Perfect! All right, welcome back to the Barclays Tech Conference. I'm Tom O'Malley, US Semiconductors and Semicap Equipment Analyst. Happy to have Bill Miller, CEO, and John Kiernan, CFO of Veeco. Thank you for being here, guys.
Great to be here, Tom.
As we do every year, we come back, and Veeco has a host of new products and wins that we need to discuss. But I think the overarching view that I kind of want to take is, you guys have managed to traverse a whole year, and had wins in data storage and some large macro challenges, and that's largely on the back of your semi business.
Mm-hmm.
Maybe you could start there. Where have you seen success in the semi business? Maybe give a recap for those who haven't been listening in carefully for the last couple of quarters. And then, what are your expectations for growth in the semi business from here on out?
Yeah, great. Thanks for the helpful question, Tom. I would say we're actually really excited about the growth we've actually executed here in 2023. We're probably going to grow about 10% in our semi segment. It's the largest segment of our business, and really, that's on the back of a few things. In the LSA business, we this year were able to win our first application and ship volume for our third front-end leading edge logic player.
And we've also now started shipping volume laser annealing equipment to our first DRAM memory maker. And so in an environment where DRAM is down, Veeco's gone from basically 0 market share in memory to now having some business. And so those are two major incremental opportunities for us.
Then I think farther out, we have shipped an evaluation system, our next generation annealing tool, that we're really excited about. And we've also shipped our first two ion beam depositions for front-end semi for low-resistance metal. So we think the future—although we did well in 2023, we think the future in 2024, 2025 and beyond is pretty exciting.
Awesome. So, you just mentioned one of the questions I was going to ask, so you're ahead of me here, is that you had a second shipment on the ion side. Is that also going to be a memory customer? And can you just talk about why that technology is gaining traction in the market today?
Sure. Yeah, we're very excited to announce that we did ship our second ion beam deposition system for low-resistance metals. So what the industry, the semi industry, both in memory, and yes, the two first tools are going to memory. I forgot to mention that, but there's also huge opportunities in logic as well.
What the industry is seeing is that as lines and spaces continue to shrink, the actual resistance of the material itself really becomes very important. The traditional PVD technology is really running out of gas in terms of its ability to reduce the overall resistance of the metal. So we've demonstrated the ability to show about a 20% lower resistance.
So when we share that with customers, they're really excited and see this as a big opportunity for them, and so we're really excited to be successful with that.
Very helpful. And then when you think about the ion beam opportunity as a standalone sliver of your business, can you talk about what you think that TAM could be for you guys? And then where do you see, you know, your growth as well? Because I think you just mentioned some of the qualities as to why you guys are being successful, but is the market going in a certain direction, whether it's kind of playing more into your hands?
Yeah, I think as lines and spaces continue to shrink, or the bit lines are such a critical part of the DRAM memory device, this ability to deposit low resistance becomes more and more important. And so we size that market in the next few years. Today, it's zero for us. We size that at about $250 million incremental opportunity for us to serve the market.
Okay, so you've got the ion beam side, 250. Now, I always have to recall here, and you guys have updated it a couple of times, so let me know if I'm right, and please update. So you guys originally talked about the LSA business. You guys had talked about a $300 million opportunity with a couple of customers, and then memory maybe doubling that. Can you talk about, one, where you are with LSA today, and then recently, and I think as recently as last quarter, you talked about NSA.
Mm-hmm.
What is that opportunity? Explain to people that don't know, you know, what part of... You know, why is that different than LSA?
Right.
Let's start from the base.
Okay.
Yeah.
So yeah, we size that market today, including logic and memory, at about $500 million with today's laser annealing-
Got it
Tools. Most of that business today for us is in logic, and there's opportunity to grow our market share in LSA, memory. Nanosecond annealing, the tools we're shipping for eval this quarter, are really. They actually heat and cool the wafer 1,000 times faster than our laser annealing. And so that allows only the very top surface of the wafer or the device to be heated, and that layers below that are unaffected by the energy that you're putting in at the surface.
And so that opens up opportunities like backside power distribution networks, where you want to anneal the back of a wafer, but not affect the front of the wafer. Also, for material modification, where you're building a stack and that you want to anneal this top surface, but not the surfaces below it.
There's a lot of opportunities that we see for potential in the future.
And when you think about additional TAM-
Mm-hmm
For your existing LSA business, how big do you think the NSA opportunity would be?
So I would say if we flash out, say at a 2027 time frame, we're sizing the total laser annealing market at $900 million, and we think $600 million will be our traditional laser annealing, from 50% logic, 50% memory, and an incremental $300 million for the nanosecond annealing. Also, probably about 50/50 logic memory.
So help me understand, as a non-engineer, I think it's very useful, the way you described just the wafer. But when I think about existing technologies, you're saying 1,000 times better, right?
Right.
The first question we're going to ask is, why isn't this cannibalistic? Why is this not cutting into your current tool opportunity? Are they playing at different steps? Help me understand why those are both able to sell into the market at the same time.
Yeah, I think, laser annealing will probably be the larger market because as you know, back 30 or 50 years ago, people pushed wafers into the center of an oven and then looked at a stopwatch and pulled it out, right? So now, with flash technology, we can heat it up for half a second, a millisecond kind of time frame.
And today, we can heat the wafer up for 150, 200 microseconds, and so that goes through the whole wafer. By heating, using a nanosecond wavelength laser, which is what we're doing, we can heat the wafer up and move the laser very quickly and heat it up only for a very short amount of time, and just the amount of energy you're putting in only stays at the surface.
There really are two distinct classes of problems. You're heating through the wafer and then only heating on the surface of the wafer.
When I look at where the opportunities lie for you guys, you mentioned backside power, and I know that you clearly have some traction. You used to talk about different steps at logic customers. There were, you know, 10 steps. You were in 3 of them, and you sized those steps, and I think you kind of backed away, like, "Look, like, that's the way we originally talked about it. We don't talk about it that way anymore."
But how about with the memory side of things? You know, is there a way to be more granular in your opportunity and say, look, ultimately, the TAM can be about the same size, right, logic? Is there a way that we can get a little more deeper, like unique wafer starts per foundry? Is there, like...
Help me understand a little better how that opportunity plays out.
Yeah. Yeah. So the laser annealing opportunity that we've just recently won, we originally were really only in their highest performing memory, high bandwidth memory applications. We've since learned that we now are qualified more broadly for their most advanced application. That's one application.
So right now, we have one customer, one application. There are obviously three major memory players, so there's opportunities for us to grow there. But also, our customer, we're working with our customers in product development, trying to qualify other applications. So that's, you know, it, it's... The steps actually are in some ways similar. There's some void removal steps, there are some direct annealing steps, very similar to logic, some are a little different.
But yeah, I think I'm hesitant to go in and start naming all these steps because it was really good when we were first starting out. I've got one customer, and I've got one application. It's called contact, and then we could walk through it with everybody. But now, customers are using a tool for all of these applications. So it's very hard for us to segment how many passes this application is versus that. So it's a good opportunity.
It's good to have too many steps to name-
Yeah.
In a given situation.
Yes.
That's good. So, when I think about how you scale at a given customer, is it generally the number of machines on a given line increases, or normally do you scale when more lines? So I guess a better way to go about it is, is it technology evolution or is it just throughput that helps you scale more?
I would say, obviously, it's technology that gets us in the door, but, you know, if we're qualified at an application step, you know, we're running 15,000 wafers per month, or 30, or 60. A lot of that is scale, what you call.
Then, I think a consistent theme at this conference is that, you know, this far has been AI.
Mm-hmm.
When you look at those two opportunities that we just talked about, LSA and ion beam deposition, which will benefit more from the AI side? Because I know in your deck, you have a very handy, and those in the crowd should spend some time there as well, a very handy graphic showing where play-
Mm-hmm.
with some GPUs and some HBM sitting next to it.
Yep. Mm-hmm.
Which of those do you think is a more lever driver to have?
I would say today, we have the process tool of record in GPUs with today's LSA systems for logic, as well as any GPUs that are made with EUV. We sell the equipment to make the EUV mask blanks. That's a real process tool of record. And HBM, as I just mentioned, we just talked about, we've now won one memory customer for their HBM opportunity, so we're process tool of record there.
What's actually as exciting is, we're working very hard in both the GPU and the high bandwidth memory to position our nanosecond annealing tool and our ion beam deposition system, both of which are under evaluation. And we think this could be really quite an enabling thing for us in the future.
Something that we've talked about a lot over the past year, and you guys have spent time on the call talking about, is just the trade-off in your LSA business from a geography perspective, where you have, you know, varying geographic exposures moving up and down throughout the year, and I think that that is something that many companies saw.
Mm-hmm.
Could you talk about what you think your run rate normalized geographic exposures in your LSD business, and what happened with China and why it became such a big portion of that mix? Do you think that that's all sorted about, or do you see China at an elevated rate into next year as well?
Let me start, and then John can fill in some more color here. I would say historically, we spend a lot of time talking about our leading-edge wins because that's where our R&D investment's going, that's where the focus of land and expanding our strategy is. And trailing edge, we treat more opportunistically.
So I would say historically, two-thirds or maybe a little bit more of our business was leading edge, and one-third was trailing edge. I would say this year it's probably flipped, be two-thirds trailing, one-third leading. And specifically on China, John, maybe you can add more.
Yeah. A lot of that trailing business did come from, you know, short node applications in China this year. As a company, we saw about 30% of our revenue come from China this year. Last year, it was about 20% of the company's revenue. As Bill mentioned, as we look out, you know, further, you know, our expectation with these new technologies that Bill just spoke of, whether it's nanosecond annealing, whether it's expanding, you know, laser annealing with more memory applications.
We've penetrated one customer so far. We're looking to, you know, penetrate another customer there. We would see that, you know, 1/3, 2/3 is sort of put back more towards our expectation over a longer period of time to move more towards the leading edge. I would say that if we look at, you know, China right now as a percentage of the business and what our visibility is, it's hard to, you know, predict next year at this point, will China business be up, will it be down, will it be flat if we look at the whole year next year?
I would say where we are right now is that interactions, engagements with customers are still very strong. Customers are still talking about investments in more node capacity, and our visibility at least into the first half of next year, you know, we should move in with a reasonable backlog there. I think it's once you get past the halfway point next year into the following year, it's really more difficult.
Yeah, it's not a phenomenon that is unique to you. You've seen it across the market with accelerated China spend. I guess the debate now is: How long does that last and the sustainability of that? So I guess the follow-up question is, you talked about what you think is your... You have, you had your China close this year.
You just gave some comments. I think you've mentioned publicly that you don't believe that any restrictions on equipment is going to impact your business. But could you talk about what you see in the moving regulatory world, and, to then anywhere, are you careful about which tools you can turn them on?
Yeah. So, I'll start with the laser, you know, annealing side. There's not specific licensing requirements that, even with these new regulations that just came out in October, that change any of the regulatory requirements from a, you know, technology perspective there. So it's not a type of technology that we would generally have to get, you know, a license to in order to sell to China.
Now, of course, it is very difficult to predict where, you know, regulations and export control may go, you know, in the future. You know, so for us, this really our engagement, these new technology engagements with customers at the leading edge, investing in these EUV programs with the customers, all meant for the long term, to drive our business more towards the leading edge, and as Bill mentioned earlier, you know, be more opportunistic.
If there's trailing edge or trailing node business that's available in China or in other places as well, yeah, we'll sell our existing equipment there, but you know, counting on it as we look out into the future.
I wanna kind of pivot to the data storage side. So it's been remarkable that you've been able to grow your business despite what has been a pretty extreme headwind in the data storage space. I think if you look at end customers, they're talking about early stages of a recovery, but you guys have already talked about seeing some positivity.
Can you talk about your expectations for that business? Are you ready to set and bottom, and we're going to improve from here? Are you setting common trajectory off the bottom?
I'll take that one as well, Tom. So what we've seen in the data storage business for us right now is we've had about two years where, you know, customers have invested a lot less. So in one sense, yes, our, you know, business is up a bit this year, low single, you know, digits in terms of growth over 2022. But 2022 is a very, you know, muted year in terms of our customers' capital spending.
You know, so your question is, have we seen, you know, bottom yet, or we've seen signs of recovery? We've not yet seen any meaningful sign of them making, you know, increased investments, you know, in capacity. So as we look at it today, you know, we don't see a significant change.
So as we start to roll into this upcoming year, and typically, we work with a longer lead time and visibility with our customers. We've been in this business with them. But what we are encouraged about is that customers continue to talk about technology investments.
They're investing in this, you know, HAMR, you know, technology, and that to the extent that they continue to invest in technology to keep the advantage over solid-state drives, so these larger, you know, format drives, I think that's good for the industry and their business.
The other area is that they continue to forecast or talk about that storage into the cloud and these large, you know, format, you know, drives, that they still see a 20%-25%, you know, CAGR going forward there. So I think the investment, you know, in new technology, growing, data stored in these large formats, give us, you know, opportunities as the market recoveries and as a leader of selling equipment into this market to see recovery.
We're not just yet calling that 20-25, from our perspective. But one thing I think is really important here, you know, the data storage, you know, business and the technology is really, you know, important, you know, to the company and the company's strategy, because as Bill mentioned, our foray into selling ion beam deposition equipment from companies that make the EUV mask blanks.
And these new tools that we're shipping out on an evaluation, these 300 millimeter ion beam tools for low-resistance metals, this comes from the decades of experience in working with our customers to make the film, you know, magnetic head here. So ion beam is important to us as a core of the technology, not just to serve the data storage market.
Helpful. Something that I've been meaning to ask for a while, that I haven't got the chance to, is clearly there's a material change in the storage business move to HAMR. You guys have always talked about how you leverage more heads and more complex heads, more passes, right? Is there a change that you can point to with HAMR in the complexity of the heads?
Because I know that there is an increased complexity with each of them that would help you or benefit you. And can you contextualize that in a number form or any way to be on?
We actually, I think in some of our investor materials, do kind of talk about an increase in the number of steps, kind of at a normalized number. It's substantially more.
Yeah.
I mean, it's a lot more deposition and etching. That's what our ion beam dep and ion beam etch tools are for. As the heads are more complex, it drives a lot more steps and a lot more passes. I think we have it, I just don't have it right off the top of my head.
Well, yeah, I think what we view is, in one sense, they go to HAMR.
Yeah.
There's greater areal density, so they're able to pack more storage onto individual, you know, you know, platters there. So you could use the same amount of platters today, in one sense, to have more storage. So that would be, you know, sort of a negative to selling more equipment to the customer.
Mm-hmm.
On the flip side of it, as Bill mentioned, the complexity of this third element in the head, of having a laser in the head that, you know, sort of heats the element, is significantly more complex. So all the analysis that we've done and our marketing teams have done, and again, we're very close, we've been working with these customers for decades, is that they kind of offset each other. So in the one sense, yes, there's more steps, there's more complexity in HAMR, but they're more...
So I think for us, it's really viewed as kind of neutral. When is the, you know, technology shift to HAMR? So I think for us, the more we continue to grow the amount of exabyte, you know, out, probably as important as an element than, is it more HAMR drives? Is it more non-HAMR drives? I think that for that, we view it as pretty neutral.
Helpful. I want to shift the conversation to compound semi. So, I think last year we talked about this as well, and we talked about an area that you wanted to move in and clearly transition in the market from more of a photonics geared market to one of more compound semi with both silicon carbide, GaN, et cetera.
Where are you in your process of catching up to the market? And, in terms of the current platforms that you have, do you think that you're capable of competing with those? Is it a marketing decision that needs to change, or customer subset, or coming to you? Just help me understand, you know, the strategy there and, and what...
We've spent this year of 2023 and end of 2022 kind of retooling our product lines. We made an acquisition of a small silicon carbide epi company the beginning of this year. We've installed that tool in our facility, in our lab in New Jersey. We're bringing the tool up, we're qualifying it. We're getting ready now to start running customer demos here in the near future.
From our discussions with a number of tier one silicon carbide players, they remain quite interested in the capability of the equipment set. We feel pretty good about that opportunity. You know, I went to this IEDM event in Europe the last two years, and I would say two years ago, it was 99% of the discussion was silicon carbide, 1% GaN on silicon. This year, I would say it may have been 70% silicon carbide, 30% GaN on silicon. So a lot of more interest in GaN on silicon.
And so what we're doing is we're positioning ourselves at 200 mm GaN on silicon, but also, we're developing a tool for 300 mm GaN on silicon, and there's a lot of tier one power companies that are very interested in GaN on silicon. And so we're working with on a number of those leaders, and our plan is to put an evaluation system out in 2024 for that.
So it takes some time to get the ball rolling, but we've been pushing the ball for a while.
Yeah.
As you know.
No, I think that something else that's going on while you tried to push the ball forward is that you've seen, at least in the last quarter, a small reset in the total market for carbide. Are you seeing that in conversations with your customers? Are people reducing capacity, or is it really just a blip on the radar in terms of we're going to continue to see this market kind of grow?
Our conversations are more of it's a blip on the radar, and their conversations with us are your tool is the right tool set for us for the future. And those type conversations aren't as impacted by current views. So, we haven't seen people not losing interest because the market is momentarily saturated.
Got it. I want to move on to the gross margins, and this is more John's world. So, you've really seen a great increase from 23, up from the low- to mid-40s. You noted the last earnings call that you will continue to make progress towards the goal of 35. Do you think there's an opportunity to reach this target model by the end of calendar year 2024?
I mean, you've got a lot of different moving pieces of the business, and again, you're not guiding 2024 yet, but a lot of what you've seen this year in terms of headwinds have been external to the business, and those should be rolling off. So tell me what percentage of the, you know, try to frame for me how much you can get just from having those.
Yeah. So, yeah, I mean, we're really happy the fact that, we're able to, increase our gross margin this year over our initial estimates at the beginning of the year. As you mentioned, we guided in 41%-42% gross margin, and this year we're now, updated our guide, so we believe that will come in full year, you know, 43%, but that's been increasing as the year, progress there.
And, you know, gross margin is an area, that we're really focused on, and we've constantly got a number of programs in place, around gross margin. So what we, what we saw this year is that really internally focused on getting more efficiencies out of our various processes in manufacturing, both to reduce the time to build tools, reduce the time on installing tools, and things of that nature to take advantage as volumes are going up, we're not increasing our costs there. So that's been, you know, favorable for us.
I think the other thing is, we've seen less of an inflationary factor this year than we thought at the beginning of the year, particularly on the material side. So that's been good, that material price increases have been less now than original expectation. You know, an area that, that many of us saw, you know, headwinds over the last couple of years was in the area of freight logistics costs have just increased so, you know, dramatically.
We've seen that, you know, you know, pull back a bit there. So all this we focus and we're able to gain some improvement this year. As we look into next year, same areas of, you know, focus. It is a focus of us to continue to make, you know, progress towards this 45%, you know, gross margin, you know, target, you know, that we have as we bring, you know, volumes up.
Because the other area, and that we saw in the last quarter in particular, where we saw a nice bump up in the margin, we did have a pretty sizable revenue increase in Q3, and we saw the impact of the volumes there. So if we can get the volumes, you know, focus on the things that we're focusing on, you know, we see the opportunity to continue to make progress towards that 45% gross margin goal.
Very helpful. I wanna—we have about a minute left here. I want to open it up to the audience in case anyone had any questions. All right, I guess I'll, I'll conclude with just as you look into 2024, traditionally, as you go throughout the year as to how the year will shape up, but within the semi business, you guys seem excited about NSA, you seem excited about IBD. Is there anything that you would want to voice to investors that you don't think is appreciated enough about what you think will happen with your business into the next year?
Yeah, I think, we've been excited about the last couple of years. We've been able to grow our semi business, grow our semi business quicker than WFE, right? And obviously, as Tom mentioned at the beginning here, WFE was a really challenging this year. So our focus is continued to try to outperform, you know, WFE by opening up these new market opportunities for us.
So we're excited about our annealing product line. We're excited about penetrating more, you know, DRAM, you know, customers and memory customers, you know, as we move forward. And we're going to continue to try to target growing our semi business quicker than WFE.
Very helpful. Thank you both for being here.
Thank you, Tom.
Have a great rest of the week, and thanks again for being.
All right. Thank you, Tom. Thank you.