Okay, good morning, welcome to the third day of JP Morgan's 54th Annual Technology, Media, and Communications Conference. My name is Harlan Sur. I'm the Semiconductor and Semiconductor Capital Equipment Analyst for the firm. Very pleased to have the NXP Semiconductors team here with us today. Bill Betz, Executive Vice President, Chief Financial Officer, Jeff Palmer, Senior Vice President of Investor Relations here with us this morning. Gentlemen, thank you for joining us this morning.
Thank you, Harlan, and thank you, JP Morgan, and of course, all our support and ownership in the room. Great to be here.
Yeah, absolutely. Let's start off with, you had your earnings call two weeks ago. Rewinding back to last year, the team and customers were still working through demand inventory dynamics in the first half of the year. After six quarters of year-over-year declines, you drove the positive inflection in the second half of last year as demand began to accelerate. Those strong trends continued into this year. Maybe just, this is a good starting point. Just talk about the industry environment and what you're seeing right now.
Yeah, let me take that, Jeff, and first start off with the signals that we can see. Internally, the demand signals remain strong. They continue to gain momentum. This shows up in the metrics that we measure, such as book-to-bill ratios or our backlogs for next quarter, the quarter after next quarter, our distribution backlogs continue to be strong and to improve. Customer escalations continue to increase. We feel pretty good across all the signals, specifically with our company-specific growth drivers.
As we mentioned a couple of weeks ago, those are intact and doing quite well, which we feel very confident, a conviction on the growth targets and the growth of more than double-digit growth for this year and next year gives us confidence there, Harlan.
Absolutely. First phase of the cyclical recovery, shipments rising to the level of consumption. Second phase is customers feeling confident on sustainability of their demand profiles and then starting to restock inventories, right? CEO Rafael said back in March that the team was not seeing signs of broad-based restocking. It's not baked into any of your forward outlooks. Some of your peers more recently have talked about seeing signs of restocking at disti customers. Harder to tell with direct, right? Is the team now seeing any signs of broad-based restocking?
At this time, no. The way to think about our business, as you know, is close more than half the business is through distribution. In distribution, we have a really good handle. We're finally back where we want to be at 11 weeks to support the strong demand that we're seeing. The other half is linked to automotive. That distribution split is probably 60% direct, 40% distribution inside auto. What we see with our Tier 1s, working capital for them are quite tight.
Yeah.
Their margins are kind of thin. They're going through in an inflationary environment, just like many companies are. We don't see pull aheads. We're not seeing anything that have them doing restocking at this point in time. No, not at this time.
Okay, that's a little bit of somewhat of a tailwind that's kind of still in front of the team when it does happen, potentially.
Yeah. Absolutely. Along with, obviously, on the top line, with the inflationary costs improving, we expect more of a price increases to offset some of those higher input costs, and that's more of a second half for NXP.
Actually, that was going to be my next question. On the pricing environment, the team came into this year with a view of low single-digit percentage type price declines in line with historical trends. Given the tight supply environment, as you just mentioned, accelerating end demand profile from your customers, we have seen some reports that competitors and NXP are raising prices more broadly in the second half of the year. It was interesting, right? At earnings, Rafael said, the team is selectively making small pricing adjustments on rising cost pressures, but nothing material in Q2. Is the team raising prices more broadly as you look at the second-half profile?
Yeah. First with pricing, obviously, the inflationary costs are out there. The first line of defense that we try to do for our customers is through productivity, operational efficiency. After that, if we still have a headwind.
Yes
We kicked off first with our distribution channel. We're working through that this quarter on selective areas, as Rafael shared several weeks ago. If it continues, then unfortunately, we'll have to do more of this with other customers, maybe broadening it. I wouldn't say we're at the level of during COVID when it was kind of broad-based. I think it's still a selective area at this time.
I see. Would it be fair to assume that with a lot of these discussions, and if some of these discussions hold, that more of the selective or broad, however it plays out, will be probably more of a second-half dynamic for the NXP team?
Yes, that's the way to think about it.
Yeah. Okay. In the January earnings call, the team was a bit reluctant to articulate the 2027 revenue and earnings profile relative to your analyst day targets. 90 days later, on the April earnings call, you were very definitive on driving, as you mentioned this morning, double-digits growth this year and double-digits growth next year, hitting your 2027 revenue margin, gross margin, operating margin targets. What was that big inflection over the past 90 days up until the last earnings season that gave the team the confidence on calling out that strong growth profile through 2027?
Yeah. I'm not sure if there was a trigger event between Q1 and Q2. I really think if you look back during Investor Day, we were convinced that our company's specific growth drivers will grow. They grew nicely last year, but it was disguised by the inventory digestion for primarily auto in our core part of the business. In Q4, they started to shine through more in the Q1 results, and now what we can see, and it gives us more confidence, is that all our company growth drivers are actually at the high end of the growth model or above.
Yes.
The way to think about our company-specific growth drivers in Auto, call them in the mid-to-high 40s. We're working to get half of that portfolio at 50%. Once at 50%, think about growing somewhere between 15%-20% of the year. On the Industrial IoT, I think we shared we're about 1/3, about 35%, growing at 30% clip, going to 40% clip. This is really about design wins that have occurred several years ago. They are now just ramping, very company specific, and they will continue to ramp. We feel very confident. We see that in all the signals. When you're ramping new products, obviously you have to be prepared for them.
Right.
You get more signals than just versus the broad base.
For the first time, and we'll go back to some of the growth driver dynamics in your various segments, but for the first time, the NXP team actually broke out your data center revenues. It was $200 million last year, expected to more than double to $500 million this year. 50% of the mix is Industrial IoT, 50% comms infrastructure, targeting applications and systems cooling, power supply, board management, control plane switching applications. You're attacking much of this with your i.MX processor family for board management, your MCU families for security and cooling, your digital networking processors. Given this new segment of growth relative to your Analyst Day, how should we think about the three-year revenue growth profile of this particular new end market?
Yeah. It's embedded, as we said, both in industrial IoT and our comm infrastructure. As you said, the networking side is in the comm infra and more of the board base on our i.MX and MCX families are on industrial IoT. The way to think about it, and it's coming through our numbers.
As we sat back and we're like, "Ooh, we now have to explain this," and it's getting sizable.
Yeah.
That's what you saw us do in this last earnings, is talk about it. I think the way to think about it post this year at this point in time, we're comfortable saying we'll grow better than the SAM of this market. I think Rafael shared that. As we ramp through it, we'll give regular updates on the opportunities of our data center exposure to you all on a quarterly basis.
As the data center opportunity, whether that's being driven by these agentic workloads or training workloads as they continue to drive a strong growth profile, it's pretty clear that compute is a key part of all of this. We all tend to think about XPUs and GPUs, but there's a lot of other compute that supports these architectures. A big part of it is switching and control plane, which is an area where NXP actually has quite a bit of leadership. Switching and control plane, that's your digital networking franchise. It's great to see the team leverage its historical leadership in compute. You've historically had a portfolio of PowerPC, your QorIQ processor family, your Arm-based Layerscape processor family. What processor families are you winning with now in data center?
Given this sort of refocusing on control plane and switching for these AI compute workloads, is the team planning on re-architecting a new data center portfolio or processor family?
Yeah, let me first start off on the networking side, it's really linked to our Layerscape platform of products that exist in the portfolio. Networking, as you mentioned, we reuse in other parts and very important for high-performance compute in our products and solutions. Our exposure there is to several hyperscalers and several OEMs. That's more of a direct.
Right.
If you think about the board compute and the cooling and so forth, think about multiple more customers in that case.
Yes.
We're supporting that both direct and through our distribution channel across the board. It is an area that we have. We have a broad portfolio, and there's opportunities to continue to invest in those areas, and we plan to do so and to be competitive in that space in the control plane, not the data plane.
Yeah. We're looking forward to seeing some of those new product announcements. Let's turn to your flagship industrial and IoT business. I'm going to buck the trend. Most people like to start off with automotive. I'm going to start off with industrial and IoT. Partially because we do like your embedded business, and we often feel like your leadership in compute is a little bit underappreciated. The other interesting thing is Rafael, who's been on board, what, for one year now as CEO, a big part of his upbringing at NXP was embedded. I believe that, again, the market underappreciates NXP's leadership in the embedded space. You're number 2 global share leader in the microcontroller market. You're number 2 global share leader in auto and industrial MCUs. You're a top 10 supplier of embedded processors, right?
You've got a three-year target growth CAGR of 8%-12%, with the accelerated growth drivers in IIoT driving 20%-30% CAGR. Right? One of the fastest-growing accelerator product categories that you guys laid out at Analyst Day, right? Looking more recently, Q1 IIoT was up 24% year-over-year, expected to be up high 30s here in the June quarter. Again, well above your 8%-12% CAGR. How much of this outsized growth is sort of the core industrial business cyclical recovery? How much is due to the accelerated growth drivers? And does the focus on physical AI bringing more intelligence to all of your focus areas actually drive a higher CAGR than what you projected at the Analyst Day?
No, it's a really great question. Let me try to parse out our industrial IoT business. First, step back, about 60% of it's industrial, 40% of it is IoT. If we break it out between our company-specific growth drivers that are growing even better than the 20%-30% that you called out.
That's approximately in Q1 about 35% of the portfolio. We believe over time, we'll get that into the 40s. The mature or core or broad-based part of the portfolio, that grew in Q1 at about a 15% clip.
related to it.
Got it. One of your strategies has been to expand your presence in the distribution channel focused on broad market and mass market segments. Right? It's this long-tail, small to medium-sized customers. More focus here will definitely help to drive a higher gross margin profile, right? The team has been adding more to its family of general-purpose, high-performance microcontrollers, like your MCX family is a good example, but also focusing on systems-level solutions. Any way to quantify the systems-level strategy in terms of pipeline expansion, dollar content uplift per new customer opportunity? Just a basic question, how large is your long-tail mass market business today?
Yeah. If we step back, industrial IoT, approximately 80% of that business goes through the channel servicing the tens of thousands of customers. As I think I mentioned in the past, gross margins with the low-volume type of customers, the margins are greater than shipping large volumes to direct customers. That area is a focus. That go-to-market is very focused for us. It's very fragmented. The system play is very important to us, and really, the next real leg of growth is that physical AI on the edge. Right. That's where we want to be a leader in. We're setting ourselves to be that, and we're really, if I use the baseball analogy, we're in spring training. We haven't even gotten to the game with AI on the edge in that space.
I would think of that type of business as margin accretive because it's fragmented, low-volume customers, higher margin.
That's right. Yeah, exactly. That's a good segue into my next question, as you focused on edge AI and physical AI, right? Because not a lot of people know this, but the NXP team was one of the first to adopt and integrate your own organically designed, we call them neural processing engines, NPU cores. These are purpose-built cores to accelerate machine learning and AI applications. Right? NXP was one of the first to embed these NPU cores into your microcontroller and microprocessor families way before GenAI, right? Before your recent Kinara acquisition. With Kinara on board, they have a standalone NPU architecture. This is purpose-built for more high-performance compute applications. Combined with your organic sort of NPU, what's the overall traction been like for edge and endpoint-based compute applications with Kinara, with your embedded NPU, MCUs, NPUs?
Is the team planning on integrating, at some point, Kinara's high-performance NPU accelerator into your processor and MCU product lines?
Sure. It's part of our strategy. AI on the edge is super important. We believe that market hasn't even played out, as I mentioned earlier. The way to think about it and the way I measure it with Rafael and many others in the company is the penetration. We call it AI enablement, that capability that we're putting in our products in specifically Industrial IoT. Eventually, it'll be auto as well.
Right.
If we look at industrial IoT, I think Rafael has shared a metric where in 2025, total industrial IoT probably had AI enablement about 5%.
Yeah.
4% or 5%. We expect that AI enablement in 2026, with the growth of Industrial IoT, be approximately about 12%. You're right, that doesn't even include Kinara-
Right
which that revenue doesn't come to be until 2027. How we're measuring Kinara, since we've acquired it over the last year, is really, if you think about it's all about engagements, which then turn into design wins and then converts into revenue. The engagements, I think we shared two quarters ago, we had a funnel of $750 million. We just surpassed $1 billion of that funnel as you work through these proof of concepts with your customers. It's a huge long tail. Obviously you focus on the ones that are really going to move the needle direct.
Right.
We also train our distribution partners to go work on the ones from a system and a solution standpoint on that long tail. It's very early innings there, but again, that pipeline, the engagement, the excitement, the interaction with our customers are quite good at this time.
Perfect. Before I move over to your automotive franchise, let's see if there's any questions from the audience. If you do have any questions, raise your hands, and we'll get a mic over to you relatively quickly. Any questions from the audience? Let's move over to your automotive franchise. 8%-12% three-year CAGR on target growth. You drove 10% year-over-year growth, MEMS divestiture in the first quarter. Your guidance implies high teens percentage year-over-year growth this quarter, suggesting acceleration. Last year, your accelerated growth drivers, S32, software-defined vehicle, radar, electrification, connectivity, drove 10% growth when your overall business was flat. Are the accelerator growth drivers in the first half of this year driving within or above that 20%-30% sort of growth target?
Yeah, the way to think about total auto, approximately today, about 45%-46% of that business is linked to those four growth drivers that you just shared. Last year, they were all growing, except for BMS was kind of sideways. That is back now to growing. I would say, as I look at the first half of this year, of those accelerator growth drivers, they are all at or above-.
the target that we shared out during Investor Day, which is obviously double-digit growth, and as we continue to mix up the portfolio and drive that to about 50% of the automotive business.
The team continues to drive very good traction with its software-defined vehicle, SDV portfolio of system-level products, strong software development platform, target of hitting $2 billion in sales by 2027 or up 2x versus 2024, right? There's several ways to implement a software-defined vehicle, right? You have central compute, leveraging your flagship sort of S32 and 5 nanometer processor. You can either implement that with a hybrid sort of domain/zonal based architecture or moving to a full-blown zonal architecture, right? Again, this is going to be moving to a full-blown zonal architecture leverages your 5 nanometer processor, but it also leverages your S32K sort of zonal MCU microcontroller family of products, which is the ideal architecture because it reduces wiring, it reduces weight, it's easy software and over-the-air updates, right?
What's the revenue outlook this year, first of all, for your S32 SDV platform? If you look at your forward pipeline, what's kind of that mix of hybrid versus full zonal implementation?
Yeah. If you step back, the software-defined vehicle, again, going back to a baseball analogy, we're in the early innings, call it inning 1, inning 2 of the real growth. What we're servicing today is the software-defined vehicle with the domains, the different type of domain infrastructures that we have. As you point out, as you look ahead, there's two extremes. There's extreme where you go all zonal, there's another extreme that you go central compute, there's many variations in between. What's the beauty about NXP is we're setting ourselves up with a total platform to support all architectures. As you mentioned, our flagship, S32N is on the central compute side of the house, another flagship that is going sampling, we hope to go to production in 2027, is our S32K5, which is the most superior product out there.
We believe that'll get more traction first into China, because China typically is going with the zonal architecture at this point in time. That's their next leg in the architecture journey.
The Westerns seem to be more going toward the central compute. We have examples of both, and there's some mix in between, of course. We're super excited. We're on target. As I mentioned, the software-defined vehicle is $1 billion, going to $2 billion. That is tracking or tracking ahead. The real growth, with the acquisition of TTTech, which accelerates our CoreRide platform, really doesn't take effect until 2027 and 2028 when we start getting meaningful revenue related to that transaction as well, which brought in a lot of software expertise and move us up the stack for supplying and the value that we provide to our customers.
That's actually a good point because it's one thing when you're developing these platform solutions, it's one thing to have the right portfolio of silicon compute. If you don't have the software, the firmware, it makes it extremely difficult, right? Maybe help us understand TTTech Auto and just the NXP teams, you call it your CoreRide platform, right? That's your integrated full platform approach. Where TTTech Auto fits into, because the software stack is very complicated. There's so many layers. The customer only cares about that top layer, which is the application layer that they can customize, help us fill in the gaps below that. How has NXP filled out that entire stack, make it easier for their customers to implement their silicon, put their customization on the core application software layer, bring it to market?
That's the idea of the CoreRide platform. Is that what is driving the leadership in this segment of the market? Is the platform and maybe less so on the silicon and some of the other building blocks?
It's the platform, it's the knowledge, it's the expertise. The biggest takeaway now in the future, you win with the engineers, you win with the OEMs, you don't win with the Tier 1s. As you get ingrained in the stickiness and you're working as a partner to develop and get ahead, and have the software ahead of the hardware.
There is a competitive advantage to that, and that's the things that our customers like. It's a partnership both ways, and once you're in, obviously, the barriers of entry are quite hard.
Yeah.
The switching costs are quite expensive, as simple terms I can put it. Jeff, maybe if you want to share a little bit more technical on TTTech itself.
Yeah. One of the things I would add also, Harlan, you alluded to this earlier, if you step back for a second and look at NXP, not from end markets, but products that we offer to market, almost 60% of our revenue is processor based.
Yes.
That's the first thing to remember. The days of being able to just walk into a customer just with a piece of silicon and be successful are in the past. You have to provide software enablement tools.
That's right.
You have to provide firmware, and we're finding more and more that the further up the stack we can go, the more sticky and more interested customers become. The TTTech acquisition actually was a middleware product called MotionWise. That is the layer that sits right below the application layer that the OEMs write to. We're silicon guys at the bottom.
That's right.
We work with a number of partners for that next layer, that OS layer above it, and then we have TTTech above that for middleware.
Yeah.
Our vision long term is to constantly push further and further up the stack. We won't become application providers to the OEMs.
Right.
They'll develop that themselves. We don't envision ourselves taking a toll, if you will, from our customers. Some of our partners have talked about that.
Yeah.
Yeah, I better say this. TTTech, obviously more marketing guys, we have renamed it to TrustMotion.
Mm-hmm. Yeah.
I think we had an announcement on that as well recently.
Sure. Name change.
Name change, but you all know what we acquired, but the branding of it is called TrustMotion.
Yeah. No, it's good to see NXP moving in the direction. What we've been saying for a while is that great semiconductor companies don't just design chips. Great semiconductor companies bring system-level solutions and a strong focus on software and firmware. This is pretty apparent as you look across NXP's entire portfolio. Good to see the team continuing to drive that strategy going forward. Before I move on to the manufacturing and financials, are there any questions from the audience? I just want to make sure we address any investor questions. If you do, raise your hand, we'll get a mic over to you. If not, let's talk about the manufacturing and capacity outlook. On the 200 millimeter consolidation plans, 300 millimeter expansion initiatives, your objective is to drive from 38% of your wafer capacity internal mix in 2024 to 20% in 2030.
The primary initiatives to achieve this are going to be the ESMC and VSMC 300 millimeter joint ventures. I believe you're targeting first advanced node wafers out of VSMC next year, first analog, power, mixed signal products out of VSMC in second half of next year. VSMC alone will drive an incremental 200 basis points of gross margin in 2028. Can you just give us an update on the build outs?
Yeah, no, everything's going really well with VSMC. I believe, like you said, it'll start to ramp in 2027. Then in 2028, how does that ramp go? We're qualifying products today. Clearly, if it ramps better, we'll get more of the 200 basis points. Do we get a partial of it? We'll have to see how that plays out with any factory that you ramp. The other good news about, and I think our partner, Vanguard, released this a couple of weeks ago, is that also that facility combined is going to be cheaper. I think it's coming in over $1 billion less than originally thought. They're doing a really nice job on the construction build, the tooling, and so forth.
Also from a cost standpoint, for something that size coming, typically, you don't have large projects that come in under budget. This one has. Typically, they're over budget. We feel really good about what's coming to be there.
VSMC, which is going to be more focused on analog, power, maybe some of your older generation MCU products, that's a very different financial profile than your average foundry engagement and versus internal, which is 200 millimeters. Have you quantified 300 millimeter VSMC cost profile advantage either relative to foundry or relative to your internal 200 millimeter capabilities?
Oh, absolutely. Today, obviously, we buy from the outside on 300 millimeter, but that's at a market price.
Right.
The beauty about VSMC, it's a cost-plus type of model. Since we're owners of it. That's part of the 200 basis points improvement.
Yes.
You take the geometry scale part of it.
Right.
You take in, obviously, we want a little bit more of that capacity up front. That's why we're paying these capacity access fees to help us to get more of that supply. That technology is TSMC proprietary technology.
Right.
You know our two largest foundries, our GlobalFoundries, as well as TSMC, great partners of ours. It's very common to, even in some cases, have dual source between TSMC and VSMC as well. Many of our customers want dual source, so it's an advantage that way. That's all contemplated into our 200 basis points improvement once it's fully ramped.
Often overlooked, I think a very strong competitive weapon for NXP is your internal assembly and test, right? It's 80% of your backend mix. The team was targeting to build out your second facility, right, ATKL in Kuala Lumpur, Malaysia. Total output, I think 6 million units per week. Has this facility come online? How does the team's internal backend cost compare to some of the big Asia-based OSATs?
Oh, yeah. Actually, I was in Malaysia last week reviewing the plans.
what we call ATKL2. That will break ground in Q2. I would say that will be up and running in line with 2028, linked to our VSMC plans. That's all on schedule, and very automated in advance. That's where I spent my time last week, making sure it's all a go, everything. Look forward to that supporting. You're right, from a competitive standpoint, you get to control the inventory, so you don't have to build all the way out to finished goods, while if you use OSATs, you have to buy finished goods.
Right.
Sometimes that becomes an obsolescence risk for us. We don't build it out until we actually have the physical order. That helps us there. Also from a cost standpoint, clearly you can do a lot more from controlling your inventory and timing it for your customers the appropriate way throughout the quarter when shipping your product. All a good go there on ATKL2, we call it.
Perfect. Good to see the execution there. On the financials, team has a long-term target of 60% gross margins. You just guided June quarter 58%, with the biggest lever to 60% from where we are today being utilizations, right? Every $1 billion of annualized revenues drives 100 basis points of gross margins. You will be in the mid-80s utilizations in the second half versus low 80s here in the first half. Mix is the other driver. This is where I kind of want to dig into, because if you look at your pipeline, we just talked about data center, we talked about software-defined vehicle, we talked about mass market, broad market, right? If you look at your pipeline of design wins that are starting to ramp now and over the next several years, what type of gross margin profile do these products have?
As you know, probably about four or five years ago, our entitlement model to our R&D spend probably back then was anything we're investing in had to drive something at 55 or higher. Think about those type of products now just ramping, coming to production, will have that kind of 55 or higher. Obviously higher, some may be a little lower that you work through process to get them where they need to be. In general, in that ballpark. Anything that we invest today has an entitlement model of 60. Think three to five years.
Yes
you overlay those type of investments.
Those type of products, and that's why we continue to move up the stack and why we really are doubled down in our investments and really excited about two things I want you to take away today, is really the growth in SDVs, early innings, and physical AI, which we're still in spring training.
That's right.
There's so many opportunities, and we feel very confident with the pipeline, the first engagements that you can see that there's a lot of traction. Now it's about execution.
Right.
That's what we talk about constantly when we get together as an MT. We spend our time talking about the next five to 10 years, and I think you'll hear more about NXP, specifically AI on the edge.
Great insights. Really appreciate the participation today. Thanks, Bill. Thanks, Jeff. Really appreciate it.
Thank you all for the support.
Thank you.