Everyone, welcome to day three of Citi Global Technology Conference. My name is Atif Malik. I cover U.S. semiconductors and semiconductor equipment stocks. It's my pleasure to welcome Bren Higgins, Chief Financial Officer and Executive Vice President of Global Operations, and Oreste Donzella, Executive Vice President Of Electronics, Packaging and Components, also known as EPC Group at KLA. I'm gonna kick it off with my questions first, and then I will open up the question for the audience. Welcome, guys.
Yeah, thank you for having us. It's great to see everyone in 3D again, right?
Yeah. Yeah. Bren, Oreste, I'm gonna take you back a couple of months to your summer investor day, which I thought was wonderfully done. There were a few interesting takeaways, at least for me, which were a bit unexpected, and I thought it's worth it for investors to know those things as well. Maybe we'll start with the EPC business first. Oreste, I think Electronics, Packaging and Components business through Orbotech acquisition has been a home run for the company. It is very well diversified across many end markets, including data center and auto, has exceeded your profitability and synergy expectations. It is expected to grow faster than the core semiconductor process control market in the next four years.
Investors are excited about EV going into next few years. Oreste, I thought you had some interesting case studies on silicon carbide and advanced packaging in your EPC presentation. Can you recap for us what is KLA's position in the auto market and how well you play in that segment?
Yes. Thank you for the question. Let me step back a little bit. Going back to 2019 was a pivotal moment for the company because this was the moment where we decided to acquire Orbotech. With Orbotech, we got really three different markets in the same transaction. We got access to the printed circuit board and advanced packaging. We got access to display, we have access to the so-called specialty semiconductor that is highly exposed to automotive market. Since then we applied the so-called KLA operating model. As Atif said, we materialize more synergy in terms of cost reduction than we were expecting at that time. Our profitability grew more than we planned ahead of our acquisition, and we are keeping the pace of a double-digit revenue growth every year.
A couple of areas where we are very excited about that became bigger exposure of KLA after the EPC formation and the Orbotech SPTS acquisition, as Atif said, one is automotive. Automotive is not only PCB. Actually, my group represents the entire KLA in the automotive ecosystem. When I say automotive ecosystem, I say the customers that we sell tools to, like the tier two suppliers of automotive industry, but also tier one automotive OEM. We created like partnership in all these automotive ecosystem to generate some new machine learning-based software solution that we call I-PAT to screen the chip reliability in fab. This was absolutely revolutionary because if you look at the automotive market right now, it's going through secular inflection. It's going to go through huge set of innovation and changes.
One of the biggest concerns of the automotive OEM right now is the reliability of the chips in cars. Because thinking about when you are driving a car in autonomous way or semi-autonomous way and the chip is going to fail, you are in trouble. They want to make sure that not only the chips are responding to the functional electrical performance, but also reliability performance. We decided to create this machine learning-based software solution that is attached to our inspection and metrology tools and now is creating a new standard in automotive industry. On top of that, with SPTS, that is part of EPC and part of the Orbotech acquisition transaction, we generate a set of new solutions for process tools, not inspection and metrology.
These tools are mostly deposition and etch solution that are customized for these particular automotive needs, in particular for the power. We have a customized solution for silicon power and also for silicon carbide power. Let me talk a little bit about silicon carbide, because it's been such a big engine for growth for us in process control and in SPTS process. Silicon carbide is a very, very tough material to deal with. Of course, silicon carbide is going to be the future of EV industry because it is a much better material in terms of thermal performance, power efficiency performance, and people are transitioning as fast as they can from silicon-based wafer to silicon carbide-based wafer. However, it comes with a lot of challenges. The first challenge is, it's very hard to control the defectivity.
It's very hard to make a process integration with the silicon carbide substrate. It's very hard to control the silicon carbide substrate, the quality itself. We developed a lot of solutions in KLA to make sure that we properly inspect and measure the silicon carbide wafer at the substrate level, at the wafer processing level. We also design a new solution for etching, for example. We have a new solution that creates like this trench etch, rounded bottom that are really, really important for the performance of silicon carbide device itself. Again, really good opportunity.
Just to give you some numbers, we don't publish revenue by end markets, but I can tell you that in the last three years, our revenue in automotive space almost tripled. From 2019, 2022 based on the current forecast, and the booking is even more than that. These are incredible growth that we see right now, and we see ahead of us. This is for automotive. I can talk about packaging as well. Packaging is the second big vector of growth coming in our EPC group and not all EPC group, but across KLA. Of course, packaging is a new enabler for semiconductor technology roadmap. Of course, with EUV insertion in a semiconductor front-end, we are seeing another leg for Moore, but it cannot be done only semiconductor front-end anymore. It's too expensive.
The packaging is really enabling the next technology roadmap in semiconductor, in particular with the new heterogeneous integration of multiple dies in a single package. We have a huge play in that area as well as a KLA, including EPC. Last year, based on VLSI Research, finally we came in the top 10 suppliers for packaging. In three years ago, we're not even mentioned as a supplier of packaging. Now we are already top 10. We created inspection technology to inspect these very, very advanced wafers in packaging for sub-100 nm inspection sensitivity right now. We are going through the roadmap to achieve this kind of sensitivity. We also have a new technology in process, like for example, plasma dicing.
The process of singulating a wafer is extremely dirty right now with the saw and lasers. We are creating a custom plasma etching for singulating the wafer in much cleaner way, keeping the strength of the die, and we are a technology market leader with SPTS division on this. Also packaging, actually the growth was quite significant right now, three times more in three years in, including what is happening from Orbotech NDI and the PCB that we call substrates.
Okay. Oreste, you talked about double-digit future growth for EPC business, and obviously auto is growing very fast and then you have maybe data center also growing pretty fast. How about other areas that are more consumer-centric, the smartphones and the PCB? What assumptions are you making for those end markets?
We are seeing weaknesses in our PCB business tied to mid-end and low-end smartphones. No question about it. We see weakness in the smartphone demand, and we see some weakness in our PCB portion tied to the conventional PCB serving the smartphone market. In that kind of area, we don't see the double-digit growth, for example, is going to be diluted in terms of growth percentage year-over-year. I see more in the single digit type of growth at the best. Again, even in PCB, we have a huge opportunity because the high-end segment of PCB, the so-called substrates, that is tied to this heterogeneous integration packaging. Again, the beauty of EPC Group is it's diversifying KLA overall, and it's diversified inside EPC because within EPC we serve multiple markets for multiple applications.
Whenever a market is slowing down a little bit, like mid and the low-end smartphone, we have automotive going up, and then we have packaging going up. We have this kind of interesting dynamic that is completely new to KLA, but is absolutely addictive to the growth of the company long term and creates a huge diversification and of course more stability to the company.
It was important for us to expand our presence where our customers are investing more, right? Obviously, we have a very strong front-end wafer fab equipment position, but we're seeing our customers move and for capability or for cost, and we're seeing more value in some of these areas. We wanted to increase the company's exposure to that. We were looking for market-leading positions where technology was evolving, more value was shifting, and we knew that we could apply the operating competencies of KLA, both in terms of just operating and trying to position those businesses for corporate-level operating leverage, which we've been able to do with the synergy reductions and other cost actions we've done there. That was important. We also wanted to make sure that we were able to enhance customer engagement.
We knew we could engage with customers at a higher level and be able to. Sometimes small suppliers have a hard time getting attention at the right level from customers. With KLA, sort of a two-way street. On one hand, you get the meetings, but you get the meetings because there's an accountability there, and we have to deliver to that. We've done that, I think, pretty effectively as well. Then the third thing, which is coming, and Oreste Donzella talked about some of the future opportunities, is how do we take KLA technology and know-how and then start to apply it to those products in future product development.
We think that there's that still to come, and it's part of the plan that we talked about in terms of new opportunities, particularly in the higher value areas, substrates being the most obvious in terms of how it integrates into the package. Also demonstrating that, you know, we can run process businesses well too, where we can differentiate and we can understand the value of that differentiation, and we can apply our the operating competencies to deliver on the profitability of those markets. We segment reports, you get some pretty good visibility to how specialty semiconductor looks in terms of revenue levels, but also gross margin.
During the last point of taking the learnings from semiconductor business and applying to EPC business, I feel like software plays a big part in the semiconductor tools. You have a higher attach rate of softwares, your gross margins speak to that. Are the auto products also have a software feature in it that will help your gross margins over time?
Yes. We have the, as I said, the I-PAT, that is a machine learning based software product. That is attached to our hardware actually, so you can use these I-PAT software available on our inspection tool and making sure that you do proper screening of potential reliability failures after test. Actually, we have done this in collaboration with not only the direct customers, but also with the automotive ecosystem as well. This particular software methodology is inside the Automotive Electronics Council guideline for automotive semiconductors. It's recommended from really the organization that is driving all the specs for automotive semiconductor. Also software is giving us other opportunities. If you look at the packaging fabs today, they are not automated at all.
What really made the semiconductor industry absolutely outstanding in the last many years is all the software that you use in the semiconductor fab to leverage data and make your process better. KLA created all this software in semiconductor front end. We are moving now to let our customers understand the importance in manipulating and analyzing the data in the packaging industry, even PCB. We are creating all this software portfolio, software solution that eventually will be deployed in the next one to three years, either in packaging and in PCB fabrication. You absolutely, your question is very well taken.
We are doing that because we believe that the ability to store and analyze the data inside the fab is absolutely critical to improve yield and the process integration of these devices.
Great. Let's talk about services. One of the reasons the semi cap companies sound more confident about next year or the resiliency in the market next year is services. It has been growing nicely for all of the companies. I noticed that your services growth forecast of 12%-14% is a bit higher than 10% number that we hear from the process equipment peers. I was curious, I mean, what is driving your services growth, and are there unique drivers for your tool set that are different from deposition and etch-type equipment?
Yeah, no, it's a great question, and our service business behaves a little bit differently than the process guys. It goes back to how customers buy process control. Since process control isn't necessarily part of production in getting outputs or yield management, good output. Customers tend to try to buy what they need and then run the tools as hard as they can, right? They buy what they need in terms of capability. They don't wanna spend any more than they have to, and then they try to keep the tools up. You take that, and you couple it with the complexity of our offerings and the need to be able to match performance across tools and so on.
It makes it very difficult for our customers to be able to do their own self-service. Also, the supply chain tends to be fairly exclusive to KLA, so the ability of our customers to go out and buy parts to try to do it on their own, really, really hard to do. As a result, we have very high contract penetration, and our contracts are tied to a certain entitlement levels that we can customize to a significant degree, by tool types, by fabs, response times, inventory levels, all of that, and allows us to kind of optimize the cost structure underneath, and then we deliver to that level of entitlement. There's a much more predictable stream that comes from the contract.
We have the sort of protection of the parts in this thing, and so customers really aren't able to do a lot of self-service. They also don't have that many of them, right? So if you have, for example, 200 etchers that aren't as complicated, you can build your own capability to go service those tools. You have 10 broadband inspectors or 12 broadband inspectors, our guys can only service one tool, let alone a customer trying to build a whole team to be able to go and build the competency to provide that service or do that service on their own. So those are some of the factors that are unique to us. It drives the contracts. The contracts are penetration about 80% in terms of recurring contracts, of the revenue stream.
That's one of the factors. We're seeing in our tools, you know, the useful life of our tools increasing. As customers and Brian Lorig, who runs our service business, did a chart and showed, I can't remember exactly the numbers, but how many of the tools that we shipped in 1990-1 995 are still in service today. It's an astonishingly high number. In fact, you can go back to, like, KLA and Tencor, the origins of the company, and there's 80% plus of the tools are still in service today. It gives you the sense of the legacy environment of these fabs, and they continue to run, and this is all depreciated equipment.
These, they kind of print money in a lot of cases to continue to build parts for toasters and washing machines and everything else. We're seeing the useful life as demand has become broader, particularly as it relates to automotive and communications infrastructure and industrial IoT, that you're seeing customers use these tools in different ways. There's also increasing and changing specs, as Oreste talked about, so that's changing how they do process control, which is in some ways and is therefore kind of translating into our service business a bit. That's a factor. Of course, we've had great relative growth over the last few years, particularly as foundry and logic and has moved the way that it has in terms of the percentage of the mix, but also the leading-edge environment.
You have more complex tools, cost more, higher ASP, higher contracts. Customers are relying on us even more as they ramp new facilities in different places to provide service. I think all those are drivers for us in terms of growth going forward. We talked about 9%-11% and outperformed it pretty significantly in terms of the through cycle growth rate over time for service. We talked about that at 2019 investor day. In some ways, adjusting to 12%-14% was just reflecting the reality of where we've been operating. Historically, service grows every year. We've had one down year in 2009 in our service business, and it was only down about 12%. Because customers, even if they pull back on investments in CapEx, they continue to run the install base they have.
Our stream allows us to, you see, the growth there. Maybe, you know, it slows down in contracting environments a little bit, but still, you still have positive growth. Then the cost optimization that happens underneath it, given the visibility we have to the stream, allows us to generate profitability levels that are higher, we believe, than the corporate averages. It's a bigger part of the company today. Back to your point of just confidence, and it's a strong anchor in terms of how we think about predictable streams and how it drives how we think about debt in the capital structure, how we think about an ongoing dividend because we have this business that performs this way.
Great. Let's talk about the target model. Bren, you presented an impressive 2026 target model with revenues of $14 billion and EPS of $38. Can you talk about the underlying assumptions you're making about wafer fab equipment in your model and the profitability goals?
Yeah, sure. It's we looked at obviously where we're at today and the mix of business going forward, and given what we've seen in terms of broader demand, foundry logic, 55%-60% of WFE investment moving forward. We've seen rising process control intensity. KLA share has also improved in this time frame. The growth that we've seen as foundry and logic with the introduction of EUV and the robust design environment that we've seen that started at 7nm has driven our customers to invest more in sort of production monitoring in a mature state because they're managing more designs, they're managing more process flows. That's been a driver in terms of our participation in production. We've always participated a lot in development and the ramp phase. Production, a little less so.
There was a period of time there where the design follow-through of second designs behind the leaders was fairly limited. The node benefits weren't all that compelling for lower volume parts. As a result of that, EUV was delaying, so customers were able to get a lot of efficiency out of their capacity by migrating it to the next node. As we've seen started at 7nm, when TSMC opened up the design libraries, and we saw the introduction of EUV and more compelling attributes of Moore's Law that drove this design activity, it has forced our customers to invest more in the current node, and they can't really leverage any because the demand is strong in the previous node. For 5nm, it's a whole new buy.
Now, it's also a harder buy because the technical requirements are changing because of some of the technical changes, mostly around EUV, but even architecture changes that are coming. All of that created this environment where we believe we'll see this, a growth rate of the process control part of the business, a more logic foundry, but also process control growing. Semiconductor revenue, lots of forecasts of 1 trillion out there in 2030 or so. You just do the math from 2021, you end up sort of in the 6%-7% kind of CAGR range. Capital intensity has been rising since 2012, 2013 in our industry, and our view is it'll continue to rise, but as a result of that, you'll see WFE grow a little faster than semiconductor revenue.
For the reasons I talked about, process control grows a little faster than that. You couple it with some modest share gain, what Oreste talked about in terms of growth of EPC, and then the service model gets us to that $14 billion. The rest of the financial model is much more of a continuation of what we've been doing. Gross margins in the, we'll call it 63% range plus or minus our gross margin. Incremental gross margin 60%-65% generally. There are parts of the business that are diluted, but we think we can maintain that at 63%.
We'll get some efficiencies in our cost structure, although our model has always been about investing in R&D and introducing products faster, going to market with a portfolio approach to compete against our competitors and deliver more value to customers. Don't think there's a lot of leverage in R&D, but there should be in our corporate infrastructure. It all translates into 42% operating margins and an ongoing model of 40%-50% incremental operating margin from there. We also announced some reaffirming actions in terms of capital structure, raised some incremental debt to fund an ASR, plus an additional amount to be executed over the next 12 months-18 months after the ASR completes. All that translates into about $38 in earnings overall.
When we looked at this plan, and we looked at the ability to put the capital and the go-forward cash flows of KLA to work, the best value opportunity was invest in the businesses we have, accreting the shareholders we have, and driving earnings per share growth. That the actions we took based on that plan, I think reflect that. Also raised the dividend, 24% as well. That was our thirteenth consecutive dividend increase. That's the wrap up of the four hours or so that we walked through a couple of months ago.
Great. Let me stop here and see if there are any questions in the audience. If you have a question, please wait for the mic to come to you.
Could you just speak to the risks to that financial model? If it's not achieved in 2026, what would the potential reasons be that that's the case?
Well, we didn't sort of think linear. I mean, obviously, the industry growth rates assumptions, we contemplated, you know, that look, we gotta have some sort of correction at some point. It translates into a $125 billion WFE level in 2026. It gives a sense of how we thought about that growth. We thought about KLA share improving a few points. We have a very strong share position, about 54% in our core market, as reported by Gartner. We gained a few points of share over the last couple of years. We see that creeping up, but if you look at the markets we're in, share is meaningfully higher than that.
While we think there's opportunities, it's hard for us to gain a lot of share, I think, just given the presence and our customer's desire to try to leverage second sources. That's how we thought. Obviously, the mix of business is also a factor. I tried to make a case for why I think foundry logic process control intensity is higher than it is in memory. That's probably the biggest factor that affects our share of WFE, if you will, in a given year. Those are the kinds of things that, potentially, you know, if it doesn't play out, you could have different results. We feel pretty good about the trajectory we're on. We aren't assuming major changes.
We're assuming a share of WFE going from 6.2% in 2021 to 7.25%+. I think given the dynamics we've talked about, given what we believe high NA EUV will do in terms of reflections of the business, we think there, you know, there's potential there. We didn't model a lot of change, but you have EUV introduction, not just in, you know, what's happening in logic foundry, but also in memory. Things like that.
Yeah. I'm recognizing there's a very big long-term opportunity.
I, you know, got a few months to go here. As you know, one of the comments I made in a series of meetings over the last couple of days is that while we've had a lot of macro, negative macro data points, and there's been some signals from customers in certain markets, and certainly the consumer markets have affected what memory customers are saying about next year. Even with all that, we continue to see a very strong demand profile, and I don't think we've had more pressure today. I think the pressure levels today are as high as they've been over the course of this year. We're continuing to see backlog grow.
At earnings, I had not only talked about the September quarter lining up as we expected, and we see that, you know, how that's playing out now and expectation for sequential growth into the December quarter. As I said at earnings, I would expect that we'll see a sustainability in factory output levels, at least as we look at the March quarter today. I don't see anything kind of falling off. We're trying to ship more. I think what we're seeing now is how much under demand the industry has been able to supply. I think as I listen to my peer companies, I think you're hearing something similar in terms of just the view of, you know, the pressure levels are as high as they've been.
My supply chain challenges are also as high as they've been. They're not necessarily getting better. Certain areas, they get a little bit better, but other areas they get, you know, worse or there's still unpredictability and a fair amount of strain there. In about next year being down, so you're hearing that. I don't expect that. I expect that they'll follow through, although in memory, it tends to be a little bit more volatile in terms of what they say and what they ultimately do, and it's very end market dependent. A lot more of supply comes from new capacity, and new capacity is bigger investments versus shrinks and doing node shrinks. Sometimes those are quick turns because of that. Our backlog remains very strong.
We tend to be a technology enabler, so we enable our customers' roadmaps through development and ramp. If technical schedules kind of hold around node introduction timeframes and so on, that tends to be a big driver for KLA. There's some resiliency in our backlog to that. It isn't as capacity-centric as maybe certain process companies would be. Logic and foundry continues to be a pretty competitive environment. I expect that to continue into next year. I would expect mainland China to be at least in line with what we saw this year, and that's mostly, you know, it's legacy investment in logic and foundry. Infrastructure investment, wafer infrastructure, mask infrastructure, and then memory. I'll call legacy non-automotive. Automotive still is short. We haven't shipped a lot to those customers.
Process control intensity isn't as high in that part of the market if it's customers that are just making a lot of the same parts. If specs are changing, that could change it. I think that part of the market, we'll continue to ship into that part of the market. I think, you know, as it relates to some of the more legacy nodes that are for other products, if you start to see utilization rates change as some of the leaders do some of the second sources to the leaders, all of a sudden you see that capacity or the demand migrate. There might be some issues there. I'm not seeing any of it today. I can, you know, if you just kind of follow through and try to connect the dots, I could see that.
A little qualitative color from what we see today.
Any questions?
As the nodes come down to 3nm, you've got to start asking the question, electron leakage, quantum effects. Where does that actually start to kick in? Is it? Are we here yet or is it a couple of years' time? Is there any use in talking to the quantum guys like Rigetti and stuff like, because they'll be actually facing similar problems. Similar, but different.
The question is for EV, when is it really going to happen? I mean, look, the investment is happening now. You see, like for example, an explosion of investment in silicon carbide based on better performance of the silicon carbide over silicon. This investment is not showing any sign of decline. Actually, it's accelerated. Every time I have a silicon carbide customer of my SPTS products, like the etch and the precision tools I mentioned before, I see always these people pulling in, pushing me to pull in orders, pull in shipments. I don't see this to change at all, because thinking about the silicon carbide, it started from Tesla. It's public domain. Tesla made the first silicon carbide electric engine for the Model 3.
Pretty much everybody now is converting because they see the value. The market is 80 million cars-90 million cars. Of course, it's more percent EV today, but a faster growing percentage tomorrow. Most of these cars will see a transition from silicon power to silicon carbide in the next three, five years. That's the reason why I see, first of all, there is no secret that more and more electric vehicles have been sold as a percentage of the total number of cars. The second is more and more electric vehicles will have silicon carbide in their cars. That's the reason why I see the demand continuously growing. We saw, 1.5 years ago, the automotive business skyrocket, right, suddenly. Of course you understand why.
There were no semiconductors anymore to make cars. Ford, the GM, everybody's halting the production lines of cars because they didn't have chips. We saw this huge sudden demand skyrocketing. Now you see the demand is still strong, but it's a different demand. It's a demand of new material, silicon carbide. It's a demand also of more advanced technology in the cars. Something completely unexpected. A few weeks ago, we had an announcement that one of our customers was introducing 5nm node inside the car. This is crazy because generally, the automotive industry was based on very well proven technology, what we call like N-3 nodes. That's because they need to be very well tested, very well proved in terms of not only functionality but also reliability. We see this changing completely.
More advanced parts, more semiconductor content, new material coming up, new reliability screening methodologies that gotta be applied. This is a really unexplored territory for automotive and for semiconductor, and we will leverage from it, because I believe KLA is the best semiconductor capital equipment supplier in this industry right now. It's the only member of AEC, as I said, Automotive Electronics Council, and we work with everybody in this industry.
With that, we'll wrap it up. Thank you, Oreste and Bren, for coming to the Citi Conference.
Thank you for having us.
Thank you.