Good. All right. Thank you, everybody. Welcome back. I'm Joe Moore, Morgan Stanley Semiconductor Research. Very happy, oh, sorry. Very happy to have with us today, Jeff Palmer, Director of Investor Relations for NXP.
Right.
So, Jeff, I think maybe if you could just start off with a little bit of an overview of the company's priorities and in the year ahead, and then we'll go into direct Q&A.
Yeah, I, I think... Ooh, this is a little loud. Excuse me. I think for the year ahead, the way we're looking at is how we navigate what clearly looks like a troughing of our, our end markets-
... in a very controlled manner. You know, over the last couple of years, we've managed our distribution channel very, very tightly, very cautiously, and over the last couple of quarters, we started to see a little bit of maybe excess inventory at some of the Tier 1s in automotive. And so we wanna navigate through that and really navigate a soft landing into the first half.
Okay, great. You know, let's see if I can speak more quietly. I'm scaring myself. Can you talk to the gross margin and operating margin trajectory of the company? You've executed really well there. You're ahead of the longer term targets that you've set. You know, how are you thinking about that going forward?
Yeah, I, I think the, the thing people have to realize is, over the last decade or so, we have flipped the fixed versus variable cost structure of the company. So when I started with the company 13 years ago, our fixed costs were about 70% and our variable costs were 30%. We have flipped that 180 degrees now.
And that goes hand in hand with the fact that we've increased the amount of wafers we buy from foundries. We're now buying 60% of all of our wafers from foundries. That's all variable costs. We run our own back ends, where we own about 85% of our capacity. So that's the first thing, is that the cost structure has changed. Secondly, for those of you who have followed the company for a number of years, if you'll remember back kinda 2019 and before, we struggled to get to about 55% gross margin, and that was 'cause we couldn't get the revenue over $10 billion. The business was kinda structured and organized to at least pour $10 billion of revenue through it. Once you get above that, you get great flow to fall through.
Great. And can you talk a little bit, and I'll get into some of the end markets, but the role of pricing. You know, one of the things I found really helpful is you guys give good data around this, and you talk about what like-for-like pricing has done, and that's a key debate now for the sort of analog microcontroller businesses. Going forward, can you talk to what you're seeing on the pricing side?
So the reason we raised prices over the last couple of years, and maybe just to level set everyone, in 2021, prices were up 2%, in 2022, they were up 14%, and then just this last year, in 2023, they were up 8%. The reason prices went up is our input costs went up, and so we made a decision that we would only pass along the gross step amount of our input cost to our customers, so we didn't pad our margins. So our current thinking is that maybe we get into 2025, we might be at a place where we may be able to offer some of our larger customers, low single digit like-for-like pricing, but off the elevated base.
We don't agree with the thinking that pricing has got to revert back to where it was in 2019.
We just don't see any reason why that would occur.
Okay, that's very helpful. And then last big-picture question. You know, manufacturing footprint is gonna be a debate here. TI, you know, very aggressively investing in capacity in the U.S. As you said, you're more of a variable cost structure, but you've also taken, with your own fabs, a very geographically diversified planning process, where you have manufacturing in each region. Can you just talk to that and talk to how much your customers prioritize that?
Sure. So, we would call our model a, a hybrid model, not a fab-lite model. So we build about 40% of all of our wafers in-house. We have about four fully owned eight inch factories, one in the Netherlands, two in Austin, Texas, one in Arizona, and then we have a joint venture with TSMC in Singapore, which we're the 62% owner, so it consolidates into our financials. All these factories are eight inch. They run proprietary mixed-signal products, all above 90 nanometer. We pushed all bulk CMOS out of our factories over the last couple of years, and it's all in the foundry network.
We would say that our own factories give us a very distributed, geographically distributed base, and then we're partnered with people like TSMC, who we have a joint venture in Dresden with. We are gonna use their factory in Arizona for their 5-nanometer process, and then we build in Taiwan and other parts of the world. Then there's GlobalFoundries, who also is very geographically diverse. We think that our model works well for us.
Okay, great. Thank you. So maybe looking at some of the end markets, starting with automotive, you know, which is your biggest market, you know, you're looking at a year-on-year decline for the first time in March. You know, and you talked about a soft landing. How, how do you feel like you navigate that, and, and can you elaborate on, you know, how, how you guys have... I feel like you guys have managed this very tightly. You saw weakness maybe when others didn't.
Yeah.
Your peak to trough is probably a little different than the company.
Yeah, we would agree. I think our peak to trough is much different than some of our peers. You have to remember that about just over 50% of our business goes through the channel, which we manage completely.
We manage what gets shipped in, we manage what inventory is there, and we get full visibility to where it goes to from a sales out perspective. We get that data daily. We review it weekly as a management team. So I think we've managed our channel very well. The other 45% or so of the business, which is direct and is very much automotive and large industrial customers, we started to sense in Q2, if you go back and read our transcripts, that there was some inventory building at some European Tier 1s-
... that then expanded to some North American Tier 1s, and we do believe that the Tier 1 auto Tier 1s are a bit over inventoried. We think it'll take us through around the first half of 2024 to enable them to burn off that inventory. And you might say, "Well, how do we enable that?" Well, we had non-cancellable, non-returnable orders with these Tier 1s. You know, they contacted us and said, "Look, if you force us to take this inventory, we're just gonna put it on the shelf.
And we made the decision that wasn't gonna help anybody. So we found business accommodations with those customers to allow them to—the kind of breathing room, if you will, to burn off that excess inventory, and we felt that's the right thing to do. That'll give us a little less growth in the short term. You saw that in second half of 2023. As you mentioned, we've had a down quarter here in Q1.
Yeah.
We think that's the right thing. It sets us up very cleanly for what we believe is the beginning, kind of, a next cycle of our industry sometime in the second half of the year.
And isn't there an argument, I mean, given how severe the automotive shortages were a couple of years ago, that your customers would continue to operate in a higher inventory, you know, having a safety stock? You know, and do you think that's the case, and we're seeing a burn-off relative to that, or just how should we think about that?
I think you're right, Joe. I think the OEMs continue to remember the pain they went through for a couple of years, and I think they, there is a lot of tension between the OEMs and the Tier 1s as to how much inventory should actually be held by the Tier 1s. But there's not a lot of levers the OEMs have on existing running business.
Okay
... to tell the Tier 1s, "You must hold this much inventory.
Okay.
There is some concern that if the Tier 1s go too far, correct too far to burning off inventory, they're just gonna sow the seeds for another shortage situation in the out years.
Yeah. I mean, the math seems fairly straightforward, that internal combustion vehicle, maybe $400 worth of components, most of which is long-tail stuff that you just said is staying, the price is staying the same. It seems like the math would say, "Hold more inventory to avoid these shortage situations.
We would agree with that.
Yeah.
I think the challenge is a lot of the Tier 1s operate on very tight working capital metrics, and they, you know, their challenges were, they were unable or couldn't renegotiate the business conditions with the OEMs directly.
Great. And one of the drivers that really matters for you guys, you know, not just idiosyncratically, but for the overall automotive ecosystem, is the migration to EVs
I feel like sentiment has shifted a little bit more negatively from investors on EVs. Our auto team is really negative on EVs, just for what it's worth. We're trying to parse all that out. Can you talk about how you think about that? Is this a movement that just continues, you know, maybe not linearly, but, you know, that we're gonna move away from internal combustion towards EVs over time?
I think long-term, EVs is the future, and various types of EVs, whether it's plug-in hybrids or full electrics. I think the angst about the transition to EV is very much a U.S. Wall Street, fixation, if you ask me.
If we look at how the MCU business is trending in China and other parts of the world, it's continuing to grow. We continue to see 2024 being a year where EVs will make up maybe about 40% of the total global SAAR, and that's still 25% year-on-year growth.
It's slowing down a little bit, but it's not going negative. And the thing that was very interesting for us when we laid out our growth drivers in 2021 was we had a view of what EVs would do for us from a powertrain perspective, primarily battery management systems, inverter control.
Yeah.
What we didn't fully contemplate was the pull-through on the rest of the portfolio: more safety features, more, you know, driver comfort features, and just higher electronic content that EVs are pulling along, and so that's been a very nice positive trend for us.
You mentioned the role of China. I feel like the deceleration is kind of a U.S.-centric viewpoint a little bit because China does continue to invest very aggressively in these technologies.
Well, that's how we see it. They continue to invest very aggressively, and a lot of the large, Chinese NEV players are looking to tap into export growth to drive their business.
Mm-hmm. Okay, great. Some of your specific drivers that you've talked about, maybe starting with battery management on the EV topic, you know, that, that is a very compelling opportunity. Obviously, doing that well, enhances the range of the car, which is one of the most important elements.
Yeah.
You know, how are you doing there? You know, Analog Devices talks about that market as well. It's-
Yeah, I'd say it is a... Currently, it is a duopoly. ADI is larger than NXP today through their acquisition of Maxim and Linear. Great company. I think the difference between our offering and ADI's offering is, our offering is more of a system solution, so it's a combination of analog devices that sit out on the battery pack, connected with a back-end processor that we run load balancing software on, that it monitors the health of the battery cell, monitors, and manages the discharge and charging rates, and so it's more of a system solution. That business we had expected to grow from about $200 million in 2021 to about $400 million in 2024, and we're above that growth rate, that kind of should value growth rate today.
Mm-hmm. Okay, and it's not one of your Analyst Day drivers, but you just alluded to it. You know, as you move to EVs, you know, you look at the invention of Tesla. They sort of rethought the whole ecosystem of the car, and there were different, you know, MCU counts and different things.
You know, are you seeing that elsewhere as well? And is that... You know, you seem to have pretty good capability in, like, low power MCUs, stuff like that-
Yes
... that should help you.
So what we're seeing is, if you think about the current architecture today of a car, it is a flat point-to-point architecture. And as more data comes into the car and more data moves around the car, that architecture is just not extensible. And so a lot of the OEMs, and this is an OEM-driven trend, not a Tier 1, is to build more of a kind of a processing hierarchy in the car. And think of it from the kind of the top to the bottom as, at the top, you'll have a vehicle computer, which is, maybe think of that as a gateway type of device, managing the input and output of data into the car. That gateway device would then manage multiple domains around the car.
Think of domains as logical representations of different parts of the car, body and comfort, in-vehicle networking, e-Cockpit, safety, things like that. Those domain processors then would manage different logical zonal processors, and zonal processors are much more physical, logical representations. They aggregate data together so that they can lower the wiring harness to the car, and the zonal processors then would manage edge node microcontrollers, things that... You know, a microcontroller that goes into the window lift or your seat recline is what you have. Our family of S32 processors enables the customers to adopt this technology and have a common software stack from the very high end all the way down to the edge, and we've done very, very well so far with that business.
Early ramps of the S32 family started at about $300 million in 2021, and we expect it to be about $600 million by the end of 2024. We're right on track. Design win rates, which we don't like to disclose, 'cause, you know, it's hard to just talk about those type of things, are about the highest they've ever been in the history of the company, the last several years.
Does that sort of shift the competition away from, like, an MCU type of domain to people like Qualcomm?
Well, I think what it does is, it's not just an MCU business, so the vehicle computers, these are MPUs.
Yeah.
These are more powerful, multi-core, 5-nanometer products. The domain processors are 16-nanometer MPUs. The zones can be a range of either microcontrollers or MPUs, and then you have the edge nodes, MCUs. And so we think we're uniquely positioned to offer that full stack to, to enable a processing fabric, if you will, in the car.
Yeah. Yeah, and then on the more traditional MCU business, which is, you know, a lot of your, your revenue, we, we've sort of heard these comments. I think one OEM talked about reducing the number of microprocessor families in a car by, like, 90%, which is sort of remarkable. You know, that could be a very good trend, you know, for you guys if you're the beneficiary of that.
Yeah.
Can you talk about, you know, that?
Well, I think what you're alluding to, Joe, is during the supply crunch, a number of the OEMs realized because the way the design awards had been historically run, where the Tier 1s would award multiple different microcontrollers to different companies all over the industry, and the OEMs would end up with a mishmash of different architectures that were incompatible. And so how do you write a software stack that can navigate those different architectures? And so you do see a lot of the OEMs saying, "Look, we want to reduce the number of vendors that we work with. We want to build more of this fabric hierarchy." And because at the end of the day, the OEMs want to develop software that it gives them life cycle management of the car.
When the car goes off the lot, they want to have at least enough headroom to add features to the car after you leave the lot.
Great, and then the third growth driver you talked about in, in auto is radar.
Yeah.
Can you give us an update on how you're doing there?
Yeah. So, on radar, we're market leader of 77 GHz radar. Again, it's a system solution, whereas there's transceiver, a back-end processing engine, and Ethernet interconnectivity, and a PMIC with it. That's our complete solution. We've got about 45, almost 50% market share of the 77 GHz radar market. We thought that business would grow from about $600 million in 2021 to about $1.2 billion in 2024. We took a little bit of a pause in 2023 as a result of some of the inventory we talked about-
... because it's a very concentrated buying market. There's only a few Tier 1s who actually buy those systems, and so we've given them a little bit of a breathing room to burn off some excess inventory, but we feel very confident in the long-term growth trends of radar.
Great. So overall, kind of wrapping up autos, I mean, you're going through a little bit of a correction now.
Yep.
You feel good about that longer term, kind of as you work through-
Yes
... those issues. And then, you know, Kurt made the comments on the call, like, I think we asked about peak to trough, and he said, "You know, that's not the right way to look at it, because we did a much better job of kind of minimizing the peak issues of that peak.
So it seems like that's a key differentiator for you guys relatively.
Well, I think if you look at last year alone, where prices were up 8%, but we only grew just under 1%, which means volumes were down about 7%.
Yeah.
Right? And clearly, last year you had global production, auto production of about 9%, and we didn't grow greater than 5% in the second half. So the math would indicate that we're under shipping end demand.
Yeah. Yeah, okay. It is interesting, because you have this wide variance of opinions that semiconductor companies have about autos in 2024, and yet the growth was actually pretty narrow range in Q4, and you were at the lower end of that. Speaking to what you're saying, is that-
... there was some conservatism. But you have company... You know, one of your competitors has been vocally negative that, of course, automotive will correct, and they were up, like, 10% year-on-year in Q4. So, like, they're all seeing it, but-
I think one of the things that Kurt, our CEO, has tried to really communicate is each of our end markets that we participate in are cyclical. We're not saying it's a non-cyclical. They are cyclical, but they're not synchronous, right? The mobile handset market is different from a cycle perspective to the auto market. We think we are going through a bit of a soft landing in autos, but we do think that long term, the growth rate and the content growth rate of autos is something that's very attractive and something we're very focused on.
Mm-hmm. Okay, great. Can you talk about trends in the industrial market? You know, and again, it's an area where you saw weakness maybe early, kind of worked your way through it.
Right.
Where are we in that process?
Yeah, so, industrial and IoT is about 20% of our total company revenue. And while we don't disclose this in our filings, about 60% of that is what we would call core industrial. About 40% of that is what we would call consumer IoT. Consumer IoT ranges from wearables all the way to smart home-type devices, and the core industrial range is primarily to factory automation, building automation, and some healthcare. Those are the kind of the end sub-markets, if you will. We did see some very disturbing trends coming out of 2022 into early 2023. We took a very cautious look that industrial and IoT business is about 80% through the channel, and so it was something we were able to control very much with our own capabilities.
And so we really think we took our trough in that business in Q1, in the first part of 2023. It has started to improve incrementally off that lower base, but it's not back to where it needs to be.
Yeah. Yeah, I mean, this looks like the most severe inventory correction industrial that I think we've ever seen in the sense of- 'cause demand is okay.
Right.
It's not great, but it's like-
Right
We're in a PMI 50 type of environment, as opposed to 2001, 2009, where we saw these-
Right
-economic meltdowns.
And I'd say, I think most industrial customers are serviced through the distribution channel for all of our peers, right? It's a very, very long tail business, tens of thousands of customers. It's more efficient in many ways other than our friends in Dallas-
Yeah
who go direct to manage those type of customers.
Yeah.
For NXP, I mean, we're not saying we're the smartest guys in the neighborhood, but we do think we took a very heavy hand on how we would manage the channel, and we did forego some revenue early on. I mean, it would've been very easy for us to ship product into the channel. You know, our target in the channel is to run about 2.5 months. We've been running about 1.5 months since mid 2022. So that's about a $500 million delta. We're using our balance sheet to kind of buffer that material, and as we see demand come in from the channel, we can quickly turn it and see sales up.
Okay, great. Maybe you could talk about... I know it's not lower priority market for you guys, but the mobile market, smartphone market, you know, what are your prospects there you've seen?
So, we're a niche player in mobile. Let's be real up front.
Yeah.
You know, we're 12%-13% of total revenue. What we do in mobile is we provide secure mobile solutions, and those solutions are, if you've used Samsung Pay or Google Pay or Apple Pay, you've used our technology. That kind of security capability has also allowed us to do things like introduce ultra-wideband solutions, which is a RF device but slaves off the security in the mobile wallet. We've introduced other type of eSIM technologies into the mobile wallet as well, for network access and personalization. We saw the Android market be very weak in the first half of 2023, like everyone.
We think we've worked through those inventory corrections, and we're certainly seeing more of a kind of a pre-COVID seasonality type of trends in that market. We're normally down in Q1, and it's usually weaker in Q2 as well, and then rebounding in Q3 and four.
Okay. Okay, great. So maybe you could talk a little bit about, you know, your uses of cash in this environment. I mean, you've, you know, you've had so far a very, you know, much better than peer kind of results, I think largely due to how you've managed it, some of the automotive issues maybe hitting later. You know, you have very good cash flow. You know, what are the priorities there between dividends?
Our capital allocation strategy is actually very simple. Our goal is to return all excess free cash flow to our owners. We have one of two major methods to do so. We have a dividend, and our goal is to have a 25% cash flow from operations payout ratio, and we're a little above that in Q4, but that's about the target we want to run at. We view the dividend like debt, so you never want to violate it, and it will grow as cash flow from ops grows. The other opportunity is to buy our shares in the open market, and our only real governor there on buying the shares is so long as our leverage ratio, net debt to trailing twelve-month adjusted EBITDA, is 2x or below, we can be in the market when we would like to.
That's our model.
Okay. I have a couple more questions. Let me see if we have any questions from the audience first. Maybe if you could talk a little bit... You know, one of the debates that we hear is about Chinese competition-
Yeah
Given that the EV market is so prevalent in China, and that China is building a lot of fabs, you know, and hasn't really talked about what they're gonna put in those fabs. You know, how do you guys... I put you guys in a higher tier of defensibility, that you probably don't have to worry about that, but just how do you think about that part of the market?
So if you think about China in two halves of a coin. First half, you mentioned the manufacturing. They're clearly building out a lot of capacity, a lot of 300 millimeter fabs. I'd say a lot of that capacity is going towards power discretes for the electric vehicle market, which we don't participate in, but there is a number of mixed signal and bulk CMOS fabs in China. We currently today don't do any manufacturing in mainland China today, but a lot of our customers in China have asked us for a manufacturing strategy where we'll build in China. They don't want us to design there-
... but they wanna have manufacturing security just like the US, just like Europe does. So we've engaged with TSMC, our major foundry partner. We will auto-qualify. They have a 300mm factory in Nanjing. That engagement to auto-qualify that has already kind of... train's left the station. We're in process on that. That's for 28, 16 and 28 nanometer products. We also have a long history with SMIC in China, but that's been primarily for our mobile and our consumer IoT business. We are engaging with SMIC to auto-qualify their lines also, and then we're also out hunting for a third 300mm mixed signal factory partner. So once that's done in about a year, we'll be able to foundry wafers in China for Chinese customers. That's front-end.
On the back end, we own a large factory, a back-end manufacturing factory, in Tianjin. We will slowly start to rotate rest of world products out of that back end and make that more of a Chinese-only back-end factory, and so within a year or so, we'll be able to have a full front-end and back-end manufacturing footprint in China. So that's kind of on the manufacturing side. On the socket design side, we fully believe that over the next decade, you will see more Chinese indigenous semiconductor companies. It would be naive to say you wouldn't. But currently today, and for the foreseeable future, we still compete with the same folks we compete with in Europe and North America, the TI, ADI, ST, Infineon. It's the same cast of characters.
We do know that there are some smaller startup microcontroller players in China, mostly at the consumer side, mostly at the low end, but we've not yet seen them come up into the automotive market. We are aware of some, what I'll call, catalog analog companies starting up in China, not really our bailiwick as much. But, you know, over time, our view is our day-to-day requirement is to compete effectively. As long as it's a playing field, meaning there's not a government oversight, we welcome competition.
Yeah. And the microcontroller business, you know, there is a lot of focus on startup MCU companies in China and spot prices of MCUs in China, which seems like there is no spot market, you know, when the shortage ends, so it's kind of an odd concept for me. Can you talk... Excuse me. But can you talk to barriers to entry in MCUs? You know, it's a broad market, there's-
Yeah
... you require a broad range of products to be successful.
I think what you have to remember on MCUs, very similar to FPGAs, these are programmable devices. When they leave our factory, they're dumb pieces of silicon. You need to provide software enablement tools for customers to write their code to. You have to provide firmware for them, and so there's a high barrier to entry for providing those tools and that capability. And that's why it also does take a longer time, I believe, from design win to revenue in microcontrollers, because there's a higher software component.
Great. And then just going back to the pricing debate, because it is something that people ask about a lot. We have a couple-
... minutes to talk about it. You know, you talked about, you know, not seeing any reason for major degradation, for giving that back. Can you talk about why? Because, you know, when prices went up, it was because foundry costs were going up.
Yep.
If those costs reverse, you know, do you see that as passing that along?
Well, currently today, and for all of 2024, we don't see that as the case.
You know, I think you do hear of pricing going down in what we would call second and third tier foundry quotes, but your kind of tier one foundry quotes, your TSMC's, your GlobalFoundries, these type of folks, they're not lowering prices. And so if our prices aren't coming down-
... how are we going to pass on lower costs to our customers, unless you, as our, our owners, are happy with compressed gross margin, which I don't think you are.
Right. And the what you're talking about in 2025, a return to maybe a modest downward rev- is that normal? Is that, like, back to normal?
Yep.
Like, down 2% or something like that?
Well, pre-COVID, we were down to about low single digit, 1%-2% annual cost concessions to customers, and that really represented our operational efficiency.
Yeah.
Right? You eke out a couple points every year, you pass on a little bit to your customer.
I mean, you, to your point, you never really saw the gross margin expansion from raised prices-
No
... that may be some other sauce, you know?
That's correct. And I think that's really what differentiates NXP, how we approached pricing and inflationary costs, was we aggregated together all of our inflationary costs, and we spread those costs across almost all of our end markets. Not all, but almost all of our end markets, in a very peanut butter fashion, because it just was too difficult for us to go out and say, "Customer A buys a product from TSMC, Customer B buys a product from GlobalFoundries." So we just basically aggregated that cost up across the board.
Okay, and as there's, we, well, actually, yeah, we do have one question from the audience, if there's time.
Hey, thanks for being here. I was just wondering what your thoughts on M&A are currently, and whether there's any technological capabilities you'd be looking to enhance in the future?
Great. So, I guess kind of part of the capital allocation strategies comment. So we do M&A sub-$100 million, $100 million dollar type opportunities. These are small opportunities, tuck-ins, design teams, IP, what have you. Large, transformative M&A, while interesting, just the environment is not conducive to that. Almost every large deal would have to go through China, and then you have a lot more, what I would think are geopolitical risks, both in the U.S. and Europe, about any sort of large M&A. But long term, if you could take the governmental oversight and put it aside for a second, I think our industry does need to continue to consolidate.
You know, it's been a long time now, but, you know, how do you guys feel about Freescale at this point? I mean, there was some ups and downs around that, but obviously, some of the core capabilities you talk about came from the Freescale side.
In the limit, it was an excellent deal.
Yeah.
Was it easy? No. I mean-
Yeah
... like, well, M&A, it looks so pretty and beautiful the day you buy it, and then you get in there, and you realize you have some things that are not exactly what you thought. But in terms of what it provided us and the complementariness of what our existing kind of pre or legacy NXP portfolio, it was a great deal. Great deal, and, and we would do it again today if it was on the table again.
Yeah, if there was a regulatory environment.
Well, there was a re-
Yeah.
Yeah, exactly.
Got it. Okay. All right, any more questions from the audience? If not, we'll wrap it up. Oh, there's one more.
Hi, Jeff.
Hi.
You, you guys have been an excellent gross margin story through time. Given that now the cost structure is such that you have 30% fixed, 70%- ... flexible, is there still a gross margin story from here, or is it just getting more difficult?
Well, I, I'm gonna quote my CFO and my boss's statement. "58% is not the destination." And the way to maybe think about this, if I could maybe put some crumbs on the ground for you, if you think about our channel, a little over half of our revenue, the channel's effectively a margin accretive channel for us. As we start to refill that channel back up to the 2.5 months, that's a margin accretive event. At a certain point, by refilling the channel, you're gonna burn off DIO on the balance sheet. At a certain point, we're gonna have to start wafers again in our internal factories and drive the utilization back up to the mid-80s. Again, another margin accretive type of tactical event.
Those two things won't be synchronized, but that's the way to kind of think of the drivers that could drive gross margin higher over the long term.
Can you also comment on operating leverage, right? When that gross margin starts to recover, or not recover, but goes towards the 58, how much additional operating leverage can you drive?
So we don't really provide, like, a gross margin leverage structure like that. Our model is 55%-58% gross. On OpEx, we run 16% R&D, 6%-7% SG&A. We have more projects than we have R&D that we're allocated to, so we're gonna continue to probably run at that 16%, because R&D for us is the lifeblood of the company. So you can kind of do the chainsaw math. As gross margins grow and revenue grows, you will get some leverage on the operating line.
Great. Any other questions? Okay, we'll wrap it up there. Jeff, thank you very much.
All right, great.
Appreciate it.
Thanks. Thank you, everyone.