Title

Presentation

Operator: Greetings, and welcome to the Veeco Q1 2024 Earnings Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. [Operator instructions]. As a reminder, this conference is being recorded. It is now my pleasure to introduce Anthony Pappone, Head of Investor Relations. Thank you. You may begin.

Anthony Pappone: Thank you, and good afternoon, everyone. Joining me on the call today are Bill Miller, Veeco’s Chief Executive Officer; and John Kiernan, our Chief Financial Officer. Today’s earnings release and slide presentation to accompany today’s webcast is available on the Veeco website. To the extent that this call discusses expectations for future revenues, future earnings, market conditions, or otherwise makes statements about the future, these forward-looking statements are based on management’s current expectations and are subject to risks and uncertainties that could cause actual results to differ materially from the statements made. These risks are discussed in detail in our Form 10-K, Annual Report and other SEC filings.

Veeco does not undertake any obligation to update any forward-looking statements, including those made on this call, to reflect future events or circumstances after the date of such statements. Unless otherwise noted, management will address non-GAAP financial results. We encourage you to refer to our reconciliation between GAAP and non-GAAP results, which you can find in our press release and at the end of the earnings presentation. With that, I will turn the call over to our CEO, Bill Miller.

Bill Miller: Thank you, Anthony. Veeco started the year with top and bottom line results toward the high end of our guidance. Revenue totaled $174 million, non-GAAP operating income $29 million, and non-GAAP EPS $0.45. Semiconductor revenue totaled a record for the second consecutive quarter, primarily driven by strong demand for our laser annealing systems. As announced in our press release earlier today, a customer placed a multi-tool laser annealing order, including a nanosecond annealing system, for their 2-nanometer gate all-around process. We’re looking forward to supporting their future ramp plans. I’d now like to provide an overview of the technologies driving the business today, our served available market expansion opportunities, and our investment strategy.

Our portfolio of technologies, highlighted by our new products, is gaining traction, and we continue to expand our footprint to new applications. New device architectures and shrinking geometries require the precision of advanced annealing solutions to increase performance. Our LSA systems continue to gain share at customers’ most advanced nodes, as evidenced by several recent wins. In ion beam deposition, our systems are the technology of choice to deposit defect-free films for EUV mask blank production, and we’re working closely with industry leaders to enable their roadmaps. Looking ahead, we’re well-positioned to serve growing demand for EUV lithography, as well as next-generation high NA lithography. Our wet processing solutions are used for advanced packaging applications, and we continue to see strong demand for high-bandwidth memory.

During the quarter, we shipped several flux-clean systems and receive follow-on orders from leading foundry and memory customers. Veeco’s strategy of investing in advanced logic and memory has enabled our semiconductor business to outperform WFE growth for three consecutive years. Moving forward, we’re investing in new technologies to expand our served available market to a broad range of new applications. Beginning with laser annealing, we expect our served available market to grow from $600 million to over $1 billion, inclusive of both our Laser Spike Annealing and Nanosecond Annealing products. We continue to receive follow-on LSA business for high-bandwidth memory, and are making progress towards winning additional memory customers. We’re equally excited to expand our nanosecond annealing SAM to a broad range of advanced logic and memory applications.

In ion beam deposition, we see opportunities for our SAM to grow to $350 million for front-end semiconductor applications where low-resistance metals are critical. And in the compound semi market, we’re focused on long-term opportunities within power electronics and photonics. Investment in our EVALuation program has been essential to our growth strategy and will remain a top priority. We have two nanosecond annealing and two ion beam deposition EVALuation systems outstanding with leading logic and memory customers. Each are progressing well and our team is highly focused on executing. We’re also targeting an LSA EVALuation system to a second leading memory customer later this year, in addition to EVALuation shipments in the compound semiconductor market.

I’d now like to take a deeper dive into two of our largest opportunities in the semiconductor market. Device scaling challenges at our customers’ most advanced nodes are driving the need for new annealing capabilities. Our nanosecond annealing technology offers a substantial opportunity to broaden adoption of laser annealing to new logic and memory applications. Due to our unique laser and architecture, our system can achieve a lower thermal budget and shorter dwell time versus today’s most advanced annealing solutions. This results in a shallow anneal that can impact only tens to hundreds of nanometers into the wafer, enabling industry inflections such as backside power delivery and 3D devices. Our NSA system can also improve performance by changing the structure and properties of the device, opening the door to several material modification steps.

As we look ahead, we see potential for initial high volume manufacturing orders from logic customers in 2025. We’re also pleased to see strong pull from memory customers excited to evaluate our system’s new capabilities. Turning now to ion beam deposition for 300-millimeter front-end semiconductor applications. Veeco is the industry leader in ion beam deposition technology, which is a key enabler in driving aerial density growth in the hard disk drive industry over decades. This core technology has direct applicability for advanced semiconductor wafer-level manufacturing by solving our customers’ high-value challenges. As device geometries continue to shrink, low-resistance metals are essential to maintaining device performance, and traditional deposition technologies are struggling to address scaling challenges.

Our ion beam deposition technology differentiates itself from incumbent technologies through its ability to achieve superior thin film properties, making it ideal for advanced applications where low-resistance films are critical. Based on Tier 1 customer data, our ion beam deposited tungsten and ruthenium films are demonstrating lower resistance compared to traditional deposition technology. In DRAM, this enables tungsten bit-line scaling while maintaining electrical performance of the device. For logic, ruthenium metallization can enable new integration schemes at future nodes. Looking ahead, we see potential for a high-volume manufacturing orders for memory customers in 2025. I’d now like to touch upon artificial intelligence and the role Veeco plays in the AI chip manufacturing process.

Growth of AI is having a profound impact on leading-edge product roadmaps requiring the most advanced technologies to manufacture higher-performance AI chips. As we look ahead, we expect several Veeco technologies to benefit from growing demand for AI chips. Our LSA systems for transistor formation and IBD systems for EUV mask blanks are production tool of record for GPUs and HBM DRAM. Equally as important, we see future opportunities for our nanosecond annealing and ion beam deposition solutions for each of these applications. With that, I’ll turn it over to John for a financial update.

John Kiernan: Thanks, Bill. Turning first to our revenue for the quarter. Revenue came in at $174 million above the midpoint of our guidance range, up 14% from the prior year, and flat sequentially. Semiconductor revenue increased 29% year-over-year and 5% sequentially to a record $120 million, comprising 69% of total revenue. Sales to the semiconductor market continue to be driven by strong demand for our laser annealing systems. In the compound semiconductor market, revenue increased from the prior quarter to $21 million, totaling 12% of revenue. Revenue from our data storage customers contributed 10%, and lastly, scientific and other made up 9%. Now turning to quarterly revenue by region. Revenue from the Asia-Pacific region excluding China totaled 42%, an increase from 34% in the prior quarter driven by semiconductor customers.

The percentage of revenue from China totaled 37% during the quarter, in line with the prior quarter, led by sales to mature node semiconductor customers. Revenue from the United States totaled 16%, followed by EMEA at 5% of revenue. Switching gears to our non-GAAP quarterly results. Gross margin totaled approximately 44% above the high end of our guidance. Operating expenses totaled $48 million in Q1, in line with guidance. Tax expense for the quarter was approximately $4 million, resulting in an effective tax rate of 13%. Lastly, net income came in at approximately $26 million, and diluted EPS was $0.45 on 60 million shares. And moving to the balance sheet and cash flow highlights. We ended the quarter with cash and short-term investments of $297 million, a sequential decline of $9 million.

From a working capital perspective, our accounts receivable increased by $4 million to $107 million. Inventory increased by $5 million to $243 million, while days of inventory declined to 218 days, and accounts payable increased by $12 million to $54 million. Customer deposits included within contract liabilities on the balance sheet declined by $25 million to $72 million. Cash flow from operations came in at $9 million, and CapEx was $6 million. Now, turning to Q2 non-GAAP guidance. Q2 revenue is expected to be between $165 million and $185 million with gross margin between 43% and 44%. We expect OpEx between $46 million and $48 million, net income between $22 million and $29 million, and diluted EPS between $0.38 and $0.48 on 61 million shares.

And now for some additional color beyond Q2. Based on our current visibility, we’re reiterating our 2024 revenue outlook between $680 million and $740 million. We also continue to target diluted non-GAAP EPS for the full year between $1.60 and $1.90 per share. With that, I’ll now turn the call over to the operator to open up the Q&A.

Operator: Thank you. Ladies and gentlemen, at this time, we’ll be conducting a question-and-answer session. [Operator Instructions] Our first question comes from the line of Charles Shi with Needham & Company. Please proceed with your question.

Charles Shi: Hi. Good afternoon. Congrats on the good results, steady execution for another quarter. Maybe the first question, I want to dig a little bit into the laser anneal product that you guys put out the press release this morning. You may have addressed this before, but I think at the top of my mind for a good amount of folks is the nanosecond anneal. Does it — is it the incremental opportunity on top of what you already have with the laser anneal on the best nodes or maybe one replaces another and maybe a little bit of cannibalization going on there? Thanks.

Bill Miller: Yes, Charles, let me give a little background. So we’ve been working, as you know, on future nodes for many years with our most advanced customers in logic. And our laser spike anneal is qualified at all leading logic customers for their gate-all-around architectures. Separately, we have NSA or nanosecond evaluations at two of the three leading logic players and have pulled from the third for an EVALuation system. So NSA is being evaluated for gate-all-around, but also backside power distribution and 3D structures. So as you mentioned in our press release during the quarter, we received both LSA and NSA orders for a new customer for their two nanometer gate-all-around project. But to get to the crux of your question regarding NSA potentially cannibalizing LSA, so historically the contact annealing step for a transistor consisted of multiple anneals followed by an LSA step.

As the customers are moving towards gate-all-around, customers are considering replacing one of the multiple anneal steps with LSA and they’re evaluating nanosecond anneal NSA for a follow-on step. Although, decisions have not been finalized with our customers, it does appear in gate-all-around that NSA does not cannibalize LSA for gate-all-around. And that being said, NSA opens up unique opportunities for the company in applications like backside power distributions and 3D structures. So I would say largely it does not cannibalize LSA.

Charles Shi: Got it. I understood. Maybe we’ll keep asking you this question, but hopefully we get a firm confirmation on this being incremental to LSA opportunities. Secondly, I want to ask about the HBM opportunity. So, specifically, I’m asking about LSA for HBM because going back a couple quarters, you guys were not so certain where the LSA was qualified is it a standard DRAM or versus HBM. But based on your prepared remarks, you did sound slightly more assertive that this is going into a high bandwidth memory. I just want to confirm, am I overthinking this or this is indeed the case? Thanks.

Bill Miller: So, Charles, I would say originally we were qualified in the logic chip in the first level of the high bandwidth memory. And now we’ve been qualified with one customer, with our one customer, for the peripheral logic on the DRAM stack, DRAM devices. So we went from having the annealing step on the logic device to having, adding the peripheral steps on all of the HBM memory stack. Does that clarify for you, Charles?

Charles Shi: Yes, yes, yes. Definitely, definitely. It sounds like it’s the baseline, right, that the HBM, that’s the first one. But the second one I want to clarify even more. You said there’s a peripheral circuit on each of the DRAM stack, but isn’t that part of the standard DRAM manufacturing also? What I’m — I’m just trying to think, how much of the opportunity is there? Is it really only limited to the second application you talked about? Is it really only limited to HBM or that applies to all the standard DRAM, DDR5 or more advanced than that DRAM there? Thanks.

Bill Miller: So that’s the area I don’t have crystal clear understanding. I think to the first order, the peripheral logic in the HBM stack and in standard DRAM are very similar, but I’m not sure of the exact details of that to answer that affirmatively yet.

Charles Shi: Thanks, Bill. This has been very helpful. I’ll hop back to the queue. Thanks.

Bill Miller: Thank you, Charles.

Operator: Our next question comes from the line of Rick Schafer with Oppenheimer. Please proceed with your question.

Wei Mok: Hi, this is Wei Mok on the line for Rick. Congrats on the results and the LSA, NSA order announcements. So my question is, considering you guys didn’t update your 2024 outlook, is it fair to say that these incremental orders start to shift in 2025? So I was curious, how do you guys think about that outlook looking into 2025?

Bill Miller: Yes. So, Wei, when we gave our guidance for the year, this business, if we shift one system at the end of this year, beginning of next year is timing, really doesn’t have a significant impact on our full year view for 2024.

Wei Mok: Okay, great. Thank you. My next question is on backlog. I think you guys have around $500 million in backlog entering the year led by semi. So considering this backlog, what is the mix of your LSA between trailing and leading edge today? And how does this mix look at the end of the year?

Bill Miller: Yes, I would say what we saw in our business in 2023 and into the beginning of 2024 was more laser annealing for the trailing node and particularly China so that we saw strength there. And that was sort of a change from if you go back prior to 2020 — 2023 where we saw typically two-thirds of the business would come from the leading edge. I think as we progress throughout 2024, the expectation is and we start to see engagements with customers that we would see a pickup in the leading edge in the second half of this year and could probably sort of exit the year with a more balanced backlog.

Wei Mok: Got it. Thank you. And maybe one last question if I may. It’s on NSA. Considering this is next generation and a more advanced technology compared to LSA, how much of an — like a pricing uplift do you expect from NSA compared to LSA? Thanks.

Bill Miller: Yes, we do expect a higher price for our NSA for the additional capabilities and value there. I would estimate the 10% to 15% higher ASPs would be our expectation.

Wei Mok: Okay, thank you.

Operator: Our next question comes from the line of Gus Richard with Northland Capital. Please proceed with your question.

Gus Richard: Yes, thanks for taking the questions. On your compound semi-business, can you give us a little bit of update on GaN power and sort of placements of EVALuation tools in that market?

Bill Miller: Yes, Gus. We’ve obviously, as you know, been working very hard in focusing our business in power electronics, namely GaN on silicon and silicon carbide. What we’ve been seeing is a transition from 6-inch to 8-inch over the last year and a half or so. But what we’re seeing is some customers are looking to leapfrog to 12-inch, particularly some of the Tier 1 silicon power electronics manufacturers are interested in introducing wide bandgap materials into their standard products. And so we actually are planning to ship a 300-millimeter EVALuation system to this customer in the coming quarter. So we’re excited about that form factor transition to 300-millimeter.

Gus Richard: Got it. And then your hard disk drive business has been kind of lagging. The drive business looks like it’s bottoming and starting to turn. Are you starting to see more activity there where your spares and services can kind of start to pick up in that market? Or is it just still looking flattish?

Bill Miller: I would characterize it, Gus, as clearly our customers have been running at very low utilizations, and clearly that’s been reading out in our service number in terms of spares and whatnot. I would say when we look at our data storage kind of run rate business, kind of weekly and monthly, it’s been running at very low levels here. And maybe we’ve seen a number of weeks, maybe a month or two of a bit of an uptick, but clearly nowhere near kind of the historical run rate. But maybe going in the right direction, I’d say.

Gus Richard: Got it. That’s helpful. Thanks. And then the last one for me is an ion beam depositions a couple quarters ago, you had a customer that was going to utilize it for EUV mask pellicles. And I was wondering if anything ever became of that if there was any interest in an additional system, any color there?

Bill Miller: Yes, we’re — they’re actually, I would say, pressuring us pretty hard to accelerate the shipment of that. That tool is going to ship here in the coming months. And we’re prepared to install it and fully support it at the customer. And assuming we have success, I wouldn’t be surprised that we have another follow-on order in a year or so.

Gus Richard: Got it. Okay. That’s it for me. Thanks so much.

Bill Miller: Thanks, Gus.

Operator: Our next question comes from the line of Dave Duley with Steelhead Securities. Please proceed with your question.

Dave Duley: Yes. Good afternoon. I was looking at one of the slides in your slide deck with the architecture of the high bandwidth memory next to the GPU die. And you talked about earlier to someone else’s question that you’re working on, I think, both the logic-based logic die and the high bandwidth memory stack itself. So I was kind of wondering, is that true you’re working on kind of two steps, both the individual high bandwidth memory and then the base logic die, if that’s the case? And then also, help us understand what sort of performance improvement the customer gets from using an LSA tool versus a flash annealing tool.

Bill Miller: I would say, just to be clear on that cartoon, I don’t have it right in front of me at the moment, but I would say the comment was not the GPU kind of left hand side of the picture, but it was really about the logic die underneath the stack.

Dave Duley: Correct.

Bill Miller: So we are qualified for that. And now, in the cartoon, you see four blue HBM DRAM die. There is actually peripheral logic on each of those four dies. So you can imagine the opportunity, and this is just a cartoon, would be multiplied by the amount of area of — times four for the HBM DRAM die. Does that answer your question?

Dave Duley: Yes, partially. And then as a follow-up, similar to in the foundry and logic space, where you’ve gone from one annealing step to two or three at a couple of these leading customers. I think there’s multiple annealing steps in this process as well. Do you think there is opportunity for you to repeat that performance you’ve seen in high-end foundry and logic, pick up more than one annealing step in the stack?

Bill Miller: It is possible. Today, we are qualified, my understanding is, for just one application step, but there are — there is potential opportunities and we actually are working with our customers’ R&D group to address that. And I would also say that, going back to your previous question on the performance, it’s very hard to attach the speed of the HBM device, but I can tell you that the customer does have very high-performance HBM. And I’m not sure it’s directly attributable to LSA, if I did, I would love to tell you but…

Dave Duley: Yes. Okay. And then, as far as, I’m assuming that you’re going to grow your HBM business with the current customer, would you expect to start to recognize revenue from the other customers? In other words, have the other customers started to — the interest levels started to increase where things are progressing with these other guys?

Bill Miller: Yes, Dave, we’ve been doing demos with the other two customers for some time, and we are really pushing to target an EVALuation system to that to a second customer by year-end or early 2025 at the latest. So that’s an internal goal that we have is to try to place that second memory EVAL system. And just like we did in logic, kind of win one customer, win one application, and then expand more applications and more customers. That’s definitely part of the plan.

Dave Duley: Okay, thank you.

Bill Miller: Thanks, Dave.

Operator: Our next question comes from the line of Mark Miller with the Benchmark Company. Please proceed with your question.

Mark Miller: Congratulations on your orders. Just wondering, what is your outlook for data storage this year? When do you feel that data storage will start feeling a growth from AI-type applications? Are we still a year away there?

John Kiernan: Yes, so thanks for the question, Mark. Maybe this is a combination of Bill and myself. I’ll start with our outlook for data storage for the year. So for the overall business, including our spare parts and service, we expect our data storage business, compared to last year, to be flat to up 10%. And we have good visibility into our systems and our systems backlog and confirm ship dates from the customer, et cetera. So I am pretty — I think that’s within a pretty tight range there. And I would say the upside to the range is if, as discussed on one of the earlier questions on the call here today, if the service business picks up a bit quicker than we’re forecasting.

Bill Miller: And to answer the – your second question on the impact of AI on the data storage industry, I mean, we’ve been speaking with our customers, and two of our largest customers really do expect the rise of generative AI to drive healthy long-term exabyte growth, which is really critical for our business. And they are projecting growth rates on – at the 20% kind of CAGR number in the kind of mid to long-term, which would definitely be the bit growth rate that would drive a healthy industry at an — a healthy equipment industry to support it.

Mark Miller: Thank you.

Bill Miller: Thank you, Mark.

Operator: There are no further questions in the queue. I’d like to hand the call back to management for closing remarks.

Bill Miller: Thank you, operator. I want to thank our customers and shareholders along with the Veeco team for their continued support as we execute our growth strategy. Have a great evening. Thank you.

Operator: Ladies and gentlemen, this does conclude today’s teleconference. Thank you for your participation. You may disconnect your lines at this time and have a wonderful day.