Great. Thank you, everybody. Welcome back. I'm Joe Moore. Very happy to have with us today the CFO of KLA, Bren Higgins. So Bren, maybe you could start out just with a little bit of an overview. We'll dig a little bit deeper into your targets and so forth. But maybe first, just give me a sense for your corporate priorities at this point.
OK. Well, Joe, thanks for having me here. Excited to be here and spend a little bit of time talking about what we think is a pretty compelling opportunity moving forward. You mentioned our 2026 financial targets, which are, for those who don't know, we have $14 million revenue and $38 in earnings per share. And when you back up and just decompose that a bit, it starts with a view of semiconductor revenue growth that's roughly 2x GDP. So we'll call it 6%-7%. Aligns to the McKinsey study of $1 trillion semiconductor revenue in 2030. Capital intensity underneath that has been climbing since about 2013, 2014. So we've seen a nice, steady increase in capital intensity. And of course, over the last few years, it's been even more elevated with some of the investment that's happened in China.
But in general, as you just look at long-term expectations for capital intensity, we would expect WFE to grow a little bit faster than semiconductor revenue. If you look at the mix of WFE, we believe that you'll see memory generally about GDP plus kind of growth rates. Memory tends to be a little more episodic and more minimum driven. It has a little more cyclicality to it. And that foundry logic will likely grow faster. And as a result of that, you'll end up with a mix that's roughly 60%, maybe a little bit higher, of logic foundry. And that matters a lot to companies like KLA within process control, because process control intensity is meaningfully higher in logic and foundry than it is in memory. So that should enable the process control part of the market to grow a little bit faster than overall WFE.
KLA's position in process control is a very strong position. We believe that the markets that will inflect over time, as we continue to see robust adoption of leading-edge design rules, will drive certain parts of the market to grow faster where we already have a very strong position. Things like optical inspection, which if you just look at 2019-2022, grew at about 1.7 times what WFE grew. A very strong growth profile driven by the adoption of EUV and ultimately more and more scaling of devices. That should enable KLA to continue to gain a little bit of share on top of our already strong position. Our service business is growing 12%-14%. Expected to grow at that level in 2024.
For very specific reasons about the growth of service, KLA's service business is a high mix, high complexity, relatively low volume business, lower redundancy for our customers. They expect high uptime. They expect matching performance. When they buy KLA process control, it's very hard, given that volume dynamic, for them to build the capability to do their own self-service, which you see with a lot of process companies. Also, the complexity of the products that we build factors into that. And as a result, our customers tend to buy contracts. These contracts are usually at least three years. 75% of the revenue stream is contract. The other 25% is billable. And so that consistency and predictability, even upturn and downturns, allows this business to grow over time.
In 2024, I would expect service to grow even faster, as the tools we shipped in 2021 and 2022 come off a warranty. Our EPC business, which is our electronics packaging and components business, should grow about 10%. We have product offerings. The strongest part of that business is specialty semiconductor, which has exposure to compound semi but also advanced packaging. We have a PCB business that also has exposure to packaging. And so we should get about 10% out of that. So you add it all up, you end up somewhere in that $14 billion range. Gross margins are in the low 60s. We have a target of 63%. Drop through in terms of operating margin, a 40%-50% incremental operating margin. Capital structure action should enable us to drive the EPS level about 1.5 times the revenue growth rate.
We're going to return about 85% of the cash flow that we generate. We think it's a pretty compelling model. The team is pretty prepared for us, as we start to see our customers start to spend again, as we move through this year and into next year, to make sure we're prepared for growth at the leading edge and what will drive and enable this model that we think is pretty compelling.
Great. So I'd like to dig into a few of those things. Maybe if we could start with the process control intensity. I guess you talked about a 1.7x from 2019 to 2022. 2023, we sort of had a year where there was a lot of focus on capacity additions in the trailing edge that probably went against you. How do you see that? It seems like we're now shifting the focus back to the cutting edge, and in particular, a lot of new technologies and things like that that would seem very process control intensive.
Yeah. So the 1.7x was for the optical pattern inspection business. So overall, if you just back up and say, OK, process control intensity in 2021 was in the low 6%. So it depends on what denominator you use. We'll call low 6%. And in 2023, it was in the high 7s. 2022 and 2023 are a little noisy because of some of the challenges that a lot of our peer companies had in terms of the timing of WFE, given some of the supply chain challenges. But in general, if you adjusted and shifted the money around, you're kind of seeing a nice trajectory of somewhere from the low 6s into the high 7s. Obviously, mix is a big factor in that. And that's why it was important for me to talk about the mix dynamics.
But I think one of the things that we've seen is we've always seen a nice level of adoption of our systems in R&D when customers are developing process. And then once they start to work through their process, then they start to ramp it. And they introduce the element of volume. And that tends to drive more inspection, because you sample and inspect and measure more. And then as you get into production, typically, historically, what customers have done is they tend to back off, because they tend to be in a more mature state. And you would see less adoption. And that generally is still the way that it flows.
But because of the design environment, because of the iterations in each process node, you're starting to see and have seen over the last couple of years more what we call HVM adoption, which is high volume production adoption. And so that's been a nice driver of incremental growth that's fueled the process control intensity that we've seen and seen that increase. And that's very different than what we saw if you go back the middle of the last decade, where Moore's Law wasn't scaling all that quickly. EUV was being delayed. There wasn't the end market adoption. Incremental scaling was fairly limited. And so customers were able to be very efficient with their capacity in terms of trying to reuse it from node to node.
That all kind of broke down at 7-nm with the proliferation of design, which we've seen continue as we moved here through five and three and expectations for two here moving forward. So we feel pretty good that where we are today in that sort of high 7s, that we'll be able to increment that over time, given some of the growth drivers we think are out there in terms of incremental scaling, changes in architecture, power distribution, and some of these things that we think will drive interesting challenges from a process point of view. And then, of course, maybe stronger competitive dynamics moving forward in terms of the drivers of leading edge.
Yeah. OK. Great. And then on the way to the path to that 2026, you guys have talked about WFE in the mid- to high $80s this year, I believe, kind of flat to modestly up. Can you talk us through your views there? And I know there's some puts and takes. Maybe talk big picture first.
So 2023, we think was somewhere around $87-$88 billion. We don't know yet. We're kind of waiting to see how all the numbers get added up. But we think it was somewhere in that ballpark. And as we look at 2024, we see it at roughly flattish and kind of high 80s kind of level. I guess if WFE ends up coming down ended up in the mid 80s, then maybe then the growth expectation is a little bit higher. A little more second half loaded in terms of our expectations. If you back up and look at leading edge, I'd expect leading edge to be up. I think we saw significant adjustments in 2023 from high levels of investment in 2021 and 2022 from our major customers.
I would expect to see the leading edge start to improve, most of it driven by 3 nm activity. On the memory side, a little bit better, mostly in DRAM. Talked about HBM and advanced DRAM as drivers, a little less so on the flash side, but maybe some modest growth in memory fueled by DRAM. Legacy investment off a little bit compared to a very strong last year overall. I think what we've seen in terms of activity levels in China to be more or less flat, some changes in the mix of that business between what we call infrastructure, which is reticle and wafer infrastructure, coming down a little bit, particularly on the wafer side, and then the logic foundry piece being a little bit higher. But overall, roughly flat.
And so translates into, we'll call it, flattish to maybe up a little bit for the year.
And then how do you think about that? I mean, so now you have a couple of years to go between that and the 2026 guidance. Ordinarily, I feel like it's clearer if we're troughing or peaking. Like right now, we have some businesses troughing and some businesses peaking at the same time. So how do you think about 2025 in that context? Is there a baseline that we should have growth next year?
Well, I think some of the bigger drivers are semiconductor consumption. We're starting to see the markets move and improve, particularly in handsets and in PCs. More data center investment. Obviously, AI is growing faster. And that's driving stronger data center, much more semiconductor intensive compute requirements. And so that's a good driver there. I think memory, as we start to see the memory business and utilization start to improve this year, and they'll do some we'll continue to see some tech investment in terms of driving shrinks. They'll start to get to a point where you need to add capacity into next year. And then if you back up and you say, OK, if you start thinking ab out the 2 nm ramp, ATAA ramp, and some of the competitive dynamics driving the leading edge, we would expect to see some meaningful capacity investments there.
So when you take it all together, while I'm not guiding quantitatively, I feel, along with a lot of others, that we'll see a step up in investment as we move into 2025 driven by these factors. I think the China business is probably flattish. It's a little hard to tell at this point. We'll see how that plays. I don't necessarily see it falling off a lot. I think the legacy, which will be weaker this year, we'll start to see that improve a little bit too as you move into 2025. I think one thing that's clear in that part of the market is that those customers have to invest more than they used to. There aren't used and excess equipment to service that part of the market. So you have a lot of new investment that's happening.
That should be good for WFE investment as well.
This China piece is a pretty big debate at the moment. And I guess to me, the bullish case would be you step back and look at a region trying to develop self-sufficiency in semis. And they're going to have to spend a lot more to do that. The cautionary side of it would be they don't necessarily have products to fill those factories with yet. So there's a little bit more of a building ahead or maybe building at the same time. So how do you think about the role of geopolitics in China? And I know you've talked about revenue declining from not very high levels of the back half of last year. But just how do you think about China in general?
Yeah. I agree with you that I think when you have a number of projects around some of the investment that's in the logic foundry space, that subscale, the customers are doing R&D, not very efficient, trying to prove out process, introduce some volume, prove to customers that they can deliver, that they've got a roadmap, and that there's a fair amount of inefficiency in some of that. Now, there's also some investment from a lot of the normal customers we've had there for a long time. I think that's more supply supported or demand supported, if you will. The infrastructure investment was a choke point, both in terms of wafer access and reticle access. So I know there's a desire to have more control over that, where a lot of the wafers were tied up in long-term contracts. Reticles, you'd have to go to the merchants.
The merchants haven't really invested. So it was a challenge. So I think there's a desire, A, to meet self-sufficiency and have a little more control from, we'll call it, wafer and reticle all the way through the process, supporting that but also supporting the export economy in China as well. In the long run, when we back up, we think generally you can have periods of what we'll call inefficiency in terms of investment. And over time, some of that will have to rationalize, whether it's because there's some consolidation or whether there's funding tied to the next increment of investment or whether there's technical milestones that have to be hit. But over time, that you would expect some normalization there, although the activity does create a little bit of inefficiency.
Going back to the way I talked about 2026 is that, look, our view in terms of how we're running the company and how we're sizing the company in terms of how we invest is back to that view that at the end of the day, we believe WFE is generally going to grow a little bit faster. It's not going to grow a lot faster. I don't know how healthy it is if it grows too much faster in terms of overall profitability. And that's how we're going to size the company.
If you end up with a permanent tier of inefficiency that exists because you just have either strategic investment or regionalization investment that spreads the footprint out, and that's a little bit harder to rationalize over time in terms of getting supply just right relative to demand drivers that have some fluidity to them, if you end up with a permanent tier, in my view, that's upside to the plan and the model. We've proven over time that we can execute against incremental opportunities. But as I back up, I kind of think about it that way. Look, in the near term, we've seen a nice cushion relative to the drop-off we've seen from our other customers. And over time, we'd expect to see some rationalization in how that plays out.
As it pertains to export controls, I'm not going to speculate on what changes the government might make. You can if you want. But you probably don't want to either. But how does that frame the conversation? Do you think Chinese companies are anticipating further controls and spending in front of that? Do you think anybody's? We're sort of assuming that we've sort of achieved the status quo and we're not going to have further tightening? I'm not asking for your prediction of what the government's going to do. But what is today's behavior predicated on?
Yeah. So I'm not going to speculate either. What I would say is that it's really hard. I hear investors ask me a lot about stockpiling. And it's like Lisa says it relates to process control equipment. You typically don't see that behavior. We're installing the tools. The tools are running. They have to be supported by us. And if they're not supported by us, it's very difficult for our customers to get value out of them. So it doesn't really make sense to buy the tools and not install them and start to use them.
Yeah. Yeah. That makes sense. But it's a very long tail investment from their standpoint. I mean, they're not investing in the cutting edge where you're obsolete in five years. We'll have a 20-year life of these fabs.
And to the point of self-sufficiency, there's a number of IoT markets. There's the electrification market. A number of the new projects are compound semi-driven projects. So yeah, I think it's a long-term plan. And it's one of those things where one thing about this industry is you have to keep spending. R&D never really stops. And if you've got a desire to move from very old design rules, even as you move through the legacy parts of the node, it's new to you. There are challenges, unique aspects of certain products that require you to continue to invest if you're going to try to get a customer to make a commitment and be committed to a roadmap over time.
OK. Great. And then just last China question, China DRAM, I think everybody had talked to there was a pretty high second half of last year because a customer came off of an export control. And so there was a little bit of an inflation. But then you have a second customer spending. I guess what do you anticipate from China DRAM going forward? Do you think there will be more customers? And even in NAND, I mean, the one company that builds NAND is on the Entity List. So I know nobody's doing business with them. But do you see?
Well, no, U.S. companies are. But.
Yeah. The non-U.S. Yeah.
Yeah. The non-US companies are.
Right.
So on the DRAM side, I think it's more or less flattish. You described the environment pretty clearly. Not a lot of sort of financial efficacy to trailing edge DRAM generally over time. So I don't think you'll see a lot of incremental investment from new suppliers there, in my mind. I think it's always been tough to make a value case for trailing edge memory. And so I think it's probably limited. And then on the one customer, they're bumping up against the technology threshold that was established. And so is there another product cycle? Maybe there is. And I think that I would expect this year to be lower than what that customer spent last year.
I mean, I take your point on that. But it may not be economically viable. But if there's enough subsidization and the government wants domestic memory, you can still do it. I mean, you can still build high quality, expensive memory with those trailing edge nodes. OK. And then on the advanced technology, it feels different to me the last two or three years, maybe just because Intel's pretty loud about their progress on these new technologies. But it feels like there's an active debate now among investors who's going to have the best node and things like that. That would seem to be a really good construct for you. And it would seem like people not to get into anyone's customers, but some of these customers had been dormant for a while from the standpoint of major investments in process control.
Now you see an arms race where really three companies are trying to duplicate all around on a timely basis. It seems like that would be a good setup for you guys.
It's a very good setup. Yeah. Really good end market interest in the nodes. 3 nm has strong interest. Our big customers indicated 2 nm has similar interest at the same level. So I think that's one of the biggest drivers for us is you've got good competitive dynamics, you've got technology transition, and you've got end market adoption. So yeah, it's a very good constructive setup for process control where we're, I don't want to say agnostic, but in some ways agnostic to how it plays out in the overall market as long as those dynamics are happening.
Yeah. But on that note of obviously, 3 nm are a really good node. And a lot of people are interested. But there doesn't seem to be the same interest in being the pipe cleaner that we saw two cycles ago. We had Huawei and Apple pushing really hard on the cutting edge. Now it's kind of like there's a lot of customer that want to move to 3 nm. But maybe nobody wants to be the first. Does that matter to you if it plays out that way?
Well, I think one of the good things that's also happened is it's not just mobile driving the leading-edge process node adoption. You're seeing more and more custom silicon in terms of high performance compute opportunities. So I think that tends to you're having to support more designs earlier in a process node life. And that can be good for process control. It puts pressure on yields as you're trying to ramp that fab. You're already seeing customers over the last few nodes tending to ramp at less mature levels than they used to before. And I think that's also been a factor in overall adoption of process control. So I think that you've got multiple markets driving these leading-edge nodes and how they're coming to market. And I think you get maybe a bigger commitment, hopefully earlier, in terms of the process node life.
It's much harder for customers to try to reuse the previous node as they're trying to ramp at a faster slope for the leading edge node. I think it's a good setup as we move beyond mobile as the principal driver.
Yeah. That makes a lot of sense. And then you saw a lot of enthusiasm from investors when ASML had a really big order book for High NA, which is something that's really beyond the roadmap that most of your customers have talked about. Can you talk about that generally, how you see High NA benefiting KLA?
Well, the good thing about High NA is that it continues the scaling roadmap that we're going to continue to see scaling. We're going to see more adoption of EUV. As you get to High NA, you start to see a pitch shrink, which enables more scaling on the reticle, which is an opportunity today as we get there. And it will continue to drive what tends to be the biggest driver for us is with scaling smaller and smaller feature sizes on wafers. It drives our high end inspection. It drives more precision film measurement. So we think it's sort of validation of the roadmap that continues to enable scaling, which has been a big driver the last few years once we saw EUV introduced. And we expect it to continue.
OK. Great. Maybe shifting gears a little to talk about your own supply chain. First of all, you mentioned that you didn't get tripped up in 2022, which actually mattered a lot, as it turned out, because a lot of the 2022- 2023 dynamic was people that had deferred revenue pushing up. But I feel like maybe you guys didn't get the credit that you deserved for managing through that so effectively in 2022. So maybe just talk about that. I mean, you have much more of a semiconductor intensive bill of materials than your competitors do. How'd you manage through that so well?
Yeah. We certainly didn't do what our customers asked us to do. We still had our challenges. And I think that the nature of our supply chain is a little bit different. You talk about it compute intensive. But actually, the longest lead time components we tend to have tend to be around optical components. And the lead time on those can be upwards of 18-24 months, depending on the equipment and the amount of capacity. So I think it validated a lot of the sort of partnership drivers we have with our suppliers in terms of our willingness to invest in their capacity or co-invest to make sure that they have what we believe we will need over time because of these lead time challenges. We tend to trade a fair amount. We're pretty good customers with our key suppliers.
We want them to work with just us on a lot of these capabilities. We're willing to invest. We make commitments. We keep them. I think it's one of the reasons why the levels of inventory we have, we're over $3 billion of inventory. It's one of the things that I am trading in terms of enabling my differentiation and my willingness or ability to serve with the fact that I'm going to make commitments. I'm going to follow through. I think these were factors in how we managed through. It's not a transactional supply chain. We don't get wrapped up with our suppliers and who carries the inventory and how do you think about payment terms and things like that. It's much more about enabling our differentiation. We think it matters. I think it's reflective in the gross margins that we're able to command.
I think we get paid for the actions we take. As a result of that, I think it played through pretty well. Our suppliers, they work with us. In a lot of cases, they don't work with others, at least around a lot of the same types of supply. And I think that was helpful through that period of time. So I think we're very well positioned in terms of our ability to ramp the business to the plan we talked about earlier, made a number of investments. I talked about Investor Day a couple of years ago that we're somewhere in excess of $150 million of investment with suppliers to ensure we have the capacity that we need. We're going to continue to do things like that.
I think I'm in a place where I can scale from $10 billion-$14 billion in a few years.
What if it's better?
With the commitments we've made.
What if it's better than that? I mean, can you respond to that within that?
So we think about flexibility and what we need beyond that and some of these relationships. And I think depending on when you get that, the clarity of when it's going to be better and how much is a factor in it. But I think our history has shown that we tend to react pretty well to it. And I think we're very well positioned without a lot of incremental investment to deliver on those commitments. We're still short on certain components. High-end optical inspection is one of those areas where optics take a long time to increase capacity. We have some capacity coming online this year, which is why I'm more bullish on some growth in that part of the market this year.
I think it's a testament to the demand for the Gen 4 product line, which is our shortest product, both in terms of it being the production workhorse where it balances the wavelengths and the sensitivity that it has as a broadband tool with cost of ownership for customers. There's a special version that has new capability for Gate-All-Around that's coming to market. It also has some deconfigured capability that we can take backwards to support automotive, as an example, or where we're able to get more revenue. Even some of the legacy nodes, we're able to get more revenue at those design rules than we were able to get back when that was the leading edge design. So the customer gets a tool with an upgrade path and more capability from an economic point of view. And we get more revenue.
I think we're in pretty good position with that product. I think we're going to be short for a little while longer because grinders and polishers for optics take a very long time to get in place and the training that's involved, the automation with our customers to be able to enable optics that have the precision that's required for what we do. So it will take some time to get there. But I think we're in a good position to see some come on this year. And then as we move into 2026 or so, we'll see another step up in supply capability.
OK. Great. And then upfront, you had mentioned advanced packaging. And obviously, the Orbotech acquisition has added a lot to your capabilities in that area as that's become much more important to everybody. So can you talk a little bit about KLA's position there and kind of what growth drivers you see going forward?
Yeah. So it's interesting. It's across multiple products across the portfolio. So we have inspection and metrology, which has been a driver, which is not related to the Orbotech acquisition. But generally, the inspection capability that we've sold, particularly around GPU heterogeneous integration and co-optimized process and so on, that we've been able to do pretty well there. Over time, you're seeing rising complexity. The back end is starting to look more and more like the front end, which is interesting here in that you start to see pitch shrinking, so the density of lines and spaces shrinking. And so scaling affects in how these packages are put together. And so that drives both current inspection. But we think over time, we'll drive higher end inspection requirements. And given the trade-off between sensitivity and throughput, what will end up happening is you'll need more capability.
The speed of the tools will slow down because of that higher capability. If you're going to keep the same sampling strategy, you'll see a unit growth opportunity too. ASP, for a higher end tool, but also more tools to support a similar level of speed and volume. As you start to think about some of the applications, there's more metrology requirements. Wafer to wafer, bonding as an example, you have overlay any time you're trying to put wafers together. You've got to make sure the wafer's clean and particle-free. It's driving a number of different products. On the process side, we have plasma dicing. We have ViaReveal. We have some RDL layer support that we do in terms of some of the deposition capability of the company.
So it's a number of applications in the package that drives the SPTS, the specialty semiconductor part of KLA, which came through Orbotech. And then you've got what's the evolving substrate part of the market that gets addressed through our PCB business. And we've got a couple of new products that will be coming to market, both for imaging but also inspection and how that substrate integrates into the package and then integrates into the PCB board. If you add it all up, about $350-$400 million of revenue in last year. And I think this year, we're going to be somewhere in that $400 million range. And I think you're looking at growth rates that are probably at least 2x WFE growth rates going forward.
I think you have an opportunity to see this scale into a billion-dollar business as we move sort of into the back half of this decade across multiple products.
Great. So maybe we could talk a little bit about memory. For DRAM, you had mentioned China. You mentioned HBM. But core DRAM CapEx seems healthy at this point. Do you think we can continue to grow from there? And I would also ask you, I know you have insight into utilization on DRAM, if you've seen that utilization come back.
Well, so in DRAM, we've seen it come back a little bit, not much. But we've seen it improve a little bit. I think customers that cut the most, we've seen them respond more than others. I think, as I said earlier, you're going to continue to see shrink investment that will happen to try to drive incremental bit supply through technology shrink. I think WFE will be up a little bit. I don't think it's going to be up a lot in that part of the market. But I do think it's probably more of a 25 effect in terms of more capacity investment. HBM has been a nice driver as it relates to process control. Of course, advanced DRAM has EUV. I think going forward, as we talked about, you're going to see more EUV layers in terms of DRAM and investing there.
And so that should be a pretty good driver as it relates to HBM. Reliability is higher. You want to make sure each of the chips or the die all function correctly before you integrate them together. So you've got advanced die. You've got this reliability challenge. They're bigger, which tends to consume some capacity. The bit trade is less for an HBM die versus an advanced DRAM die. So I think there are drivers there that are ultimately going to drive these utilizations up. We'll see pricing improve. Customers' profitability, cash flow will get better. And I think we'll start to see them invest as we move through the second half of this year and into next year.
On the NAND side, I mean, the spending has been really low. Everybody has reasons why it won't come back. But it feels like, to me, people are paying attention to some of this other stuff first, but that NAND ultimately has a pretty good chance to come back. Are you seeing that?
Maybe more into next year.
Next year. OK.
I think you're still seeing some of the shrink focus. I think you're going to see continued pressure on trying to drive layer counts higher. That creates opportunities for us in terms of wafer flatness. You're trying to inspect the shape profiles. You've got high aspect ratio inspection requirements. All those, I think, are overlay. Overlay is becoming more of a challenge there in terms of overlay registration as you do double stacks and so on. So I think it's creating some opportunities from a tech point of view. But I don't think you're going to see a lot of capacity investment. Again, I think it's going to follow a similar path into next year.
Great. So then the last thing you mentioned that I want to double click on is services. And then I'll open it up to the audience for any questions. The services piece, you've had pretty good growth rates there. And you've talked about the fact that some of these utilizations are starting to come up. So it seems like that should be helpful. And I feel like you almost have a unique view into what services actually looks like because you don't have spares. There's no risk of stockpiling. You're actually just collecting service revenue. So can you talk about the sustainability of that?
Yeah. Against our peer companies, our service business tends to be all traditional service. There's no systems in the service numbers, which can skew it in an upturn, skew it the other direction, maybe in a downturn. So it's very consistent, we'll call it sort of break and fix, in supporting the systems. I talked a little bit about the dynamics that drive it. Utilizations will improve. And so that should be good over the course of this year for more service revenue. And then we shipped a lot of tools. Those tools will come off a warranty. We're greater than 90% in terms of our attach rate on services going into contract as they come off a warranty. And so it gives us pretty good visibility into what's coming.
I've been pretty open with an expectation that in 2024, we'll see $250 million or more of incremental revenue just associated with those drivers, the utilization improvement and tools coming off of warranty. The structure we have and how they bypass control is pretty important. I think because of the complexity of what we do and the fact that it's relatively low volume, high mix, all the tools look the same in the fab. But then you start to look. You take off the white skins with the purple stripes. You start to see not only different product types, but even within the same product family as customers are mixing and matching depending on the various applications and the production process. Trying to build the ecosystem to support all that, very hard for a customer to do.
I think they'd rather outsource it to us. It allows us to set up that contract structure that gives us good visibility, I think pretty good profitability relative to the overall. It's one of the reasons why we've been able to do what we've been able to do with capital structure and capital allocation over the last decade or decade and a half is that that visibility provides a pretty clear cash flow stream that supports dividends and debt service. So I think that's increased our comfort related to those decisions, which I believe generally are kind of irrevocable decisions and that it can be supported by the service business that we have.
Great. So let me open it up to the audience. Any questions? If not, I'll keep going.
Hey.
Hi.
What does a sort of forward of your business, say, sort of resurgent Japanese industry do? A lot of your share gains have come at the expense of a couple of local players there. And is it sort of you winning in sort of e-beam package inspect? Where is the winning coming from relative to some of those Japanese guys? I'm trying to figure out the incremental opportunity as that country really steps on the gas.
Yeah. I don't want to talk about specific customers. But I'll just talk about just what drives our share is that one thing about KLA is we go to market with a portfolio of solutions that allow our customers to balance their sensitivity and technology requirements with their economic needs as they go from an R&D to a ramp and into a more mature volume production environment. Most of our competitors tend to be point product competitors. And so they have to try to figure out, how do I take a certain technology and try to figure out broader use cases where we can be fairly agnostic to which platform or capability our customers choose because we're able to offer them all. Now, we collaborate with our customers about their problems and how to get close to what we think they'll need.
It takes us a long time to develop the systems that we have. We invest. We invest more, generally, than most of our competitors. I think the portfolio and then being able to tie it together with more software capability to do review and classification and do simulations and other data analytics allows us to, I think, provide more actionable time to results than what a lot of our competitors are able to provide. Our position in process control has continued to grow. We gained a few points of share last year. In the market we're in, about 57% is recorded by Gartner. I think this year is roughly flattish or so against that number. It tends to come in increments each year necessarily, but in bigger chunks.
But we think that if we continue to invest in areas where we don't have a stronger position, we have a number of investments that are happening in some electron beam markets. We think our position in certain markets like optical inspection will inflect. And that'll drive more share opportunity as well in terms of just the relative growth rate of that part of the market versus the rest of it. That we're in a pretty good position to see it continue to grow a little bit more over time as we go forward.
One more question in the front. Can you wait for the mic?
I have a question on how you think about process control steps and attach rates within the High NA roadmap. Directionally, do you think there will be more process intensity for the same amount of capacity versus, say, Low NA multi-patterning?
I think as you continue to drive, you might see some substitution here and there as it's being adopted. Some customers and around certain layers, you may choose a High NA tool. But it has some challenges with reticle field and cost potentially and certain steps where a customer will try to leverage double patterning. I think either way, you start to have more and more EUV layers. You're usually seeing more pitch shrink. You certainly will as you get to High NA. So I think what it does, and you back up, is it tends to, I think, create a scaling roadmap. That's usually a good driver for tighter transistor density. And therefore, smaller defects matter and precision measurements required and so on. So I think we're pretty early in it.
I think there's a lot of capability and complexity in EUV masks that you'll see it play out differently than what we saw when multi-patterning was introduced at fairly mature DUV lithography levels and reticles that really weren't changing all that much. I think as you move into a High-NA environment, you're not going to see that. I think you're going to see higher levels of complexity. I think it's going to create some interesting opportunities for us.
Great. Well, that brings us up to the end of our time. We'll wrap it up there. Thanks, Bren.
Thank you for having me.