We're going to get started. Hi, I'm Timothy Arcuri. I'm the Semiconductor and Semi Equipment Analyst here at UBS, and very pleased to have Bren Higgins from KLA with us. Bren is the CFO at KLA, and I've known Bren for a long time, and thank you for joining us.
Yeah. Thanks for having us. Appreciate being here. Really quickly, let me just make a couple of statements, and I'm sure we'll dig into some of the things I say, but I'll just sort of level set for folks here. Don't know if I need the microphone in this room, but I guess for the folks on the webcast. For safe harbors, I'll be making some forward-looking statements today, and those statements are subject to risk, and you can find all our risk factors in our SEC filings. 2019's been an eventful year for the company. We started the year with a correction in the memory space, and we thought that the overall industry spend for the year would be down somewhere between 15% and 20%, maybe even closer to 20%.
As we would start the year, we had a kind of big sequential drop in the March quarter and expected sequential growth through the rest of the year, and we saw that play out. I would say, if you go back to what we got right, certainly our expectations around memory were pretty clear. There wasn't a lot of business. Certainly, there was technology migration that was happening as customers pursue their roadmaps, but no incremental capacity adds really to speak of. We saw the foundry business come back in the second half of the year. We thought it would be balanced across the year, but actually came in much stronger, and the second half based on guidance looks like it's up about over 50% versus the first half. First half is weak, second half stronger, and most of the spending really concentrated with the leading customer.
As we look at next year, we think that foundry logic overall for the year looks like it'll be sustainable in terms of the level of investment to expect next year. If you look at the end market drivers, if you look at the adoption of the 7-nanometer node, you look at the broadening, I think, competitive environment. If you look at even in some of the trailing, I'll call them trailing edge markets, which are affecting industrial and automotive and IoT markets, we do expect higher levels of investment there. Finally, our China mix of indigenous China flips to more foundry logic centric than memory, which more memory centric in 2019. For all those reasons, we have confidence in the sustainability overall.
We're excited about some of the longer-term structural drivers, with the introduction of EUV structured drivers in the leading edge of foundry logic. Over the last four or five years as EUV has delayed, linear scaling hasn't been how customers have achieved new nodes. As you introduce EUV and you start to scale again, we believe that that will introduce new demand, certainly more compelling offerings for our customer's customer, but new demand from our customers. We would expect to see spend in that area because of some of these structural changes and rising capital intensity, therefore, rise over the next few years, and see logic foundry WFE actually grow in line with semiconductor revenue. In the last four or five years, it really hasn't grown much at all. We're excited about that.
As we look at memory for next year, we think it'll be a better year for memory, more NAND centric than DRAM, not expecting a lot of growth in the DRAM market. I think that's more of a middle of the year to second half dynamic. There is some investment in the front half by one customer. In general, in terms of overall market, I would expect it to be more second half centric. I don't have a lot of unique information, just looking at what we're hearing from customers as we look at some of the industry data, what we see in utilization rates in our service business, and so on.
We closed the Orbotech transaction this year, and we're working on synergies there, excited about some of the market opportunities that are there, both from just a general sort of complexity and roadmap issue with their existing business, but also some things we think we can do with them. We had an investor day where we outlined some of our growth strategies back in September. We can talk about that a little bit more if you'd like, and then also announced some adjustments to our capital return policy , really a reiteration of what we've been doing. We raised our dividend, 10th consecutive year dividend increase. We raised it 13%, $0.10 dividend increase at the Analyst Day, and also additional authorization for share repurchase. As I said, reiteration, that balanced approach across capital returns. We've got an asset-light business. Generates a lot of cash.
We think we should be able on a go-forward basis to return at least 70% of the cash flow we generate. I think that's kind of where we're at. We're excited about where we are in 2020, we're excited about some of the new products that are coming, and I think we're excited about some of these dynamics within all of our markets that I think will increase the relevancy of process control.
Awesome, Bren. Thank you. Maybe we can start, I recently downgraded the entire sector, KLA, which was a favorite of mine in the sector, came down with the group. I just wanted to ask a question on sort of the sustainability of WFE spending today. If you look at the top five companies in the space, they are pretty consistently about 75% of total WFE. If you look at the December quarter guidance, and now pretty much across the board now we have the numbers, we can see that WFE in the fourth quarter is running close to $15 billion, maybe $14.7 billion, $15 billion. You're kind of annualizing to like $58 billion, $58.5 billion, somewhere in that range, and that's even before EUV comes in, because really ASML's guidance for Q4 is primarily DUV and not so much EUV.
That's more coming in Q1. Most of the increase has been on the foundry logic side, of course. If you look at the foundry logic math, you're sort of annualizing maybe to the mid-30s in terms of foundry logic WFE, which is like $10 billion, $12 billion higher than we've been the past couple years. I guess unless foundry logic is going to grow significantly again in 2020, it seems like the math would say that foundry logic segment has to sort of cool off a bit. Can you just talk about the sustainability? I know that it's going to sustain at a high level, but can you talk about the sustainability of it being at this level as you kind of look into the first half of next year?
Well, to your point, the math implies 10%-15% growth in the industry. I just said it looks like it's flat to modestly up. What I've been saying since earnings and what I said last week as I made some public statements. I would expect some moderation from the leading customer. We're excited about some of the structural dynamics over the long run, which I talked about. We do expect some broadening of investment. As I look at this, I think the broadening of investment, you've got on logic, I think there's foundry competition that hasn't invested much. Yes, I don't think foundry logic is going to be up 10%-15%. I do expect some moderation from the leading company in terms of how they're investing.
At the same time, as I look at how sort of certainly the first half of the year sets up, I don't see a lot of change in overall business levels. Certainly, there's some moderation. We have big integers on our tools, right? To your point, consolidate customer base, and you have a couple of tools move out that are $20 million a piece, it can have some impact on the numbers in terms of how they move around, both pulling in and pushing out. In general, I expect sort of reasonably balanced kind of business levels here through the first half. As we move to the second half, I think, the memory question becomes a bigger question, we'll see how that plays out.
Sure, yeah.
Certainly, from where we sit today, I think structurally, I think when we roll it up, we see more investment from, let's say, the other guys than the current sort of run rate levels. We're optimistic that for the year, the spend levels will be pretty consistent with what we saw in calendar 2019.
Great. Can we talk about WFE share? If we assume WFE, most companies are saying 47, 48 this year, which we were thinking 43, 44, and then the leading customer dropped a bunch of WFE into the back half of the year, so it's probably going to be 47, 48. I don't think you disagree with that number for this year, WFE.
No, everybody counts it a little bit differently, and do they include the wafer-level packaging piece or not? Generally, it's, yeah, it's in that ballpark.
On that basis, you're gaining about 30 basis points of share this year. That's really before you even see EUV coming in really in earnest next year, which you have fairly good attach to EUV. Can you talk about WFE share? Can you get back above 7% like you were back in the 28-nanometer cycle? Why or why not? Can you sort of walk through how much attach you do have to EUV as more $ migrate.
Yeah, sure.
over to EUV?
No, it's a great question. The short answer is I do believe that we've seen about a year or 1% improvement overall in the process control intensity, which process control is a percent of WFE, which translates then to your 30 basis points of overall WFE. This year, in calendar 2019, against the backdrop of the industry being down 10%-15%, our business would, you take just the midpoint of guidance, we should be modestly up versus 2018. Does give you an indicator of the strength of our exposure to foundry logic because our memory business is down 35%. We do see sort of similar kind of industry down sort of numbers in our business as the broader industry.
The strength of the 7-nanometer node, I think one of the great things about 7 nanometers, for the first time, and I think it's because of this lack of scaling, that you actually have pretty compelling transition of performance and capability and power, but also for cost. The end market adoption of 7-nanometer node has been very significant. We've seen that in our reticle business, and we've seen it overall as the customer, particularly the leading foundry customer, has added a significant amount of capacity to support. Public numbers in excess of 120 designs, and so you have, for the first time in a while, significant end market adoption. A, they're adding capacity to support a pretty big node, but also they're preparing for the 5-nanometer node.
In preparing for 5-nanometer, the ability to try to migrate and use some of the tools they bought for 7-nanometer, because usually they buy it for the leaders, and then if the follow-on investment doesn't happen, then they're able to try to reposition some of that capacity into the new fab. Not able to do that that much in this transition. It's driving investment in 5-nanometer, but also the investment to support the activity at 7-nanometer. To take that into consideration, plus, I believe, increasing process control intensity in memory. We've seen it, A, China's been good for us, but also just in general, we saw some improvements in the flash side of the business from the transition from 2D to 3D.
When you add all that up and you take into consideration what really drives our business is end market adoption and scaling, we feel like we should be able to get back to the numbers you're talking about. What happens with EUV is a couple of things. First of all, scaling, smaller defects, smaller measurements matter. Right? What you tend to have is what we expect to see is a migration to some of the higher-end capabilities in the company. With the adoption of EUV, A, you always see a migration whenever there's a transition, because there's an immaturity of the capability, customers will monitor their processes more. We expect to see a migration to some of the higher-end capability where we have stronger share and stronger margins. I think that will be one of the positives that we see from this.
There also is the challenges around the reticle and the reticle ecosystem. We believe that we're already supporting that activity now, but we'll also see more investment in that area, where that market will structurally change and become a bigger area of investment, because it's sort of critical path in protecting wafers from defectivity as they transition from the reticle to the wafer itself. For all those reasons, we feel pretty bullish about the opportunities that are afforded to us by EUV and some of these end market drivers that are driving the rest of the business.
Got it. Was there a way to maybe mathematically as people, it's sort of somewhat easier to get your head around, okay, I think there's going to be an incremental X billions of dollars spent next year on EUV. Is there some way to sort of translate that back to you and say, okay, out of that incremental, for each $1 billion that gets spent on EUV, your capture or your attach is X to that?
One way we tried to do it was, at Investor Day, we put out a plan to 2023. When we modeled our business, we said that, look, we're going to see $400 million-$500 million of incremental investment related to share and intensity improvement. They blend a little bit because some of it's impacted by EUV and so on. We thought against a base of 4%-5% baseline growth in the industry, there'd be an incremental $400 million-$500 million tied to that and $400 million-$500 million tied to share and intensity. We picked 2023 for a reason. It's when you see crossover where over 50% of the layers are EUV layers. In a device, you also have the 5G crossover in 5G mobile handsets.
Other reasons to pick it in terms of the impact on the product synergies and things like that with Orbotech, for sure. That's how we thought about it. I think if you add it up, we believe as you see some of this puts and takes, I mean, at the end of the day, customers drive down cost a little bit because there's fewer layers, so maybe fewer metrology, but higher-end needs for inspection and other metrology requirements. I think as you model it out, maybe it's a point of intensity overall in terms of improvement tied to EUV adoption, which is logic foundry, DRAM, obviously not flash.
Broad. Yeah, okay.
Yeah.
Got it. Let's now talk about China. Other companies, not you guys, but your peers are now pretty, I think, centered on the fact that there's going to be $6 billion-$6.5 billion this year that of WFE that's for indigenous domestic Chinese projects. That's going to be up in 2020 as well. It was up this year, too. A, do you see any signs of pull-ins as a result of maybe the fear that they won't be able to get equipment, number one? Number two, do you have any way to determine that as you get orders from your large foundry customer, do you have any way to go back in and triangulate and think, well, okay, I know how much this tool can produce in terms of capacity.
Is there some aspect of their customer placing orders for them to add supply because they're also fearful that there could be some kind of export control? Do you have any way of sort of like monitoring that? It's sort of like, that's asking the same question.
Right
1 level.
The first question is, we haven't seen a change in customer behavior. Really, it's been actually fairly stable business for us. If you go back, I've been pretty open that if you look at 2018, 2019, and 2020, business levels are about the same in terms of about $650 million or so of investment for KLA or revenue for KLA. As I look at 2020, I'm sort of in the same ballpark. I talked to the mix changes earlier. It's been fairly consistent and also not reliant on a spike or any significant change in it. Clearly, there's a lot on the headline front, but customer behavior's been fairly consistent. It makes sense because these tools, look, there's a certain amount of support that's required. You've got to service the tools, and our customers tend to adopt contracts.
Our service business is a little bit different in terms of the complexity of our tools and how to service them and our sort of control of parts and supply chain. Also application support, where we have people in the field that work with customers. We think it's competitive advantage for KLA to optimize how they use the tools. If they don't have that capability, both on the service and application side, hard to get that much value out of the tools. Hard for the tools even to run over the long run because they have to be serviced, their preventative maintenance, things that are required, and so on. As a result of all that, we haven't seen any change in behavior.
This year, for example, we thought one of the projects, a DRAM project, was not going to be in the year, and they're pulling back in. It's another upside factor for this year.
Pulled back in more for technical reasons than for anything related to trade. Certainly, a lot of parts are sourced outside of China for the China economy, there's that activity there. I don't know if there's a model generally of saying, okay, the way our tools are used are generally to monitor a certain amount of processes, but every customer has a different sort of sampling strategy.
and how they deploy it. Very hard for us to say, okay, they bought this tool because they're monitoring X capacity, and where is it going? What we have seen around what we'll call the trailing edge of, which is where most of the foundry logic in China is, but we've also seen it in some of the trailing edge of our other non-Chinese customers, is over the course of this year, we have seen a tightening of utilization, which affects the service business. A strength in those markets, both in terms of what the Chinese customers are supporting, but also what, let's say, for example, TSMC is also supporting, demand for 200-millimeter service and so on. I'm feeling that that's improving over the course of the year.
I think it's sort of a testament to, I think that the markets that they're serving on logic foundry, that there's some broad opportunity, whether it's IoT or industrial or automotive, and that the market's sort of behaving pretty rationally. Memory's a little bit different, right? Because leading-edge memory is where most of demand is, and memory capacity always gets evolved to the latest generation. The players in China are behind, and they're improving, but the leaders are also improving. I'm not really seeing a real sort of change in some of those dynamics.
Got it. Maybe can we talk about potential export control? It's a topic, who knows if it's going to happen or not. I guess a couple issues around that. Number one, how much of a seat at the table do you have with the government in D.C. around these conversations? How much of a risk do you consider the potential for the government to restrict, in some way, export of U.S. semi equipment into China? How much of a realistic issue do you think that that is?
Well, I'm less worried today than I was a year ago. A year ago, I think part of it was, to your point, I think that we've been very active around engagement, and engagement in Washington, educating about some of the complexities around the semiconductor industry, that it's a global industry, that there are U.S. players and there are non-U.S. players. That anything that the government does has to sort of keep in mind that there are alternatives to U.S. suppliers. There's still some of the broader questions about leading-edge capability in Taiwan and Korea that's not in the U.S., which doesn't necessarily get controlled by activities that happen within China. You could have, for example, which we haven't had any export control for our business, but if you had export control in China, it doesn't affect how China's sourcing from TSMC, for example.
Right.
I think it's one of those lower probability, higher impact kind of events. I'm not concerned about it. I think we haven't been selling leading edge tools to China. If, I guess there are certain stages where you can't ship anything, where you then you can't ship leading edge. If we couldn't ship leading edge, we're not really shipping leading edge anyway. Does that change some of the marketing and specs and maybe increase some of the administrative burden if that were to happen? Maybe. At the end of the day, I think that all along we've been shipping sort of trailing edge, and I would expect as the leaders continue to move their roadmaps, that they'll maintain the lead that they have. We'll always be selling sort of older versions of tools into that market.
Got it. Let's talk about memory. That's been a smaller portion of your revenue typically, but should we expect that to grow in the next cycle, maybe with the new 3D NAND offering? Just kind of talk about how you're attacking the memory market?
Yeah, no, it's really been a good story for us as we've transitioned from planar to 3D NAND, that the incremental metrology requirements in building the 3D structure are much higher. You have to do side wall angles and profiles of the stacks as they go vertical that are increasing the sort of intensity of metrology. You also have wafer flatness requirements, which affects our business both directly as we report X% of our business is memory or NAND, but also there's also a piece of our business that goes to the wafer manufacturers. That doesn't get lumped in with the memory spending as we report, but is being driven in a lot of cases by memory. As wafer flatness specs have increased, because you have these stacks that are getting taller and they can move and they can tip over and so on.
That's been a big driver for us. Defectivity control is a challenge, one of the things that customers have done to get around that is to buy more unpatterned inspection. What they do is they run more monitor wafers prior to processing to make sure their processes are good and clean. For all those reasons, we've seen a step-up. It's probably been a good two points or so of process control intensity. Process control as a % of the WFE, so every WFE dollar that gets spent on flash, a 2% improvement in that, or two points improvement. We also have a new product that we're introducing, that we have a couple tools in the field that we talked about at Investor Day, but we will be shipping in calendar 2020.
Would expect some revenue in calendar 2020, I think that's sort of WFE independent of incremental revenue. It's a new technology that does basically metrology for the channel hole, which is a challenge for customers, which is channel hole that goes all the way through the stack, the ability to measure that in line in non-destructive ways is a value proposition for this product with customers. Today, they do that measurement in destructive ways, so they have to actually cross-section the wafer, and then that's how they make the measurements. As you might imagine, it happens offline. It destroys product wafers. To be able to do that in an in-line non-destructive way is pretty compelling. We're excited about that. We'll have this product that will start to ship. We hope to see some revenue this next year.
I would expect that to be another sort of $100 million-$150 million type contributor to the company, which is about another point of intensity in the flash market. We're excited about that.
Is that number a 2020 number or that's just like an ultimate number?
It's a ultimate number for the product.
Got it.
Yeah. I'm just picking sort of a level of spend. I think flash is about $15 billion today or something like that.
Got it.
I think it'll take a little bit of time. 2020 will be much smaller than that, I think as we get into maybe a steady state and get the product seeded in the marketplace, as we look at 2021 and 2022, we start to see that level of contribution, maybe higher into the range as we get in that 2022, 2023 timeframe. Which is also important why we sort of picked that timeframe for Analyst Day to put out a model, is that there were a number of these products that were coming to market, seeded to market, transitions in the industry, for all those reasons, we thought that was a good point to pick.
Got it. Okay, that makes sense. Let's talk about Orbotech.
Sure
A couple of minutes. I thought it was a great deal.
Thank you.
Obviously at the time, I thought it was really smart. Let's talk about the flat panel business there. It seems like you could argue that maybe it's not really core to the company now. Maybe can you talk about sort of, A, the different segments of Orbotech, and B, focusing a little bit on flat panel. That business, sort of in my opinion, is the least exciting of the three segments, and it sort of feels like it's not going to grow very much next year or so. Can you just sort of talk about the.
Sure
strategic aspect to being in that market?
Yeah. Let me just take a step back. Orbotech, overall, basically three distinct businesses. One is specialty semiconductors, which is etch and deposition tools for specialty semiconductor markets. The RF market, the power supply market, MEMS, advanced packaging. They've been able to target some market niches that have now seen some inflections due to automotive and incremental capability and packaging. Certainly, 5G is a big driver in terms of communication infrastructure for the RF business, focus on power and so on. They've been able to carve out very good competitive positions, and if you look at our segment reporting, gross margins are 55%, so it gives you an indication of the margin profile of the business, certainly mid-50s relative to other process businesses that tend to be in the 40s.
We like that business, and certainly, we have opportunities to take that business to another level in terms of engagement with the big customers. Sometimes small customers struggle to get traction with the Intel and the Samsung and the TSMC of the world, being part of KLA certainly gets them to the table more often. It can be a double-edged sword because they have to execute, because they can hold us accountable for that, but that's why they like it. We're excited about not just the growth opportunities there, but there's channel opportunities, and I think there's engagement opportunities in particular with certain customers, and there'll be some new product offerings there that we think can enhance that position. They have a printed circuit board business where they do imaging and inspection in circuit boards.
Most of the business is imaging, people ask me what that is, the way to think about it is sort of litho for PCB. These are very high-end tools in that market, carry very high contract penetration. Service contract is about 90% of the tools that ship stay on contract. It's a nice service stream that comes with that business that doesn't have any volatility and also is well-positioned in the market. What's happening in PCB boards is you're starting to see this transition into heterogeneous packaging, where you have PCB boards you have IC substrates or IC-like substrates that are in between the package and the board. The ability to image and to do inspection in that substrate is becoming a new sort of evolving opportunity.
There's capability within KLA from an inspection perspective and what Orbotech's doing technically to address that market in a more meaningful way. We talked about revenue synergies that could come from this combination. We thought it was an incremental SAM opportunity of another $400 million for us if we can execute. There will be product offerings, not in 2020, but beyond, that will help address this market. That's that market. Very good market. Margin profile in sort of the mid-40 range on gross margins. The flat panel business, and I don't expect growth in flat panel in 2020 either, has been a fair amount of capacity on the LCD side, so I don't expect some of the new evolving opportunities are limited. Orbotech hasn't had any Samsung business historically because of some IP disputes, and we're working on improving that situation.
We have some eval tools that are out there. That could create some opportunities for us. They have a test tool that does basically electrical test of the panels and an inspection tool. Two good businesses that have a cost problem, I would say. What we're trying to do is improve the cost structure of that business. They did some manufacturing engineering in Silicon Valley. We're moving that to Israel and trying to consolidate and drive scale, but also leverage our broader footprint. To answer your questions about that is, look, I'm open-minded about the business. There's a couple products that are there that are leading the market. What's not clear is will customers pay for that differentiation, and will that relevancy ultimately drive higher margins? We figure that out.
We're going to improve the cost structure and try to put the business in a place that, look, I'm not afraid of cyclicality of a business that's 200 that cycles to 300, but I've got to be able to do it in a way where there's profitability at the bottom, and I can drive leverage through the model as it scales up. If there's through cycle that creates opportunities, certainly semiconductor-like technology, MicroLED might create some unique opportunities. We're open-minded about it, but I agree, it's less compelling of the other ones in terms of where it sits today, and it's something that we're evaluating. The first thing we have to do is get the cost structure and the footprint correct.
Got it. One of the big themes that I think is super interesting in this sector, as I've covered the sector over 20 years now, has been that service is a much greater percentage now of company's revenue. Given how much the install base has grown the past couple of years, it will continue to be a large portion of revenue. Can you just talk about your service business and maybe break it down a little for us in terms of how much is recurring and how much is non-recurring?
Sure. 70% to 75% is recurring, right? It's contract, and that tends to be very consistent every quarter. Customers renew in either one-year or three-year type service contract structures. The attach rate is pretty high. There are reasons for that. It's a little bit different than our process tool peers, in a couple of ways. Number one is the complexity of our tools is pretty high. Customers have a hard time sort of building capability, A, because of complexity. It's very hard to service our tools. Our engineers service one flavor of tool, let's say, and not multiple ones. The other thing is that all the components in the tools tend to be custom to meet our very high-end requirements. You can't just go and buy these parts off the shelf. We control the supply chain as well.
You take that, you take the complexity of the tool, and you also take the, look, it's one of those purchases that at the end of the day, it's a bit of a cost center for customers. Certainly, they have to monitor their process. They have to make sure that WIP is protected in terms of yield, and they got to get to time to results sooner. We're not actually making wafers, and so I think that they have a budget for it. They buy the elasticity of demand as such that they buy what they need, and then they run the tools very, very hard. There is no offloading of capacity. Because they've got a certain number of them, it's hard to build internal capability. They're complicated, and so they rely on us to keep the tools up.
We've done a good job over the years of increasingly monetizing that, both in terms of our offerings, but also consolidating the cost footprint around customers as they have consolidated. I think it's an underappreciated part of KLA, frankly. I think that the contract penetration is pretty compelling. Doesn't change all that much. We have the ability to vary our offerings. As we vary the offerings, we get predictability, the customer does, but we can also adjust our cost structure underneath it, and we deliver to what they ask for, and we try not to deliver to more than what they ask for because it sort of undermines a higher-end offering, but what exactly what they ask for. We've been able to improve the margins. I believe that the service margins are incremental, in terms of service.
There's always the accounting decisions, but accretive to the operating margin of the company. It doesn't have a lot of volatility. Even in 2009 when everything was falling apart, it was only down about 15%. We expect long-term growth to be somewhere between 9% and 11%. Historically, it's been about 10%. Orbotech should present at least some cost opportunities for us to drive higher levels of profitability out of their service business. Small companies struggle with the footprint to support service.
Yeah.
We think that's an opportunity for us as well in this transaction that isn't valued anywhere in terms of synergies or plans, but we believe that there'll be opportunities for us to execute there.
Great. Thank you. Let's talk about the financial model for a moment. Right now you're sort of run rating to about $6 billion right now, revenues. Gross margin 61%, give or take. The financial model, if you look out to the $7 billion-$7.5 billion range that's built into the model, there's not a lot of incremental drop-through built into the model on gross margin.
Yeah.
Is that mix? Is that having some new products hitting the P&L? Can you just talk about that? Because it seems like that's a fairly conservative number.
Yeah. It is mostly mix, and what it was intended to do was to reflect the expected mix of business when we get to that point, both in terms of potential new targeted M&A, if we do it or not. If we don't do it, then the margins will be higher. That's why we put the range out there, but also the contributions from the Orbotech business today and where they grow to. An important message in that model was that the model for the process control part of the business wasn't changing. We had a 63% sort of plus target model for process control, semiconductor process control, and that's the same model we have today.
That the focus of the model all the way through, though, so the GM was reflected by the mix, then all the way through was really built on the concept of driving incremental operating margins of 40% to 50%. If you go and you play it through, and I didn't want to go to basis points on the overall model, but if you play it through, that's how we targeted it. We believe the Orbotech business carries different margins. Not every business is a KLA-like business in terms of the margin profile, although we've proven in the past we can improve the margins of market-leading positions that we buy. We did that with ADE, and there's a number of other examples in the company. Really we could drive incremental operating margins on those businesses that were commensurate with the other parts of KLA.
That's how the model's built. Certainly, if the mix changes, we will see some variation to the sort of 60%, 60, 61% kind of model out in that time frame. It was really sort of based on reflecting the mix, but also the expectations around leverage in the business.
Got it. The model also is interesting in that I don't think I've seen a financial model that accounts for potential M&A as well. I think you're accounting for $300 million in revenue, something like that. Do you have something that is identified? Do you have an area I'm not asking you what it is. I'm just saying, do you have a particular market that's been identified? How did you come up with that to put that in the model?
We looked at the last five years or so, and we looked at how much they're contributing today, ex Orbotech, and the numbers ended up around there. We wanted to put a placeholder more for the purposes of saying, "Hey, look, there may be opportunities for us." We're going to be fairly, I think, opportunistic around it in terms of most of it likely more make versus buy type decision type M&A, technology M&A. Obviously, the alternative to that is buying back stock. The point of putting it as a placeholder is, look, our history has been that. Let's put it in to get the financial model right, the financial model I just talked about in terms of the blending. At the end of the day, the earnings number we talked about, $1,450-$1,550, we get either way because we're going to deploy the capital.
As I talked about what the alternative is, like, look, you take $300 million, and you apply a normalized sort of multiple of sales to get to a placeholder number of capital deployment to M&A, and you would take that number and buy back the stock because we were very clear about how we're going to run the company in terms of leverage targets, cash balances, and expectations around capital deployment. To your question specifically, we've done some small M&A in the company around opportunities in MRAM, for example, where you've got evolving magnetic memory opportunities that might be out there. Customers, will we develop our own product, or do we go buy a target that has a product in the market, engage with customers? We think those markets will evolve over time.
That's why I call it a make versus buy, more of a portfolio type M&A. Expect to see those kinds of deals. You never say never, but at the end of the day, I think we're pretty comfortable with the portfolio of products we have. We think we can augment it with some of these things. I don't see Orbotech-like transactions out there at this point, so in terms of size and scale.
Yeah. Got it.
That's how we did it.
Got it. I sort of always thought of you as being very forward-thinking in terms of capital return. I remember the special that you paid several years ago. That was definitely sort of an outside-the-box move, I thought. You've always been quite forward-thinking on that. Can you just review how you think about capital return now as a larger, combined company?
Yeah, sure. I don't think a lot has changed. Certainly that transaction was unique in that, given the characteristics of our business, it was pretty clear that you could have a permanent tier of debt in the capital structure in terms of how we finance the business. It was a structural change, and we ended up doing levered recap and paying out the way that we did. I think when we look at it today is, look, it's an asset-light business, and we're comfortable with, given the growth of service, that we ought to be able to return a substantial amount of cash that the company generates. I think it's a good thing, and it's important for our employees to understand, I think they've understood over time, is that it's a good thing that companies generate more cash than they need.
I think given different investor types, you need a balanced approach, and our approach, I think, is that. We've, I think, done a good job of looking at our business, about how much capital do we need on the balance sheet, how much leverage do we have, and that we're going to deploy the capital, because it doesn't get valued unless it's deployed productively. Whether it's in creative opportunities or return to shareholders, we create or sort of augment the value of an investment in KLA by doing that. The dividend is a 35% payout ratio. We grow it in line with our expected growth rate and free cash flow.
If we go back to our model, if we can drop earnings growth of one and a half the top line, 1.5x the top-line growth rate, and the top-line growth rate is 6%-8%, we should be able to drive our dividend payout between 10% and 15%. That's half of it. The other half is share repurchase. Like I said, it's balanced, and we're going to deploy the capital either way. I felt pretty comfortable that we ought to be able to return at least 70% of the cash the company will generate over time. That's how we're thinking about it today.
I'm okay with being specific and putting those sort of very specific markers out there, because I think it makes sense for the business, and it's important for investors to understand so they can model it appropriately.
Great. Maybe with the last 30 seconds we have here. With the stock where it is right now, what do you think people are still missing on the stock today? You talk to investors all the time. What's the key element of the story that you think people still miss?
Well, we talked about service, so I think that that's important for people to understand, and that's growing in a bigger piece of the company over time. We've augmented the position the company has with further diversification into markets where they have leading positions that we believe we can drive higher levels of profitability out of. We're more of a technology play. Certainly we have capacity exposure, but I think it provides a little less sort of volatility to our business levels maybe than some of the more capacity-centric peers. I think we're a little unique within our core space.
I think given the fundamental drivers we talked about earlier, that you now have through-cycle growth, I think that conservatively a 4%-5% base and then opportunities to grow that faster with share and intensity plus service, that there are growth rates expected in the future out of this business that are higher than the past. There's always volatility associated with consolidated markets. At the end of the day, I think the discipline from customers, the diversification of demand, is driving our, I think, a lower sort of longer-term sort of cyclicality of the business.
For the characteristics around assets and capital structure and all of that, I think it behaves in a lot of ways like, and I continue to say, look, it's one of the leading business models in semiconductors, and I think the business behaves a lot more like analogs and those types of companies in the space, and I think ought to be sort of valued in line with that. Certainly industrials kind of behave the same way. We talk about that in our different business with tech. One thing about semiconductors is they're foundational. Semi equipment is, look, I can't tell you what they're going to come up with, but I can tell you how they're going to do it. Since we play that role, I think the risk of technology disruption that affects semi equipment, pretty low.
We have the market position we have today, 4x our nearest competitor, 52% of the market today. We had that same market share over the last number of years, and it's 4x our nearest competitor. A lot of competitive noise, we've been able to maintain that. I think for all those reasons, it's a pretty compelling opportunity.
Great. Awesome, Bren. Thank you very much. I do appreciate it.
No, thank you. Thanks for having us. Appreciate it.