Good morning and welcome to JPMorgan's 53rd Annual Technology, Media, and Communications Conference. My name is Harlan Sur. I'm the Semiconductor and Semiconductor Capital Equipment Analyst for the firm. Very pleased to have Bren Higgins, Executive Vice President and Chief Financial Officer at KLA, here with us today. I've asked Bren to start us off with KLA's view of the wafer equipment, process control, specialty semiconductor equipment spending environment dynamics, maybe this year, KLA's growth outlook within that, and then we'll go ahead and kick off the Q&A. Bren, thank you for joining us this morning.
Yeah, thank you for having me, Harlan, and great to see everyone here. I appreciate your interest in KLA. 2025 has been a good year for KLA. A lot of leading-edge investment that's happening in the overall WFE market. What's been driving our business is mostly, and it's great that after a couple of years where a lot of investment was more in legacy as it relates to some of the activity in China, but also just broadly in legacy markets, to see a real strong growth profile as it relates to the 2-nanometer node has been a nice driver in logic. Given all the activity around high-performance computing, high-bandwidth memory has also been an area of investment. All really being driven by the same thing in a lot of ways. That's been really good for the company as well.
We've also had this increasing exposure really across the portfolio to the advanced packaging market, again, driven by high-performance compute, but really a new and evolving growth opportunity for KLA. That's been the biggest driver in terms of the WFE outlook, which is up about mid-single digits from last year. We'll say in the low hundreds from about a little less than $100 billion last year. We also have a strong service business that has a very high contract penetration, and service is a business that grows every year. We've had some access issues as it relates to some of our China fabs and export controls here in the U.S. Despite that, we're also seeing some growth in service of about 10% overall. When you look at WFE, you look at what's happening in leading edge.
I think overall in China, it's digesting a bit. Our China business is down after a couple of years of more elevated investment as our larger customers were correcting after the post-COVID investments that were happening. The rest of the legacy market is more or less flat. Really a leading-edge profile driven mostly by high-performance compute. The service business is operating well. I talked about packaging. Across the portfolio, you mentioned specialty. Part of our packaging opportunity is process control, but it's also in some of the process tools we sell that's also been a nice driver. Gross margin profile, I talked at earnings against that WFE backdrop that then translates into some soft guidance around the year that takes KLA into the, we'll call it the low double digits in terms of growth year over year.
Gross margin model of about 62.5% and operating margins in line with our higher end of our leverage target of driving incremental operating margins across the company between 40% and 50%. Translating, I think, into the financial model, working pretty well and some growth drivers that we're pretty excited about. One of the things that we've said a lot lately, you know, as we've seen this evolution over the last couple of years and this change in terms of the growth drivers for semi-revenue is that what's happening in high-performance compute, we couldn't really pick a better environment for a company like KLA. I'm sure we'll talk about why that is as we go forward here today, but I think that sort of sets things up for the rest of our discussion.
Yeah, no, let's start off at sort of a high level, right? You know, wafer equipment spending intensity overall bottomed in 2013 at 9%. It increased to 16%-17% last year as a percent of overall total overall semiconductor industry revenues. Another way to look at it is that over the past 11 years, wafer equipment WFE spend has grown at an 11% CAGR. Process control segment of the market has grown at a 12% CAGR, and KLA has grown its process control systems revenues at a 13% CAGR, right? Clear, strong outperformance, right? You know, a lot of the drivers, you know, complexity challenges, more focus on yield, more focus on cost, the need for more installed capacity to support some of the strong demand trends out there. Do you anticipate process control and KLA continuing to outgrow the wafer equipment spending dynamics?
What are going to be some of the major drivers as you look out over the next, call it, two to three years?
There's a lot to unpack there. I think that the first thing, it starts with just general capital intensity or WFE intensity for the industry. For a long time, that was a decline in WFE intensity for the industry, created an environment where we had a lot of cyclicality and very little growth. About the middle of the last decade or so, we saw all the drivers of lower capital intensity start to play out, and those were things like wafer transitions and consolidation in the industry. From that point forward, we've seen WFE intensity start to rise. We've had a continuous rise here where it's actually now been growing. You've got now structural growth of WFE that didn't exist before. You have semiconductor revenue growing, you know, over the long run, we'll call it, you know, 7% or 8%.
You have got WFE where intensity is slowly rising, growing a little bit faster than that. From about 2019 forward, the introduction of EUV happened, which brought scaling back to the industry. For a company like KLA, change and scaling is a big driver. You went for a period of time, there was not a lot of scaling, there were not a lot of end designs that were adopting new nodes. As a result of that, the level of risk that our customers were taking was fairly low. The Moore's Law attributes were not all that compelling for follow-on designs to come through after the leaders moved and the higher volume designs moved through. Once EUV came to the industry, scaling started again. We saw this dramatic increase in the number of designs.
You saw growth in WFE in the logic part of the industry, which was a negative CAGR from 2012 to 2018 or so. For process control, logic devices tend to have characteristics that drive higher level of what we call process control intensity. Process control is a percent of WFE. When the mix shifts to WFE that's logic-based, then that's favorable for our part of the industry. You had this environment where semi-revenue was growing, WFE was growing a little bit faster because of the mix shift and expectations for growth in logic foundry WFE that it creates this environment where process control now grows faster than WFE. Within process control, KLA has a portfolio approach. We're exposed to a number of the markets within process control. The markets that are growing the fastest, we have our strongest share position.
Process control is growing faster and then KLA is growing faster and gaining its share within process control. That has been the environment. We think that continues because as you see this shift to high-performance compute in terms of the composition of semi-revenue, that brings forward a bunch of new challenges and risks to our customers. They're starting to make devices that are quite large. When they're big, there are defect density challenges. There is a technology roadmap. You have EUV, but you also have now architecture changes. When you have a lot at risk, then you're able as a customer to invest more to ensure that what you're producing is working. It's working from a yield point of view. It's also working from a process or a variability point of view. Those are great for process control dynamics.
As a result of our position, we've been able to benefit from that. You take into what's also happening with memory and what's happening with DRAM and high-bandwidth memory and the challenges in building those devices, which I'm sure we'll talk about. The integration of the high-performance compute logic chip with the memory is creating this opportunity around packaging where there's still a lot of risk for our customers after they finish the front-end part of processing. They go to singulate the wafer, and then they still have a lot of processing to do to get that device to perform as it should. They do not want to do rework, particularly given the demand requirements that they have. All that's been really good with our relative performance.
As a share of WFE, we went from the low sixes to this year of WFE, and everybody adds it up differently, but we're approaching 8%. Over that timeframe, you know, really one of a select few companies who have really grown their share of the overall market. Our relevance within the wafer fab equipment market has increased pretty significantly.
Perfect. Why don't we bring it down to this year, calendar 2025? You know, we estimate that your process control franchise will outperform WFE again this year, right? As you mentioned, up low double digits versus WFE at up mid-single digits. We know that your growth outperformance has been a combination of share gains, as you mentioned, process control intensity increases. This year's demand profile appears to be advanced foundry logic, DRAM, advanced packaging. If KLA does outgrow WFE this year, as we are modeling, is it going to be more process control intensity? Is it going to be a mix of who's spending and who's not spending? Is it going to be share gains or some combination of all of that?
For the reasons I talked about, intensity is higher. That is certainly a factor. There are certain markets, though, where we feel pretty comfortable about opportunities for share gain. We talked a little bit about it at earnings with our presence in some of the electron beam markets and opportunities supporting electron beams that not only electron beam inspection, which creates opportunities for its market itself, which is more of a defect discovery and debugging market, but the ability to deploy AI in that system to be able to train and point our inspectors, which is a very strong market share position for the company. That is a share opportunity we are excited about. In advanced packaging, we think within process control, we are sort of right at more of a parity position with the leader today.
We would expect that as we move through 2025, that that's also a share opportunity. I think rising process control relevance and then some share opportunities in these markets. I think the final thing is if you look at our flagship products in the company, our Gen 4, Gen 5 optical pattern inspectors have had very strong demand even over the last few years where things slowed down, where I did not have the supply to meet the demand. As new capacity has come online, I've been able to ship more of those products. I would expect that part of the business to also grow faster than overall WFE. That is a nice tailwind as well.
Probably we'll have a share effect, not that I'm gaining share because I already have a strong share, but the relative growth rate of that market within overall process control will have a share impact. A number of areas where we're pretty bullish about the opportunities that are in front of us. We've got to execute, but we're doing well with the customers that are investing today. We see some sustainability in that.
As you step into the second half of this year, it does feel like all of what we just talked about continues to be in full effect, right? You know, advanced node momentum will continue to accelerate both in foundry logic memory. That does appear to be reflected in your confidence on this sort of $3 billion-ish plus sort of stable spending outlook for this calendar year. Additionally, leading-edge design starts, which is something that we track at 3- nanometer and 2- nanometer, those design starts are actually accelerating. Given your relatively longer lead times, your critical role in enabling these advanced technology migrations, how are customer discussions, initial forecast visibility, and outlook for calendar 2026 sort of WFE shaping up for the team?
Yeah, so for our leading customer, for sure, their supply is below the demand levels that they're facing today. The acceleration in designs and the design transition cadence is also, we think, a positive for the business over time. It does give us some confidence as we look at next year that the sustainability around this into investment and expectation of the convergence of a higher number of designs at the sort of the first stage of a node ramp to be higher than what we saw at 3 nanometers. The 2- nanometer node is setting up to be a pretty widely adopted node, and our customers are investing a lot to support it. You know, I think as we look at the second half of this year, as I said, you know, we're operating in and around this $3 billion level.
We guided $3.075 billion for the June quarter. We'll see how that plays through as we look at next year. As I sit today and look at not only what we're seeing, and of course, there's a macro backdrop that no one really knows, you know, second order effects of whether there could be changes to demand profiles or not. As we sit today and look at what our customers are telling us, what we're seeing in terms of investment this year, and you look at what I think might be a broader, more leading-edge profile next year, continuation in memory flash, which has been fairly weak for some time, that likely has some growth opportunity, that it does set up for supporting 2026 as a growth year for the industry, absent some macro effect on broader demand.
I think it could set up for much stronger growth if we start to see some growth in the mobile and PC markets as people replenish to support AI applications. You know, that would be, I think, a tailwind from how we're currently seeing it. That is certainly a factor out there that as those markets, which consume a lot of semiconductors, if we got some inflection there, then that would be an incremental positive.
You did talk about trade and tariffs. As we enter this period of uncertainty from trade and tariff-related dynamics, you know, the team noted back at earnings that you're not seeing any major changes to your customer spending plans. You have acknowledged that there is a high level, there are high levels of uncertainty. Maybe walk us through some of the potential mitigating factors as it relates to your manufacturing footprint and your value chain.
We did talk about a tariff impact of roughly 100 basis points of impact. We put together a tiger team across the company. Really, you know, once a lot of this news came out, really, you know, combed through the business to really understand some of the tariff dynamics. Tariffs have not been a big issue for us in terms of how we manage the company historically. We feel pretty good about that assessment. I think with what we heard this weekend, and again, very fluid, things could change. I would say that is probably a worst-case scenario now. We are optimistic that it does not change, but we will see. I think that is an environment where, you know, we are looking at that. Then we are looking at, you know, how we move parts around the world. Historically, like I said, tariffs have not been a big factor.
We had a design around parts and parts movement process supporting our services business. For example, we have a manufacturing footprint in California. You know, we're looking at how we can do things a little bit differently. Now it has to come with an operational motive because, as we saw this weekend, things can change very quickly. You do not want to suboptimize something operationally for what is a regulatory change. We're looking at all that. Certainly, if you assume you're operating in a higher tariff environment going forward, and I think we likely are, then there's probably some things we can do in terms of how we move parts around that will be mitigating factors to that exposure. That's really how we're thinking about it. We have the footprint in the U.S. We export most of what we ship.
Certainly, if you export things, then, you know, you round trip and have the ability to reclaim any tariff exposure you might have. You know, what stays in the U.S. structurally, if it has a cost increase, we either have to figure out a way to mitigate it or figure out how do we pass that along and work with our customers as we think about pricing strategies. We're assessing all of this. I think for now, we feel pretty good about where we're at. Some of the exposure relates to our service business. You know, we have this unique service business in the industry that is really a contract stream. We sell service contracts to customers. What we sell is we sell performance. We sell availability of the system.
As a result of that, the customer does pay us a service contract and says, "Keep this tool up, have it meet its performance requirements and its matching requirements across other tools in the fleet." Whatever's required to enable that, we manage. It provides a nice predictable stream of revenue. We can optimize cost structure underneath, but we don't necessarily, it isn't about just replacing parts. It's about maintaining the system and driving its performance. If we have to swap a part, we swap a part. Something breaks, it breaks. We control that. As a result of that, we are importing that into some of the countries. Potentially, you know, certainly as it related to China, we had some exposure to incremental tariffs.
Now, there were some exemptions that were also sort of playing through that was minimizing that. It gives, but as you look just even broadly across the world, how we move parts around that structure, which has a tremendous amount of value to KLA, both from a revenue predictability point of view and a profitability point of view, does create potentially some exposure in a higher tariff environment. We have to go and look at, okay, what can we do differently as it relates to that. Overall, I think I feel pretty good about the assessment. I think the team did a great job. Like I said, I think it's probably more of a worst-case assessment today based on what we know than what we thought a few weeks ago.
There you're talking about the 100 basis points headwind. Okay, that's good news. Before I move over to some of the product and market share dynamics, any questions from the audience? If you have any questions, raise your hand. We've got a question up here. Let's get the mic over.
Hi, this is Jordan Krawitz from Mark Asset Management. Thanks so much for taking the time and taking my question. You made a comment about deploying AI on some e-beam inspection systems. I wanted to ask more broadly, when can we expect more sophisticated ML or transformer-based AI to be embedded across the broader portfolio? What sort of quantifiable benefit could that bring to your customers?
Yeah, no, it's a good question. I focused on the electron beam systems because that's probably our most advanced system as it relates to deploying AI. We've had AI in our systems, physics-based AI in our higher-end systems for going on about a decade today. What's happened, obviously, is the cost to compute has come down. We've been able to change the architecture of our systems to a GPU-based architecture, which has had cost benefit, but also enabled more robust algorithms. We can now use it to do, I mean, the biggest challenge in our business, and particularly in shrinking design rule environment, is signal to noise. Finding signals among all the things that you can imagine you're finding with the capability of these systems and the size of the defectivity.
Being able to filter that to point the systems is how we've been leveraging this capability. We've been able to push it to all the systems. It was an investment for us to redesign and do the software to enable the GPU-based architecture. Now that we've done that, we not only are getting cost benefits, which is helping us, obviously, but also creating a much more robust environment around the algorithm work we can do. We really are a user and enabler of it. I think the opportunities are pretty significant. The other thing is that we can extend the hardware.
Historically, where you might have needed to change your optics or add more light in your sources to be able to do more advanced inspections, because we can do better analysis, if you will, by using this capability, we can extend the hardware, which obviously has an effect on the R&D intensity of the company. It is one of the factors why I feel pretty good that our R&D intensity, which is 12%-13% today, has come down from, you know, as the company has grown, it has come down from the high teens. I think, I do not know how much more it, I do not think it goes down, because I support a very broad platform and differentiation is fundamental to our model. It has been, I think it will allow me to continue to drive R&D intensity on the platforms and extend them over time.
That's another positive as well.
Any other questions out there? Right there.
Can you hear me? Oh, thank you. Thank you for taking my question. Just if you could, a quick update on the EUV inspection, mask inspection tools.
In the market for EUV lithography, there at some point in the industry, as the feature sizes, the dimensions on the masks or the reticles shrinks, you will need a new capability, which is this, what they call actinic inspection, which is really actinic to the wavelength of the scanner inspection, which is an inspector that's inspecting at the same wavelength. Today, the needs of the industry, given the feature sizes that are being printed, is leveraging optical systems. Most of your EUV reticles in production are running through KLA systems, greater than probably 90% or so. Over time, particularly as you move to more advanced lithography and those feature sizes shrink, then you will need more capability. The industry isn't there yet.
Given the optical systems meet our customer sort of trade-off between price and performance and economics, it's driving them to the decisions they're making, you know, as the industry moves forward. At some point, you'll need that capability. We're investing in that capability. We have a strategic partnership with one of our suppliers, Carl Zeiss, on that, which will allow us to bring a product to market, which we're comfortable with in terms of timing as we move into the later part of this decade. We'll also be able to leverage the synergistic benefit of aligning, you know, the optics and any changes in the optics with a litho roadmap to actually be able to reflect those changes very quickly in the scanner. We think it sort of drives extent, or in the inspector, so it drives the extendability of that market.
Investment's happening. We think the market demand in terms of production is likely in the later part of this decade. And so, we're investing in the capability.
Let's talk about your strong market share performance in calendar 2024, right? WFE market share numbers recently released by Gartner. Process control segment of the market outperformed WFE spending. KLA outgrew the process control market. You gained 30 basis points of share, six and a half times larger than your number two competitor in the segment. You are a strong number one share leader in five out of the six major subsegments in process control. You gained or held share in four segments. You lost a little bit of share in two segments, mask inspection and overlay. As you look at the dynamics here, I mean, does the team believe, number one, it's going to gain share and in what segments?
Yeah, so as we talked a little bit earlier, you know, I feel pretty excited about the opportunities that are in the electron beam markets, both inspection and review. I think what's happening in and around advanced packaging is a positive. The inflection of optical inspection is also something that will influence the overall share math. In reticle inspection, I think that part of the market, which can be a little bit lumpy because the integers can be quite large, given the cost of the systems, I would expect that business to grow faster than the overall market. We'll see how that translates in terms of share. Certainly, there is some momentum in some aspects of that market that I feel decent about. I think it's a good construct for share opportunities in 2025 for KLA.
You know, the one area, and you've mentioned this several times already, where you made significant progress from a share perspective was in e-beam pattern wafer inspection. And despite optical inspection, where you guys have 90% plus share, optical continues to dominate, you know, seven times larger market versus e-beam. But the KLA team doubled their e-beam inspection revenues last year. You gained 700 basis points of share. I think this is actually a great example of the KLA team's strategy, which has always been to introduce new solutions when the market is ready to adopt, right? It does feel like the market is ready to adopt e-beam. What's driving the incremental opportunity? Why is e-beam needed now? How is KLA's e-beam platform differentiated?
It's a market we've invested in for some time. We were in the market, then we were out of the market. It gets back to KLA's core fundamental principle of looking for opportunities where we can differentiate. We can understand the value that we add. We can share in that value with customers. We can have a position in a market that's big enough to sustain the investment requirements that are required to participate in it. The ability to drive relevancy of optical inspection has been a primary driver, but there's also very specific use cases, particularly with Gate- All- Around, that's driving opportunities for direct e-beam inspection market growth and opportunities for KLA. We have multiple products in the market. We have a single beam system that's for small physical defects.
We have a multi-beam system in the market today that does more electrical testing where you can get electrical signature. You lose some sensitivity relative to the single beam, but you're doing a different set of applications. We think that the drivers for opportunity are twofold. First, these markets and how e-beam grows over time, and then how it supports our optical market. Electron beam systems have always struggled getting into production in a meaningful way because of the speed. Very, very sensitive systems, but there's always a trade-off. We like to say there's no free lunch in physics, right? If you have significant sensitivity, you're going to have a slower system. There's always a trade-off between speed and capability. You've always used optical. You debug your process with electron beam. You deploy it here and there where you need it. Use optical in production.
The ratio of the total aggregate spending for pattern inspection is generally about 80-20, optical versus e-beam. That ratio can move a little bit, plus or minus a few points, but does not, frankly, move very much if you go back over the years. Now, because it is a faster growing market within process control, it means that both parts of the market, because I do not think the ratio changes, are also going to grow. You can end up in these periods where you cannot get contrast in certain defect types, where that e-beam system can identify that contrast and then point your inspectors and train your inspectors to look for those types of defects as you scan the entire wafer, because you are looking for defect signatures across the wafer.
There is a lot of attributes that being able to have that interoperability across the two capabilities that are positive for us, but also some unique characteristics as customers drive their technology roadmap that is driving opportunities in that part of the market as well. We think there is some good momentum there, and we think it will be reflected in the 2025 results.
Yeah, I was thinking about the gentleman's question earlier and about AI and machine learning. I was recently talking with a bunch of buddies of mine that do process control and yield improvement in some of the big mega fabs. They always brag that, you know, their platforms, their KLA platforms are the only platforms within a mega fab that have NVIDIA GPUs embedded into them. I mean, it's literally like a supercomputer, a mini supercomputer. Your optical inspection systems are almost like a mini supercomputer because of all of the analytics and image processing that has to be done. I thought that was pretty interesting.
You have sensors, cameras that run at 20-30 GP/s .
Right, that's right,
Which is the equivalent of, you know, about 2,500 iPhones taking pictures simultaneously, right? Managing that amount of data and being able to process it and go from finding hundreds of millions of defects literally to the 50 that matter and do it all within a one-hour inspection time across an entire wafer requires significant compute, which is why the changes in the compute architecture were really important for us from a cost point of view, not only to enable the robustness to support these applications, but also to manage the cost. Had we stayed on a more CPU-based architecture, we would have had, you know, the costs, you know, which has been the fastest growing part of my cogs or my build of materials has been in image computing.
Being able to change that ratio by being able to leverage the robustness of this capability has had a cost benefit, but then also obviously supports the long-term roadmaps.
Advanced packaging. $500 million in advanced packaging revenues last year. Team is set to drive $850 million this year, up 70%. I assume a combination of HBM, 2.5D, 3D advanced packaging. We know many of your next-gen 2-nanometer AI XPU customers that are designing some of the most sophisticated AI ASICs in the world are targeting the next-generation 3.5D packaging, stacked chips, more HBM stacks, overall significantly more complex versus 2.5D packaging. Help us understand where you're seeing incremental growth opportunities. How does this segment break out? Process control solutions versus your semiconductor manufacturing solutions.
Yeah, it's been an interesting market, and it's a market that has really moved to KLA. If you just look at overall advanced packaging, it was probably back in, you know, 2021 timeframe, a $3 billion-$4 billion market. Today, it's about a $10 billion market. And KLA's exposure within that market has effectively doubled over that timeframe as a % of the overall. Now, we come at it with a portfolio approach, not only of our process control tools, but we have specialty semiconductor tools. We do some chemical process control also, which is a slightly different market. We've got a number of products that are addressing these dynamics. The integration of these GPUs with a high-bandwidth memory is driving more complexity in the interposer package. And, you know, the customer has to make sure that those connections are viable and they're working.
The density of those is shrinking in terms of the lines and spaces, which is driving the need for more capability. The most momentum we have seen has been mostly in and around the logic opportunities. Memory, we're also seeing increasing momentum as some of the things you mentioned in terms of what's happening from a roadmap point of view as you look at, you know, shrinking of micro bonds and eventual move to hybrid bonding and so on, that's also creating opportunity. We're making progress there. Intensity has absolutely been growing, but what we've seen and what has driven those numbers that I talk about that it feels like every quarter I'm like, well, it was $500 last year, and then it's $750, and then it's $800. Now I said $850. It's because of some of the share momentum we've seen.
We think it continues because we have today the entire portfolio to support the sensitivity requirements in products that are already developed for KLA. The front end has moved to the back end. As a result of that, we have the scaling roadmap that can support the need for more capabilities. There will be changes in moving from wafers to panels and so on. There is some engineering around how you handle the base substrate. We have the ability, and I think over time as customers need more capability where they are running with very high sampling rates because the last thing you want to do is have to rework these packages. You want to test them to make sure before you start that processing. All that we think is good for our business moving forward. High sampling rates, the need for more capability.
If you're going to maintain sampling rates and you have more capability, the tools are likely going to run slower, so you're going to need more. I think there's an ASP element, which will translate to margin, but also potentially a volume element. This part of the market, I think, continues to grow. I think it grows faster probably over the next several years than the WFE market. It's a whole new sand of opportunity for a company like KLA. Traditional WFE, good drivers there, now advanced packaging and a share of a market opportunity there that, frankly, five years ago didn't really exist for us. We're pretty excited about it.
Perfect setup. Thank you, Bren. Appreciate the participation and look forward to the strong growth outlook from an execution by the team this year. Thank you.
Thank you, Harlan. Thanks for having us.