KLA Corporation (KLAC) Earnings Call Transcript & Summary

My name is Atif Malik. I cover U.S. semiconductor, semiconductor equipment and communication equipment stock here at Citi. It's my pleasure to welcome Bren Higgins, CFO of KLA-Tencor. We also have Kevin Kessel, VP of IR, in the audience as well. Welcome, Bren.

Yes. Thank you for having me.

I'm going to kick it off with my questions first and then open it up to the audience to ask their questions. If you have a question, please raise your hand, and we'll send the mic over to you. Bren, I'm going to start with where we are kind of in the cycle, when investors look at KLA, you guys have the highest quality gross margins, very high dividend rate, a very defensive, very steady eddy and you've been an outperformer relative to WFE over the last few years. So from my perspective, investors are kind of mostly focused on how you're going to maintain that outperformance relative to WFE? And let's maybe start with the backlog. It has been always a relative area of strength for KLA in historical cycles. You talked about 6.5 to 7 months. It sounded like on the last call that the lead times are still quite extended. What is driving above-average backlog levels this cycle?

Okay. First, it's a good thing, Rick is in here. You said KLA-Tencor, right? KLA now since 2019.

My bad.

But we're -- we had multiple decades. So you get -- it's okay. Sure, backlog is -- it's extended and it has been for a couple of years. It's really, I think, focused on certain product types where despite the cycles in the industry, we still have struggled to ship relative to demand, our high-end optical inspection being one area, reticle inspection also being another area where the lead time for optical components is extended, takes a long time for that capacity to come online. And there, with the introduction of EUV and the strength of leading-edge logic and foundry, it's been a really strong market for us. So we have a very strong position. And I think it's probably the fastest growing market in WFE, certainly of larger markets. And in fact, this year, even with WFE down, what could be somewhere around 20%, this business will be probably flattish year-over-year. So that's a big part of it. We also have customers that wanted to get into the queue, given the demand over the last couple of years, lots of new projects out there. And so part of that is to control the timing of when they receive tools needed to give us orders. About 40% to 50% of the backlog was shippable beyond 12 months. And so there's a number of projects were -- are currently under construction. And so that backlog will be actioned as we move out a year or so. But for what's more current, it tends to be a lot of these products. And so it gives us a little bit of confidence moving forward about just the stability message, which has been an important one for us here that while re-acceleration in demand in a sustainable way is less clear from a timing point of view. Where we're operating today, we feel pretty good about the business levels we're at, and it feels, as we said at earnings, that things are stabilizing from here.

Stability is good. And then you guys have talked about WFE down 20% this year; memory, 40%; foundry logic, 10%. And then the question we get from investors is, first of all, on the memory side, we continue to hear about incremental production cuts at the memory makers. And the question is, have we seen kind of the trough of memory investments for you guys? Or there's more downside?

Well, so from a new equipment point of view, it looks like it's fairly weak this quarter as we -- in the last couple of quarters and so, we're seeing our customers make meaningful reductions and new equipment purchases. From a utilization point of view, we've seen utilizations, and they're not declining anymore. So they're not really improving. They've been fairly steady over the last several months or so in terms of the installed base. So I think the correction is happening in that part of the market. But the biggest question we get from investors as you may think about timing, I tend to focus on the cash flow of our customers. And we'll need to see the fundamentals improve, I think, in a meaningful way to start to drive the improvement in the business and ultimately, the cash flow. Obviously, the financing environment is more prohibitive as well before but I think these are going to be the factors that are ultimately going to determine the timing of when we start to see a recovery in memory spend. I don't think it's worse next year because it's pretty low now. But in terms of timing of when it gets better, I think it's probably predicated on pricing environment, lower inventories, all the usual things that drive ultimately improvement in the business and the resulting cash flow from that.

Makes sense. And on the logic side, we've seen kind of a more balanced mix of mature logic investments over the last year or so versus leading edge. And then the question is in terms of the process control intensity and how does that change in mix, if you assume mix remains -- I think you guys have talked about maybe 60-40 or something in that neighborhood for you. But broadly for WFEs, if the mix remains 50-50, mature logic versus leading edge, how does that impact the process control intensity for you guys?

Sure. It's a good question. And over the last couple of quarters, our normal business mix tends to be leading edge, 75% more leading edge, 25% lagging. And we defined it at the 28-nanometer production node and some people talk about it differently, so I think it's important to put that out there. But the last couple of quarters has been closer to 60%. If you look at what we see today versus at the beginning of the year, I mean, the leading-edge environment has been worse than what we had originally. So the overall WFE number hasn't changed. Leading edge is an area that has been a little worse than what we thought. And the legacy parts of the market have been stronger, legacy both China and non-China legacy. So process control intensity, it really depends on what our customers are doing. If they're adding incremental capacity to support existing parts, and there's not a lot of change happening there that the process control intensity tends to be lower. If they're changing design rules or driving new specs and certainly in automotive, specs around reliability are becoming much more stringent, and that's changing process control buying behavior. So it really depends on what they're doing. Generally, across the legacy parts of the market, it tends to have a process control intensity profile that's more like memory. So if process control as a percent of wafer fab equipment, in memory is somewhere around 10%. It tends to be in that ballpark, whereas a high-mix leading-edge foundry would be somewhere in the mid -- maybe high teens to just give you some perspective. Now again, if they're making fundamental changes, design rule shrinks, reliability, spec changes and so on, you'll see that tick up. In the China part of the market, certainly as it relates to new customers and new projects where they're doing more development investment, you tend to see it weighted higher and ultimately -- and those facilities, a lot of them are subscale and will be subscale for a while. So over time, as you see more maturity, you start to see the process control intensity level off, but it's certainly more concentrated earlier on as customers are developing and working out a new process, and then they introduce volume and then try to ramp that process. Now if you're a new customer, it may be a legacy node, but it's a new, it's not legacy to you necessarily. And so it doesn't always behave the same as it would for a much more established customer. But generally, those are the dynamics in terms of how they play out.

Good. Let's just stay on the logic side, Bren. On the leading edge, there's a perception among investors that as we go from FinFET to get all around maybe the process deposition or etch companies are going to benefit more incrementally. But historically, when we move to FinFET, you guys saw a decent demand because of the nature of the device is changing. And so can you walk us through your expectations on process control intensity as you migrate from FinFET to get all around device?

Yes. We're excited about it. First, even on the litho side, you're more -- the introduction of more EUV layers, which tends to be a driver for inspection. And we certainly have seen that as the smaller defects matter more and the ability to find them is a benefit. And then you couple it now with changes in transistor architecture that carry more process steps, you're depositing more layers. And so as a result of that, those steps are going to create new defect mechanisms, new metrology requirements. Obviously, depositing more layers is going to require film measurement tools that we have that are more capacity-centric. On the inspection side, because of the structure of the device, subsurface defectivity becomes a challenge. In fact, we're introducing a version of Gen 4, that's a non-upgradable platform that has a longer wavelength that allows us to find service defects, but also do more subsurface defectivity where typically shorter wavelength tools tend to have higher levels of sensitivity, but struggled to penetrate through opaque film structures. So there's an example where there's a completely new application for Gen 4. So it's all the same drivers in terms of inspection from shrink, but also a new driver related to transistor architecture. I talked about metrology. You also have the power distribution systems with backside power, and that creates a second wafer, wafer-to-wafer bonding. So then it creates overlay challenges. It consumes a wafer. So the wafer inspection and metrology business, specs on those wafers and then additional films as you connect. So there's a lot of opportunity there. I think we're pretty excited about what it means, and you take that and couple it with a very robust demand environment, changes in terms of what's driving leading-edge demand, not just leading-edge mobile, but now AI and customization around GPUs. It's creating, I think, a pretty good environment that's good for process control generally, but then also you've got these opportunities related to these major changes that customers are going to do in chip architecture.

Good. And on the mature or the trailing edge, investors are concerned about sustainability of demand, particularly in China. Is it being kind of restrictions-driven or there are rail demand drivers? So when you talk to your customers in China, particularly the domestic Chinese chip makers, do you get a sense that they're adding the supply kind of strategically to meet the long-term demand or they're more worried about restrictions?

So let me break it up into two pieces because I think you have the non-China legacy, which is serving markets that you've seen semiconductor content rise. You've got automotive obviously being the biggest one. Most of our customers there have talked about multiyear investment plans, both for serving those markets, more in-sourcing, given some of the challenges with supply over the last couple of years, different inventory management structures and so on. So that's part of it. And I think that you'll continue to see that demand be strong. It's been strong for the last couple of years. As we look at next year, my baselining has it is generally flattish. But we'll see how that plays out. But again, a wide variety of markets that they're serving. In the China side, so if you just thought about overall China WFE levels, about $15 billion of WFE more or less, right? And so you've got, we'll call $4 billion or $5 billion of that related to memory with YMTC and CXMT being the biggest players there. You've got a, we'll call it, about $1.5 billion of that is related to infrastructure -- infrastructure for wafers -- silicon wafers, which we participate in, but also infrastructure for reticles. Reticle or wafer supply was a problem over the last couple of years given the nature of long-term contracts with the non-Chinese providers. And then reticle, no captive mass shops, so we have to go to merchants, limited investment from merchants over the last few years, so a desire to have more capability. And for reticles, it's a business for us that behaves like a process business. Every reticle has to be inspected. There are multiple sets per designs. Every design has -- and so you see this proliferation of design, so you have more reticle requirements. That's roughly $1.5 billion to $2 billion, and that investment is happening. So then you get down to, we'll call it, about $10 billion of WFE that's logic/foundry. I'd say roughly about half of it's the established customers, the HH groups, the SMICs, the next chips, which they have defined processes. They have customers. They're serving established markets and have for a long period of time. And then there's a piece that is new project-centric. Obviously, some of it is covering the specialty parts of semiconductor, too. And you've got customers that are trying to develop processes, establish positions, win customers and so on. So in that investment, there's a lot of projects. There's a lot of development effort starting to ramp limited capacity. And it's like anything else, it's something that takes a very long period of time, and you have to continue to invest. Tools are being installed, they're running. And that's the part of the market where I get the questions about sustainability, and we'll see over time how that generally plays out. But because I'm a believer that it's good for the industry to have long-term supply/demand relatively in balance and we'll see how that sort of rationalizes. But I think you have to think about that mix of what's happening and understand what's happening within the various buckets that I described. So, so far, it's good. I would think it continues into next year. I have $900 million plus of deposits. I have backlog and customers that want to maintain credibility with us and prioritization in terms of slots and slots planning. So I would expect them to perform. In a lot of cases, it's some of the long-extended backlog where it's tied to new facilities. So I don't think things are going to change much as we move forward. I don't know how much China grows next year, but I don't think it's going to shrink.

Great. And then earlier on, you mentioned AI and GPUs. And in general, leading edge design rules are a bigger die in GPU architecture should drive higher inspection intensity as same defect intensity has a bigger impact on yield. Can you talk about maybe the volume is not big enough for GPUs right now, but any kind of green shoots or indications you're seeing as more GPUs and custom compute chips get taped out at foundries? Are you seeing any signs that the leading edge demand would recover next year?

Well, I think it's going to pick up. Just think of it over the next few years. And yes, today, roughly maybe 10,000 wafer starts or so in that ballpark, given the percent of businesses out there, but a pretty strong growth rate. So if you just fast-forward a few years, then you're talking about another driver of leading-edge nodes that is consuming meaningful amounts of wafers. You just go out a few years and you're up to 40 or so wafer starts, that you -- if you compare it today, you typically be the leading like N3, it's roughly 40,000 or 50,000. So in some ways, you have another driver at a commensurate level of leading-edge capacity requirements. And I think that, that's encouraging for KLA. You talked about the opportunities that exist typically with the larger dies and defect density creating a bigger yield challenge and also a much more valuable die. So the inspection -- the desire to inspect, to protect and ensure that yield is good and that the wafers are functioning is important. So I think that's a good driver for us. You're also starting to bump up against the reticle field limits given the size of the dies, you have single die reticles, which is a challenge in our -- for reticle inspection business, a very strong part of KLA. And so you do that, and then you've got a whole bunch of designs, which generally then creates challenges as designs, test design rules in different ways, different process variants also too. So if you take that and extend it out even further and you think about the heterogeneous part and you start to test the reticle field, so then you've got to put things together with packaging, you've got high bandwidth memory, which tends to be a larger chip. You have EUV and DRAM that create even more logic circuitry and that kind of chip. So those creates opportunities for us. You've got the interconnector or the interposer connecting between the memory in the logic part of the device, that is -- it does create opportunities in a lot of different places. So we're pretty excited about it, both in terms of how it drives leading-edge demand that you've got other drivers of leading demand at the front end of any node migration. And then the general process control challenges that come with it. So we're encouraged by what we see.

Great. And I know on the last earnings call, you talked about upside in advanced packaging, maybe just for DRAM. Some of your peers, the smaller peers, they've been talking about orders for HPM packaging, whether it's onto or the Camtek. I mean you guys are too big to put press releases on orders. But I'm just curious if you are seeing the advanced packaging demand also kind of pick up from these applications.

I did say it was an upside factor. Obviously, what is purchased for inspection for advanced packaging is not Gen 4, Gen 5 level tools and the ASPs that go with them. But we did have some incremental demand that we saw related to this pop in the AI business. And it's encouraging in that space, particularly as you start to see lines and spaces shrinking as it relates to the interposer and connecting layers, it is creating, I think, opportunities that, over time, play to the differentiated capability we have with the platform that we offer into that market. It's a market that's segmented and to a low end and to a higher end and different tools have served the different parts of the market pretty well. It's always been the challenge as trying to design a tool that the market isn't big enough to support an R&D effort for products to serve each individual parts. But we're seeing more and more of it start to shift into the higher complexity requirement. And we saw it play out in terms of the overall macro market, not just because of packaging. Automotive was also a factor in it. But if you look at the third-party market share results on macro inspection, in the last year, for the first time, we actually have the number of positions in the market. So driven by a number of factors, but certainly, it's an upside factor there related to just how much our customers are going to spend to ensure that this now very valuable package isn't failing because of a connecting layer somewhere. We also do have a small business that we acquired that does some chemical process control. When you think about the TSVs that are in the -- that connect some of these layers together, will we monitor from a metrology point of view, the chemical composition. And so that's been a nice business for KLA as well and new money. So not big, but part of this evolving part of the market, roughly $300 million or so -- $300 million to $400 million of business for us in the last year and pretty flat year-to-year. So we're encouraged by what we have there, but also growth. And there's also the process part of advanced packaging, too, where we serve that part of the market with our specialty semiconductor business. That's part of that $300 million to $400 million. So we come at that market with a number of different products and encouraged by the potential growth that's there.

Great. Then power devices remains a secular investment theme for investors outside AI. And you guys are uniquely positioned there with inspection and metrology equipment across multiple platforms. You've talked about $700 million annual sales with 50% on silicon carbide. What is the company doing to expand new products in power devices? And how meaningful could silicon carbide applications be longer term for you?

Yes. I mean, look, the electrification is becoming a much bigger opportunity and higher semiconductor content as a result. Controlling the substrate is not easy, the quality of the substrate. So that's been part of inspection that we've seen and we've benefited from. And we have products in development to further enhance that position on the inspection side. We also have the process tools through specialty semiconductor. That so when you take that 350, it's a combination of inspection, mostly macro, but also some additional inspections, but also process tools. Over -- in the total amount serving the automotive market, a few years ago, it became pretty clear that we were going to see stronger demand in this part of legacy generally, but automotive being the biggest driver. And so we have done some customization of certain platforms rather than selling the old tools into these markets where we are taking new capability or converted -- reconfigured capability on some of the existing platforms that we've customized for these markets. We also have some software that we're also selling that allows our customers to connect and get better correlation between what's actually happening in the fab from a defect point of view and also electrical test. And so it's creating some opportunities. In automotive, you increased semiconductor content if you have given the margin structure in the market, the cost of recall is pretty expensive. And so you need to make sure that you don't end up with semiconductors that are failing. But at the same time, you also need to make sure that you're not overly sorting out the good die. So the ability to make sure you don't want to under-kill chip failure but you don't want to overkill either because it affects margins. So there's a lot of opportunity for us to take these systems that we were effectively driving better margins for us, higher revenue levels and lower cost of ownership for our customers rather than selling the older versions of these tools. And then also providing a new capability that helps address some of these reliability challenges as well.

Great. And let's talk about services. It's somewhat of a game changer for you guys given how strongly it's growing. When I look at your services sales as well as the growth, it does appear that you guys are outperforming some of your process peers. And I'm curious in terms of your attach rate, I mean the software portfolio, what is so different about process control tools versus maybe the deposition tools that helped you give service at a higher pace?

The services businesses are different across the companies in this space. There's no equipment in our service business. It's all traditional service. There's really no software in it either, maybe some software for maintenance here and there that shows up in terms of how we do the accounting for service. But our tools also don't have consumables. And so when customers buy process control, they tend to -- they have very stringent uptime requirements. Our tools are 90%, 95% uptime. And so they tend to go out and buy what they need and then rely on us to provide the service. So it's a high complexity given the ASPs and the technology involved. It's high mix, customers don't have many and then a lower volume business. And while it creates some operational challenges in terms of how we manage it, it does provide meaningful barriers to entry to third parties or customers to do their own service. So they rely on us to do it. We have some control over supply chain in terms of customization of the parts. And because of that and the matching performance across the tools that they generally buy a contract stream, and we can build the contract in a way that establishes an entitlement level, entitlement determined by response time, coverage levels and inventory stocking strategies and so on and then we delivered to that entitlement. So the attach rate, 90% plus as a percent of ASP, closer to somewhere between 4% and 5%, and we've seen that grow over time. We're seeing the percent of tool revenue we get out of service over time as tools are in service for longer periods of time. We're servicing tools that we shipped in the '90s as an example, 80% of KLA and Tencor historically have shipped is still in service in some way out in the field. So process control tools have a very long life. Like I said, it creates some operational challenges, but does provide a fairly protected contract revenue stream that we can optimize our cost structure underneath. And when WFE goes into weaker periods like we are this year, customers still invest and run the installed base. And they over -- they sample more. They want to make sure they get the efficiency out of the capacity that they're operating at the time. And so it does create this -- I think, this pretty protected sort of stream over time. 12% to 14% is our long-term target rate. We're slightly below that this year. Obviously, utilization rates are a little bit lower. We had some headwinds related to China restrictions. I would expect actually service to grow next year as we start to see the tools that we're under warranty in '21 and '22 that we shipped. We had 46% growth in our semi process control business, in '21 and '22, it's 36%. So all those tools are going to start to come off of warranty and go into contract and with a 90%-plus attach rate, I would expect that we'll see probably growth over the next year or so, that's ahead of the long-term target. So it's a very resilient business and one down year in the last 25 years, that was in 2009. And even then, it was -- when the equipment went to 0 effectively, it was down 10% or so.

It's probably outperformance against WFE that has also contributed to outperformance and services as a tool that come out of warranty?

Fair way to think about it. Yes.

Let me pause here and see if there's any questions in the audience. If you have a question, please raise your hand. All right. Bren, we didn't spend much time on EPC. Obviously, it's a bit of a tough year for the PCB business this year, consumer end markets are down, smartphones, PCs and so in terms of the SPTS business has been stable, so as you look into next year, how do you feel about EPC?

Yes. So specialty semiconductor, given what we talked about earlier, I would expect it to continue to grow. I don't think we're going to see that turnover given what's happening in those markets. So I feel pretty encouraged by the offerings there. There's also hopefully some new product traction with an expansion of the portfolio doing particularly around packaging where you do plasma dicing where they've used historically mechanical saws, to singulate wafers and now using a chemical process, a plasma process to -- because you end up with micro fractures and other cracks that can come from the process itself. So it's creating some opportunities there. On the PCB side, to your point, and even seeing it on the component inspection, very mobile-sensitive. And so once we saw mobile start to slow down in a meaningful way back in '22, it's affected that business. It's more capacity-centric. It's more closer to consumers, shorter lead time. So it could be one that could turn fairly quickly, both directions. I don't see it being worse. And I think the bigger question is, once we start to see some recovery in the mobile markets as you move into next year, I would expect to see that pick up. One of the opportunities in PCB is related to substrate and how the substrate integrates into the package. And as you start to move from litho into direct imaging on the substrate to do finer features, and so it is creating some opportunities for us in the future. It was certainly a part of how we saw the Orbotech acquisition that was attractive as we're seeing more and more of the PCB become integrated into the package through substrate and so on. So we'll see. I'm not -- it's a tougher one to predict as we move into next year. And I think it's much more tied to a stronger consumer recovery in those very consumer-centric markets.

All right. Couple of questions on the model. On the last earnings call, you mentioned that gross margins for the calendar year expected to trend near 61%, which is in line with the target model set back at the Investor Day. In terms of expectations moving forward, we went through big supply chain disruption during COVID, are you guys kind of at normal kind of cost levels and freight and all that stuff? Are you kind of caught up or there's more room for margins to expand as things get normalized?

Yes. So in those areas, it's gotten better. We're not at pre-pandemic levels, freight as an example, but certainly, it's lower than where it was. We've got some headwinds related to the capacity we added. Now this is more of a factory and output statement than in our operating expenses. But the capacity we added to support the near term, but also to support our longer-term plan. I think that as we start to see an acceleration recovery in demand in a sustainable way, our status will start to flip from a headwind into a tailwind. I think we're very well positioned from a facilities point of view, from a supply point of view. Obviously, given the complexity of what we do, we have to be very careful about how we manage our resources in our factories, long training cycles to train our folks. And so those are all things that we're fairly careful with. Our model allows for it whenever we have a softening period. But I think we're pretty well positioned moving forward. So we laid out a plan and said, "Hey, look, we're 61% today, 63% given our mix expectations is our expectations for gross margin." And I feel like we're really well positioned to -- mix can move around, but generally well positioned to get back to those levels. Our leverage expectations of 40% to 50% drop-through is -- drives the company in terms of setting expectations about what -- how much of every revenue dollar needs to fall to the bottom line. And I think it keeps a whole fair amount of discipline in how we size the company. We've seen the cost level flatten out over the last several quarters. And we've had some incremental upside as we move through the year in terms of quarterly revenue. And so of course, against that backdrop, we see, given our gross margin profile, meaningful drop-through. That concept has driven most of the estimate today, increases we've seen over the last few quarters. So I think we're very well positioned to execute against the model. And I think at the '26 plan, 63%. We said 41% to 43%. I think we're well positioned for that.

Great. And then the capital allocation, we saw the big dividend hike earlier this week. Can you just walk us through the priorities of cash and if there's any M&A in the pipeline?

Yes. So obviously, our number one priority is to invest in our business organically and make sure that we can continue to drive the KLA model that we've had for a number of years. It's a business model that's built on differentiated solutions, discipline and understanding the value we delivered to customers, sharing in that value in terms of how we price and then reinvesting to ensure that we can sustain the portfolio, which is important in terms of how we go to market and have the resources in the field to be able to work with our customers to get full value out of the tools. So that's first and foremost, we've done some M&A. Obviously, we wanted to get into what we call the more and more part of the market away from not the semiconductor front end, which we have a strong position and we're excited about, but also exposure to some of these areas we've been talking about that are growing over time. We've also done some smaller tuck-in a number of software acquisitions. We don't talk about it much, but I have at least $150 million in revenue -- software revenue today. I know you go back, it was a few years, that was less than $100 million. So we're -- it's not WFE-dependent, but we're seeing more and more software contribution that we're selling independent of the systems. So that's been an area of focus. And really around simulation, correlation, other data analytic methods to help customers just speed time to results, understand their processes better. And it reinforces the strength of the portfolio and how we connect different tool types to each other and the data, more importantly, that comes out of them. The challenge, there's obviously a challenge from a regulatory point of view in terms of how you think about what might be possible. We're pretty happy with the businesses we have and the growth prospects for what we have. Anytime we look at opportunities, we compare it against the alternative of putting the cash back to work, investing in ourselves, buying back our stock. And so it does create an alternative that we have to evaluate. When we laid out the plan at Investor Day last year was an important message, we talked about 85% of the cash flow we generate, we would return to shareholders that we like the businesses we're in. We think there's compelling value in our plan of $14 billion in revenue and $38 of earnings. That's a plan that we think has significant value, particularly relative to where we are trading and that the best place we could put our capital is to reinvest it back in ourselves. So that's really the underlying philosophy. We're big believers in being aggressive capital allocators to get the full value of the business, valued in. And so we're going to put it to work with your back into the business or along the priorities that we talked about. The dividend increase was the 14th consecutive annual dividend increase. We're going to grow, given our expectations around topline drop-through and earnings per share growth, we think we can grow earnings at 15%. We think we can grow cash flow at 15%, so we can grow the dividend at 15%. So this was consistent with our long-term CAGR. And additional authorization to enable from our Board of shares to enable the execution of the plan that I described. So in some ways, a continuation of what we've been pretty consistent and explicit with. And I think we're going to continue just to do our thing here and execute along the way.

Great. We're almost out of time. Thank you, Bren, for coming to the Citi conference.

Yes. Thank you for having us.