KLA Corporation (KLAC) Earnings Call Transcript & Summary

[indiscernible] to be that some of our peers aren't as exposed. And it's -- there's a lot of legacy projects in China, and so they need wafers, they need reticles and being able to source them from the merchants out there has been somewhat problematic over the last few years. So that's incremental investment also for KLA that I would expect to see. So long story short, when you add up all the puts and takes, and we'll see how that plays out. I think we're operating from a new level, and I feel pretty good about our ability to maintain this kind of mid-7s moving forward.

And you think even as we look at -- again, this is out into '24. So who knows what happens then. But there's this school of thought that there's a bunch of this deferred WFE that your peers get that is part of WFE this year that your peers get that then won't get repeated for them the following year. So you think even as you look out into '24, you think that your share should sort of stay in the mid-7.

I think so. Look, if you have a huge growth year in memory investment, and logic relative to logic and foundry, that puts pressure on process control intensity. I think some of these puts and takes will play out, though. So I think as we said at Investor Day, I think 7.25% plus is what we thought we'd be driving to moving forward. We're sort of in that mid 7s now. And I feel pretty good about our ability to maintain that going forward through a combination of a lot of things to talked about.

Great. Let's talk about China. And maybe we can start by talking about how to handicap exactly what is lost from the restrictions. And how you think about some of the puts and takes are around this whole concept of gross versus net? Because obviously, there could be some add back. There could be some licenses granted. There could be some companies that are in the red zone that can move their process node into a green zone. So can you just talk about that?

Yes. I think it's going to be very fluid as a result of that, right, because the predictability even if that can happen in our customers are actively trying to work through some of those issues in terms of engaging with not just us -- also engaging with the government about really, really doing from a technology roadmap point of view. The challenge will be predictability of -- if there's a change and an opportunity to get licenses as it relates to some of low cap adjustments when that might happen. We tried to size [indiscernible] when you just add up the actual direct impact of in '23 of all the systems, a portion of that, about 10% of business. So a bit of a service had to that up in China. As I was saying earlier, when you look at what -- where most of the investment has been and if you look at this infrastructure investment, my native China business overall, I actually think it's likely exclusive of the impact of some of the government regs. I think it's likely down less than the overall market based on what I see today. So it gives you a sense of some of these other things that are happening. It could be those swing factors, if you backed your point about back relatively less for a company like KLA than maybe a more profit metric peer company that maybe has the benefit of deferred, but a bigger loss in China, so back to some of these moving parts as you look at '23 and overall share of WFE. Obviously, we actively engage with the government were full compliance with the rules as they're stated, and we'll see moving forward. Certain tools will be able to reallocate to other customers. But then that assumes that eventually as you keep pulling stuff forward that if you're able to continue to do that, then you're driving WFE up, right? So I think if that happens, then yes, you'll be able to reallocate the 1 or the higher WFE number. So in terms of direct impact, it's really hard to beyond the way I described it to get into more details on that. Obviously, there is -- there could be a desire to use non-U.S. suppliers. And generally, KLA as a differentiated supplier, we tend to have less exposure to that, but there's always that exposure if you're a customer and you're trying to get parts that are products that you can't get from somebody that's domiciled in the U.S.

Got it. Can we talk about process controlling intensity generally? And it used to be that process control intensity and foundry logic was about 50% [ higher ] than memory. And memory process control sense has been actually ticking up recently. And you come out with some new tools to address that. And you do -- if you look at your share of memory WFE, your share of memory WFE has actually gone up a little bit. So can you talk about process control intensity and maybe specifically, what's happening in memory that's causing rate process control intensity to pick up a bit?

Sure. No, it's a great question and [ some ] of spend a lot of time on. The transition from 2D NAND to 3D NAND was a good inflection for process control intensity. Now it does, to your point, it is about half of what logic and foundry is for specific reasons, right? You have more commodity parts, you have the payer redundancy memory -- and so for those reasons, customers don't -- in fact, they measure as frequently as they do in a logic environment. As you move from 2D to 3D, you move to vertical structures, those structures required more [indiscernible] think about our different angles you're measuring [ trench devs ] to measure [indiscernible]. You also saw more inspection or process tool monitoring to try to make sure that process tools are very clean before they start to process. We've introduced new capability that we've been trying to drive into production over the last few years of some new X-ray technology that can do high aspect ratio measurement and channel hole measurement through the stack, which has -- we've been able to drive from an R&D point of view, but trying to get it into production and therefore, higher volume purchases has been a bit elusive for us. So that's what we've seen in memory and in flash, and we've seen that continue to -- we saw that step function up and we're seeing it kind of drifting up a little bit. In DRAM, as I mentioned earlier, you've had the introduction of EUV and the DRAM [ solution ] infrastructure around inspecting and qualifying and ensuring that the reticle fidelity is maintained through the process. And so customers have had to purchase some of that capability from us. And so that was a contributor also in the last year or so. Certainly, as you see more design shrinks, you see more EUV layers, it does create the opportunity for us in memory to see a little bit more increase. But I think it's never going to be like a high mix logic foundry intensity would be. But there's some opportunities. And if we can solve some of these unique problems, it does great opportunities for.

Great. Can we talk about EPC? And maybe can you just talk about some of the macro drivers that are driving the EPC business? I think there's a lot of volatility inside of FPD and inside of the PCB businesses. But maybe just look out a little bit and sort of is this a double-digit grower, EPC over time? And what are some of the biggest drivers for the growth of EPC.

Yes. No, it's a great question. And if I just take a step back and go back to the decision to acquire Orbotech, which is the biggest part of PCB with the other parts of KLA that were merged into that group was that could we take. We wanted to move in and look for opportunities where our customers were doing a lot in the front end was getting prohibitively expensive to continue to drive most of their innovation and capability there. And so you were seeing this evolution in the packaging space. You're also seeing this broadening in specialty semiconductor that made the Orbotech acquisition in [ UCaaS ]. We got 3 businesses. But what we're looking to do is because we take businesses that expose us to a broader SAM in an area where our customers were going. That we had market-leading positions, there was technology evolve over time, and we could take that market-leading position to get more out of our operating competencies to process our KLA operating model. We went, we drove meaningful synergies out of the business, so about 2x, and we talked about this in our Investor Day about 2x of what we had originally targeted and tried to position the business such that we could grow on the revenue growth, we could grow the profits at KLA corporate-level incremental operating margins. And we've been able to do that over the last couple of years too that we -- not every business carries the same margin profile, but if we can find the ability to differentiate market-leading positions with evolving technology that will drive the incrementals various [ numbers ] for the rest of the KLA. So we -- I think we've been able to do that. And like I said, we got the synergies. I think we've done pretty well from an operating performance point of view. The PCB business has had some weakness in the mid-end of PCB. There's the low-end commodity with less participation for KLA in the mid-end, it's just principally around mobile -- mobile devices that's where we've seen some weakening. And then you have the advanced part of PCB, where you start to get in the substrate part of it where it integrates into advanced packaging, which was the most interesting part of that business for KLA because we believe that as you start to see pitches and density shrinks and density of lines and spaces shrink in the boards and the integration in the package that it moves towards a higher technology solution where we can leverage other capability within the company. So we're moving down that road. We have some new products that will come out at the end of this year and into early '24 that is encouraging from an advanced packaging point of view. Specialty Semiconductor is -- has also been a nice growth business this year compared to last year and we see the reports so you can see some of those data as we report it. It was a -- we acquired some niche process businesses that enabled us to run a; the process business in a way that we wanted to, but also be able to engage customers at a higher level and differentiate the market we can do that. And they designed specifically for some of these markets around RF, power and advanced packaging. And we've been able to go in and execute there, and those markets have inflected. And so it's been a very good business overall for KLA. We have some new products in the packaging space like plasma dicing as an example, where they've used mechanical or laser saws, which kind of create fractures in the package substrates that this is a cleaner way of singulating wafers. So there's opportunities there. Obviously, automotive is a big part of that, working with specialty substrates like silicon carbide and gallium nitride. And there's parts of the business that there are process tools that do etch and deposition for those substrates. Automotive, generally, we've been able to then take the channel within EPC, engage with customers, drive some standards into the automotive ecosystem, leveraging both inspection metrology with some unique machine learning that we've attached to those tools to not only look at electrical parameters and performance but also reliability, which is a big problem in auto then coupled with the specialty semiconductor part of the business. So we've seen our auto business tripled in the last 3 years. It's about $700 million across specialty, but also the inspection metrology parts of it. Flat panel is a more challenged business, not as much value for yield management there. We have seen some growth. We tend to be more in the technology by area of flat panel display. Over time, we would expect, as you move to more semiconductor like processing and like micro-LED, for example, that might create an opportunity for us to engage in a higher-value area. What really drove our acquisition was really to think about those opportunities that I talked about. And we did what we needed to do to ensure the flat panel business is positioned for those inflections, but at the same time, not necessarily a negative for the company in terms of profitability.

And EPC grew about 10% this year.

Yes, should. Give or take.

Next year, I mean I have it modeled fairly growing, but what's your outlook for your...

I think it's flat, maybe it's down a little bit. I think it's probably down -- a lot of it will be dependent on this issue and how quickly we see PCB 1 hand. We felt the pressure in that business earlier, so we might actually come out a little bit later. It's also predicated on some of these product launches that I mentioned. Specialty semi should grow next year. So that part of EPC should be positive. But yes, I think it's less than the declines in WFE, but whether the whole thing grows or not sort of flat, maybe flat to down a little bit.

And let's talk about the balance sheet for a second. So you've always had 1 of the most optimized balance sheet around. You did the special years ago. We're now [indiscernible] but you tend to think or you've tended to think a little differently about the optimization of the balance sheet, I think your peers have. And you have a lot of cash. You've got $9 million worth of cash. How do you think about optimization of the balance sheet from here? What is the minimum cash balance? Would you be comfortable taking cash lower? How should we think about capital return over the next few years?

Well, I appreciate the compliments earlier. I think sometimes companies don't realize how much value gets capped and how you finance the business and how you allocate the capital that it generates. And so we [ like ] to get all the capital the business generates valued is we've got to deploy it well and be very explicit about how we do it. Certainly, there were reasons for us to take a permanent tier debt which we upped a little bit to basically get back into our target range given the growth of the business back in June to finance the ASR. But our belief that [indiscernible] leverage given the level of business that are given the type of business we have, characteristics, service being one of those drivers that provides a fairly predictable and sticky revenue stream predictable earning stream that gets you comfortable with not only debt service, but also dividend levels. And then it's about putting the cash to work either back into the business, pursuing strategic opportunities, all that juxtaposed against the alternative of returning the capital to shareholders. And that sort of guides our decision-making process at KLA. And I think that we have pursued M&A opportunities, that's the alternative and that's how we think about it or we put the capital back to work in terms of a steady return strategy. We've raised the dividend 13 years in a row. So we think a balanced approach is the appropriate way to think about it and growing the dividend in line with the growth rate expectations of our cash flow. So if we believe that we can grow the top line 9% to 11% over time, we can grow our earnings 1.5x that revenue growth rate, we ought to be able to grow the dividend 10% to 15% per year, and that's how we generally think about it over time. And that you have to be committed to growing it and have an explicit strategy to get in value. We adjusted our cash balance up from $2 billion to $2.5 billion to $2.5 billion to $3 billion. We're managing the business [indiscernible] how we're managing across different markets. Some of what we have acquired is less mature in terms of some of the working capital optimization going on, and we're working through all that. But we think that's an appropriate amount of cash for us to pursue any of our strategic alternatives as they may present themselves beyond, we'll call it $1 billion or so [indiscernible] and that would require us to go out and raise additional capital, which we can do have [indiscernible] the revolver that's partially drawn, but not much there. So we think that, that -- given the business and where it operates today is a reasonable, albeit conservative view of what's required to run the business, what kind of cushion do we need to ensure that we can rightsize in any environment and have enough capital to pursue other strategic options and we're going to run our leverage in that target range and flex up as appropriate as strategic options present themselves. But we like the businesses we're in, right? Most of what we've done outside of -- from an inorganic point of view, has been more strategic purchases, some supply chain stuff. Things that are more make versus buy to enable our -- the growth of our portfolio, our positioning in WFE and to support the business.

Great. I had a question about bookings, and I know that you don't want to tell us bookings, but you had $13.5 billion worth of performance obligations that was in -- that was in the queue. So that's not backlog, but it's backlog without a timeframe with not a 12-month limit on it. So we can basically [ look ] at your bookings, they were down 35% sequentially. Book-to-bill was about 1.2 give-or-take. But 15 months' worth of backlog is a lot. And you see some examples of customers -- and this is not just a question of you. This is a broad question that I think may or may not be actually happening. But I get the sense that some customers are placing bookings simply because there is so much backlog because lead times are so long. Do you -- in your customer discussions, do you get any sense that because you go to them and say, "Hey, I have 18 months' worth of performance obligations, you can't get a machine until 1.5 year from today and they say, "Well, I don't know what the world's going to look like 1.5 year from today. So I'll just place bookings is just to have something in the queue. How much is the -- just to have something in the queue, how much is that a factor for your quarterly bookings.

No, it's an interesting question. The quarterly bookings is always a tricky one. And you've also got generally fairly consolidated spend and you've got high ASP. So -- sometimes it comes together in large deals in a given quarter. I don't know if it still implies anything when you have a lot of backlog about the next quarter and beyond. But it's -- it provides some visibility in terms of expectations going forward. Look, every customer runs through the process when they give an order as a CFO, look, there has to be a budget, there has to -- there's certain control aspects of approving things. But if you're a customer, not 1 of our leading customers, 1 of our secondary customers or a new customer, and you absolutely will have to give an order to get slots. And you're right, with extended lead times, at some point, customers -- look, I run through our process, I believe, based on my business, I will need a tool in a certain timeframe or they're building a new facility. And so they will plan for that. There's always the issue as anyone, you don't necessarily know what your business conditions are going to be at that point. Around certain products, if you run the risk of getting out of line, right, which is always a challenge for customers if they were to say -- if I want to delay, and I either lose that slot? When do I get another one? Does it open up the contract negotiation around pricing and things like that. So there's always those calculations out there that customers have to do ultimately when they get to point. But it certainly gives you some visibility that there was a strong intention to take this business in certain timeframes. As I said earlier, we're enabling so a big chunk of our products are for these enabling products, some of them are slotted way out in the future because it's just a fundamental gap between what customers want our ability to supply. And the lead times to get optical components equipment to make them and the raw materials for them is very extended, and we're not the only 1 in the industry with that challenge. Around those products, those customers are very strongly committed, I think, to those slots and it would take a very significant change in the business that would affect kind of their long-term planning around technology roadmaps for them to try out to -- or to get out of line for those kinds of systems. So there's always, I think, some movement there in terms of how to think about that backlog. But I -- like I said, I think if you look at what we have, there's a lot more of it that is more tied to products that are more resilient through periods of softness. And we just have to see. We also have some customers that are big customers that we work very closely -- we've never actually built some orders or they give us orders in a quarterly timeframe, right? So I book and ship in the same quarter, so it never goes into RPO. So there are some nuances in there. What we did report was the $13.5 billion or so and that 45% to 55% of it was deliverable outside of 12 months.

Yes. Can we talk about China? I'm getting a question about localization efforts in China and where the local vendors stand on process control. I mean KLA in as much as a software company exists in the semi equipment growth you're really more software company. That's where the deal expertise is, and that's where the secret sauce is. I think it's going to be very challenging for China to replicate what you do. You're very, very low on the list of what China can ultimately replicate. But can you talk about some of those localization efforts. Do you see any examples of a Chinese companies picking off, maybe lag into your business that you would have gotten otherwise?

No, it's a good question. And so far, given the lead that we have. We have a number of competitors competing us with full access to a worldwide supply chain and so on for years that have struggled on certain markets. I've got a lot of hardware engineers to disagree with you about our hardware capability versus software. I think one of the things is as part of software is our algo and AI teams where you can really harness the capabilities of the hardware and the system integration that we ship. And one of the big benefits of the Gen 4, Gen 5 product is we've been able to -- I would get a lot of questions from investors about e-beam versus optical and so on. And I think we've seen over the number of the last number of years because of some of the fundamental limitations and throughput and speed of being that it's a complementary product and can really meet production use case. Our big issues at KLA was, well, you're going to run out of light. Light hasn't necessarily been our problem sort of harnessing and finding and filtering what's real from what's not. What matters to signal noise and inspection has been our biggest challenge. We work on a fair amount of some of this capability. With these product lines, we've been developing and incrementing for years, we have a very tight and in a lot of cases, exclusive supply chain -- so obviously, there's supply IP that's part of what we're able to do, along with what we do with KLA. We have the scale to have our own internal design teams. So it makes us a fairly formidable competitor. We also compete with a portfolio of products that allows our customers to balance technology and economic motives and not necessarily have one tool that meets all their needs across. And if you're a point product competitor, you have to say, okay, this tool can do everything or there's ways that you can avoid doing that. And we can be a little agnostic here in how we do it. Customers can confuse to depend on where they are in terms of process maturation on their particular nodes. It makes us fairly formidable. We understand the value. We can differentiate we can share in the value with customers that drives the gross margin which I believe is the best indication of differentiation. So it just makes it very hard for, I think, any upstart competitor to break into the space just given the legacy of -- in history of the position that we have in the market.

And [indiscernible] we have about 2 more minutes left. Can we -- can you just talk about -- there's always discussion about legacy process nodes becoming a bigger portion of WFE every time. If you listen to Applied, they'll say that legacy nodes have gone from 1/3 and half of WFE, of course, it depends on what you call legacy node versus a non-legacy node. But -- can you characterize the expression intensity for you in the legacy node versus the leading nodes? Is it that different?

Yes. Actually, it is very different. And we -- when we characterize it, we talked about 28-nanometer and above the legacy and anything less than 28-nanometer is edge. Now most of our business is 75% to 80% of foundry/logic is at leading edge and it's in the 3, 5, 7 nanometers, there's a little bit in that interim space between 28 and with some specialized offerings from different customers. But most of it has the most advanced nodes. If you're building a legacy fab and you're seeing some of your markets in [ collector ] around industrial or automotive or communications infrastructure of those [indiscernible] in the market, and you're building a part you've been building for a while and you've got a fairly mature process. Then you're going to add capacity, the process control requirements are fairly limited, it's working. You just deploy what you've been doing before. If you're doing a design rule shouldn't change for different capabilities or are there changes around specific thing as it relates to the liability, automotive as an example, and that's causing a spec change in those parts. And that does change how those customers look at their process potential strategies. So exposed to it, except in those situations where customers may be a change in how those are the particular market. Some of the logic/foundry legacy investment in China has been more upstart and more scale. So you're seeing a heavier level of investment there. But when I aggregate the whole legacy environment, it's more process control intensity is more like a memory fab where an established and established example I talked about is going to be somewhere, we'll call it 8% or 9%. The China piece a little bit higher we'll call it, 10% or 11%, you start to blend it out. You're kind of in that 10-ish percent versus a high mix foundry, leading-edge foundry is more in the mid-teens or higher teens.

Got it. Well, we've run out of time, but that -- thank you, Bren.

Thank you, Tim. I appreciate you having us.