KLA Corporation (KLAC) Earnings Call Transcript & Summary

So I'm Joe Quatrochi, the semicap analyst here for Wells Fargo. Excited to have KLA here this morning, CFO, Bren Higgins; as well as Brian Lorig, the EVP of Global Support and Services to discuss KLA's services business, which is something that we think is an underappreciated part of semicap and as well as KLA's business. So maybe first, let's pick on Bren for a little bit, ask a few questions, and then we'll kind of move the discussion to Brian and talk more broadly about the services business.

Bren, let's start. You guys just reported earnings a few weeks ago, but '24 outlook. Kind of maybe help us understand the demand visibility looking into next year. Can you talk about just the different drivers of demand, how you're thinking about recovering leading-edge, foundry logic, stability of domestic China, memory, et cetera?

Okay. So thank you for having us and really happy to have Brian with us today. I've been getting a lot more investor interest in the service business of KLA. I believe it's pretty differentiating relative to the other peers in the industry for very specific reasons. And so hopefully, we'll get into some of that and shed some more information and light on what's a really compelling part of KLA. As far as the overall market, we talked about a demand environment that was more or less stabilizing here, and we've been saying that for a couple of quarters. And we've been at slightly higher levels. When we first started in the June quarter, we were roughly $2.25 billion when we came out with that commentary and we just guided [ $2.45 billion ]. So it's been a little bit higher. But generally, across the industry, we've seen different mixes of different customers and different customer types, but generally operating about this level of business. And our commentary was more that we don't really see things changing all that much in the first half of '24. So while it doesn't seem like it's getting any worse, it's not clear where we would expect to see a meaningful resumption in sustainable demand growth. '23 has been an interesting year in that it's been a little bit of a confusing one in terms of some of the supply chain challenges in '22 and how that affected '22 performance for a lot of companies, but also the '23 WFE where there's some recognition of some of the revenue that probably shipped or at least portions of it shipped last year that shows up this year. So aligning on what that number is for this year is less clear. We talked about 80-ish, $80-or-so billion. It could be $85 billion when it would all shakes out based on the most recent views of performance and guidance. If you look at -- if you go back to the beginning of the year, whether we think -- we thought that memory would be weak, and it has been weak. We had a little bit of upside that we saw in China related to some clarifications around technology thresholds for DRAM there. So that was a factor that was positive. But in general, it's been weaker than we expected earlier in the year. We thought it might be slightly higher than where we ended up and then we had this other benefit. I think the leading edge has been worse. We've seen project time lines generally hold but capacity adjustments. And so customers are continuing to do road maps and do development work, but in terms of the capacity plans have been moderated. And so that's been a downside factor this year. The legacy parts of the market have been stronger, both China and non-China, I'm sure we'll talk a little bit more about China. Some of the business in legacy overall is really that the slots have now become available in '21 and '22, our largest customers -- most strategic customers consumed most of the supply that we had. Demand was very strong. And so while we had orders, we were prioritizing in a way to support our best customers, to support the industry because it was important to support those customers that were providing semiconductor to us and to our peers because, ultimately, if they can't get others' tools, they're not going to take ours. And so we were prioritized in those ways. And so this year, as we've seen some of those adjustments, we've seen some of the business now pull forward, and we've shipped against it. So normally, KLA is -- 75% of our business tends to be in logic/foundry, tends to be leading edge. And this year, it's been meaningfully less than that, which is interesting that how it's affected our relative performance because normally, you would think with that carryover effect, which didn't really affect us this year, but also less investment at the leading edge, the fact that we think we're going to be at least in line with the industry, maybe better, I think is a testament to some of the changing requirements and demand that's happening in the legacy nodes. So we've seen that improve both, like I said, in China and non-China. And then finally, we benefited from -- and I don't know if this was fully known by investors, but we do participate in what we call the infrastructure part of the market where customers are doing silicon wafers, and we're a meaningful part of the CapEx for silicon wafers, but also for reticles. And so both in China, but outside of China, there's been investments supporting those parts of the industry. And we have a very strong position and we're a meaningful part of the CapEx. And so that's been a nice upside factor for us in '23 as well. So I think as we look forward into '24, we're, I think, encouraged by what we're hearing from customers in terms of improvements in their business. On the logic/foundry side, the leading edge, we've seen utilizations improve. Memory, not so much. So we'll see how that plays out over time. We'd expect to see those improve as we go through '24. And I think there's some of those adjustments that happen on the leading edge in '23, hopefully, we'll start to see that show up in '24 as well. But at least through the first half, based on what we see today, I think it looks pretty consistent with where we're currently operating.

And I think one of the things I've been continuing to be surprised or just kind of trying to maybe reconcile, your lead time still for optical inspection tools are very long, right? And so I mean, how do you -- when you think about the second half, I'm not asking you to guide second half of '24. But like when you think about looking into second half '24, given those lead times and where your RPO or backlog is today, like do you need to see like a very material kind of reacceleration of bookings to think about growth in that second half or for the full year '24? I guess how do you think about lead times kind of factoring into that in the backlog?

Well, so if you look at the backlog, it's at an elevated level, but you have to segment it a bit. I mean, there's a lot of backlog that's tied to some of these projects, these legacy projects in China, and you've got a lot of new customers. And these new customers have to give us orders that tie to greenfield project schedules for new factories, but also have to perform as new customers to get commitments from us in terms of resource allocation, slots and so on. So they give us orders, give us deposits and that. So that's a portion of it. Then there's -- there are portions of the backlog that are tied to things like optical inspection, where we have a fundamental, particularly around our Gen 4 product, imbalance between what the market demands and our ability to supply. And the ability to add certain optics, components to add that capacity, it takes a very long time. There's equipment that's required, there's facilities. There's automation. There's a lot of training. And in this case, there's one supplier in the world that does this effectively. And so it is something that will just take some time. I would expect -- I would clearly have shipped more this year and -- if I had the supply. And I do have more coming on next year, so I do have some confidence in growth in that business next year. But I still look at how much backlog we have and how much capacity. And I think there's another increment that's ultimately going to be required as we move out over the next couple of years or so. So I expect that to continue to be strong. There's been a lot of demand for products that support the legacy markets, some of our, call it, our N minus 1 or N minus 2 products. And so we've seen some shortages there. And so we've taken supply chain from a very low level for a while and then ramping them up. It's hard to restart. And so we've got some supply imbalance there. And then finally, the reticle and wafer industries, again, as those have ramped and what those are driving in terms of the product types, that has also put some pressure. So there's -- the classic capacity parts of the backlog where we got capacity orders, those are -- those have some fluidity to them. Customers can push that and move those around. Those are tied process tool schedules, film measurement tools. You've got to take the films tools. You don't need the films tools, you don't really need the measurement tools, so they behave like capacity products in some ways. So those areas are a little softer, but in some of these other areas, firmer and decent visibility. Orders tend to be lumpy. And I think it's an indicator in some ways over time, over multiple quarters of kind of the current view of the market and the urgency in the market. So I think as I look forward and start to see some of that, maybe that will be more of an indicator of confidence of some improvement as we get into the second half and so on. But right now, we've been consuming about 500 million or so of backlog per quarter for the last year or so. I think that trend probably continues. But we've got some big integers, too. So orders can be lumpy because you have -- you don't have that many customers. But also when you have tools with ASPs that approach $20 million, and depending on those schedules, it can skew those numbers a bit.

That's helpful. Maybe just kind of one last question and then we'll shift to services. Maybe a year ago, roughly a little over a year ago, you had an Analyst Day, you talked about 2026 target model. If WFE is flat in '24, obviously, we're down this year, how do you think about just like the trajectory to hit that target model and your confidence level?

Yes. My confidence level is pretty good, and I took our Board through this about a month ago at our Board meeting. And then if you look at just historically, you go back and say, okay, where we're down years of WFE and how much growth had to happen between a down year plus 3 years, right? And if you look at -- to enable the WFE number, we think that makes that target model possible of $14 billion, plus or minus $500 million in revenue, if we require the industry to grow about 15%. Well, '19 to '22, the industry grew over 20%. So I think we're in position given the investments we made over the last few years in supply chain to scale the business, I think we're in a position -- I'm not worried about being able to scale at those kind of growth rates. So we've adjusted the WFE number. There are 2 kind of key assumptions. If you back up and look at the broader assumptions is the semiconductor revenue grows 6% to 7% online with this trajectory to $1 trillion by 2030. The capital intensity will rise and grow a little bit faster than overall semiconductor revenue that because of the mix of the business being more logic-foundry centric, that would drive process control to probably grow a little bit faster because process control intensity is higher in logic/foundry versus memory. And then if you look at where we expect the markets to inflect around things like optical inspection and our relative position in those markets that, that creates an opportunity for KLA to grow a little faster than that. Service will grow 12% to 14%. Brian will talk about that. And our EPC business will grow somewhere in the high single digits or so. We took a little bit of a haircut on the WFE number because of some of the export control dynamics that have materialized over the last 18 months. But our share of WFE is also higher. I think we're ahead of our plans in terms of expectations about our share of market. And if you look at some of the drivers, AI-like drivers that will materialize over the next few years as it becomes a bigger component of capacity, that those tend to play to KLA strengths in terms of larger die, more complexity at the leading edge. More value in those die, higher defect densities, which affect the yield with larger chips. So all those are, I think, positive factors that drive higher levels of sampling and play to more process control and particularly process control around some of the KLA's key products.

Perfect. Maybe let's shift gears a little bit to the services. Brian, just let's frame the discussion. Maybe can you first just give us some key metric statistics or just kind of an overview of the services business and just kind of what is made up of that organization?

Yes. Thanks, Joe, and thanks for having me today. So yes, our services business is a, first off, a customer-focused organization. We're an integral part of our customers' fab operations, so we get an opportunity to earn a lot of business. We make up about 23% of KLA's overall revenue. And it's a very predictable, durable revenue stream. And part of the reason is because it's more than 75% of that revenue comes from subscription-like service contracts. And those service contracts, the average duration is greater than 3 years, and we have about a 95% renewal rate. So it's a very, very sticky business and one that provides a lot of value for our customers and one that's a great business for KLA.

Perfect. I think at a kind of broad level, right, people think about services as kind of maintenance or break/fix, replacement of spare parts. But can you help us kind of understand, your tools are a little different than, say, like a chamber-based tool from Applied or Lam. The services that you provide for your customers and just kind of how that's different than maybe the kind of traditional break, fix?

Yes, absolutely. I think the best way to think about it is through our install base, the makeup of our install base. We've got more than 48,000 tools. And we're a little different than some of our process peers in that we're a high mix, high complexity, relatively low volume or lower redundancy and long-lived installed base. So if we kind of double click on that from a high mix perspective, if you were to go into one of our customer fabs, you'd see all our peer tools and many of those would be the same chambers. And you'd see many, many KLA tools, white panels with purple stripes that all look the same. But if you look up at the product name, you'd see more than a dozen different product lines. And if you look inside of those product lines, we have more than a dozen products inside every one of those product lines. So it's a very, very high mix which creates a challenge when you think about idling some of that equipment in a lower utilization environment, and that's where you get to this lower volume or lower -- or excuse me, lower relative volume or lower redundancy. We got thousands of chambers. Utilization goes down, I can idle some of that equipment when wafer starts reduce. When you're talking about inspection and metrology equipment, you don't have that many of any given product inside the fab. And so the ability to actually idle that equipment even when wafer starts decrease is very challenging, one, because your yield is always -- has an ROI. And so even in down markets, you're still trying to improve your yield. You may increase your sampling rate to utilize that tool. But then again, the other part is, again, this -- I don't have redundancy. There isn't another tool to move some of that production to. So this mix and lower redundancy is a big component. The complexity piece, as our customers continue to push technology, our tools are getting more and more complex. That means the need for more and more specialized labor to support those tools as well as all the subassemblies, hardware required to keep the tools running. And those are, again, very specialized subassemblies that we supply. So the complexity piece, and then this long-lived piece, we have tools that are in the installed base that have been out there for more than 30 years. And so -- and our tools, they continue to provide a long value for our customers. And so this creates another characteristic of our installed base. And so you push all those together, and it's a very different model than some of our peer companies where it's a little more volume-based. And that's why we're able to maintain growth in a down year like this.

The fact, we've only had one down year since the -- as far as I went back was KLA-Tencor merger in the late '90s, and that was in 2009. And even in that year, was down a little over 10%. So it tends to be resilient in downturns because customers, even though they'll lower their utilizations, they tend to keep utilizations higher for the reasons Brian talked about. And when capital is tight, the best thing they can do is drive higher yields. They don't want to start any more wafers, don't want to commit any more capital to inventory than they need to, to meet their customer requirements. So sampling tends to pick up. And the demands on our tools are pretty high. They don't have many. They have high uptime expectations. Have to keep the tools up 90% part of the time. They have to match performance across the different tools in terms of what we're identifying on the wafer, whether it's measurements or defects. So really hard for the customer to manage all that on their own and do self-service. We don't really sell consumables. And so -- and then there's all the complexity in the parts and the ability to procure those parts that we have a pretty tight grip around. And so I think for all those reasons, it tends to drive some resiliency in the model. And customers tend to buy the contract. I don't want to deal with it, you guys take it and we have to perform against whatever entitlement that gets established, which we can then customize. So it's an important aspect of KLA, and it carries a nice profit stream as well. It has a predictability to it. It certainly has been a factor over time that has influenced how we think about the capital structure of the company, how we think about capital allocation to have this very solid, predictable business, 75% that's subscription-based that we can then optimize cost structure underneath and deliver in a profit stream in a pretty predictable way.

Yes. I mean, I assume that you, obviously, don't just disclose the profitability, but fair to assume that, that profitability of that business is fairly mapped with like your dividend?

Yes, yes. With dividend and debt service, more or less.

Yes. I mean, next year, you guys have talked about the reacceleration of growth back towards the long-term growth of the 12% to 14% range. Can you just kind of maybe like rank order list or talk about the different drivers of what's driving that reacceleration of revenue growth looking to next year?

Yes. So we're going to finish up 2023 with high-single-digits growth on the business, which, again, is -- we think is pretty good performance given the down WFE year. And that's primarily because our service business really isn't a WFE business, much more of tied to revenue, assuming revenue and OpEx budgets. So high single digits this year. We think we return to the 12% to 14% inside of -- in 2024 and beyond. And really 3 big components. One is, of course, install base growth. Our service business does grow faster than our install base. We had the huge years in '21, '22, and even this year, healthy installed base growth. So we see continued opportunities to service our new tools, which, again, with the increased complexity, is increased service intensity. Then we have this long-lived aspect that I talked about, about the existing install base and the fact that many of our tools don't retire. So you don't see a fall off -- while you're adding new, you don't see a falloff of the existing. And that's partly driven by the continued demand in legacy. So we have tools out there, as I said, 30 years, many in 10 or 20 years old. Our customers have long depreciated those assets. And so there -- it's really good ROI for them. We have dedicated engineering teams that work on engineering obsolescence solutions and/or performance improvements on some of that older equipment. So there's definitely a value creation there that we share with our customers. So that's a growth driver for us. And then last is sort of just value services. And there are lots of aspects to the value services. Bren mentioned our mission and goal alignment with customers is to help maximize the value of their asset. They spend millions and millions of dollars on our tools. The only way that it works is if it's available predictably and if it is at performance levels. And so if you think about those on 2 axis, there's lots of different potential combinations of different availability levels and different performance levels, depending on our customers' use case. And so we can tailor offerings for each one of those in order to meet our customers' needs and ultimately drive, again, increased value.

That's perfect. Kind of along that line, I mean, the analyst community, investors obviously very focused on the increased capital intensity of just equipment, right, across WFE. What about the services intensity? Talk about -- can you talk about just the rising cost and the increasing importance of inspection of just how your customers are approaching that relative to maybe 10 years ago?

Yes. I think there are a couple of drivers, specifically for service intensity. One is just the complexity. And so with that, that complexity increase, both on the customers' operations as well as on our tool, that means they're going to lean more heavily on our service to ensure that those tools are available. So that means shipping the tools on time, getting them installed on time, qualifying them, getting them into production and maintaining them in production. And then you also have this other driver around this geographic expansion. And when you start to build fabs in new locations, it puts a lot of pressure on our customers' operations, where they don't have that ecosystem inside of their home country. And so they're now depending on us in a much larger way to ensure that we bring up those -- that factory initially and then maintain it as they continue to add wafer starts. So I think all that creates an opportunity for us to continue to increase service intensity as we look forward.

Yes, there's definitely a new value vector on this regionalization dynamic as customers are building fabs in different parts of the world is that for scale players and the ability to have the infrastructure to support at the same level, no matter where you are, I think it's a competitive advantage. Also, with the team spread out, I think there's an opportunity for them to rely on us more to help, and that's a service statement, but it's also an application statement. We have about 1,300 engineers, in some cases that have PhDs that are out working with customers in fabs every day to drive -- deliver value out of the tools, recipe, optimization and so on. So that, coupled with service, provides that infrastructure that should be a benefit to us moving forward in the form of more value opportunities within service, but also potentially share overall. So we're encouraged by -- look, I think there's a lot of -- at the end of the day, supply/demand over time generally gets balanced, but there's probably some inefficiency that goes with the broader footprint and trying to manage across it. We deal with our own versions of that within the company. But then also, there's opportunities in this area as well that we haven't even really -- we're probably investing in a little bit today to get ready for it and having to make those calls and when to make them given training lead times and so on, is a challenge. But I think that's still to come in terms of an opportunity out there for us.

In that vein, maybe like how do you think about -- I think one of the things -- when I think about the regionalization of chip production capacity, the battle for talent, right, of -- there's only so many potential STEM workers, right, that -- how do you think about that competition for just getting more service engineers?

Yes. Certainly a challenge, especially when we go into some of these new regions. I mean, fortunately, for KLA, once we acquire the talent, we have a very, very high retention rate because it is a little more of a you're more of system engineer, servicing our equipment than you are just tactical replacement of maybe repetitive tasks, consumable parts and so on. So we don't necessarily create holes for ourself. And then just trying to, again, market the fact that we do have this -- it's a little bit of a different service model. You get to work on more challenging problems and this high-mix, high-complexity component can create some attraction. But it's certainly a challenge for us and one that we work on. We go early. So when our customers announce, and that requires a lot of investment, we look at all this geographic expansion. We have to invest early, like about the same time that our customers are placing POs for new equipment. We have to be thinking about how do we set up the depot? How do we ensure that we've got the supply chain? How do we make sure that we've hired and trained people? And so being early on also helps ensuring that we acquire the appropriate talent to support the customers.

We even move people to -- from the home country to seed the team and both from a capability culture point of view and that provides some comfort to the customer as well that familiar faces in terms of how we're ramping it. But it's a fundamental aspect of the operating model of KLA. I mean, Brian and I, every 6 weeks, we review headcount plans and where fabs are and where capacity plans are and contract rates and utilization by individuals and tools, and we try to make these decisions and get the recs out so we can have as much lead time as we can. Training lead times are at least a year, in some cases -- in a lot of cases, longer. And our engineers tend to support one type of tool. They don't service all. I'm a KLA engineer, I service all of them -- nope, it's the overlay tools or the film measurement tool. So the requirements are higher, which is interesting to certain people, but it also makes it more challenging because there aren't as many of them out there, to your point.

Helpful. We've got a few minutes left. I don't think you thought you were going to get out of here without at least one question on China. Maybe -- everyone's right, the investor question that I constantly am getting is the sustainability looking into next year, obviously, like we're spending a lot of money there now adding capacity. How do you think about the sustainability of that investment, the real actual value that's being added there in terms of building out that semiconductor industry? And just how do you think about that opportunity looking into next year and even beyond?

Yes, it's not a question that I haven't been getting it in my meetings. But when you look at it, like I said earlier, when we're talking about backlog and deposits and so on is that a lot of this business that's been in the queue for some period of time, and we're shipping it now. When you break it down, there's a portion of it that's memory. There's a portion of it that I'll call established legacy provider that's -- the China providers aren't much different than the non-China providers in that realm. There's the infrastructure investment that I mentioned. And in China, that's been an important -- it's been a real bottleneck and choke point to be able to get wafers through the growth period of '20 through '22, a lot of long-term contracts from the wafer providers with other customers. And so in terms of self-sufficiency and to be vertically integrated, wafer was a challenging area. And so there's been investments there for the domestic requirements, but also to support the export economy for electronics. Reticle, in the same realm. Most of our customers -- larger customers have captive mask shops where they make their reticles. These customers would have to buy from the merchants and the merchants were -- haven't invested a lot over the years. So I think there were some shortages there from the merchant mass providers in Japan, the U.S. provider as well. So a desire to kind of control your own destiny a bit and so that investment has been happening and were a meaningful part of that CapEx. And it's for legacy designs, but every design has multiple reticle sets and there's a 30-plus reticles in every set. And so it is a demand area that needs to be addressed. Then you have a number of projects that are a lot of new customers spread around the world where you have -- and some of them are silicon carbide, some are silicon, where you have customers that have funding and different arrangements, either central or regional and are developing R&D and some amount of production, trying to prove out a process, develop -- start somewhere, develop capability, demonstrate credibility with customers. And slowly sort of engage and start to move forward to be able to provide supply. Not providing us a lot of supply today, but are starting that process. And that -- it takes time and R&D always continues, and there's a desire to keep moving forward. So I would expect that to continue. So when you back up and you look at the demand that we have in terms of the backlog, you look at the deposits, I feel like when I look at '24, and this may be a little bit more specific to KLA, I see the numbers generally being sort of flattish overall. And maybe the infrastructure piece is something that gives me a little more confidence than maybe some of my peers because we heard some different numbers from others. Over time, I would expect there's always rationalization. Supply/demand generally aligns. Things will be milestone-based. The next round of fundings will be based on milestones and return expectations, even if return expectations are lower. So I think we'll see -- over time, we'll see how this plays out. But at least as I see it into next year, I don't see it falling off all that much. They're all greenfields, so they tend to buy a lot of process control because even if you're running very limited production, you tend to manage the process very tightly. You need the whole suite of products, even if you only have a few of certain process tools because you're not really doing a lot of production. So we tend to participate, like we do everywhere, more in R&D and volume ramps. In mature production, a little less so, although that's changing as more and more designs kind of are in production, more process flows and so on. So that could be a factor as well. It's sort of KLA-unique. But I don't -- at least as far as next year, I just don't see it changing all that much. Construction schedules notwithstanding, right? Anyone who's built a home or something like that knows that construction schedules can move around. I mean, that could be a factor because there are a lot of greenfield projects. But generally, as a general case, I feel like it's -- looks pretty flat at the baseline to me.

Is the risk more fab readiness? Something that was brought up in a few meetings yesterday was more also consolidation of we kind of have almost a VC approach, right, of a lot of guys trying to do the same thing that eventually, maybe they get kind of rolled up into a larger organization. Is that -- how do you think about like...

I mean, typically, that's kind of what you've seen. You've either seen that or ambitious plans that, over time, are moderated and adjusted based on performance and achievement. And so -- but I think it will take a while because there's a lot of this that's just starting up and will take some time before you can determine what success really looks like, right?

Yes. Brian, any sort of difference in the services attach rate or just intensity in China relative to the broader? I know obviously, over 75% are long-term contracts, so maybe not that much different, but how do you think about that?

Yes. On all the key metrics, it's a little higher in China simply because it's -- they're still very much in the learning phase. Generally, sort of -- we talked about that geographic expansion. That's what's happening in China. It's a new location every time, and so they're trying to ramp up. And so they certainly rely on us even more heavily. So it's a little better in each one of those metrics in China.

Perfect. Maybe in the minute we've got left, Bren, when you take a step back and you look at the KLA story, what do you think investors maybe don't appreciate enough? Obviously, we talked about services, but what are -- what do you think investors are missing from the story that we should be thinking about more that we don't?

Well, look, I think we've talked a lot about some of the opportunities moving forward. We think we're encouraged by the demand drivers, the broad-based demand across semiconductor revenue, the investment that's happening, the competitive dynamics at the leading edge, which tends to drive our business. Dynamics around mix of leading edge, both from mobile that drives it, but you also -- we believe will have more of the AI drivers that will ultimately come in. And that will drive need for, over time, more leading-edge capacity sooner, which creates an opportunity. There's unique attributes of those kinds of chips that are valuable to KLA services and the value that we add. I think we've got a proven business model. Best in the industry. We think we're in a position to scale it. We're excited about the opportunities we talked about earlier. We have a service business that's resilient. We have -- we think we finance the business in the right way. We put every dollar to work in terms of the capital we generate. And we think we're in a period of time with structural growth and leverage moving forward that's meaningfully different than it has been over the longer-term history of the company. So excited about all of that.

Perfect. Thanks for joining us.

Thank you for having us.