Great. We'd like to get started. Good morning, everyone, thank you for coming. I'm Toshiya Hari. I cover the semiconductor and semi-cap equipment space at Goldman Sachs. Very excited to have Bren Higgins, Executive Vice President and Chief Financial Officer from KLA. With that, Bren, kick us off.
Thank you, Toshiya. Thanks for having us here today. Looking forward to our discussion. Just quickly, from a safe harbor perspective, I'll make some forward-looking statements today, those are all subject to risk. Plenty of risk factors in our SEC filings. I'd encourage you to take a look at them, you can get them on our website, kla.com. Not KLA-Tencor anymore, kla.com.
KLA-Tencor, by the way.
Did you?
KLA-Tencor.
Anyway. Calendar 2018 was a very strong year for the company. We had a very nice finish in the December quarter. Revenue was in excess of $4.3 billion. Operating margins greater than 39%. A very strong memory investment year, 70% of the spend in our core industry was memory. For us to do as well as we did in a memory environment, like 13% growth year-over-year, where the industry was somewhere in the mid-single-digits, is, I think, a testament to a few aspects of our business that are doing very well. First, obviously, memory intensity has been something that's improved for us. We've certainly seen more metrology opportunities in vertical NAND, which has been good. We've had controlling defectivity by utilizing our bare wafer inspection.
New specs that are affecting bare wafer metrology has been good for our business, not just with the memory ICs, but also with the wafer manufacturers. Reticle inspection has been inflecting as a business for the first time in a while. There was a lot of capacity, but as we've seen more and more multi-patterning, it's consuming that capacity. On the foundry side, you're seeing a lot of new tape-outs at 7nm, which is, in contrast to the last couple of nodes, you're seeing much broader market adoption of this node. The level of tape-outs there has driven tape-outs, reflects reticles and number of designs. It's an encouraging part of our business that I think has momentum into this year. Certainly, the strategic investments in China have been good.
There was more memory last year of that, and those fabs basically starting from scratch in smaller facilities, there was an opportunity to provide a higher level of support and engagement, and certainly our market position is when you're starting from the beginning, you tend to want to buy from the market leaders. That was really good for us as well. A very strong finish to the year. As we look at calendar 2019, after three years for the company in mid-teen growth rates, as we look at this year, this year seems to be a year of digestion for the memory space, as we talked about in earnings just a couple of weeks ago.
We see a first half that's weaker than the second half for us, as we look at foundry logic being growth in the year, which is encouraging to see, and that being relatively balanced for us through the year, that we would see this drop-off in the March quarter, where we guided about a $920 million revenue level and no change to any of the guidance that we gave, both from shipments, revenue, EPS. See second half or revenue growth, sequential growth resume in the beginning of the June quarter. A little bit stronger second half than the first half. Industry environment we believe is pretty healthy, pretty disciplined spending. Foundry logic investment, which is encouraging. A lot of the momentum I talked about earlier that we think continues. In a pretty good position here as we head through this year.
Great. Thanks for that. To level set us, Bren, can you talk a little bit more about the overall environment in 2019? I think quite a few of your peers are pointing to a mid-to-high teens decline for overall WFE. Where are you guys today, and how do you think about the different applications?
Yeah. I guess from where we each sit, we see things a little bit differently, and we're exposed to different things, right? ASML has a capacity sort of they have capacity, but they also technology enable. They have EUV investment that's happening, and that's kind of independent of supply and demand. You have players like Lam are more memory centric, more of a capacity player, their view is perhaps a little bit different. As a technology enabler, we kind of see things, foundry logic is a stronger market for us. I think we all see things a little bit differently. If I were to size it, and I think given the visibility this year, certainly in the first half of this year, it's probably down 10%-15%. 15%-20% or whatever seems a little bit weaker than we see it.
I think we probably had different views this time last year, too. We are in February. I wouldn't say our views are that different. I think maybe it's a risk profile as much as anything. When we look at the business, we look at the way we see it, I describe the trajectory through the year that we see. It feels like it's lining up in that sort of down 10%-15% range.
Okay, great. Then specific to KLA, you guys talked about 2019, your revenue being down less than the overall market, obviously. Is that purely a function of the bigger declines in memory and that sort of working in your favor from a mix perspective, or are there product cycles or share gains at play as well?
Yeah. No, it's a good question. Look, foundry logic is growing for the first time in a number of years. Foundry logic, because of the nature of the mix in the fab, the complexity of the devices tends to spend more on process control. It's not exactly twice as much, but it's more. If memory's in the sort of 10% of their wafer fab investment gets invested in process control, plus or minus, foundry logic is in the mid to high teens. As a result of that, when you have a shift and you see growth in those segments, there's going to be a higher level of spending and process control. I think that's one factor. The other factor is I talked about this tape-out activity driving reticle inspection investment. I would expect that business to grow year-over-year for us.
Despite the overall industry, it's in a down year, down mid-double digit, if you will, 10%-15%, we're going to see growth in that business. The bare wafer inspection and metrology business looks to be flattish year-over-year. I think a combination of the diversification, the exposure we have to those other opportunities, and it's all part of WFE, but it's a little different than the way the IC guys invest, plus the exposure to foundry and logic is driving that view for us. Certainly, service is part of it. Our service business is growing. It grew over 10% or 11% last year to almost $1 billion, and we would expect that to continue to grow in that growth rate range of 9%-11%. Service will continue to grow next year at its historical trajectory, if you will.
I think because of those factors is why we have a more confident view of a better than industry performance in 2019.
Okay, great. In memory, you talked about vertical NAND driving higher process control intensity. Just given the fact that your customers have progressed towards 3D NAND for its most part, do you think that growth in intensity is something of the past, or is there sustainability going forward as the layer counts continue to grow at your customer base?
No, it's a good question. Certainly, we've seen the metrology requirements intensify. There's a lot of defect challenges, and we've been trying to solve those. We talked publicly about our ability to try to solve those problems, and it's kind of a challenging one to do defect inspection in stacks because it's a different defect problem, trying to penetrate the more opaque film structure and identify the Z coordinate within the stack. There's a couple of ways customers are working around that problem. They're buying more bare wafer inspection to run more monitor wafers to make sure the processes are cleaner. We've seen that business inflect.
The metrology I talked about, we're spending a lot of money on product development to do new metrology capabilities that are some of the challenges in the industry around channel hole measurements, where destructive techniques are used to try to do those metrology measurements are opportunities for us. I bring that up because as we think about the go forward, I think these are sustainable opportunities for us to maintain that increase that we've seen. If we can solve that defect problem, probably even more. I think it continues as long as customers are continuing to drive more layer counts and progress their technology curves, then it should be good for KLA, and if we can develop the right products to solve some of these problems, and that creates perhaps even a little bit more opportunity for us.
Okay, great. Logic and foundry you mentioned as obviously an area of strength for the industry as well as KLA. Is it primarily the node transitions that are occurring at your leading microprocessor customer as well as the foundry, or is there anything beyond that?
I think that's part of it is we're seeing that investment both in the mainstream node plus the new node and as they're moving forward here. That was a delay in the second half of last year that brought that second half view down for the industry. We're seeing that come through this year, and I think that momentum's consistent. I talked about what's happening in 7nm. Seeing a lot of 5nm development activity happening with plans to try to do risk production on that into next year. We're seeing more activity from the leading-edge foundries as well. I talked about the tape-out. That should drive a certain amount of capacity at 7nm as customers design more there, and that should be a bigger node. Those factors.
The other thing is you're still seeing pretty healthy levels of demand at the lagging edge and the more of the legacy segments. Certainly, automotive and IoT type end markets are driving that business. If I went back 10 years ago, and I gave a mix of foundry logic leading edge, it would have been, of our orders, let's say, 75% or so. If I gave you the same metric today, probably 60%. There's a fair amount of activity. There's $27 billion-
$27 billion
of WFE that gets spent with those customers every year, it's a pretty healthy business level on there. It's driving some unique requirements as they start to try to deal with higher levels of reliability. For example, in automotive, it gets demanded from their customers. These other applications are trying to run the install bases they have. It's trailing-edge stuff, but they're looking for upgrades to that capacity. We've restarted older products to support some of that. We sell used equipment into those markets. The service demands on those markets is also much stronger, too. That's a piece of the business that's continuing to maintain some momentum here as we move through this year. Obviously, the biggest part are the two things we talked about up front, but this is a meaningful part as well.
Okay. You've been a beneficiary of EUV adoption. Based on what you know and based on what customers are telling you today, where is the overall industry from an EUV insertion standpoint, and where do you see adoption going over the next two to three years?
Well, I don't think things have changed all that much. Yes, in and around controlling reticle fidelity has been the biggest opportunity for us, both in terms of in reticle inspection itself, and we have some features on our current product that we sell that help customers with that to do all their work. Then they're also using our Gen 5 wafer inspectors to do process qualification and reticle fidelity qualification as they print the wafers and use inspection, full wafer inspection to validate fidelity reticles. So all that's been pretty good for us. There's a timeline around a few layer adoption in foundry towards the 7nm plus maybe end of this year, then you'll see 5nm have a higher level of adoption.
I don't think you see crossover to 50% of the layers, greater than 50% are EUV layers probably until 2022, 2023 timeframe, just given the throughput requirements and some of the other challenges that are there. I would say over the last year or two, we've seen that roadmap hold pretty constant. You can go back for years.
Decades
always constantly pushing, right?
Right.
Decades, yeah. Customers are taking deliveries, so they have the tools, they're going to use them. I think that that's why we've seen that roadmap hold pretty solidly.
Did you see EUV proliferating across segments, just not foundry, but also in logic and DRAM over the next couple of years, or?
I think you will. There's some challenges around reticles and pellicles and some of the other things that are driving the pace of how people adopt. Then DRAM, maybe they're seeing some contact layers and things like that, so I think that probably comes after. You'll see it in foundry first, and according to the kind of rollout that I talked about. I think as you move into high volume production, it's a big opportunity for KLA, both in terms of anytime we've seen significant transitions in technology, and it's probably the biggest one the industry's ever gone through. It's been good for process control. Processes are mature. We tend to participate in that because customers sample more. Then there are certain parts of the market that become even more challenging in and around the litho ecosystem, if you will.
We spend a lot of our development dollars to try to have the right products positioned for that when it goes into production. It's one of the, I think, few opportunities that we think there's a sustainable opportunity for us to see process control intensity actually start to increase, and it hasn't increased much over the last number of years. It's gotten better in memory, but it's maybe gotten a little weaker as roadmaps have extended out in foundry logic. I actually think as we move forward, there's an opportunity to see some increase there. There, plus increasing relevance around memory with some of the issues we talked about.
Right. From a regional perspective, China continues to be a big focus, I think on your call, you guys talked about calendar 2019 being flat to down or down year-over-year.
Down, yeah.
Which is slightly different from some of the process tool companies' comments. Is that purely a function of you guys seeing it early in 2018 and just your customers taking a breather in 2019 from a process control perspective, or is there something more to it?
It is probably part of it. Certainly, we participated. We had two very strong years in China, both 2017 and 2018. In 2016, there was a startup, so it was a meaningful level of business there. As I said earlier, they are starting up these fabs. There was a lot of engagement early on. It is still a healthy level of business for us. It is down maybe 15% year-over-year. If you go back just a few years, and we were sizing the opportunity, and we went through, there were a lot of announcements and a lot of big numbers thrown around, but we sort of worked our way through the different projects. We saw over a five- or six-year timeframe a $500 million-$700 million opportunity for KLA or for process control.
We did not know exactly how it would fall in the years, but I think we are kind of operating at that level. It was closer to in excess of $600 million last year. This year still seems like it is around $500 million, back down to 15% or so. I think lead time, and lead time is probably a factor in some of the WFE forecasts that people were talking about, too, right? Some of the discrepancies in terms of a few points here or there could be related to just timing of certain products. The first phases have happened in a lot of cases. Second phases are what we kind of thought would happen, where you would see milestone and funding and other things become in progress against technology. Roadmaps would be factors in the next phases of investment. We have certainly seen that, I think, play out.
It's a little weaker than we thought it would be. I think it's related to those things more so than anything else.
I guess collectively, we're all trying to figure out when your local Chinese memory customers start to impact global supply demand, both in NAND and DRAM. Based on what you know, what you see, if you had to guess, how much more time do you think they need to start?
Yeah, the focus right now is on technology development. There's a risk production component that can we produce it? Can we actually get to the nodes? Can we ramp it? Can we manage defectivity and all of that? Then there's then movement to the next node. There's no meaningful supply that's coming out of those customers now, and I think on the memory side, given where they are relative to where leading-edge memory is, I think there's probably a ways to go. I think we're a few years out from any sort of meaningful contribution to supply, from my perspective.
Okay, great. The services business, it's a great business for you guys. I think you guys talked about 2019 revenue exceeding $1 billion, you threw out a long-term growth rate, I believe 9%-11%.
Yep.
Can you kind of walk us through the math that gets you to that 9% to 11% long-term growth rate?
Yeah. As the install base continues to grow, a couple dynamics have played out, a metric I started throwing out, I heard in one of our recent reviews is, about 87% of the tools that KLA-Tencor has shipped from its beginnings are still in service. It gives you a view of the useful life of these tools. They get run for a very long time. Customers, in contrast to some of the process tools, first of all, there's no consumables, so it's all mechanical parts. The parts, generally, given the complexity of what we do, is mostly custom, so you can't go buy them. Our suppliers can't sell them to our customers. We usually have, at least in the key components, exclusivity kind of structures. We're able to sort of manage through that. The other thing is they can't offload capacity.
Generally, they don't buy a lot of extra tools. They buy what they need, and they need to monitor their processes closely so they can keep their fabs moving. In contrast to process, if they have tools go down, they might be able to offload onto other tools. When tools go down, they need to keep them up, they tend to buy contracts. 75% of the revenue is service contracts, and it varies whether it's tool specific or fab wide, and it varies in response levels in terms of stocking levels of parts and all those kinds of things, and different pricing increments related to all of that. It provides a pretty predictable stream. I have a pretty good idea what service is going to be. Irrespective of what they're spending in CapEx, they still invest in keeping the fabs up.
The OpEx part of it is still solid. It doesn't really matter what's happening at the overall market level. If you look at the business, in 2009, it went down about 10%. You had the entire industry falling apart, right? Significant reductions, the service was only down about 10%.
That was probably the worst year.
The worst year ever, right? The only down year we've had. It's a very resilient business. It has the contract base, which allows us to optimize around it. Consolidation of customers and smaller geographic footprints is another area for efficiency, and we've been able to leverage that as well. I think it's a part of the business that it's a meaningful part today, and it's a profit stream that's accreted to the company average. As it grows, I think it provides this anchor that, irrespective of what happens in the CapEx environment, that a solid contributor of cash flow and predictability to the business. That's pretty good to have now. I think if you go back years ago, where we didn't have that, where the industry was much more cyclical, service was a smaller part of it.
Now it's a much bigger part, so you have $1 billion that I wouldn't say is guaranteed, but is pretty predictable that it'll be there and growing at a rate that's faster than the systems business.
Given the point about exclusivity from a margin perspective, it's positive for your business.
It is. We believe it is, as we do the accounting for it. It helps with that, also, the capability of our people to service this stuff, it's very hard for our customers to develop their own internal capability. They don't have enough sort of mass across number of tools to make that investment. It's a significant investment, to keep those resources as they tend to rely on us to keep the tools up, the demands are pretty significant, getting that sort of predictable outcome and result from us is pretty important to them.
Okay. Can you talk a little bit about the competitive landscape? Your market share and process control holistically has been very stable. You've got big companies, small companies, coming after you, they've been coming after you for a long time, you've held on to share. What's been sort of the secret sauce? How paranoid are you about competition?
We're very paranoid about it, you look at the way we invest in the company, we invest to make sure that we can protect our position. Process control is one of those businesses that it's an ROI for customers, right? That we don't make wafers, they've got to get a return on it. They've got to find defects and relevancy of that make measurements and so on to be able to ensure that they can keep their production lines moving, they can advance their roadmaps. As a result of that, we invest to support it. We have a broad portfolio because you need to be able to leverage different technologies at different phases of a production ramp, also as architecture and design changes, materials change, some products become more meaningful than others.
Having the wide portfolio of products is also a competitive advantage for the company. We introduce products at a pretty fast rate. Sometimes they're new platforms or significant new products, also upgrades to products, we can be a moving target to competitors. We can also provide either new capability or economic improvement for our customers, which drives customers to want to buy new products and not reuse what they have. I think all those are factors for us. You look back over time, our market share is 4x our nearest competitor. I think it's probably an underappreciated part of the market position of KLA or the value proposition of KLA as an investment. We've been able to maintain 52% of the markets where Gartner has it, actually grown it over the years, despite entry from larger competitors.
You've got smaller competitors, you've got larger competitors, but we've been able to maintain our position, and I think it comes back to those dynamics I talked about. We're pretty disciplined about understanding that value and holding to it. We share it with customers, and it's worked out for the company.
Okay. Anything on Orbotech?
Expect to close it this quarter. In the March quarter. The conversations we've had with the regulatory authorities in China, which is the last remaining region, has moved along. It's taken longer than we thought it would, but we expect to get through it this quarter. We'll have a lot more to say about it after we close the transaction in terms of strategy and where those businesses are and how we're thinking about them. For now, we're excited about it. I think it's an opportunity for us where you've got market-leading positions, you've got technology that's becoming more complex in their end markets. Those are businesses that we know how to run, and we think we can drive better and higher levels of profitability out of them over time. Gives us another stream of cash flow that we can deploy. Exposes us to the broader electronics market.
We can leverage the infrastructure of the company around service and support. We think there's a lot of opportunities for us to take this and contribute to our through-cycle growth rate targets of 6% to 8% on top line and drive operating leverage greater than 10% on that revenue growth. Excited to get there, and I'll have more to say about it once we're closed.
You sort of touched on my next question, but it's been a while since you published your most recent long-term financial model. I think since then you've executed really well on gross margins. You've continued to grow OpEx. What are your thoughts on overall profitability of the business today and where you can take it from here?
Back to your competitive question. If there's ever the one metric that stands out about whether you differentiate or not, you can look at the gross margins. To be able, despite pressures in the market and consolidation and everything else, to see gross margins grow over the last few years, I think reflects the strength of our competitive position in the market. We set these models up to give investors a perspective on performance at different revenue levels, and we've outperformed where we were over the last couple of years in the expansion environment. If you look at where consensus numbers are right now, it's probably a little bit below that model. We think about through-cycle expectations. I look at it like, over time, that's the right model to have.
There will be periods of time where I'll be on it, periods of times I'm outperforming it, maybe periods of time where in an industry's contraction, I might be under it. Generally, where it's at. The mix factor is probably the biggest factor that affects the gross margin profile of the company. Depending on the types of products we're shipping, we have a big portfolio of products, 18 different products or so, and the margin profile is not exactly the same across all of them. Then service, growing as a % and growing faster is a dilutive factor to that.
I gave guidance in 2019, given these expectations around sequential growth, second half, first half, and so on, that we'd probably be around 63% ± 50 basis points, as I look at things bottom-up and look at where we were sized to in terms of factory utilization and so on. I don't see any dynamics that's necessarily changing our view on where gross margins should be. I think the mix of our business will be probably the factor that drives it, with variability around the percentages I talked about.
Okay. Got it. We've got about three minutes left. I wanted to see if you have any questions from the audience. Where from?
Matthew. When you zero in on the competition, wrestling metrology for vertical NAND wafers, are you guys happy with your position, or it sounds like maybe that's a big focus.
There's an opportunity for us to get customers to invest more in metrology. We do have competitors in vertical NAND in metrology. The competitive environment in metrology is probably higher. Problems are becoming more complex. We've seen a nice improvement in the process control intensity, how much is spent on process control for metrology in vertical NAND. I do think there are opportunities. There's overlay, but then there's also this in-channel or channel-hole metrology opportunity, that if we can solve that problem, it creates another tailwind for us. It's a competitive environment, and we have to compete for this business. There are a lot of competitors that like these opportunities and tailwinds for growth of the SAM, but also the profitability of this business as well. I like our position.
Feel pretty good about it, and I think the history has shown that we've been able to maintain it and grow it.
Ma'am, one in the front. Ma'am.
Just back to EUV. Light for light lithography, 10 years' time versus now, you say EUV's good for you. How do you think about the TAM for you in lithography, just for lithography alone, in a fully EUV world versus a double patterning world?
The TAM for litho? The TAM for process control?
The TAM for process control.
Yeah, I think it's higher. One thing about multi-patterning is you had more multi-patterning steps and more process steps, but repetitive steps of similar process, right? In an EUV environment, in HVM EUV environment, anytime we've seen significant transitions of litho technology, and this is the biggest one, that it's been good for process control. The challenges, in particular, around the reticle create, I think, a unique opportunity for us to see some expansion in our TAM, and we're investing in that. I feel like it's good for process control. Certainly, as litho replaces some of the multi-patterning that was happening, it affects other markets differently. I think process control is at minimum neutral, and some of the opportunities around the reticle create some opportunities to see it grow.
Thank you.
I am confident we can do that. ASML's become a bigger competitor over time as they've entered the space, but the competitive dynamics have been fairly stable over the last four or five years. Given our product roadmaps, where I think that opportunity is and our current position in that market and some of the unique capability we bring to it, I feel very good about our position when we get there.
Okay. We're out of time. Thank you all for coming, and Bren-
Yeah. Absolutely. Thanks for having me.