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

Welcome, everybody. Good to see you guys again. I've actually been coming here eight years. It's the first time I've been in this room, so somehow -- two rooms. I've only been upstairs. Anyway, thanks for coming. Very happy to have with us today the management team of NXP, Kurt Sievers, CEO; and Bill Betz, CFO. So maybe we could just start big picture on the kind of cycle.

You guys have been pretty proactively managing this, taking channel inventory down. And as a result of that, maybe you've had some of the best results in the kind of broad market space that we have. So just talk generally about how you're thinking about this particular cycle and how you've been managing through it.

Yes. Thanks, Joe, and good morning, everybody. Thanks for having us. This cycle, indeed, is probably a little different than past cycles, given that we had a much more massive supply crisis heading into it, that some of the segments in the semi market demand was driven by working from home during COVID, which was kind of different. But most of all, I'd say the price increases, which we have seen in non-commodity markets, is clearly a differentiator of this cycle to past cycles. Now we had clearly made mistakes, like many others, in past cycles. And probably, the biggest mistake we had made in the past was over-shipment into the markets, which we serve either through the channel, through distribution or by shipping too much against the over-ordering behavior of our direct customers. So this time, indeed, we decided and tried to do differently in terms of avoiding over-shipments in the first place. And that has 2 sides. The one side is our distribution business, which is somewhat more than 50% of NXP's total revenue. There, I'd say, about 6 quarters ago, when we were just coming out of the supply shortages, we had a very low channel inventory, sitting at 1.5 to 1.6 months, which is a full month below our long-term target. And then Bill and I and the management team, we decided to not refill the channel to the normal target in anticipation of the downcycle, which would come at some point. And we kept that discipline for the past 6 quarters, which means, indeed, by today, we still sit at 1.5 to 1.6 months of channel inventory, which is 4 weeks below our long-term target. So we can be sure we have not stuffed the channel this time, and we will also not do so until the market picks up momentum. Similarly, but a bit harder to trace and track, on the direct customer side, which is the lower 50% side of the company, where the situation is actually such that because of the NCNR orders, which we had put out and got confirmed in October 2022 for calendar '23, for most of our automotive customers and most of our industrial direct customers, because of these NCNR orders, direct customers came to us early this year in quarter 1, when and where they started to see that probably the obligation of these NCNR orders would drive them into over-inventory. So we got a good feedback loop through this where, eventually, we would go too high. And after some forth and back also here, we clearly decided to not enforce NCNRs with our direct customers because we felt that would not create any additional demand. I mean not a single car more is being built just because we would enforce an order into a Tier 1 customer in automotive. Now of course, we couldn't give it away for free so what we try to do is find commercial alternatives in all cases, which would have been, for example, asking for new design wins, new platform wins awarded to NXP in return for relieving some of these NCNR orders, some price-ups and some washing away of other obligations, which we have in the business relationship between us and them. So in a nutshell, Joe, the story is really one where we try to avoid over-shipments in both channels through which we are serving the market, which makes us believe that we are entering next year with a comparably good inventory position out there. It is definitely very good on the channel. And with the direct customers, I believe we should have worked through over-inventory situations latest by the mid of next year. And through that, our profile is under-shipping demand this year and through that, avoiding that cliff, which some of our peers have actually seen into the fourth quarter or first quarter of next year.

That's a great answer. We've recently shifted to a more optimistic posture on these kind of broad markets that you guys serve. And probably, the biggest pushback we've gotten is on this element of price. You saw price increases during the shortage. There's a kind of a negative view that you have to give some of that back, which it isn't how -- it's historically worked, but we haven't historically had this level of increases either. So maybe you could just talk a little bit more to that and some of the push and pull of, okay, we don't have the shortages anymore, but you also do have customers very focused on reshoring and doing things that are probably going to cost you money that you need to pass through.

Yes. Look, Joe, first of all, it's important to note that our product portfolio is, to the very largest extent, application specific. So it's not commodity catalog-type of product. Secondly, all of the price, indeed, increases, which we've done through the past 3 years, have been following a very clear but also very transparent policy to our customers, which is only -- not more and not less -- only taking the input cost increase, passing it on to our customers such that the gross profit percentage is protected. So we have not padded our margins, but we also didn't want to lose gross profit percentage because of the higher input cost. The result, indeed, just to recap, was that we have increased for the entire NXP in 2021 by 2%. So the price increase for the whole corporation, 2%. In 2022, it was 14%. And this year, once we finish the year -- and we will give you the exact number in Q1 of next year. This year, it will have been somewhere between the 2% and 14%. So absolutely, I confirm, 3 years of strong price increase. For next year, by the way, we also kind of soft guided and gave a preview here. We think next year pricing is going to be flat for NXP over this year, which, again, is just mirroring the input cost situation. We have a slight increase in input cost for next year. And with some productivity, which we can overlay to this, we can go ahead with a flat pricing for next year, which is then again protecting our gross profit percentage. Now the big question which you are asking is, is this ever going to fall back to where it came from back in 2020? My clear answer is no. Because that whole industry has raised prices -- had to raise prices because of the additional supply which we had to organize. Just let's remind ourselves that in the years before 2020, we've had enormous luxury as an industry that for the mature nodes, we could use depreciated factories. That luxury has vanished away because the demand for mature technologies in automotive and industrial has so much gone up that the availability of these depreciated factories, which used to be leading-edge factories in the past -- that availability is not enough. So we have to build additional factories as an industry, including NXP and that just raises the cost, which has to be passed on to customers. Now there is, unfortunately, an additional complication with this, which is further increasing the cost, which is the geopolitical turmoil, which actually forces the whole industry to fragment supply chains. I remember 2 years ago, it started in the U.S. when U.S. customers clearly said the only way you can ship to us and be awarded new design wins is that you have a 100% capability to manufacture the wafers in the U.S.; Europe followed. And since mid of this year, at least I am personally very much aware that now China has exactly the same requirement. So China wants us to manufacture the product for Chinese customers in China. Now given our hybrid manufacturing model, we are in a good position to follow that. So I think it's a big competitive advantage of NXP compared to some of the more IDM-type of competitors, that we can deal with these requirements. Our foundry partners have been very busy doing it. You saw our joint venture with TSMC in Europe. You know that TSMC is building in the U.S. You know GlobalFoundries is expanding in the U.S., in Germany and then in Singapore. TSMC is building in Japan. So I think there's a lot of action to support this trend and this requirement. The only little downside is that that fragmented supply chain takes away the manufacturing scale, which we used to have in the past, which makes it more costly, which is just unfortunately another argument why cost in future is not going to come down, which also means the pricing cannot come down.

That's very helpful. And I wonder if you could hone in on that a little bit. So you see one of your competitors, TI, spending a lot more money in the U.S. to have domestic fabs. And clearly, they're going to emphasize that. But to your point, they don't have fabs in Europe. They don't have fabs in China. What's your ability with hybrid manufacturing to deal with that? And do you work with foundry partners? Do you end up having to spend more on capital yourselves? How do you focus on this?

We -- the definition of our hybrid model is indeed that anything below 90 nanometers, we do with foundry partners. We do not have a 300-millimeter factory ourselves. We also do not plan to have one. And the only thing we manufacture internally is those technologies and processes which are proprietary to NXP, where we own the recipe, where we own the technology. Anything which is standard CMOS, we do with foundries. And I'm an absolute believer that, that is the better model. And the simple reason is the scale, the geographic diversity and flexibility and the access to capacity is just better. And by the way, I think the best proof point was probably the year 2021, which was the worst year of the supply crisis, where we have outgrown everybody, especially the IDM peers. And that was not because we are a better company, it was simply because we had better access to capacity. Because if your own factory is full, you cannot build a new factory in a week. But you can have access to the wider factory network of your partners. So that model is, in my view, anyway, strong. But given the geographic diversification requirements, I think it has become far superior. So to the example which you mentioned, I mean, that's great for the U.S., but what do you do in China? What do you do in Europe? What do you do in Japan? You cannot follow the requirements of local manufacturing, which -- I can at least say that for China, there is this whole angst in the investor base currently about Chinese competitors coming up. Our experience is a different one. The big, big requirement in China is to be able to manufacture locally. That makes us look Chinese, and that is what satisfies them. They want us to manufacture in China. And I'm just glad we can do this because TSMC has the Nanjing factory with 28 and 16 nanometers in China. We have been working for many years with SMIC. And we can tap into these massive investments of the Chinese industry into new wafer fabs, which are coming up also. So therefore, I have a very clear view that this model, at least for the type of company we are -- again, every company has a different portfolio, have different product. But for the type of product and portfolio we have, it's definitely the right model.

I mean that's a great point because I don't know if people think of NXP as a company that benefits from Chinese reshoring. But you actually do have a more flexible model to deal with that. Maybe just because it comes up so frequently, can you talk to the notion of -- I mean, China wafer fab equipment orders are really, really high relative to the size of their production. They're planning on building something, but it doesn't seem like it's centrally coordinated. It seems like its province is competing with one another and things like that. What are they going to build? Are they going to build stuff? Are they going to -- because I would think areas like automotive microcontrollers might be an area that they try to think about. How do you think about that?

Well, anything I can say here is, of course, also anecdotal because it's just based on what I'm hearing and that I was just traveling for 10 days in China, and I've seen some of these factories. I see indeed 2 clusters of capital build-outs. One is clearly in the discrete power space. So that's in the fields of silicon carbide, IGBTs, et cetera, which is very logic because that stuff is not export controlled by the U.S. It is comparably simple in doing it from a technical perspective, and it has massive demand from the local EV industry, so very, very clear. For us, we are indifferent to this because we don't do that type of product. We are not in discrete power. So I'm watching that. And actually, I'm happy because it takes some attention away from the stuff we do. Let them do it. So that's one big part. The other part is, indeed, what we would probably call mature node factories, which is 300-millimeter factories which can do, I'd say, 16- to 19-nanometer type of technologies. So analog mixed signal and logic in that space, which is very much what we need and which is very much also fitting our requirements where -- first of all, I'd say we are happy to see that happening because it facilitates what I said earlier. It helps us to require -- to fulfill the requirement for local manufacturing at a competitive cost. Because I cannot take product from a U.S. or European factory, which is so much more expensive, and try and compete in China. I can only do that with local manufacturing, which is then equal to the start-ups which are doing the same thing in China. So that works. Now the question of overcapacity, Joe, you kind of said it. It's like mushrooming here and there. I would say probably for the whole world, that is becoming a bigger issue. Because this industry has already failed in the past to match demand and supply, where supply was global. Now we have supply at least in 3, probably 4 different supply chains. I see a Japanese supply chain coming up, a Chinese one, a European one and a U.S. one. So I dare to say, Joe, unfortunately, you will be probably right. I think the industry will fail even more in getting capacity and demand, right? But that's what it is. There's just nothing we can do about it.

I mean the idea of Chinese capacity building a fab is 10% of building these businesses up, particularly in higher value-added areas. So maybe we could hone in on automotive, which is obviously one of the exposures that you guys have that's bigger than peers. It's actually interesting. You guys are optimistic, and there's a sense that others are pessimistic. To be honest, everybody is kind of seeing and saying the same stuff with a different spin. But TI has been cautious, but their business has been good. ADI is cautious overall, but they're talking about automotive growth next year. So maybe talk about the automotive market. And specifically, is there a reaction to these shortages that results in an inventory correction down the road in the automotive space?

Well, so there is a couple of things here. The one is the most important underlying trend for growth in automotive the past years, this year and the next years is really content increases. And we see no reason why that would go away. It actually remains strong. And by the way, this is not only EVs. I come back to the EV debate, which we currently have. It is EVs, but it is also ADAS and a lot of other technologies, which are just pervasive and more and more pervasive in automotive. So I'd say fundamentally, there is secular growth drivers, which we think will hold for many more years. Now we have this extreme supply crisis, extreme over-ordering. You might remember, Joe, I absolutely resisted to ever answer in the last 3 years on backlog size. We had a couple of peers pounding their chests all the time about how big their backlog is, which was nonsense, of course, all the time because that backlog was over-ordering. And indeed, some of that over ordering has to be kind of paid back now. Now there is a few other things at work. One is the SAAR this year, has grown far above any expectations, if you ask me. I see a SAAR growth this year of probably 8% to kind of, I think, 88 million, 89 million cars, which is much more than people thought in the beginning of the year. At the same time, the SAAR next year is forecast to be flat. So the macro is actually leading to a flat car production next year, which is, say from a [ slope ] perspective, worse than this year. We had big growth this year on the SAAR. Next year, it's flat. Underneath, however, the real thing is the EV penetration. And the EV penetration, in spite of all the negative sentiment you are hearing lately, it has come -- and it's just factual -- from a, I think, 6% penetration in 2019. So 2019, 6% of the global car production was hybrid and/or battery electric. This year, it's going to be 33%. I mean that's a stunning growth from 6% to 33%. Now we are sitting here in London where it's a bit easier to make this argument. If I say this in the U.S., people just don't get it. Because the U.S. market is actually that 1 market geographically which is far behind that number. But it's also the smallest. So in a way, therefore, there is some confusion in the U.S. about this. Now next year, the debate is, where is it going to go from the 33% to what? Now, I don't know. S&P says 40%. So they say it moves further up from 33% to 40%, which is another steep growth. I don't know. Maybe it moves only to 38% or 37% or 41%. But it will grow, and it will grow on the basis of China. That's really the important element here. It is China and Tesla which is driving this. It is not Ford and GM, which have now delayed programs. It is not Volkswagen, which is failing in the EV market. They all don't matter because it is the Chinese OEMs and Tesla which are driving that growth. So we are optimistic on this. And with that, we see a lot of other systems also penetrating further because we see that EVs are also carrying other tech features to a higher extent. So for us, we have 3 growth drivers. We have the radar business, which actually is not EV specific. It goes across the board, which is a business which we said we would grow from $600 million in '21 to $1.2 billion next year. And we are totally on track to achieve this. So that's a 20% to 25% CAGR over 3 years, working out very well. The second one is our battery management and inverter electronics for the car, which is buff plan. I think we had a 70% CAGR starting from $200 million in '21. We hit the target for '24 already last -- already this year. So it's just going very, very well, which is, of course, a consequence of the EV penetration, which is far better than we had assumed in '21. And finally, the one which probably has the biggest impact long term is what we call our S32 processing platform, which is a whole family of processors for the software-defined vehicle. That is a business which is also doubling from '21 to '24, still on a relatively small level. But for the software-defined vehicle in the future, it is probably the biggest growth driver which NXP will have in the second half of this decade. And it is about offering processing solutions all the way from vehicle computers in 5 nanometers, which we taped out in summer, which is currently sampling -- so that's very high-end microprocessors, over domain computers, zonal computers, all the way to edge nodes. All of them software compatible. And that is the element where we have no competitor in the whole market, which can match our offering with a software-compatible offering of processing solutions, which all fit into the same operating system into the car. The reason why that pulls a lot of growth in the future is the ASP increases. Just think about a vehicle computer in 5-nanometer, which will go into the market in 2026 from NXP, is close to $50 ASP, which is 10x of these microcontrollers, which are sold today. So how we drive -- this is not only to offer and go in an innovation partnership with OEMs for the full system, but is also pushing the envelope on the high end when it comes to ASPs.

That's very helpful. On the S32, are those going to be -- are you going to announce big customer platform types of wins? Or is that something that just evolves a little bit more quietly? Can you talk to that? And who do you see as a competition? I mean, you hear increasingly companies like Qualcomm talking a lot about this that haven't historically moved outside of infotainment. So can you talk about those dynamics?

Yes. So we have actually indirectly announced big wins already. I think last summer, we had a significant announcement about an OEM win in that world. We, very seldom, can speak about customer names. But it's important to understand these big wins are with OEMs. They are not with Tier 1s. It's a completely different design and model. It goes top down from the OEM, and the Tier 1s are being forced to use these solutions going forward because of compliance with the software compatibility. From a size perspective, Joe, this is just like the backlog numbers. We will not use design win sizes because I found out that every peer has a different definition. I mean is it 5 years run rate? Is it 10 years? Is it 15 years? Those wins are typically 10-plus years. They're huge. And what I can tell you is that this year and last year in automotive, we have had, by far, the biggest historic design win achievement in this company ever. I mean -- and not just like 10% more than the years before, way more, which is because of these wins, but it's also the type how it's been won. Again, in the past, we did win piecemeal with Tier 1s. Now we win at the OEM the whole thing, which is just a different way of doing it. Competitors, no, we are never alone, Joe. That's -- this is not a smartphone business where it's a one-hits-all kind of platform. We currently are the #1 in automotive processing. So there's -- any market research confirms that we are with about 30%, the largest supplier of processors, which is microcontrollers and microprocessors, into the car. I think we have every opportunity now to expand that share, but it will not be 80% or so. I mean from 30%, I don't want to hit a number, but it's going to go up. We do see competition on the very, very high end in IVI, indeed, from Qualcomm and in some central compute for autonomous cars, which we still believe, are many years out in any reasonable volume from NVIDIA. So there is an NVIDIA and Qualcomm game out there, which doesn't touch us. It's actually complementary to what we do. So we don't try to compete there. It's kind of next to each other, which is fine. And of course, we have from the traditional microcontroller suppliers in the mid- to lower end, some competition. I mean, the likes of Infineon, like being limited to microcontrollers. They don't do microprocessors, so they are limited to the lower end of this [ whole discussion ]. They are great companies. Renesas has, in principle, a similar portfolio to us. ST is also there. So it's not like we are alone. But this approach of a large, broad portfolio of software-compatible solutions, which we work in close hand-in-hand partnership with OEMs, I think this is where we have, by far, the best offering so far. But I mean, others see this. Others will try to follow. It's just that I think we are creating more and more distance to the traditional MCU suppliers. They win in MCUs, but we'd rather go to $50 MPUs.

Great. So maybe touch on the financials a little bit and see if there's a question in the audience if there's time. But Bill, maybe you could talk to gross margin. You're exceeding now the targets from the last Analyst Day. Can you just talk to what are the gross margin drivers going forward? And do you need to think about a higher margin over the long term than what you've been talking about?

Sure. Thank you, Joe. Great to be here. Gross margin, as we said and guided toward next year, will stay at the high end of the model. And this really has to do with a couple of things. One, you heard Kurt talk about our company-specific growth drivers, and these carry higher margins. We're doing quite well in those areas. Another factor we talked about is inventory, controlling of inventory, staging it appropriately. And in the past, we would have probably followed our peers over shipped. We would have peaked that gross margin and then it came down similar to our peers. And we learned a lesson, a very hard lesson back in 2016. And so we changed that approach, tapped the brakes internally and really kept inventory in check, both in the distribution, as Kurt talked about -- we're at record low levels -- as well as we start to tap the brakes with our direct customers over -- starting about in Q1 of this past year. So really doing a nice job there. And we want to -- this is all about navigating and having that soft landing through a very difficult market that you can never predict when you come out with your Analyst Day targets. And then another factor is the operational performance. In the past, we would jerk our factories around from high 90s going down to the low 50s, trying to correct inventory too quick. And as you know, Joe, we basically started to take our internal factories that represent about 40%, much lower than it used to be. We used to run probably 70% in-house, now about 40%. And this is a big factor to do with your fixed cost. So today, our fixed costs represent about 30%. And when I joined the company 10 years ago, it was 70% fixed. So clearly, a much more variable model, which is beneficial, which reduces the swings in gross margin. But going forward, looking into the short and medium term, why we feel confident of holding toward that high end of the model into next year despite what the macro has to throw at us, is driven by the fact that we have utilizations that are low, mid-70s. Eventually, that will come back and become a tailwind. Another one is holding. As you all know, our distribution product has accretive margin compared to the corporate average. So eventually, when we restock this $500 million worth of revenue, that is -- becomes a tailwind from a timing perspective. And then more longer term, I would say there's 2 other factors. I think the team at NXP is we want to penetrate more customers in that long tail for distribution. And so we've kicked off a lot of work over the last year or so by creating additional design wins in that area that will yield a benefit for us. And then finally, as you mentioned, is our new product introductions. Again, the investments we made 3 or 4 years ago, that revenue is being -- we're delivering that revenue today, which had a hurdle rate back then of 55%. Same thing, in 3 or 4 years from now, the hurdle rate of -- all our investments today have the corporate average today. So therefore, they have to yield a higher margin into the future. It just takes time with the 2 type of markets that we're targeting, both industrial and automotive.

And just adding to what Bill says, this is really the reason why we try to navigate the soft landing, which I explained in the beginning of not over-shipping. Because the biggest drama which we had seen in the past that the gross margin actually dipped so deep when we had to reduce revenue because of the inventory digestion. And that performance to the high end of the gross margin funnel is the most desired outcome of what we do. So we would really say, if we can stick around this 58% next year, while correcting, then we had it right this time. I mean that's the target. And I know we have 4 quarters to go to prove that. But that is really the thesis. So what Bill explained here for the short term is directly correlated to this strategy I explained in the beginning. And on the long run, honestly, no reason why we have to stick to the 58%. I mean there is a couple of good arguments why it could actually edge higher. We will do a next Capital Markets Day at the end of next year, and we'll update the model then.

Great. Well, unfortunately, we're out of time. I got about half my questions in. Thanks so much for all your helpfulness and transparency. Thanks, guys.

Thank you.

Thank you.