All right. Again, welcome to the second day of J.P. Morgan's 47th Annual Technology, Media, and Communications Conference. My name is Harlan Sur. I'm the semiconductor and semiconductor capital equipment analyst from the firm. Very pleased to have Bren Higgins, Chief Financial Officer of KLA Corporation, here with us. I've asked Bren to kick us off with a few opening remarks. Then we'll go ahead and start the Q&A. With that, Bren, thank you very much for joining us this afternoon.
Thank you, Harlan. Thank you for having me here today. What I'll do is I'll spend a little bit of time on some of our comments about where we are today and the industry and outlook and so on. Then we'll dive into some of the questions. I will make some forward-looking statements today. Those are subject to risk. You can find those risk factors in our SEC filings. I'd encourage you to take a look at them. A good quarter for the March quarter. We had the inclusion of Orbotech for a stub period for about 40 days with the closure of that deal in late February. We did some financing to support the transaction, and of course, had some shares in the consideration for the transaction.
We had some moving parts in our model. We updated guidance. Against the guidance that we provided for the March quarter, we had a strong finish to the quarter, beating revenue and EPS margins by solid amount, or midpoints by pretty solid amounts. In the core business, our view on industry was very consistent with what we had said back in January, that of course, we see the WFE environment down 15%-20% or so from what turned out to be a much stronger finish to calendar 2018. That puts the WFE levels somewhere in the mid-$40 range. DRAM down a lot, 30%+, flash down too, not as much as DRAM. Some growth in logic and foundry, which has been encouraging. We're excited about what's in front of us. We see that relatively balanced through the year.
For the company, we saw the March quarter as the bottom for us and would expect sequential growth in June, which is consistent with our guidance that we provided for the June quarter. Sequential growth continuing through the second half of the year. No changes from what we said in January and where we are today versus last week. Gartner market share data came out for 2018. The market share for the company was consistent with the prior year, just above 50%. We don't compete in the entire process control market. In the markets we compete in, our market share is greater than 60%, but for the total market, about 52%-53% or so. No real change in the competitive dynamics out there. KLA is about 4x our nearest competitor. It's been fairly stable from that perspective. Excited about the acquisition of Orbotech.
Brought Orbotech in this quarter, as I said. Orbotech, for those who don't know, have three businesses. Not exactly equal size, pretty close. There's a specialty semiconductor business, which is a really strong business, had strong growth last year and looks to grow this year in a WFE environment that is down like we talked about. It has a printed circuit board business, which is mostly around imaging technology, where KLA has a lot of imaging capability and IP in those areas. It presents high-end PCB boards are starting to go through transition from PCB boards, IC substrates to advanced packaging, which creates potentially some opportunities for us moving forward. We're excited about that. It has a flat panel business, which after 6 years of growth, flat panel business is softer this year, but it's a smaller part of the Orbotech business.
In calendar 2018, Orbotech was up about 16%. They weren't filing publicly, a lot of that data wasn't available. We filed their annual results when we did the debt offering back in early March. Strong year of up about 16% last year, this year, on a pro forma basis, including the period where we didn't own them, the 45 or 50 days or so of the quarter we didn't own them, they should have a year that's roughly flat, maybe a little bit down. On a relative basis against a broader electronics industry, advanced technology market environment, a pretty good relative performance. Strong growth in the specialty semi piece, which I mentioned. Some flat environment in the PCB business, the flat panel business down year-to-year.
See the puts and takes there, business holding up pretty well, we're excited about the opportunities. We think there's about $50 million of synergies that we're starting to do the work on now in terms of planning. I think we'll see that materialize over the next 12-18 months or so. Excited about the opportunities for growth, the opportunities to leverage some of KLA's IP into those markets, positioning the company to do some interesting things out there. We're excited about it. I think the environment is pretty stable and consistent with what we thought. We're watching the moving parts. There's still concerns in the memory space, at the end of the day, as we've gone through and looked from a risk-adjusted basis, we feel pretty confident with the guidance we've provided. With that, I will take your questions, Harlan.
Thanks.
Yeah, I'll kick off the first few questions. I think what I want to focus on is more of the sort of share statistics that Gartner just recently put out, because it actually also does not only drive home the point of your dominance in the process control segment of the market, like you said, number 1 in process control, greater than 50% market share, four times larger than your nearest competitor. In some of the areas, as you mentioned, which you have clearly a strong leadership position in, for example, I won't throw out some statistics here, in your flagship pattern wafer inspection business, I think you outgrew your market segment there in PWI by 500 basis points. You were 13 times larger than your number 2 competitor.
Bare wafer inspection, another flagship segment for KLA, you grew 88%, you were 26 times larger than your nearest competitor. In EUV and advanced mask inspection, you had 55% market share, two times greater than your number 2 competitor. The interesting thing here is that process control, KLA both grew 16% year-over-year, which is exactly the same type of growth that the overall wafer equipment spending market grew. 2018 was much more of a memory-biased year. Typically, in memory-biased years, we don't see process control and KLA actually performing in line with the market. What drove the higher process control intensity in 2018 that is responsible for that dynamic? Are these trends kind of sustainable on a go-forward basis?
Well, you're right. I think one of the things that you have to understand about KLA is that not every WFE dollar is created equal in terms of how they invest in process control. If you look at memory investment, where it tends to be lower mix, more commodity, more redundancy in chips, more repair done, and so on, they just don't invest as much in process control as you see in logic and foundry. Typically, when memory is spending a lot, it does affect our ability to perform as fast as the market, and most of the time we underperform by a little bit. We have seen some changes, and we're pretty excited about it. The market share numbers are, I think, pretty strong and consistent.
Sometimes I don't think people really understand, over time, they haven't really changed much.
Right.
It's usually been about 52% for the last three, four years.
That's right.
It's been hovering around 50% plus or minus a couple % going back maybe 10 years or so. We have seen in 3D NAND, it's been an opportunity for us. We've seen increasing requirements for metrology, where we're doing more sidewall angle measurements. You're doing more than just top-down measurements in a device. As layer counts are going, you've got to be able to make measurements all the way down the stack. You've seen customers starting to move to double-stack structures, overlay becomes increasingly important. There's all the films, you see the film measurement business that has scaled well there, too. On the inspection side, defects continue to be a problem for the industry, and they do a lot of destructive techniques to find defects.
One way they can try to control for defectivity is to make sure the process tools are running nice and clean. They buy more unpatterned inspection run monitor wafers to make sure before they start a process, to make sure the tools are running nice and clean, so they don't introduce any new defectivity in the process. That's driven the bare wafer inspection business. Now, the growth in memory has also driven the need for more wafers. That's what drives wafer count.
Yep.
In the wafer house, but also in the wafer fab, you're seeing more capacity investment to be able to do incoming quality control or outgoing from a wafer house perspective on bare wafers. There's wafer metrology in that you have to do the flatness of wafers, and the specs around those flatness requirements is increasing as layer counts go up. There's a metrology tool we sell into that space as well, that's been pretty positive. When you add it all up, you get a couple of points of improvement in intensity. Historically, in a 3D NAND environment, it was 8% or 9% of WFE was spent on process control, and I think today you're probably in the 10%, 11% range, maybe a little bit better than that.
If we can solve that defect problem, there are some efforts in the company to try to do that. If we can do that creates additional opportunity for us. We are investing a lot in a new capability to do channel hole metrology with the channel hole that goes down the stack that is measured currently in a destructive way. If we can do that in-line and very complicated X-ray technology that we are using to solve that problem, if we can do that in an in-line process, that creates another opportunity for us. A lot of focus here, given the level of investment that is happening in this space, and we are making some progress. I feel pretty good about our product positioning and the opportunities that exist in memory over the coming years.
Yeah. Thank you for that. When we think about the profile that you laid out for WFE spending this year, down about 15%-20%. Within that, the areas where you do have very good exposure, foundry and logic, are actually doing reasonably well. For example, WFE down 15%-20%, I think we have core KLA, ex-Orbotech, I think we have you guys down only about sort of 5%-6% this calendar year, which is consistent with the higher process control intensity there. Help us think about domestic China spending and how is that impacting WFE this year, and how do you think about China domestic WFE spending looking out over the next several years?
Our exposure to China has been pretty good over the last few years. Starting from scratch and small scale fabs, trying to ramp very quickly, they have relied on a lot of process control capability, so we have helped our customers navigate through that. That has been good for us and fairly consistent. One thing about our business, in contrast to some of our more capacity-centric peers, is there is always a level of investment that is happening as long as customers are pursuing technology transition and pushing their roadmaps. There is always something that is happening. Obviously, when they go and start to scale capacity, that is a bigger investment. There is always something. We have had a continuous amount of exposure, and I think it has been pretty good for us, probably in the 25% or 30% of our total revenue.
This year, China is down about 10%-15% versus the last couple of years, so about $500 million or so in revenue levels for us. It is more foundry-centric this year. Some of the slowdown we have seen in memory is affecting the pace of investment there too. You have also got milestone progression that is affecting the pace of that investment. I feel pretty good about the guidance we gave. Again, I am sort of in that ballpark. There has been some puts and takes. I have seen some improvement in some of the memory investments, at the same time, I have seen some of the wafer capacity planning start to change a little bit and push out towards the end of the year as the slowdown in memory is affecting the capacity planning of the rest of the-
Right
wafer supply out in the market. Puts and takes, but generally in basically the same place. It's been good. I think what you're seeing on the foundry logic side is sensor and IoT opportunities, which are driving a lot of that investment. On the memory side, it's been much more about advancing technology. They're considerably behind relative to the leading edge of memory, and proving a certain amount of capacity and trying to move on to the next node has been the behavior. It's steady, down a little bit. I think a little bit of digestion, which is probably good, and I think in terms of the impact on global supply, certainly in the memory space, I think that we're probably a few years out before there's meaningful supply coming out of there at the leading edge, let's say.
I don't think there's a lot of opportunity at the trailing edge generally in memory. I think that's probably the point we should be watching, but so far, I think that progress is okay, but we're probably a few years away from that. On the foundry logic side, look, between automotive requirements and IoT and communication infrastructure and some of those opportunities in industrial automation and so on, I think there are markets for those products to ship into. That feels reasonably balanced in terms of the supply and the markets they're targeting.
Before I get into some of the product-specific discussions, do we have any questions from the audience? In the market share statistics for 2018, I threw out some of the share gains and the strong performance for some of your core beachhead areas. One area of the market, though, which was a bit challenged, was in metrology, both thin film and overlay. The question for you is, what is the KLA team doing in the area of metrology in general to sort of reverse some of the recent share losses here?
In films metrology and CD metrology, it tends to be a little customer specific. When you think about KLA's broader franchise, we generally are number one in just about all the markets we serve. There are a couple of metrology markets where we share the market, film measurement being one. There's a lot of effort in the company to try to leverage not just data more and machine learning on our tools, but also tying the capability and data flow that comes off of the broader portfolio of tools to try to speed time to results for our customers. You see that in film measurement, but you also see it in overlay, where overlay error can happen in litho process, but can happen in non-litho process.
If you're exposed to the non-litho process, you can try to leverage the information you have to influence the correctables that actually happen in litho to improve overlay. I don't think overlay has been a fairly stable market. We really compete with ASML selling a direct solution there. Our share of the market is still a majority share in that market, and we address it with two technologies. We also invested and have a lot of this effort that I'm talking about in terms of additional design-based capability and non-litho process capability to try to improve our competitive position there. I think when they entered that market in 2013, we saw about 30%-35% of the market or so move that way.
Yep.
It's been fairly stable since then. I don't think much has really changed. I think we address it the way that I described.
On the Orbotech acquisition, closed in Q1, obviously, actually brought a lot of diversification to KLA, right? Flat panel, as you mentioned, advanced packaging, specialty semiconductor manufacturing. The profile is actually quite similar from a leadership perspective. They were the absolute number 1 de facto category killer in flat panel in the areas that they competed in, AOI, test, and so on. In PCB, same thing, automated optical inspection, automated optical repair, the DI segments of the market. Then in their specialty semiconductor business, very strong position in things like RF semiconductors, MEMS, advanced packaging, right? Very similar sort of category killer, like dominant position in other markets. As you've had them under your belt now for the last, let's say, quarter or so, two questions there. How are you thinking about the integration process for Orbotech? What are their customers telling you?
It's a good question, I think, the understanding of Orbotech in terms of the different business segments and where they're exposed and how they go to market is not as understood as maybe we assumed it might be. Especially semiconductor business, for example, is a very differentiated business that's designed specifically for the markets you described and has seen those markets inflect. Those are markets where Exposes us less to the front-end leading-edge process node shrinks and other aspects of WFE on the More than Moore, we call it at KLA. You've seen nice growth there last year and growth this year, and you're talking about a margin profile in a process business that's in the mid-50s, which gives you an indication of the differentiation they have there. We have seen already engagement with customers where they're asking for new capability and incremental capabilities.
As we start to work with them, I think that takes a while for product cycles to play out and create those opportunities, but certainly it's something we're looking at. We also think that we, as part of KLA, can get them to the table more often in certain competitive situations where small companies might not be as credible, potentially, as a larger player in the space, and certainly our customer relationships are very strong in those areas. From a cost perspective, that's the top-line opportunity.
Yeah.
I think from a cost perspective, we're supporting the same customers, we move parts around, we store parts, we manage a service infrastructure for those customers and have the scale that maybe a small player couldn't have. When we look at that business, we see lots of opportunity, a connection back to KLA in terms of go-to-market.
Understanding, also opportunities for us to improve the operating profile of a business that is already performing very well. We talked about PCB, I think the one thing to keep in mind about all these markets is that they're market-leading positions.
Yeah.
Those are companies or positions that we know how to execute at KLA, they all have different margin profiles, but at the same time, their businesses we're in that had margins.
10 years ago that were 10 points lower than they are today.
Right.
We do have a pretty good track record of grinding it out and incrementing product by product, incrementing the margin profile of these businesses, as long as we're in a position that we can differentiate. In PCB, a strong position in imaging, strong service model, 40% of the revenue in the PCB business is service, 90% contract penetration. It provides a recurring stream that we can leverage in the cost structure, which will be good for the company. As I mentioned, I think advanced packaging is becoming a bigger and bigger opportunity.
Absolutely.
We're doing some things, they're doing some things, as we look at how we put that together. Our customers are pursuing cost and capability in a lot of ways. They do front-end process node shrinks, which we know very well and involved in, they're doing things in packaging. There's PCB board technology. There's a lot of new and other ways that customers are trying to squeeze out more, this exposes us to more of that. Flat panel, we'll have to see. Again, a market-leading position, we'll see what we can do to enhance those positions. A lot of similar technologies. We'll see if it's applicable and what we can do with it. After six years, it's at a lower level now. Because they're more leveraged to technology transitions, the volatility in the business is not what the broader industry is.
The broader industry is down about 50%, they're down about 20%-25% in that business. I do think it's a little less exposed, more exposed to transitions, less exposed to broader capacity. We're already starting to engage there to see what we can do with that. When we look at it all, we look at the synergy opportunity we think that exists as we move forward here. We think there's incremental opportunities for us to do things for top line for sure, also on the cost side. We're excited about it. It changed the model, changed how people think about KLA. Overall, from a ratio perspective, probably diluted, while it enables us to grow, I think they've got growth rates in those segments. Overall, I think the combined will grow faster than our base core semi business.
Dilutes our margin profile a couple of points, but enables more growth. If we can drive more leverage out of the business over time and do some of the things I talked about from a product perspective, we feel pretty good about it. There's a financing component, so the OIE went up. We were able to borrow, I think, pretty attractive borrowing conditions. We borrowed about $1.2 billion at a blended rate of about 4.4%, including some 30-year notes. I think we were able to finance a transaction with very attractive debt. We have historic profit position that helps with taxes as well. I think when we added it all up, we saw an opportunity for financial return that bested the alternative uses of cash for the company. Good transaction. We're looking forward to the opportunities here moving forward.
Let's talk about the services business for KLA. It's a great story. It accounts for roughly about 25% of your total revenues. You actually put up your latest investor presentation on your website. If you go and you look at this, it's just really nice, sort of stable growth business through cycle, right? Bad times and good times, the services business kind of continues to grow. As far as we could tell, it was growing at roughly a 9%, sort of 9%-10% CAGR since 2015. 75% of the customers elect to go on service contracts after the warranty period expires. That's, I think, much, much higher than some of your peers out there. Taking everything into consideration, how do we think about services from a growth perspective over the next few years?
It's a great business, and I don't think it's maybe fully appreciated. With the growth of the install base of the company and now what we're seeing extended life of the tools and service in the field. Greater than 80% of the tools that KLA has shipped over the last 40 years are still in service in the field, to give you some idea of the useful life out there. The growth rate, if you go back even 18-20 years, has been consistently around that 10%, plus or minus a point a given year. In upturns, they run the capacity pretty hot, and that drives maybe faster growth. In downturns, they back off on the tools a little bit, and so it brings the lower end. It's pretty consistent.
The general manager who runs that business comes in every year and says, "I had a record this year." It's like, "You're supposed to," right? The install base is growing. It's supposed to grow every year. It had one down year. 2009 was the only down year we've had in the last 20, and it was down about 10%. You go back to that environment we were operating in, where systems business was down 70%, service was only down 10%. Why is it contract? Is because customers only buy what they need at process control, and it drives how we price it, how we think about the returns of what we generate for our customers and the discipline that we have around that. They want to keep the tools up, and they're complicated, and all the parts are fairly custom, so they can't service them themselves.
Because they don't have a big fleet of tools, they can't build internal capability to do it. If you have a big fleet of etchers or PVD tools, you can build internal capability to provide that service. Very hard to do with the tools that we sell, and the complexity of them is pretty significant. I think for those reasons, the need to have them up, the need to have the information that comes off them, customers invest in a contract model that has been very consistent, right around 75%.
Yep
of show the revenue stream and moves around in terms of how they do it. On newer tools, it's per tool. On older tools, it comes down to fab-wide contracts. There's different entitlement levels of service and support, part stocking levels, and we can customize our offerings to be able to meet different customer operating and expense requirements to keep them on contract, and then we can optimize the resources around it.
Yeah.
It's an accretive stream, we believe, to the overall company average in terms of profit, and has a nice stability to the growth.
Service operating margins do sit slightly above corporate average?
Yes.
Okay.
We believe it does, yeah. Depends on how you're allocating costs and so on.
Right.
As you go through the way we've done it and done it consistently over the years, it's an accretive stream.
When you think about it from a free cash flow perspective, is it also free cash flow accretive?
Yes.
Okay.
Yeah.
Got it.
Consistent. For the longest time, it was accessible, right?
Right.
Less of an issue now.
Right
From a tax perspective. Yeah. It's certainly given us confidence around the capital return strategy of the company, particularly the dividend, having the growing service stream. The lack of volatility in that business and the expectation for profitability gives us more comfort around the ability to grow that over time. It's a factor in how we think about it, for sure.
Certainly good to have that stable backbone being a quarter of your business. Maybe more near term, there's been some confusion around the gross margin guidance in the June quarter and the full-year outlook. Maybe first, why are the core KLA gross margins depressed in the June quarter? Then, embedded within your full-year guidance, where do the core KLA gross margins snap back to in the second half of the year?
Well, coming out of the Lam transaction, we put a public model out there that at these revenue levels, we would see gross margins between 62%-63%. Certainly, in an expanding environment in calendar 2017 and calendar 2018, we saw our margin profile above that, even above 64%. There were dynamics in our business that I think created some tailwinds for sure in terms of, A, the leverage on the growth, but also the maturity of the platforms we were shipping, the incremental margins in service, and so on. I think where we are now, A, we're in a contracting environment. Because we believe that we're looking at digestion more than downturn, we've been very careful with how we've managed resources, and certainly that's a factor. The biggest factor is probably the transition that we have to new products.
It's been a historic dynamic within KLA. This is nothing different than what we've seen before. When you go from very mature platforms that have very optimized cost structures and stable designs or stable designs within the platform itself, we're able to drive higher margins in these markets.
When you introduce a new platform, it has a higher price. Generally, the focus from the engineering teams is about use cases, application, and ultimately driving adoption. We will iterate as we go, which drives cost improvement and price inflation over time. If you look at just our broadband plasma business, where we have Gen 4 and Gen 5 products we talk about.
Yes
We're on the fifth iteration of Gen 4. It has a higher margin than the second iteration of Gen 5, which we're just starting today. Over time, we will see that improve, and usually, we end up getting the margin profiles to be basically about the same. Now, the mix of business in any given quarter will affect our results. I think as we move through this, the June quarter has a mix, and under 606, what you ship is what you revenue.
Right.
Depending on the mix is what's out there. We'll see it improve in the core business in the second half of the year. I think it's in that ballpark of 63% or so. I think we need to get an expanding environment if we're going to test the upper limits like we saw in 2018. I don't have any changes really fundamentally in the business that is affecting the margin performance. As I said, there'll be periods when we're in an expanding environment, we'll outperform our model, and periods that in a contract environment, we'll underperform, and I think that's kind of where we're at right now. Generally, in the ballpark of that 62%-63% I mentioned.
In the second half of the year?
Yeah.
Okay, got it.
I provided yearly guidance overall for the year, blended obviously with Orbotech.
Right
of 59.5%-60.5% for the year, inclusive of March and June, which when you work through the math-
I think that's what maybe confused the market a little bit because it was blended, right? We're all used to pre-Orbotech acquisition, you entered that right into December quarter with 63% gross margins. Trying to sort of unravel all of that, given that your core margins are a little bit depressed in the June quarter because of some big opportunities that you had there. Good to see that it is returning to that sort of 62%-63% in the second half of the year.
Yeah.
Your target model combined with Orbotech is gross margins greater than 61%, op margins greater than 36%. We have you exiting this year kind of 61% gross margins, 34% operating margins. If I take into account the synergies, I can get to 35% op margin. What is it going to take to get you guys to that sort of 36%+ operating margin levels beyond just synergies? Is it just growth in the business? Is it mix? What are some of the dynamics there that are going to get you to that kind of 36% op margin level?
Yeah, no, I think we're probably in that, when I said, as we think about next year from a pro forma perspective, which includes some amount of synergy plus some expectations for modest growth in our core business and the Orbotech 2020 plan. We end up in that 60%-61% type margin level-
Right
mid 30s operating margin level.
I think, where you are is a reasonable way to think about where we're at as a company. Some of the factors I talked about will play out, we'll see how those play out over the next 12-18 months. Some of the opportunities in COGS from a synergy perspective with Orbotech are related to sort of structural things that you'll do. We have internal supply for wafer handling-
Yep
as an example. They have a semiconductor business. We have to do internal design to sort of meet those requirements, those are savings, but take a little bit of time to get there. We have the ability to move some products into our global manufacturing operation that we can drive some scale there. Things that we have to go execute on that take a little time, but ultimately will add value back to the business. I think where you're modeling, and again, I don't fully understand what your top-line assumptions are about WFE growth and all that into next year, but I feel like the dilution that comes from Orbotech overall, probably about two points to the base model. Where we were at four and a half billion for base KLA at 38% or so. 36%, 35% sounds about right.
The growth of the semi business, which is a strong business-
Yeah
for Orbotech could potentially be a factor in that as well.
Once you reach that sort of 36% sort of operating margin target, how do we think about free cash flow margins at that level? Is it kind of like 32%, 31%, 32% type free cash flow margins at target?
We've been, for the most part, in calendar 2018, we were right around 30%. I think it's probably a little less than that.
Part of it is also driven by the fact, from a CapEx perspective, we're investing in facilities now.
Yes.
We're expanding our fourth floor in our Singapore operation. We're building a new site in Michigan. Our tools have gotten enormous. We need more space in Milpitas. We're under construction now in a new building in Milpitas. We're expanding one of the Orbotech buildings as well in Wales. There is a level of CapEx investment that would expect to be an incremental $50 million or so a year for the next couple of years. Maybe longer than a couple years, two or three years, for all this facilities work. It will position us for growth moving forward. The last time we had a cycle like this, you got to go back to 2012, 2013. We're in one of those periods where we're going to invest.
I think because of that, the CapEx part of free cash flow will be a little bit higher than before. I think we're in that high 20s, and given the profitability of Orbotech, harder for me to see that we're going to be able to be that far above 30. I think we're in that ballpark.
KLA has one of the strongest capital return programs. You've obviously made a commitment of 75% free cash flow return. 40% of that dividends, 35% share repurchase. Does that mix in your mind kind of change any time on a go forward basis?
We take a long-term perspective, so I don't see the mix changing, although we're spending a lot more on share repurchases right now, including a lot more in the last week.
Right
than we normally do. We target 35%-40% on the dividend payout. We govern it two ways. A, to pay out at 35%-40% so we can be committed to growing it over time, and we've grown it for the last nine years, that's part of sort of our internal culture is to be able, at the company, to be able to maintain growth in that. We govern it that way. We also govern it that we should be able to grow it. If we can grow our operating income greater than 10%, we should be able to grow the dividend payout greater than 10% as well. The rest is share repurchase.
It ought to be, I think at a minimum, 70%-75% of the cash flow we ought to be able to return, and I think over the next couple of years we're going to be above that. We are firm believers at KLA that cash doesn't get valued unless it's deployed productively. Letting it sit on the balance sheet doesn't create any value for anybody. We'll put it to work, and we're 1.5 to 2 times gross leverage. We're running the cash balance between $1.5 billion and $2 billion, we're meeting the objectives we just talked about. That's how we're thinking about capital allocation.
Well, keep up the great execution, thank you for joining us this morning.
Thank you, I appreciate it. Thank you for the time.