Guest today, Gary Dickerson, the President and CEO of Applied Materials. Before I start, I want to mention, if you have questions that you'd like to get asked during the presentation, you should have a link to the pigeonhole form where you can submit those. We'll have time for Q&A at the end. As many of you know, semi-cap is a sort of an area that's sort of near and dear to my heart. In a prior life, that's what I did. I love this space, and I think Gary loves it too, you can tell. The industry has really been enjoying a renaissance, I think, over the last, you know, five-plus years. I think it's really coming into its own now. It's still cyclical, we all know that, but I think both the underlying industry growth and it seems industry capital intensity have inflected materially higher.
I think the contributions from companies like AMAT have become more important than ever as materials-driven innovation moves to the forefront of process technology development. To that end, the semi-cap space really has been top of mind for many of my clients. You know, there are all the questions on near-term, you know, lagging versus leading-edge demand and China dynamics and competition, and now it is tariffs and geopolitics. I am also increasingly getting people looking at the longer-term potential of this industry as they start to view it more in secular rather than purely cyclical terms and looking at those new technology inflection points where companies like AMAT are key enablers and beneficiaries. To tell us all about it, it gives me great pleasure to welcome Gary. Thank you so much for being here.
Thank you, Stacey. I want to say I do love this. I mean, this is so much fun. What we do in this industry, we always make the nearly possible possible, and what we do is just like magic. I tell everybody, in your smartphones, you have an application processor chip that has tens of billions of transistors and 60 mi of wiring in that one single chip. You think about it, a wire is 60 mi long, you're moving data lightning fast with almost no resistance. That's pretty damn close to magic.
I mean, I always say, I'm amazed any of this stuff works at all. Like, it really should. That's why it keeps me getting up in the morning.
Look, Gary, I want to talk, I want to start maybe at a high level. You know, everybody's always concerned about, like, what's WFE outlook and this and that, but I'd say over the last couple of years, things have shaped up far better than anticipated for the industry. Everybody used to have this sort of aspirational $100 billion kind of WFE. We're there, we're kind of there, and things look pretty good. I was just wondering if you could maybe talk a little about, over the last several years, some of the historical contributors that have taken this industry to where it is.
I mean, it wasn't that long ago, we were talking through numbers that could have been $20 billion or $30 billion.
You know, we're clearly way above that now. I mean, like, what's been driving it, whether it's leading-edge versus innovations, lagging-edge demand, China, packaging, you know, memory inflections, and how AMAT has actually been contributing to those and to, I guess, to help those WFE numbers drive higher?
Yeah, I think for Applied, you know, we've actually grown revenue five consecutive years coming into 2025. The first three quarters of our FY 2025 year, we're up about 7% year on year. Our semiconductor business in the quarter we guided is up 10%. We're growing our service business at a double-digit compound annual growth rate. That's a $6 billion business today that's growing at a decent rate. I think that if I look at my time over the semiconductor industry, I've been fortunate to be at three different companies that all achieved incredible growth. There's been a major change. You went from mainframes to PCs to mobile social media. Everybody has a PC or a computer and camera in your pocket. You know, where we're going with AI and robotics, these are the biggest productivity drivers of our lifetimes.
If you think about aging workforces around the world, U.S. productivity has been growing at about 1% over the last several years. Again, AI is by far and away, but robotics, I would say, is another one that are going to drive tremendous growth going forward, significant increase in computing demand, and computing innovation. We'll talk about that later. What I've seen is just this increase in pervasiveness of semiconductors in every aspect of our lives. Really, where we're going is going to be way more significant than anything we've ever seen. I think TSMC is one of our largest customers. They talked about 40% compound annual growth rate over the next five years in the AI data center. Data center wafer starts will pass PCs and smartphones as a driver. Those are very large chips. And then AI.
When do you think that happens, by the way?
That's in a few years.
Very soon.
It's going to happen. AI is really driving. There are four key technologies: the leading-edge foundry logic, compute memory with DRAM, high bandwidth memory and packaging, and Applied is the leader in all of those different areas, significantly larger than our next closest competitor in all of those different areas. We've been gaining share in all of those different areas. You know, we can talk about, you know, what is needed for computing innovation going forward is more significant in performance and power than anything we've seen in the past. I think, Stacey, it's really this pervasiveness of semiconductors in everything around us.
That's true. You know, everybody sort of talks about this, you know, trillion dollar semiconductor by 2030. I'll be honest, I don't know if we get there by 2030, but we will get there. We're not that far off. I think we did $630 billion or something last year in overall semis. Have you guys ever sort of given, you talk about semi-cap, I think, outgrowing the market. You talk about AMAT outgrowing semi-cap. I was wondering if you could maybe give us a little more color on how you view those drivers of outperformance. Like, where do you see capital intensity actually? We're probably in the upper teens right now.
Yeah, I'd say mid-teens. I think if you have, you know, people forecast the semi market to be a trillion or higher by 2030. I really do believe if you look at computing demand, again, these drivers are more pervasive all the way from the AI data center with the 40% compound annual growth rate. Everything around us to feed the AI, you know, there's more innovation, more content, thousands of chips in an automobile, you know, all of that is driving significant innovation going forward. Really, more and more of that innovation is coming from materials and materials innovation. I would say kind of capital intensity, mid-teens is kind of the right way to think about it. When we're working with our customers, I've met four of the top five customer CEOs in the last six weeks.
We're in very deep relationships that are very different than the way they worked, you know, five or ten years ago. It is really focused on high velocity co-innovation. For our customers, what they care about are design wins for those future architecture inflections. For design wins, all of their customers will run test chips. They care about performance, yield, and reliability all together. If they do not have that, cost or price does not matter. If you look at the value, I mean, everyone understands that, you know, AI is consuming so much power. There is tremendous value in the race for AI leadership for power and performance. You know, we could talk more about that. Materials intensity is rising.
If you look at big inflections like Gate All Around, the next generation transistor to process the data, or backside power distribution, where you take the power lines from the front of the wafer to the, or the front of the chip to the back of the chip, you get 30% area savings, up to 30%, better power, better performance, no change in feature size, same feature size. In DRAM, compute memory, that's growing. I think the top three leading companies were up something like 40% growth. So the compute memory is growing very, very quickly. The next architecture inflection there, 4F squared, or some people call a vertical channel transistor, again, it's much more materials innovation rich versus lithography. So, you know, the percentage of spending for materials innovation is going up in all of those different markets.
In advanced packaging, where Applied is also the leader, it's a $1.7 billion business for us. We said that's going to more than double over the next few years. I think that materials intensity will remain in that mid-teens range. I think the great thing for Applied is that when we're working with our customers, we're working a decade out in the future. We have integration innovators inside Applied that are working with those customers to shape those future architectures in power and performance. For every one of those technology nodes, there's something like five or ten absolutely crucial innovations for those architectures. Applied is the leader in those key innovations. That puts us in a really great position.
Got it. No, that's great. Let's dig into each of those. You listed three or four kind of things that you seem really excited about. Now, let's start with like leading-edge logic and Gate All Around. You've given some numbers for how big that was and how big that's going and where it has been. I'm curious with that. I'm also curious just what are the innovations within like Gate All Around and leading-edge logic that AMAT is bringing to the table right now that others are not?
When you're building these, whether it's a new transistor or wiring, there are really hundreds of process steps that are needed. As I mentioned, you know, Applied has by far, if you look at Applied versus the next closest company, by far the largest percentage of those steps, and especially the critical enabling steps. What's happening as you go to these future innovations, you know, some of these films are a few atoms thick. It's really amazing. You have to control the interfaces. There's interface engineering. There's innovation in the materials. There's innovations in how you shape the structures. There's innovations where we have billions of dollars in materials modification. Again, those are areas where Applied has clear leadership.
We are very unique that, you know, when we're working on those innovations with customers, you know, we're in with their integration teams shaping those innovations in super high velocity relationships. I mentioned in the wiring on the backside power, Applied has very, very high share in wiring. We have one platform that Applied, about 30% of our revenue comes from integrated technologies where we have multiple capabilities on a single platform. That 50% reduction in wiring resistance, there are seven technologies: selective ALD, PVD, CVD, copper reflow, metrology, a number of different things under vacuum. The reason you have it under vacuum is because when you go to air, the interface will oxidize. Again, some of these films are a few atoms. This is increasing. All of this co-optimization as you move through these process flows is harder and harder, more complex.
That means we have deeper relationships with these customers. It is very different than the way the industry worked in the past. Really, our goal is to work with every one of our customers, whether it is in logic, DRAM, NAND, ICAPS, packaging, to enable their competitiveness. It puts us in a very differentiated strategic relationship with all of those different companies.
Got it. You're going to force me to jump the gun on something, but something you said makes me want to go there now. We can come back to some of this stuff in a minute. On the integrated processing, there are two things I heard there. One is clearly you're selling a value-based solution, right? You're not selling a process. You're selling, you know, I want to make a film that's this thick with this interface and out of these and seven processes under vacuum. Two questions. One is, how do you price for that sort of value add? You said 30% of your revenue is coming from these integrated types of processes. How do you price for that? Are you capturing, think about capturing the integrated value from that?
The second question, just in general on how you work with the customers as these things are getting more and more complex. It brings up like the Epic Center that you're building and some of the other stuff you're building around the world, like to do this like customer co-development and customer integration. I'd love you to touch on those points.
Sure. Yeah. Again, you talked about how you love this. I absolutely love this. I work every single day with our teams on all of these different kinds of innovations. Those integrated platforms are accretive to our margins. I think that even beyond that, if you're enabling power and performance and design wins, whether it's in memory or high performance logic or power electronics and ICAPS, it puts you in a different strategic relationship with the customer. You know, we're able to capture more value because these are the key tipping points for design wins for them. It also enables us so that when we're having discussions with customers, you know, we're allocating our innovations, our innovators for them. It enables you to win ties in many other parts of your business.
We're innovating with these enabling technologies that are absolutely crucial for these future architectures. We're also driving, and there's billions and billions of dollars of new products in our pipeline that strengthen our ability to enable the architectures, expand our total available market. We have products in the pipeline that are significant cost innovations so that I'm in the zip code where I can win the ties, you know, for those other parts of our portfolio. To the Epic Center in Silicon Valley, the equipment process innovation center, the concept there is all of these businesses are a race for leadership. The race for AI is the biggest race of our lifetimes. Everyone understands that power and performance is absolutely crucial.
We have been, as I mentioned, over the last, I'd say, seven years, building capability inside Applied where we have integration innovators across all device segments that are very unique, world-class, and we're co-innovating with all of our different customers. A real important strategy we're driving is high velocity co-innovation where, you know, we have our innovators co-located with customer innovators so that we move at higher velocity in parallel versus a serial type of process.
Your facilities are in theirs.
We do it, you know, both ways. Today in our current technology center, we do have customers coming, whether it's in Silicon Valley or we have the really only full-flow packaging line, R&D line in the world in Singapore. We have leading customers coming to those locations. The concept with Epic is really to innovate the way we innovate, to drive even higher velocity because that's what determines winning and losing in any of those different markets. We will have top innovators from our leading customers located inside the Epic Center. We have all of the capability. When we're in, you know, in those types of facilities, we're running short loops where you have test vehicles for transistors, test vehicles for wiring, test vehicles for new patterning applications, test vehicles for new DRAM architectures.
We run at very high velocity and it's all about learning rate and the customers are co-located and will be co-located inside that facility. The great thing for them is that it accelerates their time to market for design wins in power and performance. They're spending an enormous amount of money for every R&D technology node. It makes that if they can bring the device to market faster, it gives them higher R&D efficiencies. Of course, for Applied, it enables us to design in their new architectures, our equipment, and our advanced services. You know, when we came up with this concept for Epic, this was really in alignment with our leading customers and helping to shape that platform. We'll have innovators there. We have customer space where we could have certain key innovations that we're co-developing with certain customers.
We have other partners that are part of that innovation network with Epic. The focus is innovate the way we innovate, go to even higher velocity, and improve R&D efficiencies.
How far ahead are the customers running in, say, Epic or in the Epic Center versus when they're actually delivering those products to market?
We're working with people really maybe four technology nodes out in the future. You know, obviously there's an enormous amount of focus on N plus one and N plus two because that's life and death. You know, when they're thinking about design wins and power and performance for every one of those technology nodes, we're engaged with leading customers every week. We have, again, a list of prioritized innovations that we're driving. Those are the critical tipping point innovations for them. You know, when you're thinking about materials innovations, three or four technology nodes out in the future, CFET technology, for instance, where you're going vertical, again, more materials intensive.
Stacking transistors on top of each other.
Stacking the transistors vertically. Again, there's tremendous innovations. This stuff is so complex in creating those materials, shaping those structures, modifying those structures to end up with the right electrical performance and power. You know, it takes time. It takes a long time to bring those innovations to market. The earlier you start working, you know, the earlier you're going to bring those technologies to market. I mean, again, Applied is coming up with innovations that people don't know are possible. We're doing the same thing in power electronics. Power electronics are really important for electric vehicles, renewable energy, the grid. We're bringing to market major new products that expand our TAM in power and also innovating in new power architectures.
This is going out several years into the future to bring those technologies to market faster for the innovation and then also the commercialization because just because you have the power and performance, if you can't yield or you don't have the reliability or you can't build it at the right cost, you can't insert that technology.
Got it. You know, maybe your comments on power there are a good segue into your ICAPS business. ICAPS is Industrial, Communications, Automotive, Power, and Sensors.
It's ICAPS, Communication, Auto, Power, Services.
IOP, got it. And you know, investors tend to think of that as just your trailing node or lagging edge, although I think it's a lot more complex than that. And you're even still, you're still innovating in there. It's not just like it's old. Maybe you could talk about, this has clearly been a key focus area for AMAT. You know, it's in a bit of a trough right now after some very strong years and there's some of the China dynamics that we could talk about. Maybe just talk a little bit about like why that was such an important initiative for AMAT. What got you focused there and where are you focused now?
If you look at the business, about one third is leading edge foundry logic. One third is memory, more weighted towards compute memory versus storage memory. One third of the market is ICAPS.
This is the market, not AMAT necessarily.
Yeah, it's the market overall. The wafer fab equipment market is split in those different categories. April 12th, 2019, I still remember, was the day we formed our ICAPS group. You know, we could see that this market was going to be a significant market. You know, if you think about it, I mentioned how, you know, AI, robotics, industrial automation, all of these things are going to be transformative, drive dramatic productivity for every country, will be at the foundation for competition for every industry. The content, the computing content is increasing in all of those AI edge devices going forward. Again, we knew this was going to grow at a very high rate going forward. Just like we have on the leading edge, we brought into the company some outstanding integration innovators for Applied.
We have, again, this is really very largely enabled by materials innovation in this particular space. There is tremendous innovation still. I mentioned in power, if you look at what is happening in automotive or, you know, what is happening in the grid, there is going to be tremendous innovation that is happening there. In CMOS image sensors and a number of these different markets, this has been a focus for us for a number of years. We have, again, just like we do on leading edge, these deep strategic relationships with customers. Since we formed the group, we brought 20 major new products to market. We have gained several points of share in that ICAPS space. I would say that our innovation pipeline going forward is actually stronger than ever.
You know, we're continuing to bring new technologies to market that will expand, significantly expand our total available market in ICAPS, enable new chip architectures. Again, in terms of that strategy of winning the ties, we're also driving significant cost innovations in major products within that space. This year, if you look at ICAPS, it is lower, you know, as a percentage of our total revenue being offset. You know, we're still growing with the leading edge foundry logic and memory business. Going forward, we believe that ICAPS will grow. Again, this is all the edge computing types of applications at a 5% or mid to high single digit growth rate going forward. For us, how we compete inside of that market, whether it's inside China or outside China, again, same playbook. We're innovating with new capabilities.
We're expanding our total available market with significant new products that don't exist today. We're bringing those cost innovation products to market and innovating with new chip architectures inside of ICAPS.
Let's talk about China. There are two areas I want to hit. One is, you know, clearly the geopolitical environment and export controls and tariffs and everything else. What has been the impact of that, number one? Number two, I really want to get your opinion on some of the local Chinese players who are taking share because they have no choice, right? I mean, you're impaired in terms of what you can sell and what they can buy. How do you view, I guess, the state of local Chinese competition? How do you deal with that in the wake of the, you know, the regulatory headwinds that for better or for worse are on you?
A lot inside of that question. Let me start with export controls. The export controls have.
Several waves, by the way. Yeah.
Yeah. Several waves have been largely focused on leading edge technologies and especially technologies that are focused on winning the race for AI. You know, that has restricted some part of our ability to compete inside China. If you look at leading edge foundry logic, you know, we really can't sell into any of those technologies. In memory, you know, we really can't also sell into the memory market inside China. Again, despite the really significant decline in memory in China, we offset that by significant growth in leading companies outside of China. Those restrictions certainly have helped the local Chinese companies because there we just can't compete in those segments of the market. Again, you know, our strategy in every one of these device verticals is to really work. I always want to run the fastest with the leading companies.
I want to know who's moving fastest on any innovation, and then I want to partner and be ahead of them. The same thing is true with ICAPS. We have very deep partnerships. We will be expanding some of those partnerships soon in terms of ICAPS innovations and then bringing these new products to market, cost innovations and architecture innovations in those businesses. I have high confidence, you know, ICAPS will grow at a kind of mid to high single-digit growth rate. I have high confidence in being able to compete where we can compete. You know, in terms of the Chinese companies, again, they are filling the gaps in the areas where we can't compete. They're coming out with products, not so much in the critical enabling technologies, more in non-critical technologies.
Again, that's why we have a playbook where we have what is absolutely crucial for companies to achieve design wins in their business, crucial to how they compete. That enables us, along with cost innovations, to compete across our entire portfolio. That's the strategy there. Relative to tariffs, you know, during the COVID timeframe, I kind of felt like I was the Chief Supply Chain Officer in Applied because even though we made the equipment that makes the chips, we also had trouble getting chips into our supply chain. I was calling many of those companies, CEOs, accelerating chips for our equipment. During that time period, we focused on really innovating in our operations, in our supply chain, at all tiers within our supply chain so that we could remain the most trusted partner for our customers and have supply chain resilience.
Applied has a platform where in our operations, we can manufacture in the United States, we can manufacture in Asia. Regardless of what happens with tariffs, the key thing for us is to really optimize our agility so that we can remain a trusted partner for all of our customers. There are lots of things we've driven. We're driving sustainable cost innovations that is helping us to drive our margins to the highest level in 25 years, but also tremendous innovation or tremendous agility in our supply chain so that, you know, I can't anticipate exactly how all of this is going to play out. It seems like it changes on a pretty rapid basis. Yeah. The main thing for us is we want to be agile no matter what scenarios emerge.
Got it. You know, on the margin front, like I ask you this question every year, but you know, is there any reason your margins have to start with a four forever? I always wonder why.
I would say the answer to that is no.
I'm not asking for a timeframe or anything, but you know,
last quarter we hit $49.2 billion, which was the highest since 2000 within Applied Materials. We are driving sustainable cost innovation inside the company. Also, from a value capture standpoint, again, as we're bringing these enabling technologies to market, you know, our customers are in a position where these innovations that they're delivering really shape competitiveness of industries. They shape overall productivity growth, and they're highly valuable. You know, we are focused on being the most valuable partner for all of those customers. That enables us to capture more value with those companies. It is really two things: sustainable innovations in cost and then innovations in value capture. I do believe that's sustainable.
I do believe there's no reason that the margins won't start with a five going forward.
Got it. I want to ask about DRAM. This is a market that's been really good. I can't remember, 10 points of share over the last 10 years or whatever the timeframe is.
Yeah, 10 points in a little over a decade. We've gained in terms of DRAM share, overall spending from DRAM customers.
Can you tell us a little bit, like, how has that happened and what gets you excited in that space going forward? Is it all about, you know, compute memory and HBM? Is there, like, what are the innovations?
Yeah, I think there's a difference between compute memory and storage memory. If you look at AI servers and the demand drivers for compute memory, you know, I mentioned, you know, from our leading three customers, our business is up dramatically. Strong double digit growth year over year. When your high bandwidth memory is an increasing part of the wafer starts in DRAM, you need to start three or four times more wafers for HBM because chip sizes are larger, yields are lower. We think that that part of the market is going to have, along with high performance logic, very healthy growth rates. For us, we have in the memory cell a lot of strength in enabling customers to continue to scale those technologies. Also, high speed IO for memory chips is incredibly important, especially for AI.
You know, you're moving the data on and off those chips, and you want to move that as fast as you possibly can. DRAM has been adopting logic processors in the periphery of those chips. We have, for transistors, tremendous leadership. We have leadership in wiring. How you're moving the data on and off the chip is an area where Applied has clear leadership. That's enabled us to grow our share. As you go forward, they're going to continue to innovate in the CMOS, the IO portion of their chips. They're going to go to this new architecture, 4F squared, in just a few years. That architecture is more materials intensive. What we've said for the next architecture, we'll capture over 50% share. We'll gain even more share as they go to the higher speed IO, the new vertical channel transistor.
You know, that positions us for additional growth there. In HBM, you know, that's another part of compute memory where Applied is the clear leader and we have clear strength. You're right. A little over a decade, we've gained 10 points of share, positioned for many points of share gains going forward to enable those architecture inflections.
Are we still to talk about 3D DRAM or?
Yeah, I think 4F squared is the next big architecture inflection. That'll happen in a few years. Typically, what happens is that when someone brings a new architecture to market, you go at least two or three technology nodes. 3D DRAM would be after that. And that really, it's different than when you went from 2D to 3D NAND because the materials are not dielectric materials. They're materials that where Applied has clear leadership when you move to that structure. There are incredible innovations. This is where I talk about, you know, working a decade out in the future. You know, some of these processes to enable that chip architecture, you have to improve by an order of magnitude. And it is, it's nearly impossible. I'm highly confident we will make it happen. But again, it'll be after the 4F squared inflection.
Got it. And you know, just to touch on the storage memory on the NAND space, you know, you do not play as much in like in dielectric edge, which is more important in that space. At the same time, like that market in the near term is starting to pick up. It looks like more upgrades, maybe all upgrades versus capacity. What are your thoughts on that NAND upgrade cycle and where does AMAT benefit? Because you are actually benefiting more than I even expected you to. You had a pretty, at least in the near term, a pretty sizable inflection in NAND this quarter on the back of that.
Yeah, I think, you know, in DRAM, also in conductor edge, we've grown a lot. We've grown in, you know, many different areas. I talked about that. In the patterning for NAND, and then also, you know, you have the CMOS bonded to the array. Those are packaging process steps where Applied, again, is the leader in packaging technologies. Those are areas where we are benefiting. You're right. I think that storage memory doesn't have the same drivers that you have with compute memory. When we think about what the industry looks like going forward and their % spending, we think memory still remains about a third of the total spending, but more heavily weighted towards compute memory versus storage memory.
You know, the areas for us, that business did grow for us, is growing for us this year, a fair amount, but it is off a very low base of where it was before. In NAND, you do not have the big architecture inflection like you have in some of these other markets. The industry went many years ago from 2D NAND, where again, it was LISO enabled scaling, but you ran out of electrons in the cell. That was hitting both technical and cost barriers. You went vertical, just like you are going vertical in a lot of these other segments going forward. As you said, the most enabling technology is how you drill that memory hole in dielectric edge. We made a decision years ago to focus more on conductor edge.
We did not play as much in that part of the market. The patterning, the hard mask films, packaging for CMOS under array, those are the areas where we have strength.
Got it. Got it. Being remiss if we did not touch on services, right? You talked about it a little bit. How do you view growth in that business? In particular, I would like to hear your thoughts on, like you talk a lot about the move to, like, subscription contracts with customers, service contracts. I guess how does that both, I guess, drive growth and maybe smooth volatility as well as, like, what does it mean for margins?
Yeah, so service, our core service business is growing at a double-digit compound annual growth rate. Since I've joined Applied a little over a decade, we went from about 40% of our service revenue coming from long-term agreements, subscription revenue to about two-thirds today.
Two-thirds.
That's still growing. I'm actually very optimistic that we're going to continue strong double-digit growth in services going forward. If you think about these architecture inflections, you know, I mentioned that, you know, performance, power, yield, reliability, you know, for customers to ramp these very complex integrated platforms to ramp these new architectures, we have tremendous service innovation that enables them to ramp to high yield faster with our service technologies. You know, when you're developing a process, it's incredibly complex in terms of the optimization. Then you want to transfer that to hundreds of chambers in high-volume manufacturing. Being able to go from one golden chamber to all of them golden is very, very challenging, very difficult. Time to ramp is important. How you can hit the high yield, high performance, low cost, there's tremendous service innovation in our pipeline.
We have increased the number of tools that we have connected remotely. We have thousands of tools today connected remotely so that we can deliver these advanced services immediately versus putting someone on a plane and going to a customer site. That is growing at a significant rate. Off of these tools, there is a tremendous amount of data. We are innovating with sensor technologies. All of this will enable us to help customers ramp these new technologies at much higher velocity, higher yields, and better performance than what they have had in the past.
Got it. Wow, we're under 10 minutes left. Do you want to go to the lightning round?
I'm happy to do it. Whatever you want.
Got it. Here's a good.
All semi-cap companies always say that they're going to outgrow WFE.
I'm surprised. How, I guess, is that? I mean, you believe the same for Applied Materials.
I would say I was wondering about that.
I did take advanced math. I don't think it's possible for everybody to outgrow, you know, in terms of a market share standpoint. I think the key thing for us is that we have this unique and connected portfolio. I mentioned that we're four technology nodes out. We have deep engagements with customers. We have extremely high visibility on those architectures, enabling the architectures we're designed in with our equipment and our services. You know, we certainly look across the entire portfolio and all the innovations that we're driving. As I said, the thing that is really different than the way the industry worked a few years ago is that the way we engage with customers for them is life and death to bring those architectures to market ahead of others.
Materials intensity is going to increase going forward. That puts us in a good position. We have by far the biggest engagement and most strategic engagements with our customers. I do not know about the other people, but I have high confidence that we will outperform going forward.
Got it. Among the technology areas you're investing in today, which have the greatest potential to open entirely new product categories for Applied Materials?
You know, I have a lot of excitement around a number of different areas. AI, I would say, is the biggest driver overall. I mentioned in high performance logic, you know, that business is going to grow at a very fast pace. If you look at the combination of Gate All Around and backside power, we said we'd capture over 50% of those inflections. That will grow our total available market about 30% or our revenue, 30% for an equivalent number of wafer starts. We're really well positioned there. DRAM, you know, compute memory, I said that's another area along with HBM. We're really well positioned there to outperform as we go forward. There are incredible innovations that are critical for 4F squared and future generations of high-bandwidth memory that are going to be needed for AI. I'm also extremely focused on packaging.
You know, if you think about an AI server, you have 72 GPUs connected together that consume about one megawatt of power. There is going to be tremendous innovation in how you connect all of those compute components together. And Applied, I said that is a $1.7 billion business. We will double that in a few years. It is going to keep going because the way you move the data at high speed and low power is absolutely crucial to the way you connect all of those chip technologies together. Applied has a number of innovations that are in the pipeline that strengthen our ability to enable those architecture inflections, expand our total available market, and really reshape some of these architectures.
How do you think about a natural ceiling for service intensity?
I think, you know, we've grown our subscription revenue from about 40% to two-thirds. It's going to keep going. You know, I think that.
Your services revenue per tool has been going.
Our service revenue per tool, that's a key metric that we look at. Our ability to connect these tools, that's going to keep growing at a very, very high rate. We have several thousand that are connected today. Our ability to create value for service, again, ramping these architectures really hard to high yield, maintaining high yield, you know, and high volume manufacturing, that's also very difficult. We have been growing at a double-digit compound annual growth rate. I'm actually even more optimistic with what we're doing going forward. I mentioned also AI. We're using AI in a number of different areas within site Applied. There are some tasks within our service organization we can improve that consume a lot of hours, a lot of manpower. We can improve an order of magnitude.
That gives us the ability to deliver higher value services at much higher productivity. I have high confidence in that double digit compound annual growth rate.
I'm going to ask this one anyways. Maybe I shouldn't, but I'm going to ask it. You recently personally purchased.
Be brave, Stacey. Go ahead.
You recently personally purchased $7 million of shares. What does this equate to your conviction level?
We purchased, oh, me personally.
You personally.
Yeah. You know, look, I think Applied is in a great position. You know, I look at the market opportunity. We talked about computing, computing demand. We're at a point, and I look at my career over decades in the semiconductor industry, you know, computing demand going forward is really at the foundation of how every company, every industry competes. I am highly confident that this will be the biggest productivity accelerator with AI that we've ever seen. It's really important for competitiveness of every single country. I am very optimistic on edge AI and, you know, content growth in all of those different kinds of applications. You know, again, I'm in deep partnerships with all of these different companies. I mentioned I met four of the top five CEOs in the last six weeks.
We're working a decade out in the future to enable their new chip architectures, which, you know, are absolutely crucial for them in how they compete. I can see very clearly the market growth. We can see very clearly our position in those markets. You know, that gives me conviction to put my own money into Applied stock because I think it's going to be a great return.
Got it. Maybe time for one more. Why did you need to invest equity in Besi and Kokusai? What do you get for those? What do those investments provide for you?
I think, you know, I talked about these integrated platforms and co-optimizing for architectures. You know, when you're innovating with these new architectures, it's incredibly difficult. You know, that process flow, you have to co-optimize. The interface engineering from one step to subsequent steps is very important. Besi, they're the world leader in hybrid bonding. Again, we created this partnership with them more than five years ago. This technology is going to be really important for power and performance going forward. Again, you increase the IO density, you shrink the length of the wiring, that improves power, that improves performance. We have an integrated platform with their bonder with five Applied technologies into this integrated platform. It's really taking the best of what they bring to the market with the best of Applied Materials.
In a different world, if we could do M&A, we would have likely done that. But, you know, that's not possible. So creating this innovation network and the unique and connected capabilities, we have the most unique, the most connected portfolio inside the entire industry. But, you know, we're not stopping there. That is why we're looking at strategic partnerships with companies like Besi, like Kokusai, and there are others also that we're working together with because innovating across this flow is incredibly critical. It's unique. So when our customers are trying to bring these new innovations to market, you know, Applied is their most strategic partner. We want to strengthen that.
I should have asked the $7 million question last because I usually ask finally, why should this crowd buy AMAT stock? But I think you've answered that one already in space. We're out of time. I think we'll close it out there.
All right. Thank you, guys. Thank you, Stacey.