All right. Morning, everyone. Welcome to the 27th annual Needham Growth Conference. Again, this is the first in-person day. My name is Charles Shi. I'm the semi-cap analyst at Needham. Joining me here is Ultra Clean. Thrilled to have Jim Scholhamer, CEO, here on the stage. Really, thank you for joining us today. You've been coming to this conference for quite a few years, right? Great pleasure to host you again.
Oh.
All right. So let's kick off the discussion. I want to start with a topic that's gathered a lot of investor attention recently, and it's about China. It's about December 2024 new U.S. export controls that put nearly every Chinese semi-cap company on the Entity List. And mind if we start from there, because you guys do have some direct China business. And can you elaborate, or maybe remind investors again, who are your Chinese semi-cap customers, and how much of the business you have with them, and what's your view about the latest export controls, whether that would have an impact on that part of the business?
Okay. So let me start with a thank you for having me. It's my 10th conference here. Yeah, let me start with a punchline, which is, we see no impact to any of the latest regulations, and we believe we're well insulated from future regulations that will come from BIS or other departments. So we've did a lot of analysis, cross-functional team, outside legal, inside legal, as well as we have a trade compliance department, and so on. And the strategy that we've employed is in China, for China. So we have a factory in Shanghai has about 500-700 employees. They are all Chinese employees. There are no Westerners there. All the engineering is done there. There's no IP, no engineering. It's completely isolated, insulated from the Western countries.
And that facility we've had that site for 20 years, and it has a 20-year relationship with the Chinese equipment makers as well, who are mainly AMEC and Piotech are the main ones. And then BEST, we have a small position with BEST who has some ownership of Mattson, which we have a long relationship with Mattson as well. So we've had these relationships. Many of them were former Applied people like you and I, Applied Materials. We for a long time, I was just there recently celebrating AMEC's 20-year anniversary. But given the unique nature, we believe we have a unique capability and a unique position to be able to continue to support the Chinese equipment makers who have been slowly growing over time. And now, as you can imagine, that growth has expanded pretty dramatically.
There are other parts about the whole China strategy besides just the you know the domestic equipment makers. There's you know the overall supply chain, which is Chinese suppliers which are supplying other sites that we have in Asia, like you know our Malaysian, our Singapore, and our Philippines facilities. So there's a supply chain duplication for business continuity planning that we're doing in order to make sure that any new regulations or any counter-regulations or any counteractions from China doesn't cripple our Asian supply chain. So there's another piece to that as well. There's of course the indirect China impact. So we are making many of the modules and components for our OEM customers, like Applied, Lam, ASM, etc., which then also could go to China.
So we have to look at the China WFE spend because that impacts our customers' business, where we think that's, that's an okay condition. And then there's a China, then there's a, our China factory also is a sub-supplier to our other factories where they're making components and sub-modules to support our Malaysia factory and so on. So we're creating duplication in these other factories so that we have a business continuity planning in case there's any continued measure and countermeasure coming back between the U.S. and China. So we feel really good about how we're positioned. We think we have a unique, unique capability that most suppliers don't have, and so we've been able to benefit from, from the equipment growth of the Chinese OEM companies.
Yeah. So maybe I, Jim, I do wanna have a quick follow-up here, right, about your direct China business. So, you know, the non-Chinese equipment makers over the last three months then have more or less taken down their 2025 expectation, right? And obviously that to you, that's an indirect China business, right?
Mm-hmm.
Chinese semi-cap companies, we just heard ACM this morning pre-announced a pretty solid growth into 2025. There seems to be some dispersion between what they see that the environment is going. But just wanna ask, what your Chinese customers are telling you, what did they see and what did you hear about this year?
Yeah.
2025.
You know, they, our Chinese customers, and we were very open about this. When we saw the revenue from them last in 2024 double and then double again, you know, it's now 10% of our revenue is coming from direct Chinese OEM customers. So we obviously asked a lot of questions. I spoke with the CEO of AMEC just at their opening ceremony, had dinner with them the night before. They were very open when we asked them a few months ago that a lot of what they ordered in 2024, well, 40%-50% was actually stockpiling, not going directly to their customers. So we always knew there was some risk there. They indicated that they intended to continue to do that through 2025.
But I think, you know, it's, you might intend one thing and then the market and, and other things happen. So it is a, it is an unknown, whether that could continue at the level of 2024 or not. It's 10% of our revenue. If it goes down to 5%, we think, you know, we think we can more than make that up with, with the investments will happen in other places of the world. So we're, we don't think it's a material thing to be concerned about. Even, even if there is a, you know, a sudden drop in the Chinese OEMs, it's, we think it's, it's not gonna be a huge impact to us because it's, it's a small part of our business still, even though it's grown a lot.
Got it. Thanks for the transparency there then. So I wanna switch gears a little bit. We all know that the environment, WFE environment, right, that since, let's say, 2022, has been nothing like normal, right? So a few things are very abnormal over the last couple of years, right? But one of the things that I think that matters for UCT is memory is pretty depressed as a percentage of WFE, especially NAND. And I mean, normally I would say you would think it's a 60/40 logic to memory. It has been, it's probably 70/30 last year, probably was even worse in 2023. I mean, by worse, I mean it's more due to logic. It's probably not worse in that sense.
Can you help us understand, right, what UCT's breakdown looks like in a normal environment? I know you have a breakdown every quarter, but in a normal environment, let's say, you know, 60/40 type of WFE mix between logic and memory, what would you see is your business mix between logic and memory?
Yeah. It does move around dramatically. We ship out hundreds of thousands, if not millions, of parts and components and modules and sub-modules a year. So we don't know where they're all going. So from a bottoms up, in some cases we do when the customer tells us, like the AI business we're doing in the Czech where we had a very clear line of sight.
Yeah. We'll get to that.
Yeah. Oh, okay. But we can't really do a bottom-up like this is going to a NAND is going to DRAM, is going to logic, is going to foundry from the massive amounts of shipments that we do. We are sending it to our customers' collection points, which where then it goes on to our customer, the customer. So we don't know. So we do a top-down. So we look at how our major customers, you know, what their breakout is between those sub-segments of WFE, and then we look at our footprint within those and do that. So that's a kind of an end around of the top-down estimate. So we estimate around 40% typically is in memory.
Now, when NAND was going through, you know, the layer transitions from a few layers up to, you know, 180, and now the next is in the 200s or 300s, when that rapid movement through multiple layer NAND was happening, it was even more memory. It was above 40%, but flying around, you know, moving around, you know, 40%+ or - 20%, depending on what's happening in memory, and then logic foundry between 30% and 35%, and then the remainder for us is in service across all fabs, non-semi, and trailing edge applications like, you know, ICAPS and other types of, you know, non-logic memory foundry business, but, you know, silicon carbide, all the array of other chip types.
So that's roughly how it averages around those, but there's some pretty dramatic swings up and down depending on the investment cycle.
Okay. Feels like it's hard to say you are over-indexed to any specific WFE segments. It sounds like you're pretty neutral in that perspective.
We have a really good, nice footprint. You know, we're heavy in Dep and Etch , so that definitely benefits us the most, but we make a lot of the CMP tool for the number one provider, CMP, same with implant, same with ALD, a great position ALD, a growing position in litho, and a great position in the wet chemistry tools as well, so we have a very diverse portfolio, but the Dep and Etch that memory drives in NAND is if we could say where we would prefer the investment, that would be a spot.
Mm-hmm. Mm-hmm. Got it. So, I wanna dive into the products business, but gonna, I'm gonna ask the question probably in a different way than I planned. I do wanna start with the comparison of your product business with Ichor's business. I think this is a very common question, I mean, okay, what's the difference between you two? Are you the same? So, but the other thing is, probably a lot of people have been under the impression that you and Ichor are in the gas panel business. So, mind if you would start from there.
Sure.
But then you kind of give us an overview of what your product business is about.
Sure. I, you know, it's changed dramatically over time. So I think people do recall, you know, for 20 years, both our companies were mainly gas panel businesses. When I started, UCT in 2015, we were doing like one fifth of the revenue we're doing now. It was almost all in gas panel business. It was 90% of our business was gas panel, and it was the same with Ichor. Since that time, our strategies have taken us in two different directions. So one, we got rid of most of the non-semi business that we were doing, which wasn't making money. We focused on semiconductor. This is back when WFE was $30 billion, and we said, it's gonna grow, let's double down on semi and get rid of making machines for Tesla's factory, and so we did that.
Over time, I think the overlap now is more like 30%. But I don't think the investment community has completely adjusted to the fact that our overlap is relatively small now. We've gone into diversifying into other parts, you know, making the vacuum transfer modules, making the factory interface module. We, instead of just making the gas panel only, we've diversified into other modules. We've grown and diversified into new developments with ASML. We've expanded in our CMP footprint. We bought a company in 2015 to do the wet tools for the industry leader in that space. So you know, very strong footprint in implant and ALD with ASM. And so we've gone into more of an integration, and then we've also bought a lot of sub-component companies. We've even gone into service.
We've gone into the sub-fab space. So these are all areas that are not in parallel with Ichor. They're completely different spaces, and Ichor is focused more on precision machining, components either for their own use or for selling to other suppliers, including us. You know, there's actually a lot of commerce that goes on where we have 100% market share on thermal solutions, heaters, and so on. We sell to Ichor and they sell us some of their high precision machine components. So they've kind of gone in a different direction. So we're very different companies now than we were 10 years ago, and the overlap is small. We consider our biggest competitor is our customers.
There's still a significant amount of operations manufacturing going on in our customer's factory that have been over the decades that they've been outsourcing and increasing the outsourcing over time. As we win new share, it's usually not Ichor and UCT battling it out to take share from each other. That's very rare. We're going after the share of the additional outsourcing that happens every year, and we've been doing very well picking that up, especially at our second biggest customer.
Mm-hmm. Mm-hmm. Mm-hmm. So I recall that your second biggest customer, maybe it was the last, through the last cycle, there was some pretty meaningful new outsourcing opportunity, right?
Yeah.
That you captured that. Going forward though, what do you see as the biggest incremental outsourcing opportunity that could happen?
It's still there with, you know, we're the number one supplier by spend at both Applied and Lam. And I think we're pretty high level at, you know, ASM as well. You know, we have a footprint in KLA and everyone else that's pretty small, but, you know, the order of $40 million-$50 million a year. But the outsourcing, you know, we do roughly with Applied roughly 60%-70% of the revenue that we do with Lam. And Applied, of course, is a bigger company and so on. So we see continued opportunity to, and we've discussed this at, you know, very high levels within Applied to really partner on a lot of the new developments, the new tools coming out.
So we continue to see over the next several years dramatic growth with that account. If you look at the comparisons of the size of the two companies and how far we penetrated within Lam, you could see, you know, a significant opportunity at Applied continues. So I think we'll continue along that vein. We'll continue to grow our market share at ASML as well.
Got it. Speaking of that, right, you already mentioned that you have two 10% customers. I mean, maybe as of the last quarter, maybe, right. But I think getting more customers, right, a third 10% customer has always been on, I mean, on your strategy. And when should we expect to see a third 10% customer?
There's two variables in that. We've been growing our revenue with ASML pretty strongly, especially we got into their new tools. ASML is very close with their suppliers, and they tend not to take business away from tools that are already released from a supplier. It's a very sticky relationship with them, very much of a partnership. Over the last two, three years, we've been winning the new modules for the new tools that were in development, the EUV tools basically coming out. We've been growing that pretty steadily, and the third quarter of last year, we actually hit 7%, which was a new high water mark for them, for us, for them.
And so, you know, their forecast is still growing through 2025 for us, but it's not as steep as it was before they put a halt to their. They don't call it build to stock or stockpiling. They call it some pre-order or something. They were doing a more aggressive pre-order, until they realized that just made their customers more lazy about ordering if they were gonna take the responsibility to build ahead of time and put an end to that. So we saw it come down a little bit, but it's still growing. The other part where, you know, if you look at the revenue we're doing there now, it was close to where we at our three-year strategic plan where we thought we would be really close to 10% by now, but the denominator grew.
So the revenue across other areas grew, really, you know, just as fast, almost as fast as our revenue percent at ASML. So we're going up as a percent, but we're fighting the fact that we're winning business in other areas, which is a good problem to have, right? But, so the win is very difficult, at this rate, you know, it's not gonna be in 2025, but, you know, I would hope by, you know, sometime later in 2026 that we could get there.
Thank you. That's good to hear. So maybe another question related to Ichor, since they actually had probably slightly more bullish view on where the business is trending in 2025. It sounds like they were expecting a pickup of their business, I think in first half 2025. Obviously, everybody saw their pre-announcement. So the kind of WFE environment we are in, I don't know what your view about the environment, but what's your comment on their view and maybe the pre-announcement they made and how should we think about how does that, how do we read across to your business?
Yeah. As far as their own particular, because they're doing some, like I said, precision machining and other things, I really can't speak to why they think their revenue might go up in the first quarter. As I could speak to, you know, to the industry, our view is that the first few quarters, the first half is, we're not gonna see an upturn in the first half of 2025. And we'll just see it bounce around, you know, the Q4 for us, we'll see it bounce around. And I think the industry too bounce around this level, where we are right now. Now, it might be down a little bit or up a little bit, but you shouldn't take the last two data points and draw a line through it.
I think we're gonna bounce. There's gonna be some instability, but it'll roughly average around the first and second quarter average, roughly around the level we're at. We do think that the second half we'll start to see as the some of the especially in NAND the development of the 200- 300 you know layers as they start to add a little bit of volume with hopefully a price correction and supply and demand correction comes into play. And they're gonna wanna ramp up the you know the more advanced you know NAND chips as well. So right now we're seeing early business in that space as they're doing the development using CMP and ALD and implant tools to do development for yield improvement and so on and other devices. So we expect the second half is where we start to see increases.
Our position right now is a roughly 5% increase in WFE from 2025 over 2024.
Got it. Mid-single digit growth for the environment. Okay.
But mostly happening in the second quarter, second half.
Second half. You think it's more second half?
Yeah.
Okay. Another topic, AI. I cannot not have an AI question in any of the fireside chats today. For you, the exposure sounds like it's mostly around high bandwidth memory, HBM. I think you've done this, you that you gave some good color, AI or HBM driving some of the incremental upside for quite a few quarters, I think, or throughout last year, right?
Mm-hmm.
Can you tell us like, where you have been, where you have seen that part of the business have been trending and, what's the expectation for 2025? By the way, please do give people a little bit of a background information what you do, related to HBM and AI.
Okay. Yeah. I think that there's two pieces to it. You know, there's one where we don't have very good transparency when we make the modules for a Dep or Etch tool, that could go to logic, it could go to DRAM, it could go to NAND, and it could go to any of those chips being used for AI. So that part of the mainstream, we don't know, you know, SABRE 3D is a part of that and we don't, you know, from Lam and we have a great position there. And so we know a lot of that goes into the AI, but we can't quantify that because we don't know where it goes.
Where we have more clear vision is where, in the Czech Republic, we have a site, a company we bought in 2015, which does the wet chemistry tools for plating. And we saw that business double and double again. When we bought that business in 2015, it was like $26 million a year total revenue. And now we're at a $200 million a year run rate, with that business. It's grown dramatically. And in the last year or so, a lot of that has been to the AI impact. So what that is, that's after the hole is drilled with the Edge tools that we probably made a lot of the modules for the through silicon via, then the plating tool is filling those. And that's for the packaging and the multilevel packaging that's used for the AI.
So it's kind of on the packaging side of the high bandwidth memory. So that we can quantify because the customer said, you know, hair's on fire, you know, we need to increase capacity and here's why. And actually we got visibility all the way through on what was driving this dramatic increase in business. And so that was an unusual visibility that we have for us. So that we continue to see through 2025, continuing, maybe slightly ticking up. I mean, I think when you look at the customer's announcements, you know, Applied, Lam, et cetera, they're talking about, you know, 3x-6x more in that space, which will benefit us. But these are 3x-6x from various, you know, we'll still have relatively small numbers, but still it's a nice growth. So we think that could continue to grow.
I mean, obviously AI is gonna continue to grow.
Yeah. Yeah. $200 million run rate.
Per year. Yeah. We did, yeah. So we're at $50 million a quarter, or more. And when we bought the business, it was, you know, $6 million a quarter. So it's a pretty dramatic improvement. We had, luckily we had just built a new state-of-the-art inline, you know, lean factory there that we could handle the extra capacity.
Got it. I think this may be part of the, your story may be still a little bit underappreciated, but thanks for really providing that color again here. We talked a lot about your product business. Anything you can tell us about the service business here? How should investors think about how does that fit in your overall business and how the business have been trending recently?
Yeah. Service, our service business is, you know, just explain what it is. It's, you know, every few weeks the customer, or sometimes our people gather the dirty parts out of their tool, the coated parts, etched parts. And we take those and we atomically clean them with very special processes and then do analytics to make sure that they're like new. We kind of refurbish them back to like new, for the customer. And those are millions of parts a year as well doing that. So though that business moves along with wafer starts. Now lately in the last two downturns, wafer starts have moved around more than they normally would as the chip makers consolidated into far fewer chip makers out there.
The competition for market share kind of eased up and there were. They're better, they're better at cycling down during a downturn and trying to save their pricing, and instead of grabbing market share and then, and then moving up. So we've seen that business be more volatile than it had been in prior downturns as the customers have been more focused on supply and demand balance. But that's how it operates. It's a higher margin business for us. It's about 13%. It goes between 10%-15% of our revenue. We're heavily, you know, our number one customer is Samsung. Our number two customer is Intel. So it's been a little bit depressed 'cause it used to be around 20% of our business, but our two biggest customers have had issues with being at the forefront of technology.
The tailwind is TSMC is now going. Our strength is we're global. We're the only company which is really global with the factory. Intel liked that because they develop a process, the leading edge process in Oregon. We're right there, doing the cleaning and then they build a plant in Ireland and Israel. We have factories there, which then we copy paste the process. Global companies want a global provider so they can quickly transfer processes to their satellite operations. Samsung goes more global and then they catch up, you know, as they're building in Texas. We already have a clean facility down there. We'll pick up more share, you know, if and when they recover. Intel hopefully recovers, and then TSMC has now had to go more global as they've run out of people, water, power space.
And so, as you know, they're building that plant in Arizona where we already have a cleaning facility there too, where we upgraded it to handle. We have that there for the Intel business and now we manage. So we're growing our market share at TSMC outside of Taiwan. Taiwan is very competitive, but in the US it's. There's very little competition for us. So we're gonna be able to grow going forward our TSMC market share. And I think at this point we're well positioned to support Samsung and Intel, but we're more in the back seat and hopefully they drive forward in this.
Good. Good. That, that's probably gonna be a very nice tailwind for you guys as TSMC ramps volume in the overseas operations. All right. I wanna ask you about M&A. You have acquired six companies since February 2015. The most recent one was the acquisition of HIS in October 2023. So has been a pretty long, looks like a very consistent strategy in inorganic growth. Obviously we're not gonna dive into the details for those six acquisitions, but can you tell us what's driving those decisions, what you are looking at, what looking to buy? And maybe just tell us a little bit more about HIS because this is the most recent one.
Yeah. Okay. So first of all, they've all been very successful acquisitions. I think on our website we have that in our investor deck. We're six for six from the free throw line, we say. Now that I've bragged, I'm not gonna do another one. We have a filter process. We have a M&A group, small group that does this full time looking at things. And of course your bankers are always there too. And, you know, 10 years ago we changed that we're only looking at semiconductor acquisitions. We look at higher IP, higher margin businesses. Almost everything we bought had a significantly higher margin than our base business. We look to be creative, you know, either immediately or within a quarter or two.
We look for components and businesses that are sticky and that we're reselling to our customers and they need that. And also that we also can use it vertical integration. So we buy from ourselves rather than from the outside. And then we also look for alignment with our big customers. So several of the acquisitions were actually being pulled through by our customers. The supply chain for semiconductor at 60 years old now is very fragmented. Lots of little shops, engineering shops from 30 years ago that are still designed in. So there's thousands and thousands of suppliers, which is a huge headache for our customers. And now we're a decent size of probably the biggest pure, we are the biggest pure play semiconductor supplier.
And I don't know what's in Japan, what's outside of Japan, but so that, you know, we've often had like our customers come to us and say, you know, they, it's a company that's private. It's a, we're 100% dependent on their technology and what they provide. And the owner wants to leave or, you know, wants to cash out and leave or whatever. That's happened a lot. So several of those, at least two and probably three have been customer push, like please acquire this because it's critical to our products that this has some continuations and business continuity. So when they're sponsored by the customer, that's great because, you know, market share and pricing and then going forward is taken care of, you know?
So those are the filters that we go through. Also, I didn't mention diversification. We bought in the service space, we bought in the sub-fab space. You know, we bought in the spaces where we're, you know, growing outside of gas panels. We've also tried to diversify across the semiconductor spend. We not only tried, we did it.
All right. I guess, maybe let's leave some time for Q& A. We have a pro, yes, please.
The Malaysia facility do for you from a gross margin standpoint in the next up cycle compared to the last cycle when you really didn't benefit from that facility?
Right.
Allow me to repeat the question for the webcast. So the question's about your Malaysia facility, any of the margin benefit going into the new cycle? 'Cause, you know, looks like in the past cycle you didn't really appear to be accreting to your to the margin performance over the last cycle. Looks like, yeah. So that's the question.
I don't. Yeah, we didn't even have it in the last up cycle. We had it in time for the down cycle. Yeah, I think when we're up to the $608 million level annually, I think we're, you know, pushing $200 million a year out of there now. So when we get up to the full capacity level is when we're gonna test it and see, you know. We have a model right now, but we haven't quite tested to see how that comes out. But I think if we're up at that level, you know, depending on the revenue levels from our other sites, you know, which we don't know what that will be, you know, I think labor is typically about 7% of our cost of goods sold or 7% of revenue. It's almost the same.
But, you know, so I think, you know, at that site, we expect, you know, labor to be one to at least one, but probably more like two points lower, because of just the nature of the salaries there, so how that averages in across the overall business and the margins is dependent on volume. We've built capacity now for Duguru up to $4 billion annually, and as we fill that and leverage that vol, that fixed cost, is where we expect to see a big impact, and we typically overshoot our model on an upturn, but I think it'd be safe to say overall that, you know, point of gross margin is reasonable, but, you know, we have yet to test what that looks like when we're at full capacity.
My CFO is not glaring at me, so that must've been the right answer.
Sorry, can I just ask? You made a comment about your China customers, like stockpiling. And I didn't really know what you meant, like meaning they're buying just a ton of modules from you and like parking them on their balance sheet and not productizing them. And like, why would they do that? And if they're doing it, why wouldn't others be doing it if they're serving that geography?
Yeah. To repeat the question, gonna shorten it. So, about your Chinese semi-cap customers doing stockpiling, what's the rationale for them to do this?
Yeah. The Chinese OEM stockpiling was, and it wasn't really so much around us, but there are other things that make up their tool that they had concerns about getting. So there are other modules and other components that we're not providing, but they're getting from other suppliers. So they were concerned about the rules potentially changing down the road, and so that was the reason why they were stockpiling. So they'd have a complete set of modules and components that they could, when they get an order, put it together and ship it and not have to be concerned about any new impacts that come down. So it was kind of an insurance policy against what might come. It wasn't around dealing with the existing situation and what they're able to buy.
Like, let's get it now while we can in case something changes.
Maybe I ask one more question. Your two largest customers, right? Everybody can observe their inventory. One, I think your second largest customer seems to be doing better, but that your number one customer, the days of inventory has come down, but still well above the where historically it has been. What's your read into that situation and how does that change, how does it impact your business? Let's say, do you have, do we have to see the days of inventory to normalize to a certain level before you see a pickup of a business for your number one customer?
Yeah. I mean, I think we've seen some pickup for the last two or three quarters, regardless of the fact that some of the orders are buffered by the fact that they already have it in stock, right? So we're not seeing the direct demand yet because of the high level of inventory. I think they got up to almost $5 billion at one point. I think they typically run in the low $3 billion, and I think they've come down about halfway in between. So there's still some more work. We weren't seeing the direct demand, you know. We had that headwind last year because we didn't see all the orders because they had it there.
So hopefully in another couple quarters that most orders that they receive, you know, will no longer be in their inventory and we'll see that directly come to us. And so not only do we see the second half, you know, starting to pick up as some more volume investments come into play, we think we'll see more of that coming our way. So that's contributing to the 5%-10% overshoot of WFE that we typically do.
Okay. I think we're at the end of the session. Let's wrap it up. And, Jim, thank you so much for your time. And, everybody please enjoy the rest of the conference today. Thank you.
Thank you. Thank you. Thank you.