KLA Corporation (NASDAQ:KLAC) J.P. Morgan 52nd Annual Global Technology, Media and Communications Conference May 21, 2024 9:30 AM ET
Company Participants
Bren Higgins – EVP and CFO
Conference Call Participants
Harlan Sur – JPMorgan
Harlan Sur
Good morning. And again, welcome to the second day of J.P. Morgan’s 52nd Annual Technology Media and Communications Conference. My name is Harlan Sur, Semiconductor and Semiconductor Capital Equipment Analyst for the firm. Very pleased to have Bren Higgins, Executive Vice President and Chief Financial Officer at KLA here with us today. I’ve asked Bren to start us off with KLA’s view of the wafer equipment, process control, specialty segment, semiconductor equipment spending environment this year, KLA’s growth outlook within that, and then we can go ahead and kick off the Q&A.
So with that, Bren, thank you for joining us.
Bren Higgins
Yes. Thank you for having me. Appreciate being here how much time do I have for that one? Yes. Yes, no, it’s an interesting time. 2024 is setting up to be more of a transitional year for the industry on top of ’23, which turned out better than what people thought, certainly bolstered by investment from a number of our Chinese customers, our more established customers, really adjusted CapEx a fair amount across all the markets.
And so as we move here into ’24 where we see a WFE level that everybody does the math a little bit differently, but I would say somewhere in the low $90 billion range in 2023, and we’ll call it a flat to maybe up environment in 2024.
And we think it’s pretty constructive in terms of conversations with customers, improving business fundamentals for our customers in this year that will ultimately lead to — I call it like devices get better this year and you start to see investment in capacity as we move into 2025. Coming out of earnings, the March quarter was, we think, the bottom for the company and expect sequential growth moving forward as we progress through the year. We think the second half is high single digit growth versus the first half as we go through the year.
From a semi process control point of view, I think that the business probably grows a little bit as we move through this year. From a relative point of view, I feel pretty good about the relative performance of the company this year.
Over the last couple of years because of some of the challenges related to supply chain a couple of years ago and how that moved and translated into what happened in 2023, I think it skewed or created some noise in some of the relative numbers. But I think on a normalized basis, we feel pretty good about how we’re positioned as we — as we move forward into this year and into next year as you start to see more investment at the leading edge, which tends to drive the highest process control intensity, the part of the market where KLA tends to participate the most.
So we’re encouraged by that. Our service business is growing. Our service business is a really important part of KLA. Carries a contract stream. It’s a high complexity business, it’s relatively lower volume. It’s a high mix business. And as a result of that, our customers tend to buy what they need and then — and then have pretty high uptime requirements. They won’t need matching requirements, and so they tend to drive a higher contract penetration for us.
And we shipped a number of systems in ’21 and ’22. Those systems will come off a warranty. We have a 95% attach rate in terms of contract. We’re seeing rising utilization rates. So our service business is likely to grow somewhere between $250 million and $300 million this year, more likely towards the higher end of that range based on some of the dynamics we’re seeing today.
In our EPC business, which is electronics, packaging and components, which is where our specialty semiconductor business is, it was very power centric last year in terms of power semiconductors, translating more into advanced packaging this year in terms of some of the growth opportunities. So that business looks to continue to outperform the market as it has over the last couple of years. Some weakness in our component inspection, our PCB business, which has been more consumer electronic focus, mostly around mobility.
And so I expect those businesses to improve marginally this year. But as we start to see more capacity investment in those markets as you move later — this year and into next year, I expect that we’ll start to see more recovery there. So semi-PC grows a little bit, service grows a lot and EPC is kind of modest mid-single, maybe high-single-digit kind of growth, at least in terms of how we see it today.
Question-and-Answer Session
Q – Harlan Sur
That was a great overview. Thank you. I’m going to start off with a high-level question first, and we’ll get into the details. But long-term question. If we look at spending dynamics in the industry over the past 10 years, including last year, the downturn, wafer equipment spending has grown at a 12% CAGR versus semi-industry revenue growth CAGR at 6%. We estimate that KLA has grown its process control systems revenues at a 13% CAGR over that period of time. So clearly growing even faster than WFE.
Given the complexity challenges, more focus on yield and cost, the need for more installed capacity to support strong demand trends in your advanced foundry and logic and memory customer base. Do you anticipate WFE and Process Control continuing to outgrow semiconductor industry revenues? And what are going to be the biggest drivers?
Bren Higgins
Yes. No, that’s a great question. And there’s the industry piece of it, which is just rising capital intensity that’s been effectively happening over the last decade or so. And we’ve seen it accelerate with the introduction of EUV, which has driven scaling back into the industry. We went through a period of time there where WFE actually had a negative CAGR from 2013 to 2018. And most of the investment was happening in memory. And a big part of that was retooling from planar NAND to 3D NAND.
So there were dynamics in the industry that had the growth of WFE that was somewhere, we’ll call it GDP plus or so. And then over the last since 2019, we’ve seen this accelerant that you talked about. So I think, over time, if you think about what drove capital intensity down for a number of years in the industry, most of those issues have generally — or those opportunities have played out, there are no more wafer transitions, consolidation’s happened in the industry, the transition from fabless to foundry has happened.
So on a go-forward basis, consistent with what we’ve seen over the last few years, I think you continue to see WFE grow a little bit faster than semiconductor revenue. And of course, there’s a lot of interesting drivers that are driving semiconductor revenue to grow at a much faster rate.
As it relates to process control, as we’ve seen this in pickup and the logic foundry part of WFE, particularly from 2019 forward with the introduction of EUV, that that tends to be a more process control intensive part of the market where process control intensity can be somewhere in the mid-teens as a percent of WFE where memory can be somewhere closer to 10%. So it’s a much stronger part of the market. And so that plays well for the relative performance of the process control market relative to WFE.
If you also look within that, what you’re seeing from in terms of design starts and more compelling Moore’s Law attributes at the leading edge nodes, you’re seeing more design starts. A high mix of designs tends to be good for process control because designs come in at different times and test design rules in different ways.
So it drives customers to sample more. So that tends to be a nice driver. You’ve got process new tech, we talked about EUV, we’ve got what’s coming in terms of things like gate-all-around and buried power rail, so power distribution changing, all these dynamics create new defect mechanisms and metrology requirements.
The demand and the follow through demand that you see in terms of designs tends to translate into lower capacity reuse, which has also been a good factor for process control growth. So I think when you take all that in the aggregate, it sets up process control to grow a little faster than WFE. And then if you look at the markets that KLA is participating in, the markets that are growing the fastest are things like optical inspection.
And optical pattern inspection, because of the position that we have, so the inflection of that market or the relative growth of that part of the market within process control creates opportunities, I think, for KLA to grow a little faster than that. So we think that the setup is pretty constructive and it’s what we talked about at our last Investor Day when we laid out our ’26 plan in terms of some of the assumptions. And we feel pretty good about where we’re trending right now.
I think, moving forward, I was supply constrained, still am a little bit in terms of optical pattern inspection, but it’s getting better. We’d expect that part of our business to grow faster than market this year. And I think we’ve got a pretty good trajectory in terms of new capacity coming on to support what is I expect to be — continue to be one of the faster growing markets in WFE.
So I think it’s pretty good for WFE and pretty good for process control and KLA.
Harlan Sur
Maybe more near to mid-term. So as you step into the second half of this year, it does feel like advanced node momentum will start to accelerate, both in foundry and logic and in memory, right? And appears to be reflected in your confidence on improving spending outlook this calendar year. Additionally, we’ve got industry cycle and demand dynamics continue to improve. Given your relatively longer lead times, your critical role in enabling these advanced technology migrations, you talked about having visibility on next gen two nanometer and three nanometer, very complex logic designs. How are customer discussions, initial forecast visibility and outlook for calendar ’25 shaping up for the team?
Bren Higgins
Well, it’s a little early for me to quantify exactly how big, but I think if you — and we talked a lot about this in the earnings call is that the intensity of the conversations and discussions with our customers around slot and capacity planning has picked up.
In some ways, things are mostly what we said three months ago in terms of the shape of the year and outlook, but we’re seeing more specificity in terms of some of the requirements and a step up in terms of some of the long-term planning assumptions as we move into ’25 and ’26 for the leading edge.
One of our objectives in the company this year is about preparing — supporting customers, which we always do, driving our roadmaps for our programs, which are big bets for us to make sure we’ve got the right products to support the portfolio, but also preparing for growth at the leading edge, which is our strongest market and making sure we can ramp our most advanced solutions for customers. So we feel those conversations are improving. And in terms of specific slots and where we thought we were in a pretty good place relative to some of the demand, then our customers said, well, what would it look like if we had more advanced needs or higher level requirements?
And so we’re working closely there and would expect that we’ll start to see orders in the second half of this year to support capacity investments into next year. Also, good competitive dynamics. There’s investment that’s happening more broadly at the leading edge as well.
I think memory is — certainly DRAM has been something that has expected to improve a little bit. I think memory or DRAM has been impacted by a lot of China investment, which isn’t driven by some of these leading edge requirements that we’ve been talking about. But would expect, given the dynamics we’ve seen in memory, what’s happening in HBM, the silicon trade that’s part of high bandwidth memory and the introduction of more EUV layers into DRAM into next year that it’s encouraging from that point of view too and the discipline that we’ve seen over the years. But there’s still utilization rates are improving, but they’re still low.
So I would expect customers to do technology investment to satisfy some of the supply requirements. But as we move into next year, I think we’re set up constructively for more investment there.
And then flash, I think it’s been at very, very low levels, don’t see it changing much this year. But over time, I would expect we’ll see some investment there as it’s been — basically it’s been a couple of years at pretty low levels of investment on that front.
Harlan Sur
KLA is — we’ve written about this every single year now, if you like, for a decade now. But KLA’s number one market share position in process control continued strong, right? We just got the market share numbers for 2023. You actually expanded your lead over your number two competitor, right? Last year, I think you were 4.7 times, 4.5 times larger than your number two competitor. Obviously, you guys had a strong number one position. In ’23, you were six times larger than your number two competitor.
Process control segment outgrew WFE. This is ex-litho by about 100 basis points. KLA grew more in line with WFE. So you actually did lose slightly less than 1 percentage point of share, although your share still sits 300 basis points higher than your average share over the past 15 years. You were strong number one in five out of the eight sub segments in process control. You were number one or number two in seven out of the eight subsegments.
You did drop to a number two position in mass inspection. You also attributed some of the share loss to the China leading edge restrictions that were put in place in the second half of ’22. Just help us understand the share dynamics last year and how your position looks — looking out over the next few years?
Bren Higgins
Sure. Well, part of the position for KLA is you talked about all the markets we’re in, right? And across those markets, we have a number of different platforms within the markets that we ship to customers. And most of our competitors are more point product competitors. And so they’re trying to necessarily compete with us in ways where you try to have a one size fits all.
I think one thing that KLA does pretty effectively is, if you just think about optical inspection, we have about four or five platforms that we ship that meet, in some cases, very high end technical requirements for customers that are doing debugging and process integration work to more capacity centric systems.
So I think that our ability to offer somewhat agnostically, our customers, certain capabilities that either meet their technical requirements or their economic ones makes it — makes us pretty formidable from a competitive point of view. And the ability to have software capabilities to try to leverage the data that’s coming out of the systems with a network effect across the products helps, I think, speed that time to results that’s so important, because at the end of the day, we’re generating a lot of data, probably more data than anyone in the fab. And how do you make that data actionable across all the different tool types and understanding what’s happening randomly, but also what’s happening systematically on the wafer is a factor competitively for us.
Market access — absent some of the restrictions that we were dealing with that precluded some market access for us, we think our share would have likely been up a little bit. So that was certainly a factor. I think the constraints around supply for optical inspection. I would ship more if I had more capacity. So that was certainly a factor as well.
On the reticle front, it’s a market where we gained a lot of share last year. We lost some share this year. It’s served by a number of platforms. The integers can be pretty large, so it can skew the numbers. I mean that market is more or less a 50-50 overall in terms of our position relative to our competitors. So there was some loss in that market, we’ll see how this year translates. I expect that business to grow in 2024 faster than the overall market. So I think we’re in a pretty good position overall. But in general, I think that’s a way to think about the overall share dynamics. We have a goal in the company. We want to grow share over time about 0.5 point a year. It doesn’t obviously happen linearly like that. In 2022, we gained over 3 points a share.
But we do think that if we continue to operate offer good solutions to our customers that meet some of the technology requirements that they’re asking for can speed time to results, can give a return on investment in terms of time to ramp or improved yield that customers will spend more with us. And so there’s a trajectory here that we think we can continue to drive share higher over time if we can continue to execute on our product roadmaps.
Harlan Sur
Yes. And I think what we’ve noticed in covering KLA for over 20 years and during periods of time where some of your point solution competitors move up the share ranks couple of positions, right, and if you fast forward sort of two, three, four years out, you see the share come right back to sort of KLA over time, right? And I think part of it is a reflection of the R&D scale that the team has, right? When it comes time to put more R&D effort in certain segments, you certainly have the scale and the flexibility to do that. But strong showing on the share dynamics in ’23.
Within the deposition and etch segments of the market and your peers there, right, China domestic competitors have been making some relatively strong progress on share gains, like in deposition, the China domestic competitor moved to the number five share position last year from number eight in etch, right, the two China competitors move to the number four and number five position of share respectively. But in process control, you have one China domestic competitor, number 11 market share position in ’23 unchanged from ’22 and less than 1% of the overall process control market, so quite small.
Process control is so critical, but the systems, the software are so complex, right? So question is do you see other smaller China domestic competitors coming to the market or are the barriers to entry just too high?
Bren Higgins
Well, there are more than one. There are a number of them out there. And if you look at where most of the investment is happening, it tends to be in the legacy design rules where we still offer the best solutions for customers, but certainly a technology that could be 10 years old or more that the competitive advantage is less. So we do see within China, particularly in some of the restricted fabs some competition. Say it’s a factor in the market access point I made earlier, as is some of our competition that isn’t U.S.-based competition that’s able to serve parts of that market.
So we’re seeing some of it. And like I said, I think it’s somewhat constrained to China. I think there’s a push to particularly where there’s public money involved to support the local system there too. So I think customers are investing here and there. But as you can tell from just the overall share position of the company that our customers are relying on us to help them ramp even in older design rules, given the nature of the portfolio and some of the solutions that we can offer and engagement we can offer is that I think that we’re winning the lion’s share of that business and we continue to expect to do that.
We have a lot of our customers are people that we’ve worked with in the past. And so used to buying our equipment and working closely with our teams. So those are factors out there. I think absent, if we had access, we would compete. I think it would be less of an issue. It’s a small issue today. But I don’t think — outside of the market access comments earlier, I don’t think it changes the overall numbers all that much, at least as it relates to direct domestic competition.
Harlan Sur
Do we have any questions from the audience. If there are any questions, just raise your hand, we’ll get a mic over to you as quickly as possible. No, okay. Something — there’s an interesting dynamic that’s happening in the memory sector, right? In the memory sector, for those of us that have known KLA for a long time, memory has always had a lower process control intensity. But there are dynamics that are pushing more advanced logic like manufacturing dynamics into memory. For example, in HPM DRAM requires this advanced logic chip that sits at the very bottom of the memory stack, right? That’s using advanced CMOS logic technology.
And then in NAND, the peripheral controller chip, which is logic is being processed now separately from the NAND stack and then sort of bonded together, right? And additionally, this whole bonded technology process appears to be just a nightmare from a manufacturability and a yield perspective, right? And so are these dynamics and any others sort of slowly driving memory process control intensity higher over the next few years.
Bren Higgins
Yes. No, you mentioned some of those the embedded logic part of the chip, which is now creating more complexity. And in classic memory, you tended to have more commodity devices and redundancy. But now with the inclusion of logic. And if you look at more advanced like high bandwidth memory stacks where you’re looking at a larger dump, which tends to have more of a yield challenge, you’re stacking multiple die, it’s eight die or so together. So you want to make sure all those die are working. So you tend to have higher sampling rates. You have the introduction of EUV into DRAM.
Harlan Sur
That’s right. That’s right.
Bren Higgins
That’s going to — and additional layers that are coming there, which creates some unique opportunities. We have seen process control intensity start to skew up favorably. And I think that historically is to say, DRAM is probably 9% or 10%. I think it’s probably closer to 10% to 11% today.
On the flash side, you’re increasing layer counts. So you have overlay challenges. You have wafer to wafer bonding, which drives new issues with wafer quality and wafer flatness. You have more metrology requirements given the nature of the taller stacks.
So I think there are opportunities on the flash side as well. I think it’s improving, but the investment levels have been so low for a little bit of time, I’d like to try to understand some of those dynamics a little bit better. But generally, these trends are we think going to be positive for our business. And so I think the memory, while it won’t ever look like logic and foundry, I do think there are opportunities that we’ll see in memory moving forward.
Harlan Sur
On the product front, we talked about optical pattern wafer inspection. Despite the team’s continued advancements on optical or broadband plasma pattern wafer inspection and superior throughput sensitivity in high volume manufacturing, it’s interesting, we still get the questions on sort of the threat of e-beam pattern wafer inspection.
But over the past five years, including last year right optical inspection has outgrown — has grown at a 23% CAGR, e-beam has grown at a 9% CAGR, right. And over the past three years on average optical inspection market has been five times larger than e-beam. So optical inspection clearly continues to grow quite strongly. You’ve got 90% market share in a $3.5 billion optical inspection market. These are your Gen 4, Gen 5 broadband plasma inspection platforms. I know the team is working on its next generation broadband plasma light source. But does the team see optical inspection scalability well past the two nanometer node?
Bren Higgins
Oh, yes. I mean if you look at where we are today, even with Gen 4 is out-shipping Gen 5 in terms of the demand in the marketplace. So one of the challenges or one of the thesis is that it’s proven to be an incorrect one is that somehow the extendibility of optical would get challenged and you’d have to flip to e-beam.
And if you understand the technologies there’s a significant compromise in sensitivity, which, across all of our platforms, you’re always trading sensitivity against throughput or volume, but an e-beam solution is incredibly slow. So you can have higher levels of sensitivity, but it’s so slow that you can’t inspect the full wafer and it doesn’t make sense from a production perspective. You would need fabs to be twice as big to accommodate all the tools you would need to be able to implement an effective e-beam strategy.
So what do customers do? There are complementary technologies, they tend to use optical to — they do defect discovery with e-beam systems. And then as you ramp a fab, you tend to use optical systems to monitor for defectivity in the fab. And customers will use the e-beam systems as reference tools to check what the optical tools are finding.
So we find the broadband systems, so the broadband systems, which means that you have wavelengths that are quite large and you can actually then scale and see different kinds of defects depending on the different wavelengths you use. And so it’s unique to the industry where most of our competition are all our competition is a single wavelength solution. So it tends to allow them a lot of flexibility in terms of how they manage defectivity.
And then again, you can use e-beam tools as a reference point for your inspectors. We are investing in e-beam and trying to leverage our AI capabilities and then this network across the tools to point our inspectors better. It’s an investment that makes a lot of sense for KLA because of the optical franchise that if we can increase the relevancy of optical and speed time to results by leveraging e-beam capability, then it makes a lot of sense for us.
Overall, it’s about 80% to 85% of the inspection dollars gets in optical solutions. I made the comment about Gen 4 versus Gen 5, you mentioned Gen 6 in terms of light source, but also optics that we’re investing in. So we feel very good about the optical inspection roadmap for the next decade or so.
Harlan Sur
One of the new dynamics of demand that have been sort of injected into WFE spend has been this focus on advanced packaging, whether it’s HBM, DRAM, we had the Micron team here today, they are thinking like 50% HBM sort of bit growth CAGR for the next few years, right? And then you have 2.5D packaging, 3D packaging technologies. All of this seems to be driving incremental WFE spend for you and for your peers.
Team has a strong position — KLA team has a strong position in advanced packaging, both in processing, right, like for the redistribution layer or inspection for both packaging and substrate inspection. Good to see that you’ll be run rating at a $400 million sort of annualized level this year. What does the multiyear revenue opportunity look like for advanced packaging? And more importantly, once the team focuses on a particular segment, we typically do see a lot of product innovation from KLA. Wondering if you can maybe give us a sneak peek, some insights into the new product pipeline around advanced packaging.
Bren Higgins
Yes. No, it’s a great opportunity and something I think grows faster than meaningfully faster than WFE over the next few years. We were about $300 million last year. It’s about $400 million this year. And we come at the market with a number of products, both inspection and metrology products, but also in our process business. So if you look at — on the process, you mentioned the redistribution layer, but you also have barrier seed and plasma dicing. You have TSV via reveal. So there’s a number of markets that we can compete there. So if you think about the $400 million, about $200 million of it is in process tools.
And we have tools that are uniquely designed for some of these specialty markets. And as those markets have inflected, we’ve been able to drive good opportunity and I think differentiate relative to some of our competitors. We also monitor chemical properties and plating applications when you’re depositing metal films and into the TSV trenches that we acquired through an acquisition called ECI a few years ago. So that’s a nice business that we participate in as well.
On the inspection metrology side, over time, you’re going to start to see more complexity in these connecting layers on the logic side. And as you do that and you don’t — as you increase the value of these heterogeneously assembled chips, you’re going to see probably an increase in overall sampling, but you start to shrink the feature sizes you need more capability. I talked about the trade-off between sensitivity and throughput earlier.
And so where we think there’s, A, the overall market growth as an opportunity, but B, that as more advanced solutions are deployed that there will be a throughput hit. And so for customers to maintain sampling rates, then they have to have to buy more tools.
So you have an ASP effect, but you also have potentially a volume effect because of that trade-off. There are some unique requirements in packaging as it relates to handling and as it relates to operating and environmental conditions. So that’s driving some engineering requirements. But we think we can do versions of tools that are already in our front-end portfolio.
What’s been so interesting is how we’re engaging with a lot of front end people now on back end issues, which used to never happen, right? There was clear segmentation and our relationships were very different. Our customers are pushing us towards one sort of channel approach in terms of how we engage with some of our biggest customers to have a consistent kind of strategy across these opportunities.
So we think, going forward, it’s a big opportunity for us and we’ll evolve. And what we have seen historically as complexity increases, it tends to drive more towards higher value solutions with higher complexity we can differentiate and run the KLA model, which is, I think, pretty well understood by investors.
Harlan Sur
Your services business, right, 20% — 25% of your overall revenues has historically been a great buffer against overall WFE volatility, falling only once in the last, I don’t know, 20, 21 years, growing through the last five downturns, including last year where it grew 7%. What drove the resilient growth profile last year amidst a pretty bad downturn when utilizations were at record lows for your customers and more importantly, what’s driving the reacceleration back to your target sort of 12% to 14% growth CAGR in services over the next sort of few years?
Bren Higgins
Yes. So it was below our long-term target last year. But utilization rates dropped off. And when they drop off for the industry, they drop off for us, but typically not as much. If you think about in a capital constrained environment, being able to drive yield is — tends to be something our customers will do and only start the amount of wafers that they need to start. And so we tend to see our tools continue to be used although it does have an effect from a utilization overall.
So we were sort of 7% or 8%, whatever it was, something like that last year. And I think that’s one of the factors. I talked earlier about the high complexity nature of our systems. And so our customers typically rely on us. They don’t have a lot of volume. They can’t really put a lot of their own resources into servicing equipment. It’s really hard to do given the complexity of both the parts we have but also the nature of how the systems operate. And so that tends to drive that contract structure.
We can do a lot from a cost point of view to try to be as efficient as we can, understanding what we have to serve to with our customers from an entitlement point of view and what they pay for in their contract. And so it creates a resiliency in the overall service business that provides a nice level of predictability and an anchor for us despite what’s happening on the CapEx side of the business.
I mentioned earlier that systems are coming off warranty, utilization rates are rising, but tools are living longer now. We’ve seen — if you would have asked me five years ago, I’ve seen the number of the average years go up from 13 into the mid to higher teens now in terms of years of service. We have tools out there that we’re servicing still today that we shipped in the 90s. And so the extending life of the systems is another factor. So the growth in the number of systems extending life and the amount of revenue we’re now getting over time, lifetime out of the systems where it would have been 40% to 50%, 10 years ago, we’re now somewhere above 70% to 80% today. So we think those trends play out over time.
And I think the dynamics as it relates to our business that are a little bit different than a consumable centric business like some of the process tool guys deal with creates a little bit more visibility in terms of what to expect and then we can — I think do a lot to optimize and drive leverage out of that revenue stream.
Harlan Sur
Last question, I know we’re out of time. Last Investor Day, financial targets for calendar ’26, $14 billion in revenues, which implies mid-teens growth from here, 63% gross margins, 41%, 43% up margins, $38 in earnings power. This is of course, before the current downturn.
Did the team factor in the downturn in your forward modeling assumptions and your confidence on driving to those 2026 targets?
Bren Higgins
Yes, we did. I mean, look, if you look at WFE growth from ’19 to ’22, it was about 21%, right? So in an industry, even if semiconductor revenue is growing, 6%, 7%, 8%, hard to see how you’re going to have WFE maintain that. So we knew that there’d be some correction somewhere in there, which is why we had a four-year model.
Obviously, it’s predicated on the amount of WFE spend, But I think that there’s encouragement about what we expect over the next couple of years and some of the dynamics we’ve talked about that will drive WFE meaningfully into the $110 billion to $120 billion range by ’26. And so we feel pretty good about that. Our share of WFE, our original plan was 7.5%. We’re trending in the mid 7s now, maybe a little bit higher.
So we think our share of market is ahead of the plan. Service is probably ahead of the plan. EPC is probably a little behind the plan because of the post-COVID downturn in some of the electronics markets. But we think we’re pretty well positioned. We’ve made investments. I think the leverage will be there to drive the margins and the EPS that you mentioned. So that’s — we feel very good about the plan and we’re driving the company towards that.
Harlan Sur
Great insights. Thank you for joining us today, Bren.
Bren Higgins
Yes. Thank you. I appreciate being here.
Harlan Sur
Yes. Thanks, Bren.