Hey, good morning. Good afternoon, everyone. Thank you very much for joining us for our annual technology and internet conference. My name is Toshiya Hari, I cover the Semiconductor and Semiconductor Capital Equipment space here at Goldman Sachs. We're very excited and very honored to have the team from KLA with us this afternoon. Bren Higgins, Executive Vice President and Chief Financial Officer, and Ahmad Khan, President from the Semiconductor Process Control business. This fireside chat will be about 40 minutes long. I'll go through a list of questions that I put together, but also do my best to weave in any questions that we may get through the webcast. With that, I'd like to get started. Bren, Ahmad, first of all, thank you very much for joining us. I know it's a busy time, appreciate you carving out time for us. Bren, before going into specific questions, I was hoping you could kick us off by reflecting on 2020. It was clearly a very challenging year in so many ways, but at the same time, as we were sort of talking before the session, it was a very strong year for your business as well. If you can kind of reflect on 2020, that would be great. Also sort of talk about 2021, your expectations, key goals for the broader KLA team. Sure. Thanks, Toshiya, and thank you for having us. It's interesting year for 2020. I was talking to Toshiya before we got started here, that the last time we attended a sell-side conference was this one, at least the sell-side conference live back in early 2020. For a year that was awful and challenging in a lot of ways, it was very good for our business. Before I get started, I'll make a quick safe harbor statement that I'll make some forward-looking statements, and those are subject to risk, and you can find those risk factors in our SEC filings, which you can access through our website. It was a year where we saw a pretty good demand profile as we moved through the year, and as we got into the second half of the year, it strengthened. A lot of compelling demand drivers, really across the whole spectrum of in demand from leading edge to trailing edge, was encouraging to see. We had really good performance. Like I said, it started to strengthen in the second half of the year. We finished 2020 with a 15% top-line growth and really good leverage in our business in terms of operating income growth at 28%, EPS growth at 31%, and maybe most importantly, free cash flow growth, 44%. Very strong finish to 2020, and in a very strong backlog position and sales funnel as we start 2021, that gives us increasing confidence of the strength of this year. Last week at earnings, we talked a lot about the overall demand environment and our expectations both for WFE growth this year in our Semiconductor Process Control business, expectations for the overall company, both those being into double digits as we move through this year. From a WFE point of view, 2019, we saw it somewhere between $52 billion and $53 billion, 2020, still some companies still to report, we think somewhere between $59 billion and $60 billion. Our early feeling on 2021 is somewhere around mid-teen growth, plus or minus a few points. Certainly, there are some indicators in the market that it's more likely a plus scenario, more so than a negative scenario overall. We feel against that backdrop, that it's a pretty good environment for this part of our business. Breaking it down, certainly foundry/l ogic investment, we would expect to grow kind of in line, maybe a little bit faster than the overall market. The DRAM business is probably growing faster than the overall market. Flash, we're a little bit more cautious on flash. I know that there's been some incremental commentary that's been more positive. Our general view is just given the pricing environment, we've been more conservative in our views of that market. We'll certainly see how that plays out moving forward. Against that backdrop, I provided further guidance around the operating model. We think the operating model will be strong, will have gross margins at the 61.5%-62% range, will drive overall operating margin performance that's ahead of the public model that we've set. I think it's a pretty good setup overall in that part of the business. The service business continues to hum, lots of activities, and utilization rates are strong in the install base, and that's driving top end of our revenue growth targets of 9% to 11% for our service business. We're on the top end of that range, and we're really encouraged by what we're seeing in our new EPC business, which is the electronics components and packaging part of the company, which is mostly the Orbotech acquisition, but also inclusive of our finished component inspection business. I would expect that business to also grow double digits this year. We're, I think encouraged by what we see and with what's in the sales funnel, and excited to go challenge a whole bunch of execution problems here to deliver to, and challenges really to deliver to our customer demand. With that, happy to take your questions. We're excited to have Ahmad here. It'd be great to get Ahmad's perspective on products and technology in the space. Great. Thank you so much for that overview, Bren. I guess to start, wanted to follow up on your 2021 WFE commentary. I guess, the baseline assumption is up mid-teens for the year, plus or minus, but you did say that there could be a bias toward the upside. When you think about the bigger swing factors for the year across logic/f oundry, DRAM, and NAND, what are some of the things that sort of you think about as potential swing factors? I think that, certainly, the leading foundry customer was very public with their announcement around CapEx growth, and that's a big part of expectations for this year. I think that's pretty well understood. One thing that we saw at the end of last year and continuing into this year is a breadth of investment in the foundry logic space. I think that depending on what we see there in terms of others investing is a factor this year. Sometimes those investment patterns are tied to customer wins, the visibility to that business tends to be a little bit less. There's a lot of activity in the trailing edge, both in China, but also even beyond, related to automotive and comms infrastructure and IoT and so on. There's a lot that's in the funnel related to those activities as well. I mentioned flash, that could be a swing factor as well. What happens in leading-edge logic, I think, probably is another swing factor. We'll have to see how that plays out. That could be a factor as well. In general, if you just take a step back and look at all of the markets, I think all segments are driving good growth, and I think that there's, as I said earlier, I think there's probably opportunities for some upside here and there as we move through the year. We'll see how it plays out. Got it. Bren, I guess, given the strength that we saw across the broader WFE market last year and given the robust outlook for 2021, there are some investors out there who are marginally cautious a little bit. I wouldn't say negative, but a little concerned about the outlook. Any signs that you're picking up from your customers that would indicate or could indicate that this is overheating and a little too much from a spending perspective? When you just take a step back and you look underneath and you look at the utilization rates of the install base, you look at the demand drivers, and what you're seeing from a, whether it's AI and data center investment, whether it's 5G investment, and then, of course, some of the things that are happening at the trailing edge, which is pretty well chronicled in terms of shortages, is that semiconductor content's rising in a number of end markets, and so that's pretty encouraging for us to see. At the end of the day, the way we think about it generally is that capital intensity is flat to rising over time, and that the baseline industry ought to grow in line with semiconductor revenue. I think as semiconductor revenue continues to grow, I think historically it was always a GDP plus, I think it's stronger than that. I would think that WFE, given some of the technology transitions, some of the other investment that's happening that may be more strategic, probably enables WFE to grow a little bit faster overall than semi revenue. The way I think about it is on a through-cycle basis, it should be tethered ultimately to semiconductor revenue, we ought to see growth there. I think one of the other signs is if you look with the lack of design starts at the leading edge and with the slow adoption of EUV at the leading edge, that there wasn't a lot of growth in logic/ foundry WFE for a number of years. In the last few years, we've seen accelerated growth, but that's on top of a period where there wasn't much growth. I think that there's some catch-up. I think it's more compelling. You're seeing tremendous design start activity, and for the first time, you're seeing, even at the leading edge, you're seeing investments by the leaders at multiple nodes. I think that's encouraging and reflective of a pretty good demand environment as well. I think I'm pretty encouraged by all of that. Certainly double-digit growth. I'm not so sure whether the forecast for semiconductor revenue is to have double-digit growth every year, but I think what's driving the market today looks pretty good and sustainable, and given expectations around the 3 nm node into next year, some of the inflections related to architecture changes potentially coming beyond that, we feel, and then I think memory just starting to spend after a couple of years of, I think, pretty disciplined behavior. It's a pretty constructive setup, I think, as we move forward here through 2021 and beyond. Great. Makes a lot of sense. Bren, the other topic that we tend to get quite a few questions on is China. It's a region that's been growing very nicely for you guys specifically and for the market broadly, I guess both on the multinational side, but more so on kind of the native China side. How was 2020 relative to your expectations from a local China spend perspective, and what's the outlook into 2021? If you can kind of differentiate between memory and foundry, that would be helpful. 2020 ended up pretty much as we had expected, and so it was pretty logic heavy, more or less. You also have some of the other infrastructure investments happening, like wafer houses, as they're building wafer capacity. I think as we move into 2021, you might actually see some mask/ reticle capacity get added, and I think it's more logic centric as well. Different players investing overall. As we look at 2021, we characterize the environment in China as flat to up. Certainly, the funnel looks pretty good. You do have some new projects, and always with new projects, there's always some timing factors, just as they're facilitizing new projects, and sometimes it's just infrastructure, and that could lead to timing issues. My bias is I think it's flat is probably the worst case in terms of expectations for China next year. Again, as I said, I think you'll see more DRAM investment from China. Last year was a little more flash heavy. Then lots of activity around trailing edge logic investment. Like I said, different players investing, but overall, I think it's flat to up, and I would expect our business to be pretty consistent with that. Got it. Super helpful. Then sort of related to that, you had the restrictions on SMIC, and you along with other players in the market, applied for licenses. It seems like you guys haven't heard back from the government. It'd be great if you could confirm that. I guess more importantly, given the geopolitical tensions and given the shortages, I think many countries, many regions are talking about localizing production of semiconductors and TSMC in Arizona and Samsung apparently potentially in Texas and so many other things going on. How do you see that positively impacting your business in the broader market over the next couple of years? Okay, two questions there, I guess. The first on SMIC, our practice is not to comment on specific customers, certainly customers that are smaller than 10%, and this customer is significantly smaller than that. Just given the commentary we gave back in October, is we said that we thought that the business from that customer, given their CapEx capacity or cadence, was going to be not material to our 2021 plan. As we sit today, that is still true, and that's how we see it. We are in active dialogue with the Department of Commerce related to the guidance that we have gotten and where the directives are. We're actively working through that, on that front directly, also through trade associations in terms of just making sure we have a clear understanding of the requirements. We will, of course, be fully compliant with all the regulations and rules regarding this issue. As I said, I don't think it's material to the 2021 plan, and we're in active discussions, and that's probably all I want to say on that topic. On the second topic, the regionalization question is an interesting one. Certainly, that can create some short-term opportunities. I always think in the long run that this industry's done a pretty good job of rationalizing supply and demand. Any time you add facilities, I think one of the big drivers of efficiency in the industry, if you go back through the 2000s and early part of the 2010 decade, is the consolidation into hyperscale foundries, and that was a big driver of efficiency. As you start to add facilities, it creates an element of inefficiency in multiple facilities. At the end of the day, I think the industry is going to rationalize supply against demand. That's how it plays in the long run. There's probably some inefficiency of just facilitating additional facilities for reasons other than the general business reasons that could create some upside opportunity for the company. I always think in the long run, as this industry's done a pretty good job of rationalizing. Got it. On Process Control intensity, I think you saw intensity grow nicely in 2019. I believe it was up again in 2020. Historically, the key determinant of Process Control intensity was sort of the mix between logic, foundry and memory because you've got significantly higher intensity on the logic and foundry side. Within logic/ foundry and within memory, what kind of trends are you seeing from a Process Control intensity perspective, and what are some of the key drivers as you think about that component to your business going forward? That's a great question for Ahmad. Thanks, Bren. Yeah, I think I could share very good 2019 and 2020, as Bren pointed out, and that had to do with many factors. I'll start with the logic/ foundry, where EUV was first introduced at 5 nm. With the introduction of EUV, each time there is an inflection in the process, KLA does quite well because the defect and the metrology modes change quite a bit, and this drives us to bring new systems in to help solve our customer problems. As you know, KLA is a large participant and has a fairly large market share in this segment. We have a fairly wide portfolio that we are able to bring to the customers. Our engagement with all the top customers is very good. That enables us to collaborate with them in R&D, and also within HVM. With the EUV introduction, we saw large growth in our optical inspection portfolio, specifically BBP, which saw nearly 2x growth between Gen 4 and Gen 5 products, really to enable the EUV infrastructure, and helping with the defect reduction. We see the same types of trends in DRAM. DRAM is now adopting EUV, and the overlay requirements are very tight. This is driving process control intensity up. Then the third area, what I would say is that all things going vertical, as you know, NAND has already gone vertical, and the challenges in 3D NAND continue to get quite complex. For that reason, we're introducing new products with X-ray technology to enable measurements of very high aspect ratio structures. This is driving process control intensity up. Circling back to foundry/ logic again, once EUV is served, which will go from 5 nm - 3 nm, we see a growth in the number of EUV layers. At two nanometer or depending on which customer, the nanosheet introduction is going to happen. That's another huge inflection that changes the front end significantly, and that's going to drive new systems that we're developing now that we will be implementing in N3 and N2 timeframe, again, driving process control intensity up. Again, very good growth between 2019 and 2020, 2021 looks good simply because of N3 growth and as customers increase wafer starts. At N2 timeframe, the growth happens again because of inflections that I spoke about, both in the DRAM side and the logic side and then the 3D NAND side. I think it's several years of process control intensity heading in the right direction. Yeah, Toshiya, I think a couple things on this that's interesting is, the first is that it's heading the right direction into Ahmad's point. There's some products that we think will contribute to the company and contribute to process control intensity in a more meaningful way that are just coming to market. We're just seeing in the market. They haven't even started to contribute at a steady state kind of level, and so we're encouraged by what's to come there. I think one of the other interesting things about what's happening in the foundry space is because demand is so broad and the design starts are so significant, there for the longest time, everybody thought, well, design starts at the leading edge have to come down because cost to design is going up, and that simply hasn't been true. I think as you've introduced EUV, there's been compelling Moore's law type attributes to designs. There's a lot of activity, and that activity stresses the different process flows in different ways. The customers that have to monitor those processes to deliver to yield a die to specific windows is a dynamic that's favorable for KLA. The other thing is that because you have that much demand is that when the next node rolls around, A, there are technical challenges because of the things that Ahmad talked about. Also, it's really hard for customers to try to take capacity that was used at a previous node and move it. We have customers who've been very good at that over the years, and they still are, but because they have so much follow on demand, they're having to build these factories up with very little reuse or redeployment of capacity. That's been a driver, as well for us. Great. Thank you for that. Then as a quick follow-up, maybe for Ahmad, on EUV, how do you think about the rate of proliferation of EUV going forward? Obviously, you have the big logic/foundry guys introducing or inserting EUV to more layers as we go pretty aggressively. You've got the biggest memory guy introducing EUV to their process. There's talk about the number two, number three guys looking at EUV over the next couple of years as well. Could the rate of proliferation accelerate and in turn, positively impact the rate of growth at KLA? Should it be sort of a linear progression, if you will? Yeah. I think these are very good questions. What I would say, it all starts from the design aspect, meaning that our customers are making very careful calculations on how EUV enables them to deliver performance that they need to deliver to their customers. What we see is essentially that as you add the number of layers, with EUV, that performance continues to increase, and that's why we're clearly seeing the foundry logic trend towards EUV. Foundry specifically. We also see trends that the logic side will also turn heavier towards EUV. DRAM is heading in the right direction, as you said, with one player and the second and the third. They're all looking at EUV at this point. It really benefits us because if you're going to put one layer in EUV or 10 layers in EUV, you need an infrastructure. We come in first to really support the infrastructure, to support the start of the EUV, and then as the layers go up, the defect modes on all those layers change. I will tell you that there are specific segments that were essentially not changing. For example, mask requal. Mask requal is a segment that essentially had a holiday for maybe four or five nodes, where essentially mask would have almost the same type of pitches, because you were doing double patterning on the wafer. With the increase of EUV, those masks have completely changed in a huge way. What we see is now double-digit growth in those segments where we see new systems being purchased by customers for requalification of masks in-line in the wafer fabs. This is just an example of where the growth comes from, and I see that this is going to continue till N3 node, and then continue to N2 node. Helpful. Thank you. I guess taking a step back, wanted to ask about the competitive landscape, your market share, your competitive position in process control has been very stable, very dominant, despite the aspirational growth goals that your competitors or potential competitors have talked about. Ahmad, I guess if you can, again, take a step back and speak to the key components that sort of make up that competitive moat that KLA has around process control, that would be super helpful. I think that there are still pockets within the broader process control market where you're perhaps under-indexed relative to some of the other parts where you're very dominant. Are there any specific areas where you're aggressively investing to potentially raise your profile even more? Yeah. It's a very good question. The first thing I would say is that our R&D discipline is very, very good, meaning that we invest a large amount of R&D dollars across the board to ensure that we are able to bring in new technologies for our customers, and they're brought at the right time. The second thing I would say is that we really look at it from a portfolio perspective. Because we're not a single point product company, I'm able to move those R&D dollars from product to product and really double down and triple down on product development as a product is going into HVM and it delivers, you can remove the R&D from there and move it to another product. That has really enabled us to accelerate product growth and new platform growth that enables us to have an even larger portfolio. That's been the discipline that we've been using. That enables us to be really ready for each node when the customer needs a solution. The third thing I would say is that we have a very good discipline of component development and also system development. We're looking at inflections three to four to five years from now that the customers are going to need, and we start doing development on it well ahead of the time so that we are able to meet those needs. For example, gate-all-around development we started about four to five years ago, and we are very confident that by the time customers are going to go put gate-all-around in production, that we would have a new variant of optical inspection system and a new variant of metrology system that is going to be really focused towards making sure those defect modes are dealt with. That is kind of like the overall mode that enables us to really stay differentiated. Gen 5 is a really good example. We started the product several, several years back, and then we took about two to three years to really insert it in the marketplace where we had evaluations at each of the customers. Really, the Gen 5 started to take off, especially when EUV was introduced, because the middle end and back end of the process really changed so significantly that the defect modes that customers needed to find were really focused towards the wavelength that Gen 5 has. That's why we saw huge growth in optical inspection. We see the same trend happening now for Gen 4. Even though Gen 4 sounds like an older system, these are two concurrent products. Four and five are always going at the same time. We see that when gate-all-around comes in, Gen 4, which is really suitable for that wavelength to go after those types of defects, is going to see a huge growth. That's kind of the way I look at it. To your second question, areas where we are investing quite a bit and delivering is, there is the E-beam portfolio. We introduced a new product called eSL10. That was an important product for us. Our thinking behind introducing E-beam is really how do we take this high-resolution inspector and couple it with Gen 4 and Gen 5 to really bring a faster time to results to customers. We don't see any use case where an E-beam technology can take really full layers of inspection. The reason for that is because there's no technology today that is present that can do inspection at those speeds. The coupling of the technologies is really important, so we are investing quite a bit in that. We're investing quite a bit in data sciences to enable this collaboration so that that is seamless to our customer. They just buy two systems, and they get a better result. We're investing in that. We're investing in X-ray technology because as you know, both in NAND and DRAM and now in the future in logic, there is going to be 3D type measurements. We're investing in those areas. Those are some of the ones that come to my mind. Mask requalification is another really important segment. We're bringing in our next generation system, reticle inspection system that is going to be targeted towards 3 nm, for mask qualification. That's going to be a really important product that comes out and serves that market. Again, it's really about the portfolio approach, and it's also about bringing the product at the right time. We don't bring products well ahead of the time because the customers don't really need it, and the cost of ownership is really not going to make sense. It's really timing the market, investing the R&D dollars really well to bring the solution when the customer needs. We get that feedback from the customer because we have a very large portfolio, so the engagement from R&D to production really enables us to understand when the product is needed, what performance is needed, and how do we do a great collaboration. Toshiya, I think I got a couple of other points on just the portfolio approach and what I think it enables from just an overall competitive advantage point of view is first, is that having products across the different technologies, across all the different either metrology or inspection challenges that our customers are dealing with, enables us to get information using some of the tools that Ahmad talked about that helps our customers identify problems in certain areas that might be caused by systematic issues in other parts of the process. Overlay metrology is an example where a lot of litho challenges in overlay come from non-litho process. Point product competitors don't necessarily have that visibility, if you will, to those challenges, that creates an opportunity. The other is that because you have these different technologies that you allow customers to pick and choose depending on where they are in their process control and sampling strategy and also the maturation of whatever their process is, to pick and choose different technologies to optimize their production at the best economic point. That's another advantage that we offer from very high-end, very sensitive tools from development aspects of what they're doing, but also enabling production monitoring where they know what they're looking for, and they go for tools that have a little less sensitivity but more speed. We're able to allow them to optimize across the entire portfolio and meet their cost of ownership requirements, which ultimately is an economic decision that they're making to try to drive productivity out of these fabs. Great. Thanks for that. As a quick follow-up, I just wanted to ask a question on optical versus E-beam. Ahmad, to your earlier comment, maybe it's not optical versus E-beam, they're complementary technologies, and maybe that's the point that you guys are trying to make. One of the knocks on KLA several years ago was very strong in optical. The E-beam solution wasn't quite there, and you had a couple of competitors who were kind of speaking to that point. Obviously, since then, things have changed. Optical is significantly better. You've shown your ability to continue to innovate, and you've got an E-beam offering. As a company, how are you thinking about that E-beam versus optical debate, and how you're able to address both markets? Yeah, I think it's a great question, I personally never see it as optical versus E-beam, simply because there is so much performance in optical that there's no amount of E-beam systems that you can bring into the market that can offload what an optical system does. I have seen customers evaluate that in the past and always come to the same conclusion. The onus is really on KLA, which is to innovate and solve the most difficult customer problems, with the best cost of ownership at the fastest time to results. Our customers cannot wait eight hours to 10 hours a day, doing a full die inspection. They can do a full wafer inspection in less than an hour. It's very hard math to do to be able to justify if it's one versus the other. The second thing I would say is that in optical, we have really shown year after year, if you look at the performance of our optical portfolio in 2018, 2019, 2020, 2021, we see growth in the optical portfolio, which really shows that our customers have evaluated these tools and evaluated the upgrades that we're doing to the systems and bringing the new systems in and showing that they are actually solving the EUV challenges and the new process challenges that they're facing. I'll give you one example. We have seen two customers use the latest generation systems and back port them for a previous node. A customer who has basically essentially said that the node has matured in its yield, we brought in new features and optical tools for the next generation node, and the customer looks at that and says, "Okay, how can this help in the previous node?" Implemented that also. We see very good performance optically in the marketplace, and I continue to feel that this is for many nodes to continue. I mentioned a few minutes ago that with the nanosheet introduction, the front end is going to change, and we believe that that's going to drive Gen 4 growth because it has a slightly longer wavelength that is very friendly to nanosheet architecture. That doesn't mean that we don't want to invest in E-beam. I think E-beam has its unique signal that enables us to bring certain types of inspections to the table in R&D that we were not able to do. Again, I go back to the portfolio approach. I want to go to a customer and be able to solve all the problems so that they go to KLA, we solve all the problems, and then be able to offer them the best cost of ownership at the same time. That missing E-beam portfolio in the past did not enable me to do that. Now that we have the full portfolio, we're able to engage in R&D and look at all defect problems and then complement both optical and the E-beam where required. It's not always required, where required for very difficult layers. We enable those solutions now at the same time and drive those to HVM. Again, it's a really great investment in E-beam. We're going to continue to do that. We went with E-beam inspection. We are going to drive that platform to E-beam review, and then to other applications at the same time. Again, complementing all of our optical tools, defect inspection, metrology systems, all of those systems will get feedback from our E-beam systems to enable them even further. Got it. No, thank you for that. Shifting gears a little bit, Bren, wanted to talk a little bit about Orbotech. It's been about two years since you completed the acquisition. You've demonstrated healthy revenue growth. I think on your most recent call, you spoke to additional cost synergies. Seems like you've executed really well. But if you had to grade yourself, how would you grade yourself, and what's sort of left on the to-do list as it relates to Orbotech? Yeah, no. If you just go back to the investment thesis of why we thought that made sense for us, is we were looking for companies that have market-leading positions in markets that had good profitability structures, where technology was evolving over time, and had an opportunity to leverage some of the capabilities of KLA from a technology point of view. Also had a complex systems integration environment that we could leverage that, as well as an opportunity that we could leverage what we do in our service business. I think one of the things that we've seen as we move forward is we've been able to check the box there. We're really pleased with the product positioning, particularly in the specialty semiconductor part of the business, and in PCB, as the technology evolution in PCB is becoming very exciting in terms of just the tighter pitches on boards, the introduction of flex boards, new PCB technologies, and then the evolving sort of integration of the PCB into the heterogeneous package. All those opportunities, we think create good situations for us as we think about all the areas that our customers are pursuing to drive cost innovation capability. It's front-end process node shrinks for sure, but in this more and more opportunities, it's also creating some opportunities for us. On the specialty semiconductor side, you've had markets where they had a very good position, and those markets are inflecting. We think that the relationships that we have with customers can help enable a higher level of collaboration in those markets. We think that that fits. Their exposure to packaging is also significant. Between PCB, what we were already doing in KLA, plus, the SPTS or specialty semiconductor part of that, we've been able to put the channels together, and we're continuing to do that to enable just more collaboration, visibility, and engagement with customers. Sometimes small companies struggle with some of that just because of the fact that are you going to take a risk with a smaller company versus, KLA? We bring credibility, we bring a broader set of resources to help solve problems, and that's favorable. We always felt very good about our operating model and how we run companies, whether it's our strategic planning processes, whether it's our product development processes, resource planning. Ahmad talked about how we look at portfolios and manage resources across the company. Frankly, how we start with fully allocated P&Ls for every business, that every business stands on its own. As we put that into Orbotech, and we still have systems to go, which I think provides another leg of opportunity for us in this area. We've been able to satisfy the part of our thesis that we thought was true, that could we get more out of market-leading positions in terms of profitability? I'm less concerned about the ratios of margins. I'm much more focused on the operating leverage we can drive out of those kinds of positions, where we think we can get paid for incremental value in evolving technology markets. Yes, the synergy, classic synergy, but also just efficiency opportunity is in excess of what we expected. We talked on the call about a $50 million base target at announcement, and we're in excess of $80 million today from a net synergy point of view. As we go forward, I think there's probably another $10 million or more of opportunity out there related to the next step of systems integration, but also facilities, sometimes leases and things like that, consolidating facilities and depots and things like that can take some time. I feel very good about where we're at. I think there's more work to be done here. I think the next step for us is how do we introduce new products that take advantage of some of the things I talked about earlier, whether it's technology or deeper collaboration with customers, and so that's still to come. We haven't checked that box fully in terms of driving top-line opportunities. I'm encouraged about where we're at, and we always felt, given the deals we'd done in the past, that some of our operating model competencies could play out in M&A type transactions with the characteristics that I talked about. With Orbotech, I think it's proven at a much broader level for the company that we can go and drive that. We feel very good about the model and the implications it has on other opportunities that fit those kind of market attributes I mentioned. Great. Bren, we've got a couple of minutes left. Would love to get your thoughts on the services business and how you're thinking about capital allocation as well. On the services side, you had another very strong year. Your services business is a little unique vis-à-vis some of your peers in semi cap equipment. If you can kind of speak to that would be great. From a capital allocation standpoint, you've been very committed, and you've been returning a ton of cash to shareholders. How should we think about the balance between organic investments, capital return, and potentially M&A going forward? Okay. Yeah, no, it's a good question. On service, the strength of the demand side has driven more service opportunity for sure, where we're shipping more systems, and those systems have long shelf lives and long useful lives. As those tools come off of warranty, they create service opportunities, and they tend, most of them tend to go into contract at a very high percentage. 90%+ that end up in a service contract structure, after a warranty period. The utilization of the install base has been accelerating. I think customers are looking for getting more out of the install bases. We've seen more demand in those parts of the market. You've also seen changing reliability specs around certain products that are in that market. Automotive, for example, where customers are having to evolve how they think about process control to ensure better reliability, because recalls can be expensive. As semiconductor content is increasing, that creates an opportunity for that part of the business, not just to drive pure service, which is really when we disclose service, it's really a classic service revenue stream versus some of our peers, I know they include some upgrades, and there's other parts of that. It's a little bit harder to look at them from an apples-to-apples point of view. It drives that, but it also drives enhancements. The other aspect of our service business is that most of the revenue, 75% of it, is service contracts, so it's very subscription-like. That allows us to do a couple things. A, deliver to an entitlement that's expected by the customer, but also optimize the resources and asset commitments that we're making to support that contract. I think that drives a profitability profile on the service business that's attractive as well. We're seeing 9%- 11% type. That's our growth model for service. We're seeing that kind of growth right now, and that's our expectation moving forward, but also pretty good operating leverage on the overall business, a, from the contract dynamic, but also with consolidated customers, it creates opportunities for us to get better utilization of our own resources in those areas. I think Orbotech creates some opportunities. Orbotech's PCB business had a very good service model, and there's some opportunities for us to learn from that, where their contract penetration is even higher than ours. It gives you some sense of the importance of the imaging capability they have to those customers. We think that with other parts of it, whether it's in the flat panel business or even especially semiconductors, that there's more opportunities for us there. It's very important to the company because it's a nice, consistent, predictable cash flow stream, and you can have volatility around CapEx, but at the end of the day, customers are still going to continue to run their install bases, and that's going to drive a predictable revenue stream out of service. From a capital structure point of view, we've been, I think, pretty explicit that in our belief that capital needs to be deployed productively, and we either put it to work back in the business or through different M&A considerations. In some cases, some of our M&A is small make versus buy type things, and we've done a little bit of that over the last few years, or returning it to shareholders. We've got a balanced approach. We think that the dividend ought to. You don't have a dividend unless you're going to grow it, and our goal is to continue to grow it every year. We've grown it 11 years in a row. We govern it really with two ways, a payout ratio that we target right around 35% sort of over time. Some years it could be lower, some years it could be higher. That enables us to ensure that we can raise it every year. Then we're going to ultimately grow it with the growth in our free cash flow. Given the asset-light nature of our business, if we can grow our top-line model to 7%-9%, our bottom-line model is to drive 1.5x that revenue growth rate in terms of drop-through leverage, that we ought to be able to grow our dividend 10%-15% in a dividend payout. That's our model and how we think about that. Ultimately, the rest of the cash flow will be deployed in either share repurchase, and I think at a minimum, we're going to return about 70% in any given year. Other investments, like I talked about, notwithstanding. I think that's probably a low watermark, but that's going to be, I think, pretty consistent for us, that it's going to be at least 70%, and we're going to drive it the way that I described. Great. I could keep going, but we are out of time. Bren and Ahmad, thank you so much for joining us. Good luck with everything, hopefully we get to do this in person next year. That'd be great. Toshiya, thank you for having us. Thanks, guys. Thank you. Take care. Thank you.