KLA Corporation (KLAC) Earnings Call Transcript & Summary

Welcome to day 2 of Citi Virtual Technology Conference. My name is Atif Malik. I cover U.S. semiconductors and equipment stocks here at Citi. It's my pleasure to welcome Rick Wallace, President and CEO of KLA; Bren Higgins, CFO; and Ed and Kevin from the IR team. Rick is going to make a few opening comments, and then we'll do the fireside chat. If you have any questions for the KLA team, please e-mail them to me. That's atif.malik@citi.com. Over to you, Rick.

Thank you, Atif, and thank you so much for hosting us today. Hello, everyone. I hope you're safe, your families are safe, kids are back in school, at least virtually. Just a reminder that you need to refer to our SEC filings for safe harbor language regarding our comments today. Let me just start off by where we are. About a year ago, we laid out our 2023 plan for KLA in the financial model. And the first half of 2020 is looking very good relative to the expectations we had for that plan, some of the initiatives that we put forward there. And those of you who are at that know we talked really about the KLA operating model as the central guiding principle in the way we're running the company. And I've got to say I'm really pleased around how the operating model has performed in 2 ways. One, in terms of the integration with Orbotech and the success that we've had in the Orbotech business and how we've applied that and we'll talk to that. And the second one is how we've dealt with the obvious challenges that we've had throughout 2020 relative to COVID-19 and how we're performing. 2020 is shaping up to be a good year. And I think many of you know this, but coming in, 2020 was looked to probably have some growth, in 2020, from a WFE standpoint. KLA certainly expected that. We went through some uncertainty, but -- earlier on, but feel very confident now of our ability. And our customers continue to invest strategically. And I think that's playing out well for us. Our markets are working. Process control is working. Our leadership is working for us, and we're demonstrating our ability to gain share with some critical new products, and we're really excited about the performance that we're seeing out of there. So the way the year looks, for the rest of the year, we feel pretty good about how it's ending, although there's still a lot of uncertainty. The one comment I want to make is the obvious question people are having in the last few days is around U.S.-China trade. Obviously, we're aware of the media reports that are associated with this. And we do have an important customer that's being talked about. It's an important customer, not a particularly big customer. But we're going to, as you might imagine, comply with all the rules and regulations around what the decisions have been made. But to date, there have not been any decisions. And so we're continuing to monitor that, and we'll do whatever we need to do. So if that comes up, we can address it. But I think the other factors are much more interesting, especially as they pertain to where KLA is today and where we think the future is going to take us. We have some underlying drivers that are still in place. And you might say COVID-19 is accelerating them. And one is the 5G infrastructure build-out and the -- that's driving different parts of our business, including semiconductor. High-performance computing continues to be really important for us, and we're seeing a lot of strength from that. And we feel really good about our ability to execute. Also services, one of the things we weren't sure about earlier in this year is what would happen to utilization rates in our service business, and yet that's been quite strong, and we're seeing at that historical growth level that we talked about for the year is panning out and for the 2023 model. Also, as Atif mentioned, I'm joined by Bren Higgins, our CFO, and he's going to help me out in the fireside chat and take all the really hard questions. And so I'm glad that he's here to do that. So again, thank you all for joining, and Atif, thank you, and let you start the Q&A.

Thank you, Rick, for that background, and I'll go back and forth between you and Bren. Rick, if I look at the investor [ negative ] on KLA, I feel like it's changing. It has changed over last 1.5 years or 2 years as you have diversified and expanded into mobile and 5G markets through Orbotech. We recently hosted a call with Oreste, who runs your EPC, Electronics, Packaging and Components business, and learned that 2/3 of your PCB business and 40% of the specialty and RF business is tied to 5G. And that kind of agrees with our assessment that you guys come on top among front-end equipment suppliers being exposed to 5G. And you have very good history of transformative acquisitions. Can you talk about the shift from a pure process control semiconductor business and how EPC part of the business prepares you for the data center era?

Yes. That's a great question, and thank you for that. I think that when we went back a couple of years ago and we were laying out growth strategies for KLA, we assessed what is it that we're doing well in terms of how we're running the company, and that's where we really codified the KLA operating model. And we looked at what we were doing and how we were running these businesses. Many of our investors know, we have a lot of businesses, a lot of relatively small that, in aggregate, make a good-sized company. But we really learned how to run those businesses successfully using the operating model and had done that with acquisitions in the past. So we weren't so intimidated by going in the process, which is what SPTS is, as long as it satisfied the real important criteria for KLA, which is a market-leading position, having sustainable technology advantage and be able to support our customers in growth segments. So when we looked at Orbotech, we're pleased with what we saw there. We have brought technical synergies, and we'll continue to in parts of those business, but definitely, channel synergies. And some of the success we're seeing related to Orbotech is the fact that the operating model and the success we've had with some of these customers that are now investing more heavily in packaging is really playing out, and they view KLA as a trusted supplier. And so we're having very high level engagements. So we feel good about that. And we think, as you say, we're on -- we're really pleased with what we're seeing in 5G build-out. And I know everybody knows this, but this is early days for 5G. This is a several-year investment thesis. And we feel good about our position right now, and we have more products in the development stage to support that as it goes forward.

That's good. Rick. The company has done a great job in maintaining high market share in a fragmented market. You're #1 in 8 of the 10 served markets. You've proven many [indiscernible] wrong, including myself, in maintaining this share. What sets clearly apart -- or what is the secret sauce around holding this share?

Yes. I think that one of the -- and as you know, Atif, I've been in the company 31 years or so. And we've had this history of being able to be both paranoid and continuing to evolve and develop. So we have this model that we've looked at in the past where we really understand if we collaborate closely with our customers so we can anticipate their problems well ahead of when they're going to be needed, then we can innovate and drive that innovation and then if we can execute. But over time, we got really good, I think, at all 3 of those things, is collaborating with customers being engaged in the very leading edge, trying to understand what problems they were having and where we could provide capability. We dedicate a lot of resources, a lot of R&D to the innovation to the -- making sure that we have capability when it's needed. And then we got good over time at executing. The other thing is structurally, we've set up these businesses to have a singular focus on these market segments while at the same time having an overarching view. So we have general managers. We have engineering teams that wake up every day dedicated and worrying about how to be successful in those market segments. And they never rest. And so from the standpoint of that, we continue to innovate and invest, and I think that's proven out. We think we can scale that. We've seen it with other acquisitions where we've been able to take their positions and continue to grow on it. So that is one of the overarching principles for KLA is market share, but also differentiated products which show up as gross margin. So part of the other thing we measure is, how differentiated are our products? And can we create enough value that we can share that with our customers? And we think that shows up if you look at our gross margins.

Great. Rick, I really like the way you described the process control as a portfolio of products. You are technology-agnostic and offer a range of optical and e-beam products that are right in time for the market, for the volume introduction, offer compelling cost of ownership. Team expects to grow sales by $400 million to $500 million through 2023 on new products like Gen 5, which had, had great success, e-beam, x-ray. Can you just talk about EUV positioning and confidence on both the mask and patent wafer inspection?

Yes. Sure. This is a great question. And it goes back -- I'll do a little history and then talk about it because I was the General Manager of what was called the brightfield inspection division when we were looking at merging with Tencor many years ago. And Tencor was darkfield. And we felt like you could solve every problem with brightfield. And we came out with new products that were great and I go talk to customers, and then they were still going to buy some Tencor products. And I was trying to understand why, and the answer was they were trying to optimize the portfolio. And so when we merged with Tencor and then later as we added products, we realize the best thing we can do for our customer is offer them a choice, so that they have a choice between what fits their strategy, whether it's brightfield, whether -- or broadband plasma now or it's darkfield, laser scanning or it's macro inspection. And so we did that both in inspection, and we also did it in metrology. So when we fast forward to EUV, there are some people that thought there was only one answer for EUV reticle inspection. And our view is that's not true. What our customers have told us for years is they want to optimize the cost effectiveness of their solutions. So the real question is, what is the most cost-effective answer to solving the problems of defectivity in EUV? And it turns out that taking an optical reticle inspector and coupling it with a broadband plasma to check the print after you've printed the device is right now the most cost-effective answer as opposed to people that thought years ago, we needed to have an actinic reticle inspector. That was never going to be practical from a financial standpoint or investment. So our view is, if we provide customers with the most cost effective solution, that will solve their problems and allow us to continue to invest. We do think there'll be a role for actinic reticle inspection, but concurrent with high-NA EUV, which we think is a few years out. So EUV does create an inflection. It's increasing process control intensity simply because smaller defects matter that didn't matter before, and also there are additional challenges associated with the control, the process control around EUV. So we've seen an increase, and we think that process control intensity goes up as our customers drive EUV. But until we're really in high-volume across multiple, we won't know the extent of that. But it's definitely a tailwind for our business.

Okay. Rick, specifically if you could talk about how the eSL10 e-beam product has been received by your customer? And what is so unique about its architecture?

Well that's a good example of, one, where I go back to the -- we wanted to have differentiated products that would not be a "me too", and the way we had measured that was based on whether or not the gross margins were going to be good. We exited e-beam years ago, e-beam inspection, when we felt that we didn't have a compelling differentiation in that market. And so we stayed out of that market. We think it's an important market, but not a big one because as much as possible, people are going to offload their inspection on to the least cost solution. So you take e-beam and try to offload it to broadband plasma, try to offload a lot to laser scanning and try to offload it to macro, if you can do it. That's the progression. So -- but when we finally had an architecture for the eSL10, I have to give that team credit because they really -- we invested very heavily in that platform. But the other thing we believed, and it's proving to be true. If we coupled it to the broadband plasma, we would create a solution around those 2 so that we can make both more effective. And what we've seen with the early adoption of eSL is by having the optical tool in complement to the e-beam, the e-beam has got its own technological advances, but it's also as part of that solution, it's gaining a lot of traction with customers that, frankly, were skeptical. They were somewhat skeptical because they felt like we had abandoned e-beam, which wasn't the case. What we were waiting for is to have an entry point we thought we could differentiate. So I feel good about where we are in that product. It's still relatively early days. The team has done a great job. But we're very pleased with the customer reaction. There are always going to be some defects that are unique, uniquely found by e-beam. And that -- what percent of the market that is historically for the last 10 years, as you know, it's probably been 20% of the overall market between e-beam and optical. We don't think that actually changes that much, but we think this is an enabler to drive us to gain share, but also to continue to drive the broadband plasma.

Very helpful. Bren, switching to you. You have talked about a $55 billion WFE this year. Rick mentioned everything seems to be on track, if you ignore the kind of the macro noise. Just focusing on memory first. Can you reconcile how the elevated levels for inventory and software demand that we're hearing matches up with your view in second half of memory investments improving? I thought your view was far more realistic, and you also talked about some risk in memory because of the weaker prices. So how is memory shipping out in second half this year?

Yes. Tim, thanks for the question. And so sometimes when we talk about WFE levels, when we look at 2019, we saw it was about $52 billion. And when we talked about this year 2020 being somewhere in the mid- to high 50s, just to give sort of a range of perspective. And I don't think that's -- everybody counts it a little bit differently, but I don't think that our view is all that different for most of our peers. Obviously, foundry/logic has been a big part of this year. Second half looks like it's even modestly stronger than the first half. And so that's a good sign. And just echoing back to earnings, we feel very good about just where the business is, some of the context we provided, both in terms of guidance, but also views on the second half, and there's really no change to that. We thought the second half had stronger memory mix of business, mostly weighted to the December quarter with some breadth. And again, we haven't seen any real change in that. Now I think that's fairly modest. I mean, if you keep in mind that 2019 was down about 35% or so for the memory market. I mean, really since the middle of 2018, we've seen memory switching to this digestion phase. And so 2020 looks like it's maybe flattish or so overall for the year. And like I said, right at the end of the year, we do see some pickup, but I think it's off a pretty low base. So we think memory improves in '21, and I agree with you. I think some of the end market dynamics are pretty modest and certainly don't signal a strong recovery. But we do think as we move into '21, we'll see more strength in that business. I mean certainly for the mobility cycle that we'll see with 5G and some of the data center drivers that are there will be positive, both for DRAM for flash. You've got gaming cycles that will happen. So we think overall, the end market dynamics will improve. But I think we're still talking about operating off a fairly low base. So as we look at next year, we see next year as a modest growth year. We see foundry/logic sustainability, both in terms of breadth of customers, levels of investment as we look at next year. And there's -- and that memory does grow a little bit. But even if memory grew 10% off of the base, and that drove sort of a mid-single-digit kind of growth year for '21 is really how we're basically seeing it today. So we'll have to see how it plays out. One thing about these customers and the need for capacity to satisfy incremental bit growth beyond technology investment is that they have to make quick decisions and quick turns in the business. So I would expect if we see some end market strength that you'll see those customers respond fairly quickly with not a lot of lead time for us. So we've got to be flexible enough to respond to these scenarios and I think we're in a pretty good position to do it.

Great. And then switching to foundry. There was a concern earlier this year that the Taiwan foundry CapEx to be more like first half loaded, but it appears that other foundry/logic customers have stepped up. In second half, you talked about sustained outlook for foundry/logic in second half. But more interestingly, you see foundry/logic improving into next year. What's driving the higher foundry/logic investments from a demand perspective?

Well I think when you look at the 7-nanometer design start activity, which has been fairly strong. And then you'll have 5-nanometer ramps that will move into next year. You'll start to see 3-nanometer development. And so you see this from the leader, but you also have, as we talked about breadth, there's a fair amount of end market activity that's supporting multiple customers to invest. And so I think that, that continues into next year. You've also got a fair amount of demand stratification. There's a lot of trailing edge investment that's happening in China and elsewhere. And I would expect automotive and some of the things that have been weaker this year to be stronger into next year. So when I look at it, I think it's at least flat. I haven't been specific about percentage growth as we look at next year but I think the underlying drivers look very solid as we move forward. I think also on the logic side, I think that as those issues there start to -- on the leading edge, start to get resolved as we move into next year, I think that could be upside to the business as well. So I think that on multiple fronts, we feel pretty good about those segments.

And then how much of a headwind or a tailwind is potential microprocessor outsourcing to foundries in the future?

Well I mean, look, I think that's one of those questions that I think that's out there that it's really hard to see at this point. I mean, certainly, the behavior from the customer has been very consistent and a focus on developing leading edge. I think with the leading edge investment, I think that you can only -- you have to -- you're in it or you're out. And so I think that in not talking about specific customers, but could there be additional outsourcing? Sure. But I think that the commitments we've seen to the leading edge road map are the same and haven't changed, and we look forward to supporting the customer through that.

Great. Coming back to you, Rick, you guys are helping the machine learning AI revolution. Can you talk about the role of machine learning AI across all your products? Longer term, do you think KLA has an opportunity to use a disruptive feature like machine learning, AI across all your inspectors to offer a platform like some of your peers, ASML, talk about? Or this capability will always be closest to the customers?

The -- it's a great question. I think the notion of advanced algorithms and what we're seeing with machine learning or AI has been a big part of what KLA has done for years. I mean if you think back in time to some of the earlier days and wafer inspection, originally, we had a very straightforward align and subtract an image and the difference was a defect. That was the basic. But soon after, we started realizing if we started doing image processing that we could have more information and squeeze more technology out of the capability. So we invested heavily in algorithms and in development. And as the AI technology became more and more affordable in the last several years, and frankly, as the complexity of our tools went up, we started applying AI to our inspectors a few years ago. And recently, we've been able to expand it so much so that we put together a team inside, which develops AI algorithms and capabilities for multiple inspectors. So one of the questions you asked, eSL10 takes advantage of that, of the fact that we have investment that's being used for broadband plasma that we're able to leverage some of that capability. And over time, we'll expand it to -- and we've already started expanding it across all the inspection platforms. And then also, we do similar things in metrology. So we're already doing it on our tools. We're heavily invested in that capability. We've established centers of excellence where we're hiring people that are doing that. And we're kind of doing it on 3 fronts. One is for the inspections themselves to optimize the capability. The other one is for the setup. The setup of our systems can be incredibly complex because there are so many degrees of freedom. So how do you assist people in optimizing the inspector setup itself, and using learning for that. And then the last one we've been applying some is the logistics around service. And managing and moving a lot of the parts that we have as a company. So we kind of view it as multifaceted in terms of how it impacts. With customers, I think that there -- we're really trying to solve different problems. Because what we're trying to solve is the inspection measurement problem associated that's around our tools. What they're trying to solve is the -- how do you run the fab problem? And we have some cases where we have some software offerings where we assist in that. But our philosophy has always been to plug-in and be part of whatever system that they have an open architecture that they might have, so that we can be complementary to it as opposed to trying to replace them. So we feel good about it. We have good engagements with both the producers of the semiconductors that are used for AI. For some of them, we're their leading application in terms of the most complexity that they're dealing with. So we have some pretty good R&D developments with them. And then if you look at the challenges our systems face, we handle a tremendous amount of data. What we viewed AI as being able to squeeze out additional nodes of capability with existing hardware. So we think it's a way that we can extend the life of our tools by providing more capabilities. So we've been hiring those resources, and we feel good about the progress that we've made.

Great. Rick, a bit of a different question. How do you manage talent at KLA? I mean I -- working in the valley and I hear a lot about your flat organization structure, you've promoted the most kind of leadership roles within the company, a very high retention rate. How do you manage talent?