Good morning. Good afternoon, everyone. Welcome to the final day of Citi 2021 Global Technology Virtual Conference. My name is Atif Malik. I cover U.S. semiconductors and semiconductor equipment stocks at Citi. It's my pleasure to welcome Rick Wallace, CEO, and Bren Higgins, CFO, from KLA. The format of our discussion is fireside chat. I will go with my questions first. If you have any questions for the KLA team, feel free to text them in the box on your screen, and I'll take those questions towards the end. Welcome, Rick and Bren. Thank you, Atif. Atif, thanks for having us. Appreciate it. Always. Rick, as I was preparing for the fireside questions, I could not help notice the transformation in the company over the last two-three years. While you've always been strong on wafer and reticle inspection areas for chips, your exposure to advanced packaging, PCBs, displays, exposes you to exciting markets like 5G and auto. You're expecting SPC sales led by optical pattern wafer and reticle inspection to grow 40%, outpacing WFE this year. This came as a surprise to many investors. Talking to other equipment makers, like ASML, Nova, it sounds like some switch got turned on in 7 nm, in respect to PDC equipment. What's driving the strength in the optical pattern to $1 billion this year, and how can SPC outgrow WFE again next year? Great questions. I think it's a combination of a number of factors that all are kind of happening at the same time. Probably the biggest is that there's the need and then there's us having the solution. The need has been there in advanced technology transitions historically, but this is a big transition as people are going through advanced design roles. Especially as you think about 7 nm, 5 nm, and what's coming. The number of process changes is so significant, the costs of those nodes are so high, and so the opportunity space for us to operate is pretty large. That, coupled with the fact that we have two product generations of optical wafer inspection, the Gen4 and Gen5, both finding and enabling our customers to achieve their objectives as they deal with all the challenges of these advanced processes. As a result, Gen4, which we introduced quite a while ago, it's on several iterations, but that refers to the wavelength technology, is actually on pace for a record year, in addition to the ongoing growth that we're seeing from Gen5, driven heavily by the fact that more and more of our customers are adopting these systems at more steps in the process, inspecting more layers. All of those factors together mean that we're seeing advanced growth and Bren can talk about the demand on our capacity, but it extends at this point well into calendar 2022 in terms of the needs that our customers have. We're working with our supply chain, to make sure that we can continue to support that. In addition, though, your question about process control, it's really across the board, whether it's optical metrology, which is another business for us that continues to grow quite a bit. Reticle inspection for KLA is going to have a record year in 2021, again, driven by the advanced nodes, the advanced designs at advanced nodes, plus EUV. All those factors together are contributing to really an outstanding forecast for the rest of the year. Yeah. Atif, I'll just add that the proliferation of designs at once we got to 7 nm and even beyond, has been a big driver of our customers' investments because they have to manage more process flows, they have to manage more designs and how those tests design rules in different ways. There's more investment in process control to monitor and control that environment. Of course then they have to deliver to a pretty robust demand environment in terms of the semiconductor revenue part of it, where those designs are going into. That is also when you take that and couple it with the introduction of EUV and some of the changes in technology requirements, has in some ways forced our customers to keep a lot of capacity in place at those nodes and they haven't been able to necessarily migrate some of that capacity to the next generations. We had a period of time there where what we call the reuse component of our business was higher because EUV was delaying. There weren't that many designs at the leading edge besides the large volume guys. As the large volume players started to move to the next node, that follow-on demand wasn't materializing, and without a technical driver, without a demand driver, customers were able to move some of that capacity. Today, they don't have the ability to do that. The investments for each node are happening, and in some cases simultaneously, as we've seen over the last couple of years. Great. Bren, if I go back to 2019 Investor Day, which seems like a long time ago, as you guys flagged some new products on top of Gen5, you talked about E-beam as well as X-ray. The question is, on E-beam, is E-beam complementary or cannibalizing the optical market? You have about 10 tools in the field now, and you're targeting 20% share exiting this year. How is the customer response to your optical E-beam combo approach so far? E-beam and optical have always been complementary, and if you go back over a long period of time, if you think about overall pattern inspection, 15%-20% of the market's been served by E-beam capability, and it tends to be in the front end of engineering analysis and defect discovery applications. The 80% has been optical solutions just because of the cost of ownership, really driven by the speed of the technology, ability to inspect full wafers and all those kinds of things that drive customers to utilize optical for production environments. That drives the bulk of the spend there. We've seen in one of, I think, the competitive advantages of KLA is to be able to offer portfolio solutions to our customers. Rick talked about Gen4 and Gen4 extendibility. Part of the reason why we've been able to extend Gen4 is that we've been able to utilize more AI and machine learning capability to drive that product line. We've also been able to utilize that with E-beam systems to be able to increase the relevancy of the overall optical solutions. You can use that to train and point your inspectors in a smarter way. That's been a driver for that part of the business as well. The market has a number of players for E-beam directly, so it wasn't so much for us of looking in a market we have multiple players and saying, "Okay, given some of the dynamics I talked about earlier, does it need another supplier? Could we differentiate in a unique way?" More challenging, when you take it and couple it with the optical solutions, it does create an opportunity for the collective approach to drive our overall market share in the broader pattern inspection part of the market. Great. We're pretty pleased with where we're at, and I think the roadmap on the platform is encouraging in terms of the ability to scale it and drive more capability similar to what I described. Great. The next topic is really a topic of government spending. Equipment makers like KLA are clear beneficiaries of U.S. CHIPS Act and government-sponsored foundry spending initiatives in Europe and Japan, given your outsized foundry logics sales exposure. Some investors fear that equipment makers who have pushed the idea of consolidation in the semiconductor market and the associated lower CapEx volatility as a favorable shift for the group. The emergence of the new foundry and IDM 2.0 models could reverse all that. How do foundries sustain the spending and don't cause an oversupply down the road? That's a good question. I think that it is a matter of scale, though. I think if you look at the amount that's being proposed in the CHIPS Acts and the others, I think there are incentives, there's still going to have to be economic justification for these companies to build the fabs. As you know, the amount of money that it takes to build a fab is significant. Even if the U.S. spends $50 billion, it's going to be to a number of players to support infrastructure and so on. I think what that money will do is bridge the gap between should I build another fab in Asia, or should I build it in the U.S., or should I build it in Europe? It's really kind of a transition cost management, as opposed to introducing a bunch of volatility into the market. It will be less efficient. There's no question that it'll be an industry that's slightly less efficient if you have multiple sites that are for a company as opposed to consolidating. I don't think it's a huge inefficiency, and that's really what the government funding is going to allow them to do, is to take on a project in a new area. It's also been true for years that there have been incentives, even in places that whether they're heavily populated by semiconductor fabs now. The countries there have incentivized those players. Some of this is just trying to create a playing field that's more equal for the U.S. and Europe. I don't think it'll create a big amount of inefficiency, but I do think it'll create some new sites. I thought years ago, there's a chance there'd be very little future fab investment in the U.S. I think we're seeing, obviously, now that that's both coming from government funding, but also other factors are driving more investment in the U.S. Got to give the scale and maturity of these customers, I think that they've done a pretty good job over time of managing supply and demand dynamics. To Rick's point, the industry squeezed a lot of efficiency over the last few years, and last number of years, decades, actually. That anytime you're spread across multiple sites, it does introduce some inefficiency in the system. We think that it does, whenever you're starting to expand like that, it does create an opportunity for us as the resources and talent get spread across multiple locations. It's an opportunity for us to help those customers ramp up those new operations. From a share point of view, we think when you have the scale that a KLA can provide and an intimacy with sort of process and design and working very closely, collaborating with our customers, that it does create an opportunity for us to play a bigger role in helping ramp those new sites. Of course, we have to execute, but at the same time, it does create an opportunity for us. When are you expecting equipment contribution from some of these fabs? Is it already in your backlog or is it still to come? We haven't seen anything really in our backlog. I think it's still to come. You've got to build out the facilities before you'll see the WFE investment. Certainly, there are conversations that are happening in those areas, and we're starting to plan for resources. We're hiring people because we've got to find the people, and it's a very tight labor market out there, but also the training that's required to make sure that our support personnel is prepared to support it. We're investing in it now, and we'll start to see, I think that, the business start to show up, maybe towards the end of next year. I don't think we'll see anything before then. Right. I think that's right. I think it's a late 2022, 2023 phenomenon. If you think about how long it takes once you've got approvals to build a fab and then start outfitting it, that's about the timescale, even if they're committed as of today. There's some projects, as you know, that are in the formative stages right now that are going to take, that's probably later 2023, depending on how quickly they act. Great. Some of the major foundries have talked about $100 billion CapEx plans over the next three years, and they're asking for pricing discounts from equipment and material makers to lower CapEx burden and gross margin headwinds for them. Understand these customers have always been large for you guys, and they always get some kind of volume discount. Is there a risk to equipment gross margins from rising wafer prices, or it just gets passed down to your suppliers? Well, I don't think there's risk. I think our model kind of outlays how we think about our model going forward. I think we're managing our conversations. I think, sometimes things get misinterpreted a little bit. I think what we hear from our customers is not so much about price reduction, it's about productivity improvement. We get a lot of targets which we have for our own teams, but also for our supply chain about having productivity to remain relevant and competitive. That's frankly not a new phenomenon. That's what we've had for a number of years. Certainly, there's increased focus on it. In many ways, we're often the solution, not really the problem, because we're helping our customers get more usage and capability out of the rest of their investment. A lot of the process tools, if they can get yields up faster, if they can maintain higher yields. We actually, I think in a slightly different position in this relative to the process players. I think the only thing I'd add to that is sometimes, cost reduction and improvement in cost of ownership are different things. To Rick's point is if we can enhance productivity, and do things with our systems that can speed them up and provide more sort of faster time to results, if you will, that those are opportunities for us to improve cost of ownership. Our pricing has always been based on a value equation. I think that, over time, we're always looking to offer our customers more value at lower cost of ownership. That's been frankly part of our model and what drives sort of the pricing model over time. Yeah, that's a good point, Bren. Bren, the next topic is recurring services sales. You and your peers have been doing a much better job in helping investors understand the nature of your services business. Your services sales are approaching one-fourth of the total sales. I was positively surprised to hear 75% of your semiconductor process control tool services sales are recurring, and 90% of PCB services sales are recurring. What actions are you taking to further drive increased services for system sales, and how should we look at this business longer term? Well, we've talked about a long-term model of 9%-11% growth in that business, and one thing about service has been that it's been fairly predictable over time. Despite what might happen in the industry in terms of CapEx investment, customers have always invested in running their install base and continuing to deliver on their demand that way. Our business is a little bit unique. We report services through sort of service business that doesn't include upgrades and things like that. I think it's unique partly because our customers and how they buy process control is that they don't buy a lot of excess capacity. They buy what they need, and then they run it very hard because the results, the information they get is so important to them. The systems are very complex. They don't have a lot of volume. It's very hard for our customers to train their own people to service these tools. Even at KLA, we have service engineers that support individual products. You can't use them across multiple product lines, just to give you a sense of the complexity of the systems. There's a little bit more of a dependency on us, I think, to provide that service and to keep these tools up and running for our customers and matching performance. I think that's been a driver of the contract structure that we have, both in terms of tool contracts, but also fab-wide contracts as a fab matures. We've seen more interest from customers, given the strength of demand in a service contract structure. Obviously, that allows us to ensure that we've got the right resources to support them in terms of the commitments we're making. We've got the inventory in place. There are economic benefits for the customer, obviously, to be able to make us more responsive, but also allows us to optimize our cost structure underneath. It really is a win-win situation overall. We've seen that the trend moved. If you would've asked me one year ago, I would've said, or two years ago, I would've said 70%. Now we're talking about 75%. I think that there's potential for a higher attach rate moving forward. The customers are using their install base for longer periods of time. One of the great things about the cycles, we've seen demand so much more broadly across different design nodes. As a result of that, you're seeing more interest and more focus on running the install base out there, and that's extending the life of a lot of KLA tools, and so that's been a driver for the service business as well. I think that given the strength of demand we've seen in the WFE environment, just new capacity that's going in, plus this dynamic around the install base, that there's certainly, I think, a tailwind in terms of our expectations of the through cycle or the longer-term growth rate in this business as we move forward. Great. Let's talk about the EPC business and starting with auto. Auto end market has been strong this year, and you are uniquely positioned among front-end peers in auto. Auto software and system sales are growing 40%-50% this year, 5%-10% of full-year sales. Are the R&D requirements for auto products similar to or different from SPC products? They're different, I would say they're not as expensive, ultimately as intensive in terms of overall spend, because it's maybe more on the complete solution around some modeling as well as data management, because people are really looking for zero defects. As you know, most of the semiconductors for automotive are not leading-edge semiconductors. Although there are an increasing number of leading-edge that are going into cars, or more advanced, no, it's not leading-edge. They don't tend to be. The overall investment in terms of us bringing those solutions, the acquisition of Orbotech gave us a bigger footprint with those players because we now have process equipment as well as process control equipment to support them with. I'd say that that's given us a better presence at those companies. Obviously, as you know, there are challenges supporting the auto industry due to a number of factors with all the shortages that are showing up, because of all the players, they're the ones that hit the brakes the hardest at the beginning of the pandemic. I think that playing catch up has proven to be incredibly challenging. We're actually trying to help in that because we help drive yields up, which alleviates a little bit of the pressure, but obviously there's a lot there. Automotive is very good. I think that we're seeing, not just more opportunity with the semiconductor guys, but the people upstream of them are wanting to get our advice and guidance on how to have a strategy around their procurement in terms of both yield management and how they think about supply. Bren, PCB has been benefiting from 5G within EPC. You highlighted ICOS on advanced packaging on the last earnings call. How sustainable is the high teen growth that you're seeing this year into next year? Yeah, on the systems side, if you look at just overall EPC, it's about 20%. When you add in the service part of that business, you end up in the high teens. Systems has grown nicely. The growth within PCB has been very 5G centric. That's been a nice driver this year. I would expect that business to continue to grow as we move into 2022. If you look at just the overall mobility cycle and how it's impacted both SPTS, our specialty semiconductor business, which is also impacted by, we talked about automotive, but the power semiconductor part of automotive has been a driver for that business as well. Whether it's RF chips, which they have a unique position in that market in terms of how it supports not just handsets but also infrastructure. What we've seen in PCB, in ICOS, where you have volume coming from a lot of mobility drivers and driving finished component inspection. You're starting to see more higher value components related to data center that's been a driver for that business as well. I've made this comment before that given the strength of semi process control, we spend a lot of time talking about that. This part of our business is also growing greater than 20%. It shows you the strength of the semi piece when you're looking at this part of the business, at 20%, it's diluting the growth rate overall of the company. A lot of interesting drivers, and I think it's multi-year drivers given the 5G rollout, both from an infrastructure point of view, but also a handset point of view. Great. Let's talk about the end markets. Any changes in DRAM spending plans given your customer inventories are a bit elevated and prices are peaking? Not seeing any changes in our plans. We look at DRAM this year, DRAM, we would expect, if you think about the overall market at 35%+ for WFE, DRAM is faster than the overall market in terms of our modeling. Flash is slower than the overall market if you think about memory in the aggregate. I see strength in DRAM investment continuing through into the first half of the year. I've talked about sustainability of our business as we move into the 2022, comments about the first half being at least as strong as what we're seeing in the second half of 2021, certainly some incremental DRAM investment in the first half of next year is in our plan. We feel pretty comfortable about that. A lot of times, how we engage with our customers in this part of the market doesn't move in a short-term fashion relative to pricing. Certainly, how much customers invest over time is influenced by that as that drives their profitability. In terms of how they engage with KLA, particularly from a technology point of view, and with the introduction of EUV being a driver of technology within the DRAM space, is a driver for investment in KLA. Bren, the domestic China first half has been very strong. Have you seen any impact from credit issues from certain memory customers or U.S. China restrictions looking to second half? The market's been fairly stable. Those issues are out there, but we've seen an acceleration of demand in China. We've seen acceleration demand across all segments and all customers for the most part. We have seen the demand strengthen as we move through this year. I think it's spread across a number of different projects. As I look at 2022, how much does that continue to grow? It's not completely clear to me yet, but I see a level of sustainable investment that's happening there. I would say that what we've seen this year is probably in excess of what the overall market is expected to grow year- to- year. A little bit faster growth in China, particularly around the indigenous projects than the overall market. Rick, I have a question for you. The transition to 3D logic devices is beginning in earnest with the shift in architecture from FinFET to Gate-all-around. How is KLA positioned to benefit on the metrology products from this transition? We're in good shape. I think we have been working on it for years. As you know, it's maybe a little later than some had hoped. This transition was scheduled earlier in time, therefore, it gave us more time to work on our solutions. We feel like we're well-positioned. If you back up a little, what happens when there's a new architecture, is we have a team that will model that and will model what the impact is on the different systems that we offer, whether they're metrology or an inspection, to make sure that we're designing the optimal capability with the optimal configuration, and working with advanced customers on that. We've been at this for quite some time. It is an interesting area where Gen4, for example, happens to be ideally suited for inspection of Gate-all-around because of some of the contrast issues with materials. Another extension of that product family. Gen5 will be used, but in concert with that, not necessarily for the same things. Metrology-wise, we have solutions that we've been working on for quite some time. We feel really good about where we are, and these inflections always drive at least increases during transition of process control intensity. Often that is sustained because these are just more finicky transitions, and over time, that's part of what drives the overall process control intensity. Great. On the financial model, Bren, you will close 2021 ahead of the 2023 model and have done well with Orbotech. Are the Opex and gross margin synergies fully realized now with Orbotech and 60%-65% incremental gross margin level sustainable? Yeah, that's a great question on Orbotech. Certainly, some of the growth drivers has helped drive the incremental margins and that part of the business, the overall business as well. We're still not completely there. We certainly have line of sight to synergy expectations that are roughly double what we expected going into the transaction. We still have some integration activities, particularly around systems, that could drive some further synergy opportunities. Of course, facility footprints are also something that takes a little bit of time to work your way through. We feel very comfortable about where we are today, and if you fast-forward another 12 to 18 months or so, we should have realized the full extent of those synergies, which, as I said, is almost double what we expected going in. I think it's worked out pretty well overall. There's some cost pressure in the system. I talked about it at earnings, that when you look at incremental freight costs and some of the component costs that we're dealing with, our materials teams have done a great job of managing our suppliers through not just the demand, but also the sub-component demand and making sure that we can source that. I talked about roughly 50 basis point kind of headwind related to that in our overall gross margin model as we move into next year between freight, components, and so on. It's embedded in the guidance I provided as we move into 2022 based on top line expectations we have today, somewhere in that 63-ish% range, ±, the ± quarter- to- quarter is really a factor of the mix of our business. I think it's contemplated in the guidance levels that we provided. There is pressure out there, and we're trying to manage our way through it. I think the incremental margin model, gross margin of 60%-65%, in that range, 60%-65%, makes sense. I don't see any reason why it would look any differently from here. On the OpEx side, there is pressure around labor markets are tight, labor costs are higher. It feels like in some ways cost of everything is increasing, so why not labor as well? We do have some pressure in the system around labor costs, which we're trying to manage our way through. We've been pretty clear with the model of driving in a normalized revenue environment, we should be able to deliver 40%-50% incremental operating margins on our business. Obviously, over the last couple of years, we've exceeded that long-term target. In a normalized revenue environment, I would expect that we'll continue to operate in that range. Obviously, over the last couple of years since it's been higher, potentially we could be toward the lower end of the range as some of the costs catch up to the business. I think that the long-term sort of sustainability of the operating model is pretty consistent with the way we guided and modeled, and the mix of business will be the biggest driver of any variability in it across the overall, whether it's service EPC or semi process control. Great. The next one on capital allocation. You announced a new $2 billion share repurchase and 17% dividend hike last quarter. How should we think about capital allocation plans from here, especially M&A appetite in faster and exciting back-end packaging markets, or you're satisfied with your portfolio as it stands? Well, on the M&A front, I think we'll be consistent and prudent about how we look at it. We always look at, okay, how can we add unique value and how can we drive a return on the investment that exceeds the alternatives? The alternative being putting the capital work and buying back our stock. We have a pretty shareholder-value-centric approach to how we think about that. Certainly, there could be some opportunities for some tuck-ins out there, and we'll continue to evaluate those. We look at a lot, but at the end of the day, it's got to fit this construct, and it makes sense for us in terms of delivering either new capability or new ways that KLA can take a market-leading position in the deployment of our operating model drive higher operating leverage out of the business over time. We're big believers and have been for a long time. That capital gets valued when it gets deployed well. We need to be explicit about how we're modeling that and how we think about that. Certainly, our dividend, we're trying to drive the growth of the dividend payout over time, consistent with the growth rate of free cash flow of the company. We can maintain a consistent cadence around it. As we look at the opportunities that are out there, both in terms of the level of leverage we have, the cash reserves, we ought to be able to maintain, on a through-cycle basis, a total return in a balanced way of at least 70% of the cash flow that the company is generating over time. This year, I've provided some color that I expect it to be in excess of 85%. Through cycle, over time, I think 70% is sort of a minimum level of returns that we would drive. I think the actions that we took and the announcements we made at earnings in July are just consistent with that strategy, and there's been, I think, no changes to the overall approach the company has here. Great. Let's go to audience questions. The first one, could you ask the team to just talk generally on the differences in use of process and inspection in memory versus logic? Especially given DRAM complexity going up, curious if adoption there will start to look more like logic. Great question. There is a significant difference in terms of the amount of money people spend on process control, making memory devices versus making foundry logic. There are a couple of reasons for that, the biggest, I think, single reason, I would say, is the repeating nature of a memory device and the fact that there's repair. In other words, if you have defects on memory, in a post-processing of that device, you can fix and make up for that defectivity and repair it. Virtually every memory device is repaired. That's different than logic, where a defect will kill the die and cost you a lot of money. Inherently, there's a difference there. There's also the design rule differences, which is what this question really speaks to, where, of course, logic foundry pushes the most advanced design rules. There are aspects of memory that are pushing different aspects of design rules which drive process control intensity up. One is design rule is getting tighter on DRAM. Process complexity is increasing on 3D NAND as they add more layers. There are different kind of process control steps that have to be included to deal with that, whether it's overlay registration from layer to layer or just looking at the 3D dimensionality of the device. Process control intensity for memory will likely grow. It will also grow for foundries, I think the gap will probably still be on the same order of what it is today. We do see drivers that will drive memory process control higher than it's been historically, but not to the same levels as advanced foundry and logic. Okay. The next one, what are the key drivers for market share gains by end market in calendar 2021 and 2022? By end market. I think the market share gains that we anticipate have to do with increased adoption of some of our leading product areas. By virtue of the overall process control budget, our Gen5 optical wafer inspection is a good example. There's really no alternative that somebody else doesn't make a broadband optical wafer inspection system, so it's really that gaining share of the overall process control budget, and that tends to be mostly used on advanced designs, which are, you could argue, have to do with AI devices driving a lot of that high-performance computing, driving the foundry. I think you could, by extension, say that those were the end markets that are driving more of the optical wafer inspection market gains than others. In some areas on the lower-end optical inspection, it's probably share gains are happening in automotive semiconductors, because that's where we're selling those systems, and we have a very large share. It's not nearly as big as the front end, but the rate of change is significant in terms of how much it's growing. Really, you can kind of go to the main drivers, and those are the ones. We already talked about the question about PCB being tied to 5G, and there we have products that are tied to. In many ways, we're really tied to all the end drivers, but I would say the biggest one would be associated with high-performance computing AI as it pertains to advanced logic. Bren, anything? On the optical side, look, I've been saying for most of the quarter that if you just take a look at the growth of optical inspection over the last couple of years, arguably the fastest, one of the fastest, if not the fastest-growing market in all of WFE. The inflection there, both with leading-edge foundry logic, and we've talked about the end market design flow that's driving that, but also you have competitive dynamics that are also participating with that. Then a technical driver like EUV, that we're seeing really strong adoption in that market, and that's a very strong market for KLA. On the reticle, and with the introduction of EUV, there's multiple products we're selling to support reticle quality and fidelity, both in the mask shop and the fab. As Rick said earlier, we think that market, it will have a record year this year and likely grow faster than the overall market. Then in film metrology, it is the market, and I have made this joke lately that it is the largest market within KLA that nobody asks me about. Over the last couple of years, we have seen share gain there as a business that behaves like a process business in a lot of ways. If you are adding process tools, you have to be able to have incremental film measurement capability. So we have seen that business grow faster than the market as well. I think there are specific drivers, as Rick was talking about, that are driving certain parts of the portfolio. Having the portfolio approach allows us, I think, to have the right solutions for our customers as the roadmaps change and the markets change. We're able to meet our customers' technical requirements, but also their economic requirements, which is also very important to them as they're managing these pretty large-scale fab ramps. Great. We're almost out of time. Rick and Bren, thank you for your time and insights. Thank you. Thank you very much for having us. Take care.