Okay, good morning, everybody. Welcome to the Goldman Sachs Community Utility and Technology Conference. My name is Jim Schneider. I'm the Semiconductor Analyst here at Goldman Sachs. It's my pleasure to welcome KLA, CFO Brent Higgins, and President of Semiconductor Products and Solutions, Ahmad Khan. I just want to make sure I got that right. Welcome, guys. Thanks for being here.
Thanks for having us.
Thank you.
Brent, maybe to kick off, I think you might want to give us a brief kind of recap on the quarter and the business outlook as you see it as we start off.
Okay, and first, again, thank you for having us. Happy to have Ahmad here. Ahmad runs our systems businesses in the company and also runs our channel, so can bring some good insight to product roadmaps and customer dynamics. Happy to have him here. I think it's a treat to have him here in front of investors today. Just to, I think, level set on some of the themes from earnings, obviously the June quarter was a good quarter for the company. We had a raise into the September quarter. 2025 is shaping up more or less as we had expected and we articulated at the beginning of the year. We've seen some incremental improvement overall reflected in the results relative to guidance, but in general, the trends have been pretty consistent. The logic environment's been very compelling in terms of investment at the leading edge.
The 2 nm nodes have been great for KLA in terms of process control intensity, and we're seeing investment that's happening there. High bandwidth memory is creating lots of opportunities, and it's sure something we'll talk about in terms of rising process control intensity in DRAM. The nature of not just the devices themselves, but the performance requirements of the market are also pretty compelling for KLA. The packaging market, which is an interesting market, you just go back just a few years ago, was a few billion dollars kind of in the margin for error of WFE, and today somewhere, at least as we define advanced packaging, somewhere in that $10 to $11 billion level. We've seen that market, as complexity has increased, the opportunity for us to differentiate has become a much more compelling opportunity. We think is an increasingly compelling opportunity over time.
That's been a nice inflection in our business as well as that market has moved to us. We raised our guidance. Last year, we did a little over $500 million in terms of advanced packaging revenue, and each quarter we have continued to raise it this year, both on the strength of intensity with high sampling rates from customer, but also share opportunities. Last quarter, we talked about $925 million of opportunity this year. We'll see how we ultimately end up, but certainly a lot of momentum in that part of the space. Service, despite some of the challenges around the U.S. export controls denying access to certain fabs, our service business is still performing in double digits. I think that's a great trend that we're seeing. I think we can talk more about service later in terms of the drivers and why it's a unique business.
We have a long-term target of 12% to 14%. Slightly below that now, but given some of these constraints, I think we feel pretty good about that and our ability to continue and grow at that level over time over the next several years. Financial model is what you would come to expect from KLA. Gross margins at 62.5% for the year was part of our guidance. That's inclusive of some tariff impact that we're spending a lot of time on trying to figure out how to mitigate. We've got a number of things going on in the company, but right now we see it as at a 50 to 100 basis points of impact. Overall operating margins ahead of our long-term target model of 40% to 50% incremental operating margins on revenue growth, and we're above the top end of our model for this year.
The execution is very good. I think, from a priority point of view in the company, supporting our customers is always a priority. We're pretty excited about what's happening at the leading edge, some of the dynamics I talked about earlier in terms of opportunity. That obviously puts responsibility on us to ensure that we're collaborating, to ensure that we can support. We've got a number of really critical programs in the company in terms of investments that we need to make to ensure that we've got the products that support the roadmap. Innovation is the lifeblood of KLA, allows us to drive the differentiation that drives that business model that I talked about.
Finally, I think that, you know, we like everyone else are looking for opportunities to continue to drive leverage through our business model, to take advantage of the opportunities out there from a productivity point of view, the tools that are now emerging that will help us bend the cost curves in our non-technical areas, but even in technical areas as we get more efficient with code and so on. We think that those are the sort of priorities that we have in the nearer term. We're pretty excited about the dynamics that the growing segment of the market around high-performance computing, as that increases as a percent of the total, it creates a lot of unique opportunities, we think, for process control and for KLA. We're pretty excited about what that future holds. Why don't we stop there and go on to your question?
Yeah, perfect. You know, relative to process control intensity, can you maybe remind us about what share of WFE you were initially aiming for in 2026 from your prior analyst day, where you're currently tracking, and what's the added upside as you think about your position in advanced packaging that you just mentioned?
Sure, and I'll let Ahmad weigh in on some of the context behind it, but we laid out a plan at an investor day in 2022 to get to 7.25% plus process control or KLA share of WFE. If you look at the results this year, we think we're trending, you know, closer to 8% this year, so we're meaningfully ahead of that target model. Now, if you back up to where we were in 2021, 2022, that was down in the low 6% range. Everybody adds up the denominator a little differently, but we've seen a meaningful change in share of the overall market over the last few years. To the drivers of that, I want to let Ahmad weigh in.
Yeah, so the AI inflection is really tilted towards process control. We are benefiting quite a bit from this inflection, and one of the core reasons is AI has driven both advanced logic, but also advanced DRAM. In that, one of the inflections that we see that affects KLA positively is that the die size of individual chips is increasing at a very big rate, right? The die sizes were very small in the past, and now the die sizes are increasing. This die size increase inflection is happening at mature nodes at the 5 nm and 4 nm and 3 nm, but also happening at 2 nm advanced logic. Just imagine if you step back and you had a wafer and you had two dies on it only, and you had one killer defect that killed the die, your yield would be 50%.
If you had a wafer that had 10 dies and you had one killer defect, your yield would be 90%. As die sizes increase, you may not see litho intensity or deposition intensity because the wafer size is fixed, but process control intensity goes up because of the criticality of die size. We call that D0. We see that in logic and also in DRAM. Secondly, in DRAM or in HBM, traditionally you had only DDR devices, and now you have HBM. In HBM, you have a stacked device, a stacked device that starts with a logic-based die to control the rest of the circuitry. That is, you know, it's between 7 nm to 10 nm node logic that's going to go eventually to 3 nm. That is logic intensity.
The second is that you have stack dies on top of it, and in this stack die, you have a lot of TSVs and you have a lot of logic circuitry, which is different than the traditional DDR, which has repeating patterns. Because of this, process control intensity is going up. Now, as you're stacking these dies, the value of the die goes up as you stack number three and number four and number five, and therefore customers want to make sure that process control is done, ensuring that the previous stack is not wasted. All that happens, and then you have to put that together in a CoOS fab. You have a logic die, which comes from a front-end foundry. If you have a loss of that die while you're doing the integration on CoOS, you can get another die.
Of course, it costs you money, but if you lose HBMs, you have to write your customers a check. All that integration drives process control intensity also in packaging. We see process control intensity going up in logic, primarily because die size is going up and complexity is going up. We see it in HBM manufacturing because the logic content is high and stacking content is high, and we see intensity going up in packaging because you're taking very expensive devices and you're integrating them together. The ROI of doing that inspection is far better than losing the die. We see these trends all being positive. Longer term, as you know, if you looked at the number of transistors that were available for a GPU about five years ago, close to about 50 to 80 billion transistors, today it's full-field reticles doing 100 to 200 billion transistors.
Die sizes are increasing, the number of transistors are increasing. We see the same trend in memory, more stacks of HBM and larger dies, and this trend will continue as well.
One of the other things that's happening has really happened in the industry since about 7 nm introduction. We go back to 2019, 2020. There was a period of time where the number of designs at the leading edge was declining. You had a period of time where there was a slowing in the scaling roadmap, and that was a factor. You just didn't have, it wasn't as compelling to necessarily move to leading edge for a lot of different designs. Also, the cost of design was very high. The leading edge roadmap, as it became more compelling from a device performance point of view, was one thing, but also as the cost of design has come down, really enabled by the broader industry, enabled by TSMC, it's created this environment where you have this proliferation of designs.
That benefits KLA uniquely because managing a high mix in a fab is really challenging for our customers. You have lots of designs, lots of different process flows. You have the sort of capital optimization of you don't want to start too many. You don't want to start too few in terms of wafers. You have to deliver to a tight market window. All that really benefits KLA in a unique way. The other thing is, as we're seeing today, the N2 node has 15 or so customers doing designs at N2. You've got 10-ish high-performance compute customers doing designs. It's driving pretty rigorous performance requirements on that front. You also then have a very highly utilized N3 node. The ability for our customers to migrate capacity, try to reuse some of their capacity is more constrained.
We think the N2 node will probably be the largest over the first three years or so, node from an overall design point of view. It's pretty compelling overall with some of these other challenges from a technical aspect that the design environment creates a unique opportunity for the company as well.
Yeah, fair enough. Okay, you talked on advanced packaging before. I touched on that. Maybe how should we think about your exposure to advanced packaging and your share of that market? More broadly, can you maybe go into detail on the 30% of your revenue that's not tied to process control?
Oh, the non-process control parts. Okay. All right, so advanced packaging, I talked about some of the inflections that we're seeing overall. We can talk to some of the overall challenges and the complexity and the shrinking pitches, the density of the lines and spaces in the package, and the complexity of that, and the risk that the customer takes to the value of the package itself. You think about what a GPU costs integrated with HBM. You're seeing very high sampling rates in that part of the market. There's the need for more complex tools, and we've been able to see the backend kind of evolve in a way where the frontend capability is now required. We see our customers managing it very much like they manage the frontend requirements. That's been a nice inflection. We've seen share increase in that part of the market.
On the memory side, it's evolving a little bit differently, but we also see some momentum that's happening there to some of the dynamics that Ahmad Khan talked about earlier. I don't think packaging is going to continue to grow like it has in the last couple of years. We do think over time that given this die size challenge and the need to integrate these packages and the performance requirements of these packages, the sampling will likely stay pretty high. I don't know if it'll stay as high as it is today, but it'll stay pretty high. The complexity roadmap is increasing, and our customers will start to need more advanced capability from KLA. Over the next several years, we think it likely grows faster than wafer fab equipment (WFE) overall.
In the parts of the business that aren't process control, we have our process business, which is specialty semiconductor business. A big chunk of its exposure is in advanced packaging. It also has exposure in specialty MEMS and those parts of the market. It's a $500 to $600 million business for KLA in CNR segment reporting and has uniquely designed for some of these markets. We found that pretty attractive when we acquired Orbotech, as they've been able to design around specific requirements in those markets. Those markets have now inflected and so has created a pretty differentiating business from a financial point of view. We think we can leverage the relationships. We've absolutely been able to do that with that business and probably at a level that they were unable to do prior to the acquisition.
It's proving that the operating model of KLA can bring value in markets beyond process control. We're excited about that. There's also our PCB business, which has been under some pressure from a capacity point of view as it relates to some of the weakness in mobile markets and the lack of investment. We're starting to see some signs of light there. We'll see how that progresses over time.
Very good. Yeah.
Just to build on packaging, the main inflection that has happened in the marketplace is that you used to make a device in the front-end logic and then you did a fan-out package of that device and then you put it on a PCB. That world has changed, right? Because now it's really a system-on-silicon. That inflection is really what is driving the packaging intensity. You don't take a final wafer, cut it and put it in a package and put it on a PCB. You start building from there onwards with memory, with other things. This inflection will continue. That's why the packaging inflection is continuing. The cost of these dies are very high because repair on PCB was simpler. You would take the final package, put it on PCB, you can repair it in the event that that integration didn't go well.
The repair on a CoOS wafer is very, very difficult to do. Therefore, intensity of process control is going up because the value of the package is high and the difficulty of repair is high.
Great. That's a great perspective. Maybe shifting to WFE, just give us a sense about where you expect WFE to land in 2025, maybe your initial view on 2026, and some color on moving parts within that.
I'll just echo some of the thoughts from earnings. We did provide a little bit of context. I think 2025, our view is pretty consistent that we see growth rates in kind of the mid-single digits from last year. You do have this packaging element that's a little bit, I think, distorting some of the views of where WFE really is, the add them. I think, you know, going forward, we're going to have to start to separate these two markets out because it's not in the margin for error anymore. If you look at sort of traditional WFE, we see it trending in that mid-single digits. I talked to the dynamics that were driving that earlier. As we look at 2026, just qualitatively, we continue to see a lot of momentum at the leading edge.
We think that there's some broadening of investment at the leading edge that happens next year that certainly informs our forecast. Now, my forecast is obviously influenced a lot about my business and where we sit. I certainly have, you know, some perspectives on certain parts of the market where there's a lot of relative strength for KLA and maybe less so in other parts. We think the leading edge, we think there still continues to be momentum in DRAM, a broadening of investment likely there, supported by HBM. China is down this year, we think down roughly 10% to 15% in 2025 versus 2024. I think probably continues to correct a bit into next year. Most of the investment is in legacy nodes and, you know, the investment levels have been pretty high over the last couple of years.
I would expect some sort of consolidation or normalization of that investment. If you look at the legacy non-China markets, they've pulled back a fair amount. I don't see them getting worse. I don't really see a lot of investment, but I think it kind of holds together as we move into next year. The NAND market is, I think, continues to be pretty constructive off of pretty low levels of investment overall. I don't think that, you know, I think possibly it could be some growth there off of a pretty low level overall. I think that's generally how we see it. Again, we haven't quantified it, but, you know, it's our expectation today as we look at 2026 that we think WFE has growth to it and that we think packaging continues to grow as well.
Yeah. In terms of, you know, foundry logic as a mix within that overall envelope, do you think for you that foundry logic ever gets back to that sort of like 80% high point that you hit before? Or was that just sort of a confluence of events with NAND being incredibly weak and so on that just kind of is unlikely to repeat?
Yeah, so NAND was weaker and then obviously, you know, memory does have the cyclical dynamics that I think always continue. I think HBM is interesting in that there's a lot of customization and logic-like aspects of HBM moving forward that might change some of those dynamics. I think if you look at where the growth rates are more elevated, certainly that's on the logic side. You know, look, our models, if you back in 2022, I talked about maybe 60% mix over the kind of long run, given some of those episodic sort of cyclical dynamics in memory would be logic foundry. Yes, today I think it will probably be in excess of two-thirds. I think that memory will grow at, you know, kind of longer-term GDP plus type growth rates. I think logic foundry will be more consistent with overall semiconductor revenue given the drivers of it.
I think in that environment, I think it's very constructive for KLA given the relative process control intensity and logic versus memory. We're seeing momentum on the memory side that I think even in any environment where we're like even this year where memory is about 5% less than last year, we're seeing a very good setup for relative performance. I think on a go forward for a lot of the reasons we just talked about, we feel very good about KLA's ability to maintain not only our position but also see a continuation of some of this relative share of WFE opportunity that we've seen over the last few years.
Very good. You mentioned NAND before. You talked about if you think it's sustainable. Do you think it can still grow from here? I guess maybe give us your view on a framework to think about process control intensity stepping from HBM 3 to 4.
I don't think I said what I need to say about NAND. I don't know if there's much more I can add on that, at least as we look at next year. Obviously, that's much market sensitive to consumer dynamics, so we'll just see how that plays out. Off the level of investments, roughly, you know, $11 billion. If you just go back to 2017 or 2018, it was $18, $20 billion. It's down considerably from where it was many years ago. I think from the low level, it has, I don't think it has a lot of downside. Let's just put it that way. Do you want to talk a little bit about some of the dynamics?
I think on three going to four, there's a couple of inflections. First is, of course, the logic circuitry for four is more advanced, and some customers are thinking of very advanced nodes for that logic-based die. Today, the logic-based die is mostly very trailing edge nodes, and because of that, you use older process control systems. Intensity goes up because you're going to advanced logic die. The second area is that the I/O connections increase quite significantly. In some cases, can double. In that case, you have to go back again and increase the die size because otherwise you just won't get enough transistor density, DRAM transistor density. Die size will increase, which is beneficial to KLA, and then the number of stacks would go up as well. The other inflection is that most customers currently are using bump technology to do integration of HBM.
There is trend to go to hybrid bonding, and hybrid bonding has advantages and then disadvantages from a yield point of view. I think KLA will play a very important role in delivering hybrid bonding type tools as well, along with the ones that we're doing today, which is bump technology. All of these are positive trends that'll be beneficial to KLA.
One of the great things about IBM with memory is the performance requirements of high-performance compute are very, very tight and rigorous, unlike PC markets where you have memory leakage and your computer locks up and you go and reboot. You really can't do that with a high-performance server. You need to run at very high levels to feed the GPUs. It's creating reliability requirements around DRAM that's much higher than in other markets. Therefore, the performance of the stack is completely dependent on the weakest one. The stack performs kind of in line with the weakest DRAM in the stack. The process variability that our customer can deal with, because you can't really downsell to another market.
You have a device that doesn't necessarily meet spec, but there's another lower-end PC, for example, you can put that device in because it still functions. It's much harder to do that in an HPC environment. Those are trends that are continuing. They're pushing some of the logic circuitry. The I/O counts are increasing. All these are leading to higher performance requirements, and we think that over time is beneficial to KLA.
Fair enough. You know, we spent a lot of time talking already about process control intensity, but can you maybe talk about EPC and remind us of kind of like what long-term growth rate for that business looks like for you?
Yeah, so there's several businesses in EPC. The main message I would say is that we're shifting that entire portfolio towards packaging because packaging is growing. The second thing I would say is that there's a shift, which is the value is going from PCB to substrates and from substrates to silicon. What we are doing is shifting our entire product portfolio to deliver that. There are several businesses in it. There's a process business that was heavily leveraged towards specialty. Now it's mostly leveraged towards more than 50% of it is leveraged towards packaging. An example of that is in specialty, you know, to make credit cards, you have very, very small dies and you need to do simulation using silicon edge. That same technology can, will feature in the future, will be used for doing dicing inside the memory chips and also eventually in logic chips.
That whole transition is happening now in the process business heading towards packaging. PCB markets are moving towards substrates, so that's a positive thing. Component inspection is growing simply because semiconductor revenue used to be around $550 million. It's heading towards $1 trillion. Final component inspections are increasing and our component inspection business is increasing. We have a software business in there that essentially delivers all the coordinates and CAMs for direct write in inspection systems. That business is growing at a good click. I think all these businesses are doing generally positive, but again, the growth rates are not very similar to the front end and for packaging. The growth rates are different, but the businesses and the portfolio are heading in the right direction.
When we acquired this business back in 2019, our view that the PCB integration into the substrate and as it integrates into the package was part of our thesis, and we're slowly starting to see that play out. I talked a little bit about the dynamics in specialty and how that relates to just trying to leverage the relationships that KLA has. If we can find differentiated process businesses, those can make sense for KLA from a growth point of view. I think we've been able to check some boxes there. We ran it separately because it was kind of separate and distinct, at least from a channel point of view, to our semiconductor process control business. As we've seen the importance of packaging increase, we're seeing our customers drive it together. Back to my earlier comment about the backend becoming very frontend-like.
Because of that, we went from having this separate EPC group that operated kind of independent and independent with its own channel into a more of a consolidated view to ensure that we're representing and performing productively and strategically aligned with our customers. The market drove this kind of consolidation. EPC isn't really a relevant group within the company. Obviously, we do the segment reporting that shows the distinct business segments because they are. How we engage has really evolved a lot in the last few years. I think it has allowed us to get better alignment and to take advantage of some of these opportunities that are going to be pretty compelling. We're building the roadmap to ensure that we can support where the growth is and where we think we can differentiate, which ultimately then drives the KLA model.
Yeah, on services, a little slow growth rate in 2025. Do you think we can see a bit of a catch-up in 2026?
Yeah, as I said earlier, whenever you have a fab you're servicing and then you can't get in that fab, it's going to impact the near term. In the long term, though, our service business correlates much more to semiconductor revenue growth than it does to WFE growth. I feel very comfortable with the long-term growth rate. I don't want to guide 2026, but our long-term target model of 12% to 14%, I feel very comfortable with it. We're seeing dynamics and trends in that business. It is unique in the industry given its contract streams. 75% of the revenue is contract. Useful lifes are increasing. We're seeing the value in terms of what customers are paying for contracts slowly increase over time as the complexity increases, as our commitments to performance and availability also increase.
We're getting, and I didn't say this about DVC, but in businesses we've acquired, we're getting some growth from leveraging our infrastructure to support the service business. I think small companies sometimes struggle with the global infrastructure to monetize service to the extent we can do. That's an element of growth. In areas like packaging, where sampling rates are quite high, customers are now looking at the service model in a different way. We don't sell necessarily, you know, parts and break fix. I mean, obviously we do that. What we sell to our customers is we sell performance of the systems, which means spec performance, matching performance, and availability. Process control tools are always used. They rarely idle. If they idle, it's very modest because it's the most efficient thing to drive yields, right? To meet not only your customer requirements, but even in constrained environments to optimize capital.
Process control is purchased at a certain level. Uptime requirements are quite rigorous. Performance matters not only to continue to find what matters, but also to ensure that the performance across the fleet is very consistent. Those are unique drivers for our business. I think that they are contributing to some of the underlying trends that we're seeing, which I think drives that long-term growth rate.
Very good. I'd be remiss if I didn't sneak in a couple of financial questions here in the remaining time. Let me go do that. I love financial questions. Very good. OpEx, when you think about your financial model, what are the priorities for managing OpEx going forward? What are the key drivers? From a leverage perspective, if you grow a dollar of revenue, how much does OpEx grow going forward?
Yeah, so dollar of revenue, I mean, if you think about our models, 40 to 50% incremental operating margin, right? We're going to, you know, the drop-through will be 40 to 50%. Our incremental gross margins are 60 to 65%. You end up with, you know, for every $100 million, we'll call it $10 million plus kind of an incremental OpEx. Obviously, some of that supports volume, but really our priorities, and if you look at the spend in the company, we're a portfolio company. We think it's hugely valuable to our competitive differentiation. Most of our competitors are point product competitors. One of the most important things Ahmad does is understand the roadmap and make sure that we're allocating the R&D dollars across the products in a prudent way, understanding that we have to solve our customers' problems and have an intimacy around their roadmap many years from now.
To deliver an evolutionary product, we could spend $100 million and take three to five years. If you're talking about revolutionary products, we could spend a few hundred million dollars or more and take, you know, five to seven years or longer. These are not small bets. We have to ensure that we're picking the right problems to solve, problems that scale to production, and we can solve them uniquely. R&D sort of intensity probably stays in that 12 to 13% range on revenue growth over time because that's fundamental to the business model. We think we can continue to drive leverage through the non-technical areas of spend.
I talked about that as a priority earlier, and that's certainly something that our teams are looking at as we go through our regular planning process and how to leverage some of the tools that are out there to bend the cost curve to ensure that the leverage model continues over time. We feel pretty good about our ability to do that, and we're going to continue to make these R&D investments to support the business and ensure that we're maintaining and driving the relevancy we talked about earlier.
Great. Dividends, and we sort of map the growth of your dividends over time, do you sort of recommend that tying that to the growth of your services business, given the recurring nature of that business? Should we assume kind of like the same CAGR as we've seen historically?
Yeah, it's interesting you bring that. I mean, conceptually, that's how we got comfortable with it, and it's certainly been part of the cadence of how we've thought about increasing it. In our dividend, we raised it 16 years in a row, and we just did in April to $1.90 per share per quarter. We maintained a growth rate of about 15%. Now you go translate back to service, and you go, okay, service is growing top line somewhere at 12 to 14% given the leverage in the service model. The fact that service has grown every year except for one in the last 25, it's certainly been a factor.
It drives our comfort level about being able to have a dividend policy and being very explicit about our ability to grow it, and then have an annual cadence that you don't have a dividend unless you're growing it over time. I think if you're explicit about it, you can get it valued. Certainly the comfort around not just the dividend, but also our ability to fund, which has driven our capital structure perspective, the ability to also have service fund the after-tax cost of debt in the capital structure as well. That's fundamental to it. We have some other governors. We think about long-term over time payout ratios as a percent of free cash flow as another, but we're going to grow it at 15%. Generally, we think if the overall business can grow 10 to 12%, we can drop through 40 to 50.
We can allocate the capital prudently, grow earnings per share at one and a half times the revenue growth rate. We're a fairly cash-light business. We make investments in working capital, but that should translate into free cash flow growing in the teens, and we ought to be able to continue to maintain this trajectory of a mid, you know, kind of 15% kind of growth. We're going to return 85% of the cash flow the company generates. I talked about the dividend. That's 25 to 30% in terms of the long-term payout ratio, and then the rest is allocated through our ongoing systematic buyback program over time. We're going to then do periodic assessments around the capital structure of the company and what's the best use of the capital structure in terms of opportunities for growth. We're always going to invest in our business.
We're going to look at other opportunities to augment the portfolio. We'll have a high bar around how we looked at that because we understand the alternative. You know, how do we put the capital structure and the capital to work in an assertive way to drive the overall value of the company over time?
Yeah, just a quick word on M&A. Where's that kind of, I think you sort of addressed it there, but maybe just do you see it being sort of more on the back burner relative to your historical cadence or not?
We like the businesses we're in. We've talked a lot about the opportunities around both logic, memory, packaging, about what's out in front of us. We always look at everything juxtaposed against the returns that are available in buying back the stock in terms of how we allocate the capital. We've done a number of things to augment our portfolio. Ahmad mentioned some of our software capabilities earlier in the company. We've made some investments there. They augment and strengthen the portfolio. That's been part of our focus, small things. We've also found some markets that are within SEMI that are kind of new money or adjacent markets to process control. We got into chemical process control through an acquisition a few years ago. It's a $150 million business today inside the company. We're looking at those kinds of opportunities that we can leverage the channel.
We can leverage the operating model. We can leverage our customer relationships. There's a regulatory backdrop that we have to be sensitive to. I think that obviously informs our opinion on it. We like the businesses we're in. That'll be the primary focus of the company.
Great. I think we'll have to leave it there with our time. Thanks so much for being here. We appreciate it.
Thanks for having us.
Thank you.