Are we on? I guess we're live. All right, good. Well, we're past our start time, so let's, let's, let's get going. I'm Joe Moore, from Morgan Stanley. Very happy to have with us today, Bren Higgins, the CFO of KLA. So thanks, Bren, for, for joining.
Yeah, thanks for having me, Joe.
Maybe we could start with kind of an intermediate-term view. You know, your view of the wafer fab equipment market has, has improved. From the start of the year, you're looking at $80 billion. It feels like we're tracking to a number that's maybe even higher than that. So can you just talk about, you know, broad brushstrokes of what this environment looks like, and we can go into some of the details?
Yeah. 2023 is an interesting year because it has, you know, we had the 2022 dynamics as it relates to supply chain, and a lot of folks were struggling to get tools out the door and ship completed systems. And so, and we had a very, I mean, we had those issues, too, but I think in general, we were able to execute pretty well. And we're actually up about 36% in our semiconductor process control systems business last year against a WFE environment that was, we'll call it, high single digits. So we're really encouraged and excited about where things are from a, an overall demand point of view and where we're positioned in some of the structural things happening in the industry.
But 4x better than the industry in terms of growth rate is not what we expected. And so I think as you moved into 2023, I think the hardest part about predicting was to take all that business that some of the others were challenged with last year that is actually showing up in 2023, and it was partially shipped-
Mm.
-or incomplete or fast shipped, lots of different terminology. I think as you look today and you look backwards, we were trying to figure out how much of that there was, but when you look backwards, probably about $5 billion or so of the WFE.
Yeah.
So even though we said 80, it wouldn't surprise me if we're 85 or somewhere in that ballpark, maybe, you know, could be higher. I guess we'll see how people finish the year. But certainly, as people have re-reported and guided into the Q4 , it feels like there's upside.
Yeah
... To the number. And then you had those dynamics. Now, I felt pretty good about our ability, even despite that we weren't, we didn't have this carryover into 2023, about our ability to perform, at least in line with the market. Obviously, when the industry goes into a down year, logic foundry, as a percent of the total, tends to increase, and that tends to be good for process control. Customers always invest in roadmaps from a technology point of view, and so we help participate in that. So what changes in slower periods is that, you know, memory cycles generally more, and customers push out some of the capacity plans, but the development plans tend to hold pretty well.
We have certain markets where, like optical inspection, where we have a fundamental gap between the demand levels and supply, and the supply takes a long time to come on board. It takes equipment, it takes automation, it takes a lot of training, and so we have some new capacity coming, but that's a challenge. So for all those reasons, I still felt that we were in a pretty good position to at least perform in line with the market. I feel pretty good that if we just fast forward and look at where we are today, that given the reductions we've seen in the leading edge from where we were at the beginning of the year to where we are today, that we've been able to maintain this thesis of growth through this year.
I feel pretty good about the positioning of the company, both in terms of our exposure to what's happening in China, what's happening in infrastructure, which I don't know if it's fully understood by investors, and infrastructure being radical infrastructure and wafer infrastructure. And then some of the changes that are happening in the legacy markets themselves in terms of, of design rule shrinks and changing specs in terms of reliability, automotive being an example, some of the exposure we have to things like silicon carbide and so on. That normally our, our business is 75% leading edge. This year, it's been very different, but we've still been able, I think, to, because of these other factors, maintain a, a pretty good relative performance.
I'm encouraged as we move forward, both in terms of recovery, but some of the technology transitions that are coming that we're in a pretty good position moving forward. So we talked about an environment back in June. We said, "Hey, look, it doesn't look like it's getting worse, but it looks like things are fairly stable from here." We were at a $2.25 billion run rate. We just guided for the December quarter at the midpoint, $2.45. So we're slightly above where we thought we would be, but we characterize an environment that we're kind of stabilizing at this level, and at least we see more or less that this continues through the H1 of 2024. So we'll see about the rest of 2024.
I'm sure we'll talk about some of the moving parts here today, but, from an overall visibility point of view, hopefully, that provides some context.
Yeah.
and where we are today.
I think it's a really interesting year because I think of a downturn as being, we slow down on all the capacity additions and continue to invest in cutting-edge. But obviously, this was very different where the capacity additions were pretty strong. Maybe if you could address the cutting edge, like what you're seeing there, there's obviously a lot of focus on an arms race to get to gate all around, backside power. At the same time, there's some questions as to, you know, is there anybody big, like an Apple, that will be the first user of that capacity? So I don't know how much capacity we need, but so how does that help or hurt KLA? Can you talk about the dynamics?
Well, it's certainly the strongest market for us, and it typically is that, you know, as our customers are pushing the leading edge, and you have, since we've really, over the last few years, seen more and more EUV layers, you're talking about changes in architecture, that changes in process, changes in materials tend to be good for process control. The thing that's also been different over the last several years is that we've seen a much more robust design environment, and that's been enabled through lower cost of design and chiplets and other things. And so that's created a lot of designs at the nodes, and then as the customers start to invest in the next node, their ability to migrate some of that capacity, reuse it effectively at the next node, is harder and harder.
And so when EUV wasn't really when it was delaying, and we didn't have litho scaling for that period of time, most of the last decade, or certainly a big chunk of it, you didn't see that design activity, and you had the high volume parts moving through. The Moore's Law pressure in terms of driving scaling was fairly modest, and so our customers would satisfy their big customers and then try to migrate the capacity. Much harder to do today. So and as you look at where we are with 5nm node and expectations for three, I think there's a continuation of those dynamics. What we've seen is the R&D has continued. The slope of the capacity adds has slowed down.
And then as you look into next year, and you think about some of the competitive dynamics, you think about NVIDIA and some of the other things that are happening, you know, it does make us encouraged that the end market dynamics you talked about, but plus the changing technology, will be an opportunity for the company. I don't know how much of it sort of do we see in a sustainable way as we move into next year, certainly through the H1 , based on the early comments. But-
Mm-hmm.
But I think we're encouraged by what those dynamics imply, overall. And we are seeing more investment. Because of that, we're also, so we've always been exposed to your point about R&D and ramping and time to results as a customer ramps capacity, but we're also seeing more participation in production. They're managing a more robust design environment. A lot more of our tools are behaving like process tools, where there's actually set intervals. It isn't a strategy, set intervals around when they're going to do certain inspections. Certainly, reticle fidelity's been an area of that, where they run a certain number of shots through the scanners, and then they have to do a print check inspection just to make sure you don't have any issues there that were introduced in the production process.
So you're seeing more of that. You're seeing more optical inspection, and we're seeing more managing of a more robust design environment. When you think about the next nodes, you start to introduce more metrology capability or more metrology challenges with more layers for gate all around as an example, different defect mechanisms. So we're pretty excited that we'll see a continuation here of some challenges that hopefully will be good for process control and for KLA.
Great. Maybe you could touch on the memory markets a little bit. We sort of have NAND spending at near zero. DRAM spending is sort of high, but a lot of it's China, and a lot of it's specialty packaging. So kind of, you know, what are the prospects for memory to improve? And I know you have some, some insight into utilizations and things like that through your services business. What are you seeing there as well?
Yeah, so utilization rates aren't getting worse, but they're not really improving all that much. We're seeing them improve in the leading-edge logic.
Mm-hmm.
So in N5 and N3, we've seen some improvement there, certain parts of the legacy market. But in memory, they've been fairly, fairly flat. So not and, and our utilizations, because of the way our service business behaves, tend to be a little higher than the utilization rates on process tools, because that's where they have redundancy, so they just shut off capacity.
Mm.
Typically, process control tends to... which is why our service business is going to grow this year. It's not going to grow in line with our 12%-14% target-
Mm-hmm.
But will grow, you know, somewhere in the high single digits. Is that when capital is tight, you don't want to start any more wafers than you need. There's always a return on yield, and customers don't want to tie up any more in inventory that they don't have to. So it tends to hold together pretty well. You are seeing some pockets, although small, of bright spots. The sort of businesses are improving-
Yeah
... At least on the DRAM side. Then you've got things like high-bandwidth memory, which you got bigger DRAM devices, and you've, we've had the introduction of EUV into DRAM, which has been good for KLA. But it's still a pretty small amount of capacity. And then when you take a step back and look at the financial profiles of our memory customers, their businesses are still challenged, cash flows are still kind of weak.
Mm.
And so I would expect that we'll need to see a sustainable improvement in pricing and the overall market that ultimately drives profitability and the cash flow to drive investment. Given the lower utilization rates, you're likely to see, you know, some focus on shrink to meet incremental bit growth demand, if they have it, before you-- which is an investment, but not like a capacity investment. So I think it feels at least, and we never really know-
Yeah
... But it feels, at least right now, that it'll be, you know, fairly slow coming out of this down cycle. Now, we did get some upside in China DRAM that related to a clarification on a technology threshold, and so that created some opportunity for us here in the H2 of 2023. There's probably some continued benefit in 2024, probably not at the same level as 2023. So that has made the DRAM market look a little bit better, but it's been mostly-
Mm.
-you know, that piece was the upside. Everything else has been fairly weak, notwithstanding the strength in a small part of it, in the-
Yeah
... AI support, infrastructure.
I wonder if we could, you mentioned strength in kind of infrastructure, wafer and reticle. Can you talk about, you know, what drives that, and is it—how much of that is... You know, is there cyclicality to that business, or do you look at that as durable?
Well, there is cyclicality. It's a different cycle. And the capacity comes on. I'll talk about wafer first, but the capacity comes on fairly slowly... And over the last, and we're a meaningful part of the CapEx, and so we're probably part, given our lead times, of why that capacity comes on slowly. It was clearly a choke point over the 2019-2022 growth period: access to wafers. And so, there's been a push from our customers. Now, in China, it's a little bit different than non-China, but to add that capacity. You also have other drivers like wafer-to-wafer bonding and packaging that's driving more wafers. But also to add that capacity to be able to be in position to support the industry as it ramps.
There's a lot of long-term supply agreements that customers have. They're carrying higher inventory levels, given some of the shortages that were experienced. So in memory, when memory recovers, memory tends to drive wafers, and so those are all factors that are driving some of the investment. And this business, for us, overall, is, you know, up double digits this year compared to last year. So it does tend to have a different cycle. I wouldn't call it completely countercyclical-
Yeah.
But it comes in waves and takes a while. In China, it was a clear choke point because there's a desire for self-sufficiency and to have vertical integration really from, you know, designs and wafers all the way through packaging for semiconductors in country. So there's access to wafers was a clear choke point because a lot of the existing capacity was tied up in supply contract. And so over the last number of years, and it continues, we've seen investment to develop that domestic ecosystem, and so investment in wafer. Same thing on the reticle side in China, where most of our big customers have captive mask shops where they make their own reticles.
But a lot of customers have to go to merchant suppliers to get their reticles, and so there hasn't been a big investment cycle there. And so as we've seen more and more designs in legacy markets, as semiconductor content has risen, we've seen more demand for reticles, and so you're seeing mask shops to support these legacy markets that are now coming up in China as well. And so we're shipping, in a lot of cases, older versions of KLA products to support that market, which has its operational challenges to go from very low volumes to, you know, doubling or tripling the amount of output per quarter. So we've been dealing with that, but it's had some sustainability to it. So it's a factor in our performance.
These are strong parts, where we have very strong share and probably is, you know... I, I know that there's some different opinions of what China looks like into 2024, but I, I think because of this exposure, I have probably a, a relatively higher level of confidence, at least around KLA's, business as we move into next year. But, but these are, these are markets that we have a good position and, and, you know, we've, we've always benefited from, and, and, some of our peer companies don't have exposure to this part of the market.
Great. You mentioned supply chain as well. It's kind of interesting that you sort of managed the supply chain really well last year, and it sort of hurts your relative revenue performance this year because you didn't have that revenue getting pushed in. But can you talk to generally, you know, how you have managed that as successfully as you have? And then you mentioned bottlenecks on the optical side, you know, where are you in terms of resolving these?
Yeah. So relative performance this year, but then-
Yeah
... I think as we move into 2024, I don't have this-
Yeah
carry forward to deal with that, maybe others do.
It's just timing.
So you're right, and, and it's timing over-
Yeah
... over a multi-year period. You know, KLA's go-to-market is about delivering differentiated solutions and driving innovation. We tend to invest a lot more than any of our competitors, and as a percent of revenue, we're about 13% of revenue. We like to deliver products and bring them to market and have a fast cadence and so on. So part of that differentiation comes from the capabilities that exist in certain aspects of our supply chain and how we partner with our suppliers to develop them, where we bring a lot of design capability, and we partner with our suppliers in terms of manufacturing know-how and IP. So we have, you know, these. It starts with these differentiated components and ultimately lead to our market position.
To do that, we tend to work very closely with single sources, and, you know, our lead times are very long, and so we have a lot of capacity agreements with our customers and pretty good clarity, our suppliers, but pretty good clarity around what we'll get in volume. And then we, we basically sort of optimize the rest of the manufacturing infrastructure, both in terms of facilities, but also people, against those volume expectations. So usually, it's pretty predictable, and so we, we work really closely in these relationships. It tends to be more partner driven, less transactional. I'd like to think we're a pretty good customer to our suppliers, because, you know, it's one of those that, you know, what are you optimizing for? We're optimizing for differentiation and unique capability. What are we accepting?
We're accepting that we're going to help make investments. We're going to make commitments so they can plan. We're going to follow through on those commitments and take the parts we get. And I think as a result of that, we still had the challenges that everybody else had, but once we sort of established the max opportunity, perhaps we had a little bit more time to go chase down all the mechatronic and electric, electronic components, semiconductors, and so on, to be able to execute. A lot of our customers would say you know, we weren't able to do what they wanted us to do.
But I do think that, on a relative basis, you know, we were able to—and our business is different, but I think we were able to manage our way through it. I think we're very well positioned for when we see an acceleration in demand moving forward. I think we're very well positioned for our ability to scale the business over the next few years in terms of new capacity, both in terms of our own facilities, but also capacity in our supply chain. We talked about a plan to get to $14 billion in revenue in 2026. I think I've made the investments I need to make to be able to enable that. So I just need the end markets to recover, and I'm ready to go.
Yeah.
I think that's part of it. Now, on the optical inspection side, I talked earlier about, you know, it is a strong market that is inflecting because of scaling. One of the fastest, if not the fastest-growing markets in all of WFE, KLA has a strong position there. The capacity comes on very slowly. We have a unique relationship with one supplier to be able to deliver the capability we need, and it just takes some time. I think that business will grow next year. I have some new capacity coming on next year. It would have. I would absolutely have shipped more this year had I had capability. But I think as that starts to come on next year, we'll be in a position to grow that business.
There's another increment to come in the next couple of years beyond 2024, that I think positions us to see that market, which should scale and grow faster over time than WFE as the litho roadmap continues to execute.
Great. Maybe we could pivot a little bit and talk about China. You know, for you and your peers, we're looking at north of 40% of business coming from China. It seems like the multinationals aren't doing much, so a lot of that is the China sovereign. Surprising to me, at least, given the export controls that you've seen, that a lot of your business you couldn't ship to a year ago, you know, you're still at these levels. And I think, you know, the consensus is that the those percentages come down, but it does stay strong. Can you just talk to the activity that you're seeing there? How durable it is? You know, you've expressed confidence to me that it's not a pull forward, it's not, you know... Can you just talk to that opportunity in general?
Yeah. So part of the reason why it's a high percentage of the mix is, to your point, right, that we've seen a lot of the other stuff fall off more dramatically. If you go back to 2021, 2022, and there's been an investment pattern here, it's been generally, you know, 20-25%. So it was higher last quarter. It will probably be elevated. I don't know if it'll be as high this quarter as it was last quarter, but in this quarter. And this year, it looks like it's probably 30%-ish, so it's clearly higher and probably continues into next year based on how we're looking at the year overall. There are various components of the investment, and I talked about the infrastructure being part of it. We mentioned memory and that being another part of it.
And then you've got, you know, legacy. I'll call established legacy investment, which we've seen other parts of legacy that have been growing. And so, you know, we have a number of customers that are established customers, have been supporting the legacy markets that frankly aren't any different than other non-China customers in terms of how they support the legacy markets. So that's a piece of the investment. And then, you know, you have this piece that's about $7 billion-$8 billion or so WFE, that's related to a number of projects around the country that are in early stages in terms of development and limited production. Some of it's supporting specialty semiconductor fabs and opportunities, but also silicon. And these are projects where there's funding.
Typically, there's a partner funding source that's government, and they are doing R&D, and R&D never ends, and then trying to prove out process and capability to align with customers and demonstrate that they can deliver to those customer requirements. Those are early stages. They're subscale. When you're doing limited production, you still need the whole portfolio of KLA equipment. There's heavy sampling that's happening, like there always is in development. So we're participating there. A lot of these customers had placed orders in 2021 and 2022, and given the strength of the demand we had from our major customers and the supply imbalance we had, we weren't able to satisfy that business then. So this year we've been able to. So it's been in the queue for a while.
And if you look at what I have in terms of, you know, those customers, a lot of them are newer. You know, I have, I have backlog. They have to perform with us as customers to be credible for slots, to be credible for resources, for me to make the investments I need to support from a service point of view. We have a little over $800 million in deposits related to, to that business as well. So look, over time, I'm sure there's some sort of, there, you'll see some rationalization, but just as I look at the near term, and I look at next year, it feels like it, you know, we see a continuation of, of the behavior that we've seen, and, and that continues into next year across that part of the market, the infrastructure, and so on.
So, you know, look, over time, we'll see what happens, but at least in the near term, it looks pretty good for us from a visibility point of view.
If you talk about that going back to the more normal 30%-ish, not putting words in your mouth, but if that's kind of the—does—is there, is there any margin impact to that?
Yeah.
And I guess, you know, you're obviously expressing comfort that you can stay in this ballpark of revenue through the H1 of next year.
Well, it'd be like 30, 30-ish% that 2023, 2024. You know, maybe normal is more 25%.
Yeah.
But for KLA, our go-to-market is much more about understanding the value we bring to our customers and having a lot of discipline around that pricing. Our pricing is generally consistent around the world.
Mm-hmm.
Our market share is fairly consistent across every region. Outside of volume-related, you know, discounting, there isn't really fundamentally different margin profiles across the different regions. One of the things we've done in China, but even in the legacy market, is when we identified that we would see investment in legacy moving forward. Historically, it's been served by used equipment. There hasn't been a lot of investment. In some cases, take versions of current products that we can sell into those markets, where we can actually drive higher levels of revenue, where the customer gets a cost of ownership benefit and upgrade path over time, and we're able to get more revenue at that node than we were able to back when that node was a mainstream node 10, 15 years ago.
So, so that's been good from a margin point of view. Restarting sometimes old manufacturing, old products is not easy, and there's inflation and those pressures, so it can sometimes be a negative to margin, but I think as we've balanced it, it's been, it's made it pretty consistent. The legacy market overall is great because it allows us to limited R&D to be able to support those nodes.
So the, the operating margins or the incremental operating margins can be accretive overall, given that we've been able to continue to sell stuff that we've had for a while, and given the right pace of Moore's Law, where it isn't too fast, but it's moving fast enough that you've got at least enough compelling attributes to drive the design environment, allows us to extend the platforms and really optimize the platforms in terms of cost and capability by leveraging, you know, new compute and AI, AI-like capabilities to drive more value out of the platform. So there's been a lot of dynamics that have been good for margin, but specifically China, it's not much different than it is anywhere else.
Okay, so we have about two minutes left. I'm going to see if there's any questions from the audience first. Charles?
Hi, if I can squeeze in two questions. One is on China. Just looking at the SPE industry generally, the orders from China have been very high, the revenues from China have been very high, and one suspicion is that the Chinese are buying as much equipment as possible now because they are scared that the rules are going to tighten, and so they want, and they want some spare inventory. The second, totally different, is on silicon carbide. It seems that there are a couple of Europe-based companies, AIXTRON and LP, and LPE, addressing the deposition side to that market, and the silicon companies cannot really break in there. Are you able? Would you have a comment on that?
Mm-hmm.
Are you able to expand to span both, both silicon and compound?
Yeah. Yeah, so on the first question, so we're installing the equipment, and utilization levels are not very high, particularly for, in a lot of cases, because it's R&D and limited production. They're not at a place where they're driving, you know, full production in some of these smaller scale fabs. So the technology thresholds have been clear in terms of 14nm and below on the logic front and 18nm half pitch on DRAM and 128 layer on NAND flash. What we're shipping has been legacy to support the self-sustainability directives or desires within China, but both for self-sufficiency in semiconductors, but also to support the electronics export economy.
So, tools are being installed and are being used, and all our legacy in terms of the design rules that we're supporting. On specialty, it's a unique opportunity, a lot of yield challenges in the substrate. We come at that market multiple ways with process tools through our specialty semiconductor business, which is based here in the U.K. The inspection side, which you're not really dealing with tight design rules, which drives the macro inspections, so very high volume inspection, lower sensitivity systems. And so that's mostly, you know, that's where we're addressing, you know, that part of the market as well.
Good opportunity, a lot of growth there, and lots of challenges in the substrate, and over time, I think, you know, as you cram more and more feature sizes together tighter, it creates opportunities for us, on the inspection metrology front.
Great, thanks.
Yeah.
Great. In the interest of time, we'll wrap it up there.
Great.
Thank you very much.
Thank you, Joe. Happy holidays.