I guess it's still morning. Good morning, everyone. Thank you for joining us for this session with Applied Materials. I'm Vivek Arya from Bank of America Semiconductor and Semiconductor Equipment Research Team, and really delighted and honored to have Brice Hill, Chief Financial Officer of Applied Materials, join us. Typical fireside format. I have many of my questions, but please feel free to raise your hand if you would like to bring something up. Very warm welcome to you, Brice. Really appreciate you joining us.
Thank you, Vivek. Appreciate the invitation. Always nice to be here with you.
Thank you. Maybe as a start, Brice, just give us your state of the union. How has the demand environment kind of shaped up? I know a lot of macro cross-current, but how has it shaped up so far versus what your expectations were at the start of the year?
Yes. First of all, you know, Applied Materials, I would say we're not surprised by the dynamics of the year. We've been focused on this evolution of compute in terms of energy-efficient compute, and most importantly right now, the things that are being driven by AI. AI is pulling on a number of the elements of the semiconductor portfolio. If you think of leading-edge technologies that make GPUs and CPUs and accelerators, if you think about compute memory, the DRAM that's associated with that, strong pull for investment there. There's a special kind of DRAM, the high-bandwidth memory that's being stacked for those AI-type performance systems.
When companies put all those components together in a high-performance system, they're putting them together on novel packaging techniques, multi-chip packaging techniques that also pull on the equipment profiles that we have to build those advanced packages, the interconnects, the substrates, different types of bonding for those chips. All of these things are being pulled very strongly. Our business, you know, we've grown. This will be our sixth year of growth right now. We said we're growing about 7% this year in our business. I would say we're not surprised. We think semiconductors are a secular growth. If you look, Vivek, at the investments that we're making in the business, hopefully investors can see that we clearly think it's secular growth. We're building a large platform for innovation with our customers near our facility in California in Sunnyvale.
You know, we think that the business is growing as expected and that, you know, we're demonstrating in those fast-growing areas the share gains that we hope to demonstrate. You know, the one sort of headwind that we're all facing is, you know, 7% growth for the company. That might sound, you know, it doesn't sound fantastic. It's not like the 40% growth rate you see in HBM memory or the 40% growth rate you see in AI. What's happening? The last two years, there was heavy investment in mature logic technologies, especially in China. We knew that as we got to this year, there would be some digestion of that capacity. Still a strong, you know, mature logic market globally, but not as strong as the last two years. What you see is the leading edge is very strong.
It's accelerating. It's being offset a little bit this year by a slower mature logic business that serves the markets that we call ICAPS: IoT, Communications, Auto, Power, and sensors. That's kind of the dynamic this year. Very strong AI, a little bit slower from a mature logic perspective. No surprises, we think it's the secular growth that goes with the business.
All right. Now, you know, we talk about kind of broad industry growth and then specifically on Applied Materials. Broad industry growth, many of your peers have suggested kind of a mid-single-digit-ish kind of, you know, WFE environment, right, this year. I know Applied Materials has not given a specific number for the year, but does that seem like in the ballpark? The reason I think that question continues to come up is that, right, some of your peers are actually growing double digits, right, this year, obviously with a different mix of business. Is it that people are underestimating WFE growth for this year, or is it that, you know, some of the growth is coming in areas like packaging, which is not like classically, you know, front-end what is included in the classical definition of WFE?
Yes. I do think that we look at the overall equipment environment as growing. The way we answer that question, Vivek, in terms of, since we do not give a year guide, the way we answer that question is we have two quarters of actuals right now and one quarter of outlook, and we said the business is growing at approximately 7%. Obviously there is growth so far year over year. We think it is those dynamics that I described. It is a lot of the leading edge that pulls that way. You know, as far as the mix and the compares to other companies, you know, when you are looking at Applied Materials, we have grown for five straight years.
If there are other companies that are experiencing different growth rates, they may have been more associated with NAND, and they may have had years that were lower. The compares play into that. When we look at the overall forecast for devices, the semiconductor devices themselves, we think all the devices are mid to high single digits for as far as the eye can see. Of course, that is a kegger, an annual growth rate type forecast. People do not try to pick the one quarter that will not grow or the one year that might have a slower economy. It is typically a straight line, but that is the way we plan our business. We expect semiconductors, because of the productivity advantages, because of AI, because of the other applications that are emerging like robotics, those types of platforms.
We expect semiconductors to be a great investment area, and that's the way that we're managing the business.
All right. Now, on the role of China, right, as a group, it's the largest, right, 30%± of sales for the industry. Do you think a lot of the effect of restriction and other things that are known, obviously it's a, you know, evolving situation, but as far as, you know, you are aware of, is the effect of all those restrictions, et cetera, already kind of captured in industry expectations at this point, or is there a prospect for changes?
We definitely feel like Applied Materials is significantly de-risked from these trade changes, if you will. The reason I say that is because Applied Materials business in China today is mostly mature logic or is all essentially mature logic business. We're not selling to DRAM customers, we're not selling to NAND customers, and we're not selling to customers who might be trying to do leading edge logic. In the past couple of years, it's been 40, 45, 50, 65 nanometer type products that we've been selling into the market. As we look forward for the next couple of years, we think it's going to be heavily weighted towards 28 nanometer, which is a, you know, at this point, like a 16-year-old technology, right? We don't feel that's an area of concern in any way, and we think that we'll be able to compete in that market.
We have a great footprint in 28 nanometer. We expect to compete well. What we've said, Vivek, is as long as we're able to serve an account, and we're not restricted from that, we're actually having good results in competition. Our share is good, and we think our, you know, we think we'll do well there. You know, when we compete in China, we're not just bringing a unit tool or a single tool that offers some result. We're also bringing a world-class supply chain with experience. We're bringing our service business offerings, which can help customers ramp to top yields and keep those yields in place. We're bringing, you know, innovations and continuous improvements to those customers. We think that, you know, when we're able to compete, we'll do well there.
Got it. One last question on kind of the WFE forecast. One thing that, you know, yourself and Gary have always mentioned is that the chance for the semiconductor industry in its march towards a trillion dollars, right, which then roughly mid-teens WFE intensity, you know, leads to that $150 billion or so, adding that. In the last many years, you know, China's WFE intensity has been far higher, right, than the rest of the world. As they become a smaller part, right, of the sector, can the industry still maintain this kind of WFE intensity?
We think so. You know, one of the ways we anchor this, and I'm sure people that are listening here do the same thing, we do look at those anchor points like a trillion dollar semiconductor business in 2030 or a $1.5 trillion in 2035 estimates out there like that. You know, people use the intensity equation and say, how much equipment should be sold in that year if we use the percent of equipment relative to those semi sales as an intensity number. Intensity numbers have gone up in recent years, you know, from mid-teens to higher teens, if you will, and now they're coming back down as this China effect. Frankly, some of the pricing on semiconductors has changed that equation. We think that's still a good way to think about it.
If you look at wafer starts, wafer starts for DRAM, wafer starts for mature logic, wafer starts for leading edge, increase every single year. And the devices, you know, do not magically show up with the same capacity. When we say there is mid to high single digits on the device side, that is going to be more factories, more equipment every single year. I think that is our expectation. The one place that has been different is NAND. NAND is, you know, the advances in bit density have allowed relatively the same wafer footprint to deliver the output, but there are still upgrades in NAND, and that is what you see in that business. Going back to the overall theme, yep, we are absolutely convinced you are going to see more and more semiconductors. You are going to see the footprint increase.
If you say, is it possible that one year will grow less than others or be below trend? Sure, we, you know, it's an uneven growth rate. The last thing to add back, just to circle back, is 2023 and 2024, equipment grew faster than semis. That's a lot of capacity that went into place in China. We know, especially in ICAPS, that there's a time when equipment's going to grow slower than semis. We're in that, you know, our last quarter was like that, and we're in that now, but we're still demonstrating growth overall.
Got it. Now, within the spending pie, how should one think about, you know, the wallet share that deposition and etch, right, your main businesses, can get versus lithography? Like what can tilt wallet share one way versus another?
Yes, we've said that as we look forward to the new technologies like gate all-around technologies, the next generation of transistor beyond gate all-around, we think will be CFET. The next generation of DRAM technology, 4F ². When we look at these technologies, we expect the materials engineering, maybe not just depth and edge, deposition and edge, but materials engineering in total, so non-lithosteps. If you think of films and treatments and measurement techniques and materials, you know, change like implant, if you think of all these different types of equipment, we think all of those will be growing in those nodes. It's not uneven in terms of the growth. The way I think about it is, you know, if you're building a larger house, because these are devices with lots of transistors, right, if you're building a larger house, it's not just wood and nails that increase.
It's everything. It's windows and piping and carpet and carpentry and all of these things. If you're thinking about the devices, it's going to be a continued growing mix of all those device steps.
All right. On the final question on China, what's the state of competition there? Do you see the domestic companies becoming a lot more capable than before? Because, you know, it is still a large exposure to this one region. If most of their buying is in more mature nodes, which people might correctly or incorrectly say that is, you know, less capable technology, do you think their domestic vendors are becoming more competitive than they've been in the past?
I think, you know, China last quarter was approximately 25% of our business. We said that for our equipment business, investors should probably think that's a number in the medium term that makes sense. It can be higher for the company if you add in display. It's approximate size. When we think about competition, you know, the first dynamic is there's some companies and some end device markets that we just can't serve. They sort of have free reign to serve those accounts. We do think just by getting the reps of having to serve those accounts, that those equipment improve. That's why when we compete against those types of competitors, I think they can, you know, they are improving and they get closer to the sort of metrics at the device performance level that customers measure.
When Applied comes, you bring those other elements. It is the top-level supply chain, the best components in the world, the most proven components, plus the services, plus the experience, plus a roadmap for how to improve the process from there. There is a reason why when we look at the accounts that we are serving, we are doing well from a share perspective. We think, you know, the share is holding up there. I think everybody, Vivek, in the industry, that is just competition, and we are okay with competition. We have got a roadmap of products and a pipeline of products that we continue to develop even for the mature logic space. That is important innovation for us, and we think that is going to be an important and growing market for Applied Materials going forward.
Got it. You know, one of the places Applied Materials has been very strong, right, has been DRAM, right? A lot of share gains over the last decade or so. How is DRAM investment this year? I think we kind of get the HBM is strong and everything else is mature. When do you see the non-HBM part also start to recover, kind of both, you know, so China I know restrictions, but ex-China, are you seeing any recovery there?
Yes, I think this one's an interesting one because if you look at the DRAM market year over year for Applied Materials, then you would say, oh, it looks flat. It looks pretty strong, but it looks flat. The dynamic there is in 2024, we were shipping business to two large Chinese customers, right? In 2024, in Q1 and Q2, our Q1 and Q2, I said that there was not only our normal business to those China customers, but $500 million each quarter incremental business to those customers. Now this year, Gary in our earnings call said that if you look at the international vendors of memory in DRAM, they're growing at 40% this year. What's happening is we don't have those China sales in 2025, but the international companies have grown much faster.
I think it's both the HBM effect, that's about 16% of wafer starts now, so that's nearly doubled in the last year. It's also just that utilizations of DRAM have clicked up and they're actually in a good healthy range at this point. Companies will be continuing to make investments.
Are you able to distinguish between what is going to HBM because, you know, it's essentially the same, right?
I think we do, as far as, you know, the way we can see into that equation is what's HBM and what's not. We get a view of that from the amount of packaging, the stacking of packaging equipment that we sell to each vendor. That gives us an idea of what sort of capacity it is. We have just got third-party information and company information that says the number of wafer starts that are being devoted to HBM memory each time period.
How much of DRAM, WFE, HBM?
The WFE, what I gave you, the 16% is the wafer starts. I do not actually have a breakout of the WFE. That is harder to tell.
On leading edge, you know, it seems like as, you know, people talk about AI and leading edge, but, you know, a lot of the AI products are actually on, you know, N - 1 or N - 2 nodes right now. TSMC has very good utilization there. Do you see continued investments in those, you know, 5/4 nanometer nodes? How does your opportunity change as TSMC is going to 2 nanometer, which they think is going to be a really big move for them?
When we look at leading edge foundry equipment purchases today, and not trying to be specific about any particular customer, I would say the right way to think about it is it's almost all gate all-around node purchases. Companies are not buying equipment today for existing technologies. They're buying equipment for the next technology, which, you know, each customer has a different naming convention for a gate all-around or a gate all-around plus backside power. I think that's the right model that you should have in your thinking that that's where all of the purchases are going today.
Okay. How does your opportunity change as they step through from 2 nanometer, you know, at some point, right, to more advanced nodes? Like is there a specific way to look at how your content changes?
I would expect, and I haven't examined particular nodes, but as you know, Vivek, each node typically adds process steps. The way we would view it, and it goes back to this materials engineering intensity. I think two nanometer or a gate all-around node generally had 200 more process steps than a three nanometer node as an approximation. Those process steps, it goes back to what you were saying. Some are deposition, some are etch, some are cleans, some are, you know, Epi, some are, you know, it goes through the treatment list, goes through the list of products that are needed to create that additional density and additional performance for a smaller transistor.
Yes, as we look at each node going forward, what a foundry sells when they're selling to their customers, you've got to sell more performance, better power characteristics, and hopefully less area, a smaller transistor, so they can pack more performance into a similar size chip, you know, 100 mm squared as an example. The only way they can offer that technology and hopefully charge a little more for it is if they offer that kind of performance improvement. That requires those extra steps that we're talking about.
Got it. On packaging, right, that's been an area of very strong growth, right? It has, I think, tripled in the last, you know, four years or so. How much of that is CoWoS and how much of that is HBM? And, you know, as people start migrating to HBM4, how does that change your opportunity?
Let's see. Last, in 2024, we said HBM packaging tools were about $700 million of our $1.7 billion packaging business. This year, I think the way we've described it is that business continues to operate at approximately the same level. The way I think about it is our HBM is, we're still selling HBM at a good rate, but not quite at the same rate as we did at the tail end of 2024, because that was the sort of initial capacity buys for some of those customers. They continue to invest, continue to add capacity. It's a little bit slower rate than it was at the very end of last year. Shifting to a longer range view, this is where, you know, our CEO, Gary, describes this as one of the intense competitions across the industry.
When you think about, you know, everybody reads articles about the, will we have enough electricity in the world to serve all the data centers? The leading cloud service providers and leading data center product developers are thinking about how do we make these devices more energy efficient, get a lot more performance, and get the utilization of every single component up very high and have it consume less energy. There's going to be innovations, Vivek, on, you know, the bonding techniques, the interconnects, the substrates. There's going to be innovations across that whole stack, as well as the devices themselves, which goes back to process technology. Every single generation going forward, and Applied Materials is involved in all of those pieces. We're investing in all those pieces. Packaging is important business for us.
That is why, you know, we say that we expect it to double again over a period of time. I would say, I think 3-5 years for that.
I see. On gross margins, you know, one of your peers has set a 50% gross margin target, even though their scale is smaller. And I understand everyone has different mix of businesses, right? So it's not exactly apples to apples. But how do you think about that as a potential target for Applied Materials? Like what stands in the way of Applied Materials getting to those kinds of numbers?
On gross margins, you know, a year ago, we were 47%, mid 47s, and we had set an internal target. Can we get to 48% and get over 48% this year? We've demonstrated that so far, you know, 49.2% in the last quarter. This quarter, we're guiding in our outlook for our Q3, we're guiding 48.3%. What I would say first to, you know, investors and analysts in particular is we've got three reportable operating segments, right? Our display business, you know, we're looking at the opportunity of OLED and that business could grow. It's a nice call option on a new technology in that business. It has lower gross margins than our core equipment business. Our services business, which is our fastest growing business at low double digits, is what we typically describe that as from a growth rate perspective in the services business.
That also has lower gross margins. I am just reminding people when they look at our total gross margin, there are components there. The core equipment business, we have made improvements. That is why we are now in the 48s. We expect to continue to make improvements. We have worked hard on pricing over the last two years. We have made significant progress in pricing. We have always had good cost improvement programs. The way I think about it, Vivek, is on that equipment business, you should expect continued improvement sort of, you know, at the pace that we have been demonstrating.
I see. On the services side, right, which is a very, I think, attractive part of the semicap industry, which often gets more appreciation than it should. Talk us through what the different building blocks are of your services business. How much is tools, repairs, you know, refurbs and so forth? How much is cyclical versus how much is kind of, you know, more predictable subscription type contracts?
Good, good question. First, you know, thinking about the investor perspective, one thing that we've done is we've correlated highly, I would say, I would not say tied, but correlated highly our dividend payment with the profits of our services business. What we're thinking there is our services business is very much a recurring business. 85% of that business we put in the recurring category and 2/3 of that we say are under actual contract subscription. We have high renewal rates on those contracts. We feel very good about the recurring nature of that business. When you think about our dividend, we're talking about low double-digit growth rate for our services business over time. I'll come back to the drivers. We're going to continue to correlate that.
We had, we raised our dividend 23% a couple of years ago, then 25% last year, and now this year 15%. We are kind of lining that up to highly correlate that. That will help investors, I think. One of the reasons, Vivek, you said, yeah, it is kind of sometimes under notice. We are trying to increase the emphasis for paying attention to that business by doing that connection. The business itself, a few drivers, why is it low double digits when our equipment business is growing, you know, more slowly than that, maybe mid to high single digits over time? The reason is every time you ship a tool, you expand your, you know, sort of field of tools that are available to serve, right? Every single day we ship a tool that grows our installed base.
The second thing is that customers are more likely today to buy services partially because of the government contracts that they have or government incentives that they have where they're building in new cities or new areas. You know, just to have that experienced labor is a big help to them. Third, we're adding services per tool. We're raising the revenue per tool from a services perspective. A lot of that is AI driven, is services that come from helping the customers. We see all of our tools and the way they operate. We're able to help design the best parameters for operation for that equipment. Then we can share that with the customers so they get better performance as they bring those up.
Those are the dynamics that add together to get to the low double digit growth rate that we've articulated for the services business. If you try to say, okay, it's recurring, 85% of that business or more today is recurring. How much of that, you know, is absolutely run rate or how much of that can be volatile? There is a little bit of volatility in there because as utilizations go up or down at our customers, when utilizations are lower, they need fewer spares. You know, sometimes they idle a tool. That takes down that rate. I think, you know, that's a lesser component of the equation.
Is there a higher services component proportionally in China versus non-China?
We've said it's roughly same level, proportional, similar mix.
Okay. And then finally, you know, I think what we have seen with Applied Materials is a very strong capital return program, right? Especially the last few years, you've grown dividends, you know, I think like mid 20s %, right? In a number of those years. How do you think about, you know, are you targeting a certain dividend yield or do you think that, you know, it's this kind of consistent double digit growth and then buybacks is really, and you had a big buyback plan, right? Or addition you had in March.
Yeah, great, great question. First, let me emphasize just the strategy of the company because it informs the distribution side of the business from a capital perspective. The strategy and, you know, a lot of people will know this in the audience. We're investing in inflections and we are heavy R&D investors. I mentioned that we're building a new lab in Sunnyvale near our campus. It's about, you know, two blocks west of our campus in Santa Clara. This is a long-term collaboration lab with our customers. Vivek, the idea is that, you know, as you keep pushing the frontier of semiconductors, it's going to require more and more collaboration across the industry. These devices are more complex and take new techniques, new materials, et cetera.
We're investing in that process of bringing our customers in and working closely with us on the longer term roadmaps. If that works well, then we'll have, you know, good positions in those new processes that come up. First rule of capital allocation is fund that roadmap. We have a 25%-35% return on invested capital. When we look at our opportunities, we're looking at a lot of good opportunities to invest in and you see us making that investment. R&D, CapEx first, we're making those. You can see that. When we turn to excess profits, we've said we'll return 80%-100% of our excess profits to shareholders. We just talked about dividend. I would say there's not a target from a yield perspective. The mental model I have is we're going to grow that at that services profit level.
That's the way we're going to think about it. It should translate to that low double-digit growth rate after the 23%, 25%, and 15% of the last couple of years. The rest goes back through share buybacks. For the share buybacks, we're typically in the market every single quarter and we try to be a little bit price sensitive in the quarter. That's our philosophy.
One final question that, you know, I get a lot and would love your perspective. The semicap industry has such outstanding, you know, return metrics. Why does it trade at lower valuations than analog companies?
Wait, I should ask you that question. You know, I think there's obviously one of the things we think of, Vivek, when people ask us the question about the multiples, and that's why I would, you know, ask the audience, look at the growth rate of the business the last five, six years. Look at semiconductor equipment and semiconductor devices over the last 10 years. I always think that there's investors who may have met people in our industry or companies in our industry 20, 25 years ago when it was a lot more volatile and people come to us with sort of that perspective. I think the volatility hasn't gone away, but because there's so many more markets that are using semiconductors, you think of, you know, it's not just PCs and smartphones. In the old days, not just analog and, you know, MCUs.
Now you've got the AI data center, you've got people talking about robotics, you've got, you know, companies working on smart glasses, you've got all sorts of new things. As we know here, everything is using more AI and it's pulling on more semiconductors. It just goes back to we think there's going to be secular growth across this whole equation. I can't promise you that it won't be uneven growth and it just grows every single quarter or every year, but we're certainly investing for the growth. I would say that, yeah, people should look back at the history and make up their own minds on what the right multiple is for this type of business.
Excellent. Yep. Thank you so much, Brice. Appreciate it.
Good to see you.
Thank you.