digestion of inventory at the Western Tier 1s, is this a structural change because of geopolitical changes of competitiveness? Or is it a classic semiconductor inventory cycle? I'd say it's the latter. It's a classic semiconductor inventory cycle. If you kind of go back to '22 into '23, we all were scrambling for supply. Our customers were scrambling for supply. We all know how customers get during periods of tight supply, they overorder and they order what they think they can't get based on their current book of business. And so as their business changes, they do find themselves with sometimes the wrong mix, sometimes with the wrong amount of inventory, and it takes time to digest. And we started working with those customers spring of '23 to help them start to digest that inventory. But from our vantage point and our discussions with OEMs, and with the Tier 1s themselves, it appears to be a classical inventory digestion process, not a structural change. I think the OEMs are trying to figure out how they navigate a more competitive Chinese auto market as well as a Western market that has not embraced EVs as wholeheartedly as some of their Chinese competitors. So there's a lot of dynamics there. But at the net-net, I think it's an inventory digestion.

Any other questions? I'll keep going. So from a capital allocation standpoint, Jeff, you guys have been very disciplined. You've essentially returned the vast majority of free cash flow to shareholders in the form of a growing dividend and buybacks. You just recently announced an additional $2 billion buyback program. Going forward, should we expect more of the same from you guys? And I mean that in a positive sense? Or could M&A come into the picture? How should we think about sort of the puts and takes?

I think it's more of the same, Toshiya. I mean our capital allocation strategy is actually very simple. We aim to return all excess free cash flow to our owners through a mechanism of either dividends or buybacks. On the dividends, our target is to have a payout ratio that is 25% of our cash flow from operations. It's been a little below that primarily because inventory has kind of occupied a bit of our space in cash flow from operations. And then buy back shares -- we don't try to time the market. We have a 10b5-1 program that's in place all the time, and we just kind of methodically grind on the returns.

Got it. I have a bit of a random one that I've been asking all our companies at the conference, AI usage at the company. AI is a big topic. Everyone is talking about it. Curious, inside of NXP, do you guys leverage AI in any shape or form, whether it be chip design or R&D more broadly, the day-to-day operations? And if so, what kind of benefits have you been able to derive?

So we started -- the first area where we started to see kind of AI, as you would call it, in use and actually yielding benefits is in the operations side, very much along the area of identifying defects and wafer scanning, things like that in the manufacturing, test and measurement side of things. I think that's been there for a while and we'll continue to use it. I'd say over the last year to 18 months, I have heard that there's been more discussions with our EDA partners around AI in their tool suites. So Synopsys, Cadence, Mentor using their products and how they implement AI intelligence into their products. And then in terms of our -- just our operations, we have a whole organization that's brought together a multidisciplinary organization to do, what we call, digital transformation. And we are trying to look and see where AI can benefit the operations, but we're taking -- we're not just going to go hog-wild. We're going to look at it pragmatically and look and see where there's a benefit. And so those I'd say those are the areas where we see it today.

In the last couple of minutes, I wanted to give you the opportunity to speak to anything that we may have missed. I know you spent a lot of time with investors and analysts. You're a well-covered company. But anything that we collectively missed or underappreciate about the story?

No, I don't think so, Toshiya. I think we are all and by all, meaning NXP, our peers, we're all in that funny period where I think we can all acknowledge we're in the trough of this most recent cycle. We've all kind of come down at different rates for different reasons. And I think we're all sitting here today looking and saying what is the next move. I know investors had anticipated or hoped for a more V-shaped recovery into the second half. We all would have liked that. I don't think that's in the cards. I don't think it is a long, prolonged valley of death, if you will, but because the industry is cyclical and semiconductors are key to the things we do in our lives, especially the markets that we play in. So we do believe there will be an up cycle. How it starts, where -- what market it starts to show itself, I'd say we don't have that crystal ball, but I'd say that's the most important thing to think about.

Great. Okay. Thank you so much.

Of course. Thank you very much.

Thank you, everyone.

Thanks, everyone.