Applied Materials, Inc. (AMAT) Earnings Call Transcript & Summary

Welcome to this session. Really delighted to have the team from Applied Materials. Senior Vice President and CFO of Brice Hill, joining us this afternoon. I'll go through a list of my questions. But if you have anything you would like to bring up, please feel free to raise your hand, and I'll be sure to get you in.

Thanks for hosting, Vivek.

Welcome back. Very happy to have you here. So maybe let's start at the top. Maybe give us a state of the union, what's going well, what's different than what your assumptions were at the start of the year.

I see a lot of familiar faces. So maybe people know some of the story already. But the -- this year for Applied, it really emphasized the strength of a broad portfolio for the company. So people have known our services business in the past, helps give us breadth. Our wide portfolio selling to both memory segments, selling to leading logic. The other thing that's been driving this year is our exposure to what we call the ICAPS business. So the mature process technologies, the power, IoT, communications, auto, power sensors. And that business has been so strong this year that it's been able to help the company offset the weakness we've seen in the memory market and the leading logic market. So when we think about the dynamics, even looking forward, we expect that ICAPS business to still be strong, and we can talk more about that. It's a global phenomenon. A lot of people have asked us whether it's just a China customer phenomenon. It's global. It's public companies that are having to invest in technologies. There's multiple geographies that are growing faster than China. So that business, we feel, has real durability and it matches to underlying demand trends in the market that we think are stable and long term. But that's been the story of the year, Vivek. Is really -- since there is a good portfolio coverage across broad markets. Where we've had weakness, we've seen some offsetting strength and that's been good for the company.

Okay. Makes sense. One thing on just this 1 metric that we all kind of relate to, wafer fab equipment, WFE. Some of your peers have given very specific numerical, I guess, kind of a range, one's at low 70s, one's at kind of mid-70s. Applied has been hesitant in giving a specific point estimate, what is the reason for that?

Yes. I think there's 2 dynamics that make our business this year look different than any of the peers. I mean every company has a different perspective on the market. But for us, the strength in ICAPS may have been different from other -- than what other companies were seeing. As the company 4 years ago set up that division has been sort of focused in selling into those markets. The second thing is it's sort of bad news, good news. The bad news is we're still trying to catch up to orders from last year. And so we've had the tailwind this year, if you will, still servicing business because -- catching up the supply chain where we had orders for last year. So those 2 things sort of distort what we're seeing as revenue for the year versus maybe what other people are seeing. So the strength of ICAPS and the tailwinds of last year's business. If you annualize our first half, I think you get something in the high 80s. We're not saying that's representative of what the market is. You kind of have to sort out those impacts. And -- so we haven't sorted that out. And when we think about the full year and even into next year, we feel ICAPS will continue to be strong. We feel Display will get better. We feel the services business will continue to grow. What we can't really call is leading logic and memory because that's more exposed to macro factors and inventory digestion in those markets. So we just haven't put a number on it.

Got it. So you mentioned ICAPS a few times, right? That's certainly a very important part of Applied's portfolio. Talk to us about what makes your portfolio more unique? Is it just that as a percentage of sales, it's a higher mix? Or is this something more unique to your portfolio versus -- because many people sell, right, trailing edge tools. So what's so special about Applied's portfolio?

Well, I think for the company, they started the business group 4 years ago, ICAPS, and we've invented more than 20 tools -- new tools in the space in the course of those 4 years. So I don't know if there's anything that's uniquely beneficial for Applied other than focus. We've been focused on the market. We have focused sales teams. We have focused development plans across the different business units. And that's allowed us to, a, understand that this market should be growing and b, develop solutions specifically for the market. We do think it's accretive for us on a share perspective as that market grows.

Got it. Now this is also the place where you're capturing the need of the market extremely well because these were the places where there were the most shortages. But how are you making sure that there is the right level of utilization of these tools that some customers are not just buying and holding them and not actually putting them in use. Because I've heard of some scattered examples where they ordered the tool because there were shortages. Now they have them, but they just haven't put them in operation yet.

So that's a good question. And we've had the question, are customers warehousing those tools? Are they preordering them too far in advance? And what I can tell you is because we defer the install revenue for those tools and then we help with install, we track that for our equipment. And we do not see that. Across our tool set, we see it being installed in a reasonable amount of time. And we're not aware of any pools of equipment or warehousing of equipment that isn't being utilized. Vivek, 1 thing that we think, and this is more intuitive than data, but we think that the equipment being installed in China is likely operating at a lower yield. It takes a while to get up to world-class foundry yields for anybody that's putting in new process technologies. So we think part of what's happening is there's a long list of customers in China. In some cases, they're new to the business. It's going to take a while for them to get those yields up to world-class levels. And so there's probably some inefficiencies there in the equipment base. But going back, we don't see the hoarding or inventorying of the equipment.

That's a very good point that you can track, right, to as much real time as possible?

We do. We do.

Do you have the same install kind of attach in China also? Or is it only in the West that you have?

No, we have the same -- we have the same...

You have the same insight.

Same insight. Correct.

Okay. That makes sense. And then the second part of ICAPS, you mentioned is that you see this trend as being sustainable over the next few years. So what are kind of the leading indicators? What's giving you the confidence that it can sustain? Because this is one of the biggest pushbacks against this part of the business.

Well, this is 1 part, we're not surprised. We were surprised by the magnitude last year and this year, but we're not surprised by the trend. Because we think power -- I think 1 of our estimates is by if we have a $1 trillion market in 2030, we think 9% or 10% of the semiconductor chips are going to be power chips. So that's a significant market. Electric vehicles, electrification in general, green power solutions, all of these, plus the analog plus the build-out of IoT, which in a lot of cases, is video sensors and video processors. That's 1 of the largest runners. We understand how those are being deployed. We see the end markets growing. We know from public company documentation of why they're building factories and why they're investing. Basically, all those things fit together for us. So we do not think -- can always be surprised, but we do not think what we see in '22 and '23 as a cliff of some sort and unrepresentative of the end market.

So you would not be surprised to see the trailing edge market opportunity grow next year?

Next year is a question. We've said we think it's stable because it's grown a lot the last 2 years. So what we've said is relatively stable next year in terms of it's still that size. But if you stretch that horizon out to 10 years, all of the device types that were -- that we just talked about and that are in the ICAPS space, the forecasted CAGRs for those device types are 6%, 7%, 8%, 9% depending on the device type, by third-party estimates. And what's happened is -- the reality is there's just no longer -- companies have run out of brownfield space that they've had, that's been expandable. They run out of used equipment to buy. So they're having to put new equipment in place to serve greater and greater demand as that semi demand continues to grow. And we think where does that money come from? Because you can't have semiconductor business grow faster than GDP forever, but we still think it's providing productivity and so it's shifting real resources from other areas to semiconductors.

Got it. One thing you mentioned on the last call was some pushouts on the leading-edge side and then a few days later, we saw the large foundry mentioned that the CapEx would be at the lower. So from that perspective, I guess it wasn't that big of a surprise, right, because you've already kind of baked that in. But given all the excitement around AI, do you think that leading-edge from this point onwards is a little more sustainable? Or do you still think that AI is a lot more hype and it's not as accretive to the business in the near to medium term?

Well, we definitely think it's a tailwind for sure. I was recounting with a few people here today. I mean it's come upon the consciousness pretty recently, right? I think it was February, Meta put out their model. And then March, the OpenAI announcement was made, and then there's been tons of innovation and experimentation since then. So I think the industry is still digesting what is this going to mean from a system perspective, from a -- from the individual chips that scale into these systems. But we absolutely think it's a complex workload. It runs on leading-edge. It will drive cloud and data center. It will drive new solutions in terms of accelerators and GPUs and CPUs. There'll be a lot of innovation to optimize these workloads. So it's a positive. As far as what's the net effect over a long period of time? It's hard to say because it's a positive by itself, but it also adds a lot of productivity to semiconductors. And so in a good way, it can help make semiconductors more efficient and more productive and basically improve the benefit and lower the cost of semiconductors, if that makes sense.

Got it. So absolutely. We had the team from Synopsys in Cadence yesterday, and I think they kind of -- they made the same point that -- but does it mean that money gets taken out of the WFE pie and goes more into the upfront design side because it's more effective to do it there or it's too early...

It's a good question. It's way too early to call. And so then the question would be, well, if you make the by definition, if you make the leading-edge semis and the cloud more efficient and more productive, then do you use more of it or less, I don't know the answer to that. But I just -- I think we have to be cognizant that it's doing both things. And I would just say it's exciting. It will be a whole new path of innovation. I sort of believe in some of the end user models where you're going to get -- we're personally going to get more productivity from some of these solutions. And so you'll be willing to pay more for the software that includes those things. So I think it will be a real driver. And then whether we eventually change our $1 trillion into 2030 up or down, I don't know.

On the memory side, that's been the real, I think, problem area for your customers this year. What's the visibility right now in terms of when memory utilization can start to improve? Do you think that's even possible in '23 or 1 should defer that to '24?

Well, it's certainly possible in '23. Since we work with memory customers we track utilization, we track inventory levels, we track prices as the -- spot market prices and other things. 1 thing we would say is looking back into our last quarter that we just reported, all those trends are still moving in the wrong direction. So if we're at the bottom or if it's turning, at least it wasn't turning last month. We didn't see that yet. Now having said that, we said in our earnings call when we look forward multiple years, we think the -- and we think about the WFE market, we think the WFE market will be 2/3 foundry and logic and 1/3 memory, and it's less than that right now. So by definition, we do think memory should come back up to get into balance with foundry and logic. And we do think what will drive that even in the medium term as companies, customers will continue to upgrade their nodes in order to get the big growth that's required. One of the things we recently looked at is just the wafer starts across memory over the last number of years. Say, 5 to 7 years. Wafer starts have increased at a very low rate, maybe 1% or 2% per year. So the way there's more bit growth is those companies upgrade their processes, the processes deliver more bit density, and that's what provides enough bits in order to service the increasing -- the ever-increasing demand. And so then you say, well, if they don't need any wafer starts, then what is Applied doing? Well, the way they're getting that bit growth is they're upgrading their process technologies to the next node. And that does take a reasonable amount of capital investment and equipment investment to make those upgrades. So we think actually that dynamic has been in play for a number of years. And so we're expecting -- I guess, at the end of the day, we're expecting that market to improve. Calling the moment, it's hard to call the moment.

Got it. what's historically been the nature of the memory industry? Does it give enough of an early warning. Like, let's say, conceptually, if it were to turn 3 months from now, would you know it now? Or you find out that, oh, it's just going to turn up in the next month. Does it give you that kind of early...

I personally don't think so. And I don't have as much exposure to memory as I do to logic. But what I do know is the people in a company that are thinking about building factories and buying equipment they have a 2- to 3- to 4-year horizon in making those decisions. You, of course, it's a factor, what your current utilization is and what your current market is, but you really have to be looking down the road. And so I think it's relevant, but it's not definitive as to how they're going to make their equipment purchase decisions.

Got it. And just the last 1 on memory, DRAM versus NAND. Do you see any differences because you're more indexed to DRAM, which, by the way, has better correlation with AI and compute than NAND does?

I guess I would say, no. I just looked at the pricing curves. I just looked at the utilization. It's hard to point to -- in fact, the only difference that I can talk about or recognize is that the NAND market seems to be much more depressed than DRAM. And so I don't really understand that. I don't think there's a system architecture change or something that's really driven that. So it must be unique to the business in terms of inventory or something that happened. So we would expect it to improve and get back into balance. But hard to call the time.

Understood. Applied also got approvals to ship incrementally more to China. Can you give us more color on what those approvals were for? And importantly, is this now the new baseline of that China engagement? So whatever China WFE is you can then grow with that? Or are these like one-off type approvals?

Okay. So there's been a clarification on what process you're allowed to ship to in terms of DRAM. And that's really what that change is. And so we should have some business in the second half of the year that has been held up. So like I talked about in the first half of the year for us, we had 22 orders that we are still trying to ship and catch up to. This will sort of, I think, be the last part of this, where we had some DRAM business that we couldn't ship that now we'll be able to ship. It will take us again, a little time to catch up to that. But that's a clarification. And as far as the overall process with rules, we have a number of licenses pending. We have a number of places where we're working to clarify specific factories or specific companies, whether we can ship to them. So I think this is an ongoing process with the government. When we started the year, we said there was a $2.5 billion impact to Applied from the change in trade rules in October. We've probably got that down to $1.5 billion now by clearing customers and putting -- taking them off the block list, if you will. We do that by our own diligence. We get assurance letters from the customers. We compare notes with our peers. We share that information with the government. So it's a whole process to get there, but there's been a lot of action and a lot of improvements since it started. But I would say it's still work to do.

Got it. Now backlog levels are still quite elevated for Applied and some of your peers as well. What's driving this behavior? Because I would have assumed that your lead times would have come down as there's been this push and pull in demand. So what's keeping backlog level still quite elevated?

Well, I think it's really important to separate those 2 things. So lead times have come down. I think if customers call us now for most of our business units and ask us to order a tool fast, then we've got a shorter lead time. It's almost back to normal from that perspective. So that's a good thing. And we expected that because of that, when we were talking last quarter and the quarter before, we expected our backlog to come down based on that dynamic. It really hasn't come down. It's remained elevated. And so what's happening is customers are giving us a longer view of their demand than they gave before. So we have visibility, and we have orders that are extending longer in time period up to 2 years in some cases. So it's kind of a declining order pattern, if you will, as you roll forward each quarter, visibility into that. And we think it's just -- I don't know if that's permanent. It's probably a consequence of when we were constrained -- customers were concerned whether they were going to be able to get their equipment shipped to them. And so that has sort of leaped over into now. So it's still elevated. It's 1 of the reasons we don't provide it in our calls. We don't think that the backlog is the best indicator of the next quarter or the next business. We'll report it at the end of the year. You'll see the number, but it's just still elevated.

Applied Global Services, right? I think you -- the last I recall, you mentioned that you still expect that to grow this year. That has a 200-millimeter tools part, right? And then what is -- what 1 would describe as more kind of services and spares and support of tools. So are both those growing year-on-year? Or is it 1 growing and...

Well, the services business, we removed a significant amount of that business because of the China trade rule change in October. So we said $400 million came out of that business by tools that we're actually taking offline that we can't service anymore. And then we can sell spares, et cetera, to that. So I think under the covers, for sure, the 200-millimeter business has grown year-over-year, and it's grown pretty significantly year-over-year. I think the services business is just short of growth year-over-year. Together, it's year-over-year growth as a whole. And then now as we move forward, we'll expect the services business to move into year-over-year growth. So basically, it's overcome the removal. Its growth has overcome the removal of that business. And the fact that we've been able to clear some of those customers and get them off the block list has also helped cure some of that situation. But in general, what I would just say for people that aren't that close to the services business, it runs -- its driver is the installed base. So every day we ship a tool, even if we're having a down quarter in semi, every day we ship a tool, it grows the installed base. We have an opportunity to sell spares and subscription agreements against those tools, offer insights as to how the customer can get it at high yields. So that drives on a different driver, if you will, than the equipment WFE. And when we raised our dividend just this cycle, we raised -- the Board approved a 23% dividend increase. One of the things we were thinking about, Vivek, was our services business is much more stable growth than the WFE because of that dynamic because the installed base is constantly growing. And so we said, you know what, that can afford the entire dividend. Not only that, we could probably double the dividend over a period of years, which is our intent. And so we kind of looked at it that way. And then the WFE business is a little bit more volatile. So we said that will be primarily a buyback return model. We didn't tie them mathematically that way, but that's conceptually how we were thinking about it. And so that services business is a really good anchor for the company from a stability and profit returns, right?

No, I'm glad you brought that up. But very strong dividend announcement. I think it will help to kind of walk through why, right? Because that's even a faster growth than dividend growth than assuming what your services business is going at.

That's right. We got a lot of feedback that some investors couldn't invest in Applied because the dividend was lower than a certain level. And so -- and we think with that services business and the solidity of it, that we should have a larger dividend for it. So we try to -- we're moving those things into alignment.

Are you targeting a certain yield or a certain range of yield that you...

We haven't said that, but I would just say -- what we've said is we should -- it's an intent barring any unforeseen circumstances, but we should do 3 more increases at least in a similar range and get us to doubling the previous dividend that we had. And then we'll -- other information will come later.

I mean more on a monthly basis or...

Each year. Each year, yes.

Then on gross margins, 46% to 47%, they've actually been better than some of your peers. So what's helped maintain the gross margin? Is it something in the mix? And then what takes you to the target of 48.5%.

So first thing is we had a glide path or I shouldn't say it a glide path. We had a path to find to 48%, 48.5%. We've got hundreds of engineers working on cost improvements for the things that we can control on the supply chain. We expected some of our inflators like chip price premiums, like expedite fees because of supply chain challenges, those sorts of things. We expected them to recede. And then we're -- for the things that we couldn't drive back, permanent inflation, if you will. Those are the things we're working on price adjustments on our products for to sort of share that with the customers. And those things all take time. The engineering programs, if you will, it takes time to design a new solution and then it takes time to qualify that solution with the customer. So that takes some time to do the price increases, that takes some time and then actually to just beat back the inflation that can be beat back. So those 3 elements are kind of what I would focus on. And then when we lost that China business, that China business was accretive. So that put us basically 100 basis points behind where we thought we were. Now some of that has come back, which has helped us sort of be ahead of where we thought we would be. But if you're modeling that, I think it's going to be a slow improvement. That's still our target 48% to 48.5% in the medium term. We think we can get there. It's going to be reengineering products and it's going to be price adjustments as we go.

Got it. And then Applied announced this really interesting tool, earlier, the Sculpta pattern-shaping tool. Can you give us an update on the traction with customers? So we know that there is 1 early adopter customer. When do you see the largest foundries kind of adopt it also?

So we are shipping that tool for revenue. We do expect it to be at least $1 billion product for Applied. We think it will be in the years that it ships in the next few years we think it will be hundreds of millions. That's what we said. So it's really all the detail that we can offer. And for people that may not know, that is a tool that offers a litho-like solution, and it can be an alternative for customers to define specific design shapes in a product. And so a customer could literally decide whether using that tool or using an EUV layer is the best solution for them. And so if they decide that they can use the Sculpta tool, then they can save a significant amount of money from not using an EUV tool on that particular layer. So it's a cost advantage, and it's designed to give in its effect on the transistor design, it's designed to give exactly the same output or result that an EUV tool would, at least for that part of the project.

Can Applied achieve those kind of growth targets if the largest foundry is not yet involved in the tool?

It's a good question. I don't think we've provided any information. What we have said is we expect all of the customers to look at that as an opportunity and evaluate it. So that's as much as we said.

Of course. Then on the chip side, it was interesting that Gary mentioned that when you look at the overall benefit it's going to be somewhat more modest than I think many -- which I think is like the practical view. So it would help to understand how Applied came up with some of those assumptions. Because when you look at the large customers announcing tens to hundreds of billions of project, why won't a lot more of that flow through into WFE?

Good. We think it's location. It's -- so our internal estimates are that those CHIPS Act monies and all the monies across the world, $400 billion right now of subsidies, they're really deciding the location that customers will put their factories. We don't think anybody is going to build a factory like in advance or just in case. We just think that if they had it on the road map, instead of building it in place A, they'll build it in place B and take advantage of some of the subsidies. We do think, Vivek, that there's 3% to 7% of sort of lost economies of scale because it's in a different location than they wanted to put it. So there will be certain types of equipment where they'll have to have -- they won't be at a perfect matched set and they'll need to have an extra tool, and we think that will drive some uplift, but not significant. A place that's slightly different where we could see some benefits is on the services side to the extent that companies are putting factories in places that they're not familiar with or they don't have an established workforce, they're more likely to look at Applied services and have us help with getting the tools up to yield and help with all the transactional activities to keep that equipment running.

Got it. And then lastly, price. Applied has also announced R&D centers, materials innovation centers. How will they start to impact? I assume these are longer-term investments. But can you take advantage of some of the other provisions of the CHIPS Act, whether it's on the investment tax credit side, or on the grant side?

Absolutely, thanks for bringing that up. It's a big investment for the company, the epic design center. We had a lot of customers with us at that announcement. It's designed -- it's designed to accelerate development activity across the ecosystem. And so customers that previously might wait to get a tool from us before they have access to it, before they can experiment. Now they'll be able to be in the facility with us getting access to the tool and collaborating faster than what we would normally do. And so that's why there's so much excitement about it. The numbers that we quoted, $3 billion to $4 billion, those are gross numbers. We will be applying for CHIPS Act funding. You saw the Vice President at the event. We think we're very much aligned with what the government has in mind for CHIPS Act funding. So we're going to go through that process, and we expect to be able to get some benefits that process. And the other thing I would say is, from an investor perspective, it will be a couple of years before you see any impact on the P&L. It's a capital project. It will be capitalized as we go and it won't start depreciating until we put it into commission. So it will be a while before you see that impact other than the CapEx, whatever we end up with as net CapEx that will, of course, be our investment. But the company has grown a lot. If you look at our CapEx versus depreciation in the last 10 years, you'll see that -- I think depreciation might actually be a little bit higher than the CapEx. And so it's time we have to expand our lab space, we have to modernize some of it. So this is going to be a great investment for the company, and I think it will set us up for the next 20 years of growth.

Makes sense. Terrific. Thank you so much, Brice. Really appreciate your time.

Thank you. Very nice. Appreciate you. Thank you.