KLA Corporation (KLAC) Earnings Call Transcript & Summary

Great. Good morning. Welcome back. I'm Vivek Arya from BofA Securities semiconductor semicap equipment research team. Really delighted to have Bren Higgins, Chief Financial Officer of KLA, join us this morning. I'll go through my Q&A but please feel free to raise your hand if you would like to bring up anything during the session. But with that, a very warm welcome to you, Bren. Really appreciate you joining us.

Yes, thank you for having.

And maybe to just kick it off, give us a State of the Union. How is the demand environment now versus you thought at the start of the year? What's better? What's worse from your perspective?

Well, we had 3 years of unprecedented really growth in the industry, right? The overall industry up over 20% from '19 to '22. So as we've come into this year, our expectations around WFE composition has moved around a little bit. But in general, we saw kind of this mid-70s environment. We've seen it for a while. There have been some moving parts. At the beginning of the year, I think, I would have thought that the leading edge would have been stronger than what we've seen. Memory has probably weakened a bit as the year has gone along. We've seen an improvement overall in what I'll call the legacy part of the market, which tends to be logic/foundry, 28 nanometer and below or 20 nanometer and above. So that's where we draw the line. A lot of people look at that and say, well, why is the line there? I've always felt like, well, you start to get the 20 nanometer, you start to get into more aggressive multi-pattern, then you start to move things sub that into FinFET architectures, so 3D architecture. So it's more planar versus vertical structures and that's where we drive. But we've seen some strength in there. We've seen some strengthening, in particular, in China, where we've had a number of projects, probably more than 30, which we've had a fair amount of backlog for a while. We've had deposits. And we've had customers that probably were -- because they are newer, they've been behind our more strategic customers in terms of their position in the build plans. And so I think as we've seen some adjustments over the course of this year, we've seen some of that business pull forward. Some of it's tied to facilities, so it's not all that practical that it can pull forward versus what we're originally planning. But in some cases, it has, particularly around certain products where we continue to see very strong demand despite the weakening WFE environment, products like our optical inspection business, which I would expect that business to be probably in that flattish range this year despite WFE being down 20%. I think it's one of the faster-growing markets in overall WFE. And so that continues to be a strong market. We're supply constrained still in that market. And so there's pressure there. Our reticle inspection business is another one that we've seen some strength in. And I would expect that to be kind of flattish this year after very strong growth last year. Our capacity-centric businesses, though, are feeling some pressure. As you would expect, as the process tool capacity comes down, the films, tools that are used, those optical CD tools that are used to monitor, those process tools have fallen off a little bit. But we are seeing, as customers continue to invest in road maps and technology progression, we have seen some continuation of business there overall. We're also exposed to, I call it infrastructure but it's more around silicon wafers. And the silicon wafer industry is one that tends to invest in a different cycle, takes a while to add capacity. So in a downturn, you tend to see that business pick up a little bit. I won't exactly call it countercyclical but it definitely carries a different cycle. And so given expectations of a demand recovery as you move into at some point into next year and beyond, that capacity is coming online broadly. But it's also coming online in China, where there's been a struggle to get because of long-term contracts, supply contracts, a struggle to get wafer commitments. And so there's a desire to add that capacity within the country. You're also seeing that on the reticle side, where you have reticles for legacy design rules. And so those reticles typically come from the merchants. And so there's been a supply-demand imbalance there. And so there's been a desire to add reticle capacity. And so those are parts of WFE that we have some exposure to that have been good for our business that are classic, I'll call it, kind of IDM or IC WFE investments but are good for KLA. So when you add it all up, we talked about an environment in the second half that, that we saw sort of stabilization around current output levels relative to what we guided for June. And in general, still feel pretty good about that. We also talked about some upside potentially available to do a clarification related to where you're drawing the technology lines as it relates to some of the export controls and that clarification will enable some incremental shipments into memory. So you'll see the memory numbers improve but mostly due to some incremental memory business in China. And so we'll see that in the second half. Potentially, an upside factor to overall WFE, we'll see as we get into -- finish this quarter and we'll start to describe and look at what the second half looks like. But in general, that's how things look so far. On our EPC business, it's been more of a challenge. It's closer to consumers, more mobile-centric, in a lot of cases, very capacity levered business. And so we've seen that fall off and haven't seen any real recovery in that and started to fall off in '22. And in '23, it's continued. Within that, you have our specialty semiconductor business, which is a bright spot, driven mostly around the automotive market and power semiconductors, which has been a nice driver for that business. So that business year-over-year are likely flattish. So some nice strength there. We're excited about what that business means to us over time and the unique position they have overall. And so while EPC is down, that part of it looks pretty good. And then service, utilizations are down but tools are still shipping. And so that drives the installed base up. You have customers still running their installed base and you have pockets of weakness in some of that. But in general, you see service grew year-over-year and would expect to see that happen again this year. We talked about mid- to high single digit overall for the company's service profile this year. So we'll see growth. It's a business that's only declined once in the last couple of decades. So it tends to be pretty resilient. As new tools come online, the installed base grows. And there are dynamics within the installed base that are driving customers to extend and drive useful life higher and so on. And so that's been a good driver overall for that. That's where things stand.

Got it. Now if I rewind a year or so ago, we were talking about WFE $100 billion, right, unconstrained numbers, even higher than that. And here we are at $75 billion. So you think that $100 billion is just a very tough peak to get back to and you have to really gradually start building off the $75 billion? Or maybe another way of asking the question is that, if we try to look forward to '24, so I'm not asking for specifics but in terms of -- if I break up WFE in terms of memory and leading edge and trailing edge and kind of China sort of its own, what parts can potentially grow next year that inform us how WFE will trend next year? And what -- where is their most risk and uncertainty?

Well and it's how good are we at this also, right? If you just go back 12 months ago, what do we think '22 is going to be, $100 billion plus. We certainly had orders that were way in excess of that and that fills most of the backlog that everyone in the industry is carrying today. But we also thought '23 would be growth over that number.

Yes.

And I'm feeling that as we manage the capacity and adjust the capacity in the company to reflect that change, we've had some pressure in some inventory reserves that we've taken, which I think, reflective of the short-term environment. I think, over the long term, I think economically that's not going to be a problem for us. But it is a little bit tough to see, right, in terms of what's happening. When we modeled our '26 plan that we laid out at Investor Day, the reason why it's a '26 plan and not a '25 plan, as we say, look, the industry has grown, as I said, 20-plus percent over 3 years. Semiconductor revenue is growing 7% over the last, what, 5 or 6 years. And so you would expect an adjustment in there somewhere. And so we wanted to model that in and felt that it was prudent to do so. We're running the business like we expect the industry recover as we move into next year. I don't know when you'll see it, right? And certainly, our customers' financial position and particularly as it relates to their markets and profitability and what that means in terms of cash flow and all that might drive the timing of some of that. But our expectation is that we'd see that '24 would be a growth year and we'd see some improvement at some point in '24 as we move forward. And we still feel like the '26 plan assumptions that we laid out in terms of long-term growth of 6% to 7% for semiconductor revenue and capital intensity adjusting up a little bit and 7.7% -- 7% to 8% in terms of WFE. The overall mix of WFE being weighted more to logic and foundry at sort of 60% over time would drive process control up to grow faster than WFE, simply because the dynamics of how customers invest in process control is different between logic/foundry and memory. And that KLA's market share position, the inflection around certain product types that we would expect to continue, particularly like optical inspection, which is very levered to scaling and that we'd see KLA share improve over that. So all that translated into our $14 billion -- obviously, there was a service and an EPC component to it but translated into our $14 billion view for '26. And if I look at sort of what's underneath that, there are some puts and takes in terms of long-term WFE expectations. But it's all kind of predicated on this view of capital intensity and this view that sort of tracks to $1 trillion semiconductor revenue 2030 sort of time lines. And so if you think about the linear growth of that, we feel pretty good about the overall plan. We're ahead of it in terms of KLA share of the overall market. I think we're pretty pleased with the market share but also the relevancy of our particular part of the market and the strength around some of these product types. And so at Investor Day, I talked about low [ 7s ] as a percent of WFE -- KLA share of WFE. We're in the high [ 7s ] today. We'd expect this year, it looks like from a relative basis, we're going to continue to do better than the market. So I think we don't need as much WFE to get there, maybe export controls and some of the dynamics around China, maybe take some of that WFE out in terms of upside but feel like we're in a pretty good place to get there overall. And I think the industry is pretty well positioned to ramp. I don't think the growth rates over multiple years are much different in terms of this expectation that I just talked about and what we've seen in prior down WFE years, plus 2 or 3 years in terms of growth. And I think we're pretty well positioned with all the infrastructure we've added. And so I feel pretty good about the long term as we work through here and we'll see how '24 plays out as we move forward here.

So just to press you again on the '24 case, again, more from an industry perspective, what is the biggest X factor? Is it that when do the memory customers start to increase utilization? Or is it, whether leading-edge investments will stay? Like what do you think is the biggest unknown as you look at '24?

Well, if you look at this year, right, leading edge is softer, right? So I would expect that we'll start to see some improvement there potentially, right? I think the memory piece is probably the biggest wildcard, down a lot this year. And we'll see, as we move forward, as pricing and so on improves, what that means in terms of customer posture. I wouldn't have any insight into that. I think the legacy investment, probably there's China and non-China. I think the China part seems to have some sustainability. So I don't know if it grows but I don't see it falling off, particularly given the level of backlog I have, the deposits I have. I think these customers have generally performed pretty well in terms of managing and living to the commitments that we're making in terms of delivery schedules and so on. I think overall, legacy is investing over the last couple of years. I think there have been reasons for it. Maybe some of that falls off into next year. So I think there's -- so looking at the moving parts, I think when you aggregate it, I said -- I thought at least we're running it in terms of a view of growth into next year and we'll just -- we'll see how it plays out.

Got it. Is there -- does the industry have visibility at this point as to when memory utilization starts to pick up? Or do you think it's too early to...

I think it's too early. I haven't seen much change on that front. Like I said, that because of some of these incremental shipments, you'll see the percentages change a little bit in the company, right? I talked about a couple of hundred million dollars of upside there. But I wouldn't say that when we look at the pricing environment overall, that I don't think it's getting worse but it's not clear when we might see a recovery. And historically, memory is pivoted pretty quickly, right? You don't get as much new supply from technology node shrinks. And so it requires more capacity generally to meet more normalized bit growth demand numbers, which causes a pretty quick pivot in terms of capacity. It's a bigger bet for customers. It's a longer-term bet in terms of execution. So I think what we've seen over the last number of years is you see quick turns, both directions. And it's incumbent upon us to be as well prepared for that as we can. But I wouldn't say, I don't think it's getting worse but I can't point to say, okay, I think -- particularly when you look at just the financial realities and then the cash flows are, they have been fairly weak and so those will have to recover. And the financing environment is a little more prohibitive today than what it's been in the past. So I think those are, all factors in overall timing.

Got it. And then I have to ask this obligatory question about what are the benefits of AI for KLA. So positive, negative, neutral?

Yes, I only get this from every employee in the company as well. Look, when you back up and you look at just the high-performance compute part of the market, it's -- tends to be a pretty good driver for our business in a few ways: a, that you've got those customers pushing the leading edge, not necessarily leading edge being driven by 1 player around mobile devices. You tend to have bigger die, higher-value dies. So the wafers are more expensive. And so if you think about -- you're starting a lot of wafers and you need to deliver to a certain window, you don't want to start too many. You don't want to start too few. Our customers are generally going to monitor more. It's like you got a higher value of risk, you're going to monitor more. And that tends to be good for us from a capacity point of view. So if you've got a robust design environment, you've got a lot of activity, different designs, test design rules in different ways, different process flows, all that tends to be good for process control even in a production state. Defect density tends to be a problem if you have the same number of defects and fewer die, then you kill more die or the yields are lower. And so that tends to be a factor as well. And then you've got dynamics around more chiplet architectures, which is good for lower design costs and a more robust design environment, advanced packaging. And so those parts of the market as well. More pressure from a -- in terms of higher -- we'll see but in terms of higher mix of higher performance memory. And that tends to be a driver as well, even though less than on the logic and foundry side. So that mix shift more to high-performance, compute-centric devices from mobile devices is a positive. I think once we get a clearer sense of what it means in terms of how our customers are investing to support it, how their mix of business is changing, makes it easier for us to try to figure out the size of it. But it's clearly a driver. And it's something that within the company, we've been doing for a long time. I mean one thing about our tools and the cost of them is, the cost of compute has been high. But given the value of the tools and what we charge for it, we've been able to leverage a lot of machine learning capability in AI to get more performance out of the hardware. So we've been a user of it for some time. And certainly, of course, costs of compute has come down, so the application universe has become broader out there. But it's also -- which is good for the long term but it's also, I think, good for capability in our systems. And so I think we'll see more of it deployed in the company as well, in how we go to market, how we extend the hardware, which is good for R&D intensity and so on.

Do you think just AI is capable of driving WFE growth next year? Or do you think it is merely offsetting the kind of structural pressures on the consumer, right, PCs and phone, which also consume a lot of...

I think there's a lot of moving parts here and there's a macro backdrop. You do have a lot of semiconductor demand that's driven by consumers. And so we'll see from a macro point of view if there's some impact to that. But I'm not prepared to say that, yes, I can -- we can point to that and say this is going to drive a certain amount more WFE into next year, I'm not ready for that.

Got it. Okay. Makes sense. You run KLA's manufacturing as well. How are you preparing for this -- the current demand environment? What are you doing differently?

Yes. It's been quite a challenge. And in some ways, there's what are you optimizing for and what do you accept and I say this around the company a lot. And what we've optimized for is differentiation in our business, right? And that obviously affects our go to market and to deliver capability to the market that we can share the value in with our customers. And you can look at the company's gross margins and that gives you an indication of the differentiation that we bring. Part of that comes from the unique capabilities in our tools. Those capabilities come from technology company has but also within our supply chain. And so we manage our supply chain, I think, differently. It's a strategic relationship in a lot of cases, not transactional. I think we saw last year that when you have transactional supply chains and everybody is constrained from a supply point of view, it's hard to get the priority out of your suppliers to be able to maintain your -- meet your customer demand, maintain your shipment profile. And we didn't do as well as we wanted to do but I think we did a pretty good job overall. And I think it validated a lot of the strategies we have as a company. We have to make long-term bets in terms -- because of the lead time of certain materials and components is very long. On optics, if you decide that you want to add capacity, it takes over 2 years to do so. So you have to plan ahead. You have to make investments in these suppliers. You tend to sole source in a lot of cases, where it makes sense. And you try to be a good customer and so you don't spend a lot of time on a lot of the commercial stuff and -- but your focus is more on ensuring supply continuity, exclusivity and so on. And so I think part of some of the overhang that we've had and some of the headwinds we've had as we've been adjusting down, as we talked about earlier, we went from $100 billion to $75 billion or $100 billion plus. You had about [ $30 billion ] coming out of the market in a short period of time when you were driving the supply chain pretty hard. And so we're adjusting that and so we're carrying some parts. I make commitments to suppliers. I keep them. And I think it's what enabled -- I think these relationships enabled what we saw. And KLA had a very strong relative performance a couple of years. And so this is, in some ways, the trade-off we're making. We feel pretty good about it because in the long run, these tools are living a long time. And we also have the service business. And so ultimately, I would expect that will consume the parts. But in the short run, this adjustment we've taken has been a little bit of a headwind to gross margin. We've also added a lot of capacity to support that. And it's why I'm pretty confident about when we see a resumption of growth of our ability to drive our incremental margins in the company. We talked about a plan that gets us 63% plus and 42% type operating margin of 41% to 43% at $14 billion. I feel very good about our ability to scale given the investments we've made over the last few years. So a little bit of an overhang here at $2.2 billion, coming off of what was almost $3 billion in the December quarter. So it's come down a fair amount, [ $2.25 billion ] is where we are at today in terms of the midpoint. So I think those have been the adjustments mostly that we've had to manage our way through. I mean, obviously, we've had to adjust a little bit in terms of volume dependent resourcing, and that's the usual stuff that we do in the company when it moves both directions. But I think it's been more about just adjusting off of what was expected to be a pretty strong environment that we were driving our business towards and then dealing with the aftereffects. But in the long run, I feel pretty good about how we're positioned and our ability to scale moving forward.

What's KLA's mature node or trailing edge exposure in foundry/logic? Because we always think of KLA as a lot more exposed to the leading edge. And I ask that because there is a concern that there's a lot of trailing edge investments this year, right? They may fall off a lot next year. So what's your trailing edge exposure? And let's say if trailing edge were to fall of next year, what's the potential impact?

Yes, a couple of things. So generally, about 75% of our revenue tends to be leading-edge centric. So less exposure in the trailing edge and there are reasons for that. Although in the most recent quarter, it was about 60%. So you do see the influence of -- and I said this on the call, so you do see the influence of some of the higher level of trailing edge investment. It depends in terms of what we call process control intensity in the legacy. So legacy markets generally tend somewhere in line with memory. So as a percent of wafer fab equipment, somewhere around 10-ish percent. In China, it's higher. I'd say 2 or 3 points higher, somewhere in that ballpark. And you've got newer projects that are starting to scale up. And so they tend to be a little less efficient with that capacity. And so there's a higher level of investment there. In those markets, though, if customers are adding capacity from a mature part and that part isn't changing all that much, then the process control intensity tends to be fairly modest, right? And because it's like what's really changing, it's a pretty predictable part, pretty predictable yield, if specs are changing, if they're migrating design rooms, things like that can change, how those customers are investing. Those customers have historically been able to buy used equipment. Now they're having to buy new equipment. In a lot of cases, we're restarting old equipment to be able to meet those needs to target those nodes. So I think that's been also a driver of incremental investment. You're also seeing some desire to have more in-source capacity as everybody got cut short for a long period of time and customers not want to rely on the foundry market to supply. And so you're seeing some investment in those areas as well. If you look at the WFE, I would say it's -- we call it, [ 28 nanometer ] and below logic/foundry; WFE, probably some around the low [ 20s nanometer ], say, roughly 50-50. I think the China piece is growing year-to-year. I don't -- as I said earlier, I don't think I see that falling off at this point, as I look into next year, because I think that sustains given the nature of a number of those projects and how it's laid out in our slot planning. And you might see some other demand -- I mean you've seen some weakness maybe potentially in the other parts of that market. I think they have to invest more going forward. So the capital intensity is going to rise. But it wouldn't surprise me if you see some falloff in that as you move into next year. So we're less exposed to it, though, so I don't want to say I have a lot of visibility beyond some of the things I'm saying here. In the long run, it's a little bit different. If you think about how this industry has moved, cyclicality has been driven by a lot of investment. The leading edge tend to focus on certain core markets, whether it's PC market or the mobile market, then there would be an overshoot and then there'd be some digestion and so on and then repeat itself. This would be a little bit different. Leading edge investment and road maps, I think, continue. What's happened in the legacy doesn't have any real effect on what's happening in the leading edge, which is where KLA's business tends to be focused. And so you could end up in the legacy parts of the market with some excess capacity here or there. Customers operate at different capital intensity levels but also supporting markets where intensity is rising also or semiconductor content is rising. So I think as I look at KLA, I look at most of our business being about enabling technology transitions and speeding time to results by providing information to our customers to be able to scale to the leading edge and some of the dynamics that are driving the leading edge, which have been, I think, pretty good over the last few years. So I feel pretty good about the sustainability of that moving forward. And we'll see what this dynamic in the legacy parts of the market. It's kind of new and we'll see how it plays out over time. But hopefully, that provides a little bit more context.

Got it. One last one on memory. When I look at just sort of consensus expectations for memory sales growth next year, right? They are like 30%, 40%, 50%, 20% almost [ bit ] growth and then the rest obviously comes from pricing. In that environment, do you think that kind of growth rate is possible without investing in WFE? Like do they already have enough tools, that it's just a matter of increasing utilization on those. So even if the memory companies see a lot of sales growth, that the tool industry wouldn't necessarily share in that.

I mean, utilizations have fallen off. They're not getting worse but I haven't seen them really increase. We'll see. I don't want to lean into at this point, 6 months from now about how I think that will play out. So I don't think I have anything more to say about it.

Okay. Understood. On the CHIPS Act, right, do you think that drives incremental WFE? Have you seen any of that? Or do you think that, that's just very modest, right? It's more a kind of change of where they might have put capacity, right, as opposed to an overall lift in demand?

Yes, it's a great question. We get it from investors a lot. I think in the long run, it changes where it is, right? And it's bridging a cost delta between alternatives and where these facilities could be. It's very difficult, though. Even though you got customers, I think, who have done a pretty good job of managing supply and demand, as we saw just last year, it's hard to get it exactly right. And there's been a lot of efficiency gained by the industry with the consolidated geographic footprints in these big, large mega fabs. And so while I think in the long run, they've done a pretty good job of balancing supply/demand. I don't think you can do it as perfectly as they've been able to do. It's just very hard to do, particularly as you're scaling up new facilities, right? And there's an inefficiency level and then they scale up to full production. So there could be some timing gaps particularly as it relates to demand. I think most of -- from what we are hearing from the government in terms of how it's being administered, is more milestone-based. And so the funding will come, the capacity will -- they'll build the shelves but the capacity will come as there is business there and there are customers. So I don't think it's just going to all show up all at once and then drive -- which wouldn't be good for anybody to have a whole bunch of excess demand or supply in the system. So I -- but at the end of the day, I think these customers have done a pretty good job of managing. There's probably an element of capital inefficiency that goes with it, just because of the dynamics of our industry. And it really is about industrial policy of where chips are made. And we haven't seen any of it really yet. I think you'll start to see it next year. The fabs are being built now. And I think, look, we're not planning. Our model doesn't assume that there's a big -- or there's an uptick -- a permanent uptick related to this element of inefficiency because it's hard to measure. But at the same time, I think it logically makes sense that it might be there.

Got it. And lastly, look, KLA has had the best margins, right, in the industry and a very robust free cash flow return program. There isn't that much scope for M&A in the industry. So how do you look at use of cash going forward?

Yes. We've been pretty explicit about it. And you're right, as we look at just alternatives from an M&A point of view and some of realities of getting approvals and so on, we feel pretty good about the businesses we're in. It doesn't mean that there aren't opportunities for us from a supply or a portfolio augmentation point of view or technology, that there might be things that we might consider. But we felt pretty good when we laid out our plan of 85% free cash flow return over time that -- why isn't it 100%?

Yes, why not 100%?

Because a piece of it gets consumed in some of the things that we're talking about. I also do these -- I talked a little bit about how we manage our supply chain and there could be some investments that happen there. But in general, we feel pretty good about an 85% return. If we have a dividend, we've raised it 13 years in a row. It's grown at a CAGR of 15%. We're going to grow with the underlying cash flow. So if you back up and look at the long-term revenue model of the company, 9% to 11%, EPS, 1.5x that revenue growth rate, we should be able to continue to grow our dividend payout at that CAGR. The difference in share repurchases, we just executed a large ASR. We have additional authorization we're working our way through. Our ongoing return, because our cash is in our target range, which is $2.5 billion to $3 billion. Ongoing return is funded by the cash flow the company generates over time. But we should be kind of in line executing with both vehicles and in line with that 85% target.

Terrific. Thank you so much, Bren. Really appreciate your time.

Yes. Thank you. Appreciate it.

Thanks, everyone.