Axcelis Technologies Inc
NASDAQ:ACLS
Earnings Call Analysis
Q4-2023 Analysis
Axcelis Technologies Inc
Axcelis culminated 2023 with exceptional financial milestones, achieving record revenues of $310.3 million in the fourth quarter, contributing to $1.13 billion for the full year. This success story is buttressed by a 23% growth in revenue year-over-year, a striking performance particularly amid a considerable industry downturn. The impetus behind this growth is the robustness of the implant-intensive power device segment. Earnings per share for the fourth quarter also exceeded expectations, finishing at $2.15, surmounting the company's revised guidance and culminating the year at an impressive $7.43 per share.
Axcelis experienced vigorous sales in China, representing a 49% share of system revenue during Q4, and 46% for the year. This was a remarkable display of geographic market dominance, particularly in the power device segment. Other notable contributors included Korea at 18% and the U.S. at 11% for the quarter, while Europe, Japan, and Taiwan also played significant roles over the annual horizon. The U.S. accounted for 15%, Korea for 14%, and Europe for 11% of yearly system revenue, demonstrating a well-spread global presence.
The year 2023 saw an 88% share of Axcelis's shipped system revenue coming from mature market segments, with memory and advanced logic accounting for 10% and 2% respectively. However, projections for 2024 indicate an expectation that memory markets will gain traction towards the year's end, potentially contributing to less than 10% of total systems revenue. Looking further ahead, the NAND segment is anticipated to rebound in 2025, with both DRAM and NAND forecasted to surge during that period.
Silicon carbide has propelled Axcelis's growth trajectory in 2023, with an extensive and varied customer base showcasing faith in the company's Purion Power Series products. Axcelis's distinctive technological leadership has allowed it to capitalize on increasing adoption rates, especially with the Purion H200 and Purion XE silicon carbide systems. With three ongoing evaluation projects for the Purion H200 across different geographies, and the hallmark of being the sole ion implantation company offering complete recipe coverage for power device applications, Axcelis is set to further penetrate the market.
Axcelis has deployed targeted initiatives aiming to expand its market share in advanced logic segments and in Japan. By shipping its most sophisticated high current implanter, the Purion Dragon, to a leading advanced logic research institute and another for evaluation to a prominent logic customer, the company is reinforcing its footprint in these critical markets. In addition, ongoing engagements with several Japanese customers across different market segments bolster Axcelis's presence in Japan. These focuses are anticipated to substantially contribute to the company's revenue target of $1.3 billion by 2025, while sowing seeds for sustained growth beyond that time frame.
The company's projection for Q1 2024 estimates revenue to reach approximately $242 million with a forecasted gross margin of around 43.5%, signifying a lower revenue against Q4 2023 but a stable profit margin outlook. Anticipated operating income and earnings per share for Q1 stand at roughly $45 million and $1.22 respectively. Underpinning these estimations is the strong systems backlog and steady demand in the power segment. Axcelis is cognizant of the current market weakness yet remains poised for anticipated industry recoveries, setting the stage for aspiration towards a $1.3 billion revenue landmark in 2025.
Axcelis exhibited a harmonious balance between sustained demand for its flagship Purion products and financial prudence, resulting in fourth-quarter revenues surpassing revised guidance. System revenue reached $241.8 million, contributing to an annual systems revenue of $883.6 million. An additional $68.5 million in Q4 revenue was accrued from Customer Support and Integration (CS&I). The year closed with a resilient earnings per share of $2.15, leveraging greater-than-anticipated revenues and margins, coupled with controlled operating expenses. Despite fluctuations anticipated in quarterly revenues, CS&I is projected to ascend to approximately $260 million for 2024 and reach about $300 million in alignment with the looming $1.3 billion revenue model.
Axcelis's financial pillars stood firm with a Q4 gross margin of 44.4% and an end-of-year attainment of roughly 43%. Operating income was recorded at $68.5 million in Q4, recovering from the slight dip experienced the previous year. Q4 incurred operating expenses worth $43.6 million, with predictions for the following year ranging between 13% to 14% of revenue. The company's balance sheet strength was unquestionable, entailing an abundance of $196.5 million in cash and cash equivalents and emboldening its financial strategies including the share repurchase authorization initiative. Finally, summarizing Axcelis’s strategic outlook, the company underscored its vigilant efforts to underpin market share, particularly in the power segment.
You don't have any saved screeners yet
Good day, ladies and gentlemen, and welcome to Axcelis Technologies Call to Discuss the Company's Results for the Fourth Quarter and Full Year 2023. My name is Michelle, and I will be your coordinator for today. [Operator Instructions] Please be advised that today's conference is being recorded.I would now like to turn the presentation over to your host for today's call, Doug Lawson, Executive Vice President of Corporate Marketing and Strategy. Please proceed.
Thank you, operator. This is Doug Lawson, Executive Vice President of Corporate Marketing and Strategy. And with me today is Russell Low, President and CEO; and Jamie Coogan, Executive Vice President and CFO. If you've not seen a copy of our press release issued yesterday, it is available on our website. Playback service will also be available on our website as described in our press release.Please note that comments made today about our expectations for future revenues, profits, and other results are forward-looking statements under the SEC safe harbor provision. These forward-looking statements are based on management's current expectations and are subject to the risks inherent in our business. These risks are described in detail in our Form 10-K annual report and other SEC filings, which we urge you to review. Our actual results may differ materially from our current expectations. We do not assume any obligation to update these forward-looking statements.Now I'll turn the call over to President and CEO, Russell Low.
Good morning and thank you for joining us for our fourth quarter and year end 2023 earnings call. Axcelis delivered record revenue for the fourth quarter of $310.3 million and $1.13 billion for the full year 2023. The implant-intensive power device segment enabled Axcelis to achieve 23% year-over-year revenue growth during a significant industry downturn. Fourth quarter earnings per share of $2.15 exceeded our revised guidance, while full year 2023 earnings per share came in at $7.43.Looking at geographic mix, China continued to provide strength, especially in the power device segment. In the fourth quarter, China represented 49% of our system revenue, with Korea at 18%, Europe at 12%, the U.S. 11%, Japan 5%, and the rest of the world 5%. For the full year, China represented 46%, the U.S. 15%, Korea 14%, Europe 11%, Japan 3%, Taiwan 2%, and the rest of the world 9%.Looking at the market segment distribution for 2023, the overall mature segment represented 88% of the shipped system revenue, memory was 10%, and advanced logic was 2%. Breaking down the mature segment in more detail, power continued to lead system shipments with 59% of total systems revenue. Silicon carbide made up 34% and silicon 25% of total system revenue, respectively. The general mature segment was 26%, image sensors were 3%, and DRAM represented the entire 10% of memory systems revenue.In 2024, the systems revenue profile will be very similar to 2023 with power continuing to be an area of strength for Axcelis. Silicon carbide and silicon IGBT combined are expected to represent approximately 60% of our system shipments for the second consecutive year. The general mature process technology market is expected to start out slower in the first half of 2024 and improve in the second half, dependent on economic conditions. Currently, DRAM is expected to pick up towards the end of the year and contribute less than 10% of total systems revenue in 2024. NAND is not expected to recover until 2025, when DRAM and NAND are forecast to have a strong year. Geographically in 2024, we expect China to represent 40% to 60% of our quarterly systems revenue, with the remaining revenue spread relatively evenly across the other geographies dependent on specific customer projects.The power device segment, and in particular silicon carbide, has driven our growth in 2023. We have developed a large and diverse customer base in this market, and we continue to win business from new customers as well as expanding our product footprint with existing customers. The full portfolio of Purion Power Series products is valued by these customers. New fab projects and customers often start up by establishing a core of Purion M silicon carbide tools and then adopt the use of the Purion H200 silicon carbide and Purion XE silicon carbide systems to improve productivity, cost of ownership, and device performance. As a result, we have seen a significant increase in the adoption and great success with the Purion H200 and Purion XE silicon carbide systems.Additionally, we continue to work with customers to further increase this opportunity and currently we have 3 Purion H200 silicon carbide system evaluations underway with customers in multiple geographies. 2 of these systems are 150 mm and one is 200 mm. Customers are using these evaluation units to qualify productivity-limiting recipes as they prepare to ramp to higher volumes. Also, by utilizing the higher energy and dose capabilities of the Purion H200 silicon carbide tool, customers can begin optimization work on their devices during the evaluation period. Axcelis is the only ion implantation company that can deliver complete recipe coverage for all power device applications. We are considered the technology leader and the supplier of choice, providing the best product family and manufacturing capabilities. This means that using Axcelis tools provides the lowest risk path to high-volume manufacturing required to support aggressive fab ramp plans.We expect the memory and mature markets will recover later this year, but during this slow period, Axcelis remains close to our customers, supporting their installed base and working with them on future technology and manufacturing needs. During industry slowdowns like this, customers have more time to collaborate with Axcelis on new technologies and product capabilities. We use this opportunity to focus our R&D efforts in key areas that will be critical to customers as they enter their next phase of growth. Ultimately, this results in shipping evaluation systems to customers and joint development engagements that help us grow our market share. Currently, we have an evaluation system with customers across nearly all market segments and multiple technical customer engagements designed to improve capabilities and increase our footprint across all segments.We have focused initiatives expected to grow share in the advanced logic segment and geographically in Japan. In 2023, we shipped a Purion Dragon, our most advanced high current implanter, to a leading research institute focused on advanced logic process development. We also have another Purion Dragon under evaluation with a leading advanced logic customer. These tools and the associated technical collaboration will be critical to the customer's development of next-generation logic technology. In Japan, we have seen initial success in the power market due to the strength of the Purion Power Series, and we are engaged with multiple Japanese customers in additional market segments. We expect these efforts to increase the Purion footprint in this important and growing geography.As the industry exits this downturn, Axcelis will return to healthy growth in the mature and memory markets. This, combined with continued strength in the power segment, is expected to drive Axcelis to our $1.3 billion revenue model in 2025. Additionally, investments being made in advanced logic and Japan will help drive our continued growth beyond 2025.Now I'd like to turn it over to Jamie.
Thank you, Russell, and good morning, everyone. We are pleased with our financial results for the fourth quarter and for the full year 2023, especially with the 23% year-over-year revenue growth during this industry downturn. As we enter 2024, the industry continues to deal with market weakness, but as Russell discussed, there are also clear signs of recovery and an expectation for a strong 2025. As a result of the current market conditions, we are guiding first quarter revenue of approximately $242 million with gross margins of around 43.5%, operating income of approximately $45 million, and earnings per share of about $1.22. We expect full year 2024 revenue levels to be similar to 2023, with revenue weighted towards the second half of the year. Power is expected to remain solid throughout the year with the mature markets and memory recovering in the second half. Our strong systems backlog and the expected recovery of these markets sets us up to achieve our $1.3 billion revenue target in 2025.Looking at our fourth quarter, revenue and earnings per share finished above our revised guidance due to solid execution and continued demand for Purion, especially in the silicon carbide power market. Q4 revenue was $310.3 million with system revenue at $241.8 million and CS&I at $68.5 million. Full year revenue was $1.13 billion with systems revenue of $883.6 million and CS&I at $247 million. Q4 earnings per share of $2.15 was driven by higher-than-expected revenues and gross margin, as well as lower overall operating expenses. This performance led to full-year earnings per share of $7.43.Despite softness in the general mature and memory markets, bookings and quoting activity for systems in the power segment remained solid and continued to support our revenue expectations. Bookings in the quarter were $236 million, maintaining our backlog at $1.2 billion, a portion of which stretches into 2025. Given the increase in installed Purion systems, we expect CS&I revenue to increase in 2024 over 2023. Although revenue will fluctuate quarter to quarter, CS&I should be modeled at approximately $260 million for 2024 and approximately $300 million for our $1.3 billion revenue model. Q4 gross margin finished at 44.4% and at 43.5% for the full year. In 2024, we expect to see year-over-year improvement in gross margin. However, quarterly gross margins will fluctuate based on product mix. We remain laser focused on margin improvement and have a number of initiatives underway to lower the cost of goods sold and to drive higher sales of Purion product extensions. Execution on these initiatives will allow us to model gross margin at greater than 45% in our $1.3 billion revenue model.Turning to operating expenses. The fourth quarter ended at 19% of revenue, better than our guidance, and at 19.9% of revenue for the full year. We expect OpEx in the first quarter of 2024 to be approximately 25% of revenue. The increase as a percentage of revenue is a result of the lower sales volume in the first quarter and the incremental investments we've made to support the higher revenue loads we anticipate in the future. OpEx as a percentage of sales is expected to decline over the course of 2024 given the higher volumes expected in the second half of the year. Investments in R&D will increase in 2024 to approximately 9.5% of revenue compared to the 8.6% of revenue we invested in 2023. The incremental funding of R&D will be focused on the continued development of our Purion product extensions and upgrades. As you would expect, we will continue to tightly manage spending while continuing to support the future growth of the business by solidifying our technology advantage in the specialty markets, increasing our footprint in the memory and advanced logic markets, and most importantly, continuing to invest in our employees and infrastructure to ensure we have the necessary skills, equipment, and facilities required to achieve our financial models.Moving to our balance sheet and cash flow. We ended Q4 with $506.1 million of available cash and generated $65.6 million of cash from operations in the period and $156.9 million for the full year. We continued to execute against our share repurchase program, buying back $15 million of stock in the quarter. In total, we've returned over $185 million of cash to shareholders since 2019 through our share repurchase programs.Before turning the call over to Russell for final remarks, I wanted to remind you that we will be participating in a number of upcoming investor events, including Wolfe Research's inaugural semiconductor conference in San Francisco on February 14 and Susquehanna's 12th Annual Technology Conference virtually on March 1. In addition, we intend to host a Capital Markets Day on July 11 of this year in San Francisco in the time slot we usually hold our technical symposium. At this event, we will provide our next long range financial model, discuss our expectations for the market, review our new product innovations, and introduce the team members that will help drive Axcelis towards its next phase of growth. We will provide more details on this event in the coming months, and we look forward to seeing many of you there.With that, I will now turn the call back to Russell for his closing comments.
Thank you, Jamie. Axcelis achieved record revenue of $1.13 billion in 2023 and is targeting revenue of $1.3 billion in 2025. This growth is achievable due to the same factors discussed last quarter. First, the implant TAM has more than doubled in the last few years and is expected to continue to grow with mature market segments representing greater than 60% of the total TAM. Second, power devices, especially silicon carbide devices, are highly implant intensive, and the general mature nodes have increasing implant intensity peaking at 28-nm. Third, high-value Purion product extensions were designed to optimize power and image sensor device manufacturing, making Axcelis the only company with a product line capable of covering all implant recipes in these key markets. This uniquely positions Axcelis to benefit from high growth in the mature process technology markets. And finally, Axcelis has strong long-term customer relationships and a fundamental cultural desire to win by making our customers successful. 2023 was a record year for Axcelis, but a turbulent year for the industry. I want to thank our employees, suppliers, customers, and investors for your continued support throughout 2023 and into 2024.With that, I'd like to open it up for questions.
[Operator Instructions] The first question comes from Craig Ellis with B. Riley Securities.
Congratulations on the very strong exit to 2023, guys. I wanted to start off with a question that combines some near-term items with some intermediate-term items. So, Russell and Jamie, from your color, it sounds like as we look at the first quarter guidance and then the way calendar 2024's linearity plays out with the inflection in the second half, that mix would be fairly even within systems across mature, foundry, and memory. Can you confirm that? And then on the latter part of that, what are the things that you see that give you conviction in the second half inflection as you look at your backlog and customer engagements, et cetera?
Craig, this is Doug. I'll take the first half of that question. So mix wise, power continues to be strong. We expect for the year power to continue to represent about 60% of our total revenue and silicon carbide will be 50% of our total systems revenue. So we do expect that to continue to be strong. As we get into the second half, we expect the mature markets to recover really tied to the economy more than anything, as consumer, automotive, and industrial start to return. And then we expect DRAM to recover ahead of NAND, with NAND being more of a 2025 thing. So if you look at the updated presentation, we're expecting close to 90% of our business to come from the overall mature markets and about 25% of it from the general mature.
And regarding our conviction about the second half, Craig. So yes, we do have a strong backlog. We do have solid business in power, especially in China. And in speaking with our customers, they're looking to start ramping their businesses in the second half of the year.
That's really helpful, guys. And the second one is more for Jamie. So, Jamie, it's real impressive to see how resilient first quarter gross margins are as volumes decline. And I'm hoping what you can do is provide some color on how mix and some of the other company-specific factors are playing out. And you indicated that calendar '24 gross margins could rise year on year. Can you give us any color on the magnitude of the increase that we might see?
Yes, that's a great question, Craig. The team has done a fantastic job of putting in place some initiatives here to, one, lower our cost of goods sold for the systems and try to drive some greater efficiency without necessarily having to raise prices for some of these products, given the competitive environment. In addition to that, we have identified some service and upgrade opportunities which are providing incremental margin opportunities in our CS&I business. And given that we expect CS&I continue to grow in light of the higher installed base that we continue to build out there of the Purion product platform, that part of the mix is going to continue to contribute incremental margin opportunity over the course of the year.And then on top of that, it's also where the systems are coming from over the course of the year. And we are seeing some higher volume in 2024 of some of our higher-margin products, shifting the mix a little bit towards that, especially in light of lower memory volume year over year. As you guys know, memory market is a little bit more competitive there. Margins are not as strong on those products as they are in some of our other areas. And so with the lower memory volume, we're also seeing benefit from that.
The next question comes from Tom Diffely with D.A. Davidson.
Probably for Doug. When you look at the TAM for '24 in your slides, you have it actually going up for the full year, but you're talking about your business being flattish. Just curious what the differences are there. Are there certain sectors that you're not as strong in that are doing well?
Yes, that is exactly what's going on, Tom, is we see the TAM for implant going up; power, where we are strong, is where we're taking advantage of that; we're seeing early recovery in advanced logic, which, while less implant intensive, Axcelis has a much smaller position in, one that we expect to grow over the next years. But in 2024, it will continue to be smaller. And as the mature markets grow in the second half, then we'll benefit from the increased TAM there as well. And as we get into '25, as you can see, the TAM continues to grow, we are expecting a good year across all markets in 2025.
Okay. And then, Russell, just a general question about how the year is playing out. If you think back a quarter or 2 ago, were you expecting a dip in the first quarter, the first half of the year before second half strength, or did the book of business look more continuous a couple of quarters ago?
I guess what I'd say is the power business has stayed solid. I think it's the general mature has softened significantly. I think you've heard that from a number of our customers. So really, I'd say that it's an evolving picture, and we now have a little bit more visibility in Q1 and the rest of the year that we wouldn't have had a few months ago.
Yes, and Tom, on that point, we had very -- as you noted in the call, we had very strong bookings in the fourth quarter of this year relative to our systems revenue and our backlog, again, we maintain that backlog above $1.2 billion for the full year. And as we look where we are today, we don't really see a meaningful change in the amount of backlog that we're carrying. However, we have seen some shifting in the timing of deliveries. And as Russell noted, that's consistent with what our customer commentary has been on that.
And then, Jamie, last question. When I look at the margins going back to '22 to '23, you had really nice revenue growth, very de minimis margin expansion, and yet you're projecting pretty healthy margin expansion over the next year. Maybe just take us back to, what stopped margins from expanding in '23 versus '22, and I guess why you're confident that it accelerates here going forward?
Yes, that's a good question. A lot of that has to do with memory mix in the period and then the efforts that we are taking, right? So we say mix within the year relative to memory, and then our CS&I business related to service upgrades and other opportunity sets that we see for 2024. But on top of that, a number of the initiatives we put in place to drive incremental opportunities on cost savings, specifically on the cost of sales line item, really are multiyear benefit providers to us. So these are things like the investments that we've made in the automated logistics center where we've consolidated our footprint here in the Beverly area. And then also the continued work of the R&D team to identify new opportunities, upgrades and services on the CS&I front, which, as you guys all are aware, does provide some meaningful uplift on mix.
The next question comes from Mark Miller with The Benchmark Company.
You mentioned you had 3 evals underway for your Purion H200 and also a Dragon eval with an advanced logic customer. Are there any other evals currently underway?
Yes, there's [ 8 ] evals underway, Mark. We've got one medium current tool that's at a DRAM customer, an Purion XEmax under eval for an image sensor company, the Purion H200 silicon carbide tools you mentioned for power device, and also a Purion VXE in a power device application. And then general mature, we have a Purion H and then the Purion Dragon in advanced logic.
So many customers, many applications, and many different products across the board there.
And what was the medium current, I'm sorry, customer?
DRAM.
DRAM, okay. Wolfspeed indicated a week ago that they're seeing very strong design-ins. They're a major silicon carbide, as you know, manufacturer. I'm just curious why your first half will be weaker, given what Wolfspeed was indicating, what would appear to be very strong design-ins. And I think 75% of them are from automotive.
Yes. I think, Mark, so we see strength in silicon carbide globally. Right now, there's more strength coming from the Chinese customers. The Chinese EV market, well, getting lots of press in terms of it slowing its growth rate. When you look at the number of EV companies, the breadth of the product lines that they offer, they are very focused in China on silicon carbide, not only for internal to China, but to be a global low-cost provider. So we're seeing strong bookings and continued strong quote activity from China. Throughout the rest of the world, it slowed a little bit over the course of the last quarter, but as you comment, many of our customers are talking about that picking back up as the automakers start to settle on their exact product plans.
The next question comes from Jed Dorsheimer with William Blair.
I guess the first one. I just want to put a finer point, Russell, it sounds like to a previous question, when you preannounced positively 3 weeks ago that you had insight that Q1 would be weaker, I just want to make sure. Is that the case that you knew that, as you started, the Q1 would be off by 15% or did you see any pushouts over the last 3 weeks? And then I have a follow up.
Yes. Jed, I'll take that. We have seen, again, the shifting in those delivery requirements over the past couple of weeks, especially as our customers now are firming up their CapEx requirements and the timing of those requirements over the course of the year, Jed. So the reality is, as we thought through the guide for 2024, our historical practice has been to make sure that we can provide the most meaningful guidance to the folks relative to that. And we historically have done that on this call. So it was a combination of factors there relative to the timing of that preannouncement.
And I appreciate how fluid things and dynamic things can be. I guess along those same lines, I know I heard Doug speaking positively on China. The average utilization for fabs in China is below 50%. Most are around 30%. So I'm just wondering what gives you the confidence that those orders materialize. Typically you would not see additional CapEx spend with such low utilization unless the tools are being repurposed for something else. So I'm curious, what gives you the confidence that in that bookings that you don't see additional pushouts in the power market in China?
Yes, so I think, Jed, a lot of it is the fact that the Chinese companies, I think the Chinese government has a long-term plan for silicon carbide and EVs. And so the utilization is probably a little less of a factor in determining their investment policy over the course of the next few years. And so we see a lot of new customers in addition to the larger silicon carbide customers in China. And so there's quite a bit of activity, despite your comment on lower utilization. I think they're also preparing for the fact that there is still expected to be a significant growth in EVs over the course of the next 10 years. Most of the automakers globally, outside of China, we'll say, have changed their plans a little bit over the course of the last 6 months especially. But none are really backing away from the fact that there'll be a significant number of electric vehicles. And a lot are moving to a combination of hybrid and electric, and hybrid, of course, utilize power devices and inverters as well.
And last question for you guys, and I'm assuming it's probably in the software, but I just want to ask it anyways. Most of the equipment companies that have sold into China, have been reengineered and are now being supplied by local vendors, with one exception, which is in implant. So I'm just curious, how do you gauge that with so much exposure to a market that government subsidies are literally tied to reengineering of the tooling? How do you have confidence that won't happen with your solution?
Jed, this is Russell. So there have been a couple of domestic suppliers, probably for 20-odd years, there's a couple of them, that have been working on knockoff medium current implanters. One thing I'd say that does insulate us a little bit is that these are highly-complex technical products and the software is a huge component of it. The operation of the machine, the recipe, tuning the setup is a huge part. So I would say that it's a very difficult technology to replicate. People have been trying without too much success to date. And I think there's another couple of things that go on here as well that we are an innovator, and we keep moving faster and faster, working with our customers to make sure they have the most up-to-date solutions that make them competitive, whereas typically the domestic tool manufacturers get left behind. So once a technology starts to plateau, then that's when foreign vendors get run over. And I saw that happen in a couple of other areas.
The next question comes from David Duley with Steelhead Securities.
I was curious about the memory recovery you talked about in 2024. I think you mentioned that memory would be 10% of revenue this year. And I think historical peaks were around 20%. But that was split evenly between NAND and DRAM. And I think you're talking about 10%, it's mostly DRAM. Could you just elaborate a little bit about the breakout of revenue there? And if it is going to be 10% DRAM, that's pretty close to historical peaks, I think. And just talk about what the drivers are behind that memory business.
Yes, Dave. So the number that we've got in the presentation is it'll be under 10%. So we're monitoring that very closely as the year goes on, since it's a second half situation. The drivers for it are basically getting back to a point where we start to see wafer start additions by the memory companies on both DRAM and NAND. We expect DRAM to happen ahead of NAND, and the drivers for utilizing capacity are HBM, which is currently seeing a lot of our customers shift capacity over. We see shrinks happening that will allow them to get more bits out before they add capacity or add wafer starts, and then they'll start to respond to demand. And we expect demand drivers like all the consumer and auto stuff, as it comes back, AI PCs look like they could be a big driver of DRAM. Microsoft has said they're requiring 16 gigabytes per AI PC for Windows 12 and AI PC. So there's a lot of good indicators that we'll start to see capacity additions as we get towards the second half and end of this year driving into 2025, where we expect it to be a very good year for DRAM. NAND we don't expect to really see a lot of activity until we get into the beginning of next, and so that's going to be driven by storage both on device and in data centers.
The next question comes from Charles Shi with Needham & Company.
Maybe I want to start with some of the commentary around the expected recovery of the general mature in the second half of the year. So can you remind us what kind of customers, what kind of applications you considered as general mature and how do investors get comfortable with a second half recovery of that part of the market? Because the CapEx announcement from mature foundries or some of the larger analog mixed signal IDM, I assume the microcontroller part of their CapEx is considered as a general mature isn't very positive. And how do people get comfortable with that outlook for 2024? That's my first question.
Okay. So the general mature recovery is likely to be very much tied to the economic recovery or the perception of economic recovery, I guess. And so it's consumer products, automotive, industrial-type products; in terms of device types, microcontrollers, analog, RF, all the little widgets that go into all these devices that we buy. Another strong place for it will be on the Internet of Things. We do expect that as AI takes off, it does drive another wave of IoT devices. Since AI is a data hog, we expect that to happen. We do see activity and our customers talking about second half adding capacity and building. And I think if you listen, as we listen to our customers directly and their public announcements, most are continuing with a reasonably healthy capital plan.
So the second question is about China. I think I heard you talking about China probably contributing 40% to 60% of the revenue this year. The last year's number seems to be a little bit below that. So it almost feels like you're guiding to China revenue to be up meaningfully this year. What's driving that? And did some of the pushout by the non-China customers actually help you backfill some of the slots for the China customers whose orders may be parked a little bit further down the road, let's say, 2025? So really just want to understand the dynamics here. Is that organic underlying China demand growth this year, or there's a little bit of puts and takes in terms of the manufacturing slots going out?
Okay. So, no, there is continued strong demand, especially on the power and especially silicon carbide, in the Chinese market. And so that is where most of the activity is, especially through the first half of 2024. And then we would expect the general foundries -- general mature foundries worldwide would then start to recover, and I think that's consistent with all of their public releases over the last couple of weeks. And so it's less to do with movement creating slots or whatever for China and more the activity and the bookings level and backlog that we're seeing from the Chinese customers.
Lastly, definitely the first half numbers are expected to be a little bit lower compared with certainly the second half '23. Is the mix in the first half '24, you're expecting something still similar, like 60% power within that 60% power, I don't know, maybe somewhere between 30% to 50% of the total being silicon carbide. Any color would be great.
Yes. So, Charles, for the year, [ 60% ] of our systems revenue will be power and [ 50% ] of our total revenue, or around [ 50% ], will be silicon carbide. And the remainder is mixed between the general mature, image sensors, and DRAM, primarily. And so we would expect that we would see higher percentage of power in the first half and then we would start to see the other markets come in the second half and contribute and change the percentages.
The next question comes from Christian Schwab with Craig-Hallum Capital.
I'm just curious what you guys' thoughts are on the unintended consequences of the U.S. government limiting advanced chip production in China, which has led to extremely strong investment in mature nodes. But now they've said now they're going to look into mature node legacy chip production, because as China's meaningfully increased production of mature chips, it's leading to a possible competitive situation for U.S. based companies selling similar chips as China tries to attempt to gain market share with that. What is the risk that they come back at some point this year and start making some semblance of restrictions on legacy chip semicap equipment, which would obviously, with 40% to 60% of your revenue, could be a material risk.
Yes, Christian, we watch that very closely. We don't expect that to happen on the mature nodes, at this point, especially on the power side, which is where the strength is, especially in the first half. But it is something we watch very closely, so we can't predict the future on government actions there. So it's something we just have to monitor and react to.
On the big DRAM memory recovery in '25, we've seen that every leading memory manufacturer significantly reduce production capacity, utilization of the equipment on hand. And then in DRAM taking that equipment and moving it from DDR4 to DDR5, which is now that the chips are available, the demand for that is greater. But the 2 people in Korea lost $15 billion making memory in 2023. It's going to take quite some time to get all their money back. So we've seen an improvement in pricing because of those actions. I'm just trying to understand why you think there would be a substantial increase in DRAM memory when by that timeframe they may not have recovered all of those lost profits, which it's very difficult to make future investments if you're not making a substantial amount of money. What am I missing?
Yes. Well, I think right now, as you said, they've been throttling capacity to improve pricing, which has improved. They've been converting their capacity to HBM, which is higher ASPs and higher margin for them, and also reduces the number of chips on a wafer due to the die size change. And they are converting to the next shrink, which gives them better performance and ultimately a lower cost point. So they're doing all the things that we normally see them do as they get to the bottom of the cycle and prepare for the next turnaround.The next turnaround has demand drivers in AI that are very DRAM intensive for higher ASP type of parts and that will drive consumer products and so forth that still need, I'll call it, the lower-cost, lower-performance items. So we see it as no different than any other cycle where they're investing in the next technology right now. And then they'll add capacity to meet the growing demand of those end markets, which will be AI, consumer markets, automotive, industrial, and then moving out to IoT and edge computing type of environment. So I don't think it's any different, Christian, than any other cycle that we've seen for memory.
The next question comes from Duksan Jang with Bank of America.
I have a 2025 question. So you reiterated the $1.3 billion in sales model target and that implies a 15% year-over-year growth for the system side. You mentioned advanced logic in Japan as some of the opportunities, but what other types of visibility do you have in your core power and general mature markets in order to drive that growth?
Well, I think we see it both in terms of market trends and then directly from customers as they discuss their plans with us. And then lastly, more on the area that you discussed, expansion of our footprint. And so we do see opportunity in advanced logic, and we see opportunity in the Japanese market as 2 areas that Axcelis has a lower penetration rate right now. So we do expect to grow those.The 2025 number is probably more driven, though, by the overall market recovery in memory and general mature. We expect, as we've said, that to really start off in the second half of this year and gain significant momentum as we go into 2025. There's a lot of really good long-term trends for this industry right now that make 2025 and into '26 look like they could be very, very good years. So we have very good confidence based on the market trends and based on our customers and what they're saying and based on Axcelis position and the Purion product family right now.
Onto those growth markets that you talked about, so advanced logic. Is there any way to quantify or estimate how much growth that would be? Because you're obviously not guiding, but it's been at a low-single-digit run rate for a couple of years, and you've talked about growth in this area for a while. So how do you assure us that you do have some wins coming and growth is expected there?
Okay, good question. So just a couple of things. So I was going to add to what Doug said about the $1.3 billion. There's not a lot of advanced logic or Japan baked into the $1.3 billion. I think those of you have known us long enough that unless we can see a clear path, we are not going to actually go out and state this model. And I think we've also said, as Doug indicated, there's multiple paths to get to this. Going after advanced logic, I think, when we work with customers in advanced logic, it has to go through R&D. That is a time-consuming process to get qualified as a Design Tool of Record to then get into the Higher-Volume Process Tool of Record. We have now managed to position our Dragon tool at 2 locations now: one's an advanced institute where we think we're going to get significant learnings; and the second one is actually an advanced logic customer where we've placed our Dragon in their R&D region. So we are making penetrations. And those penetrations are always going to be [ technically ] driven. It's not going to be a cost of ownership play. It has to be a differentiation play. So we're working with a lot of different partners to work out how we can differentiate our technology in their application. And obviously the goal is to solve really valuable customer problems.
Understood. And as a follow up onto OpEx. So in Q1, I think, the implied OpEx guide is roughly $60 million. And obviously it's a little bit of an increase sequentially despite sales coming down. So how should we think about the run rate from here for the calendar '24 and onto '25 as well, because '25 at $1.3 billion, 19% of sales, that's about $245-ish million. And I think you're already at that level.
Yes. So as we think about OpEx going forward, we've talked historically, and we mentioned it in the prepared remarks today, that we are going to continue to make investments in incremental R&D. And so you'll see that some of the increase period over period has to do with incremental investments in our research and development team here. In addition to that, we're going to continue to try to hold our SG&A expenses relatively flat. We believe we've built a base here that can support the type of growth that we see coming in the future. And so we're going to tightly manage expenses around our SG&A over the course of the year to maintain those at the, we'll call it the, exit rate of 2023, absent normal salary appreciation and other types of cost changes that would flow through our process.Largely speaking, you're going to see that number as a percentage of sales come down as the volumes increase over the back half of the year. And as we move to 2025, we'll still be very judicious in terms of making sure that we've got the type of efficiency that we want out of our SG&A organization while continuing to make investments in research and development.
The next question comes from Mark Miller with The Benchmark Company.
Just a housekeeping issue. What was capital spending?
For the quarter, it was $10 million in the quarter and approximately $20 million for the full year.
Okay. And your cash from operations was $65.6 million, is that correct?
Yes. In the quarter, yes.
The next question comes from David Duley with Steelhead Securities.
A question on gross margins. You've talked about revenue being flat for the year with the second half recovery. Where should we think about gross margins exiting 2024 or just a progression throughout the year? Just trying to quantify. When you talk about margins being up, how much?
Yes. Again, we're not going to provide the direct guidance on where the expectations are for margin over the course of the year. Right now, what we're forecasting, given the contribution of higher CS&I volumes over the period, as well as some system mixes, we do see it being up over 2023's gross margin overall.
Okay. And what was the reason behind the strong bookings in the quarter? I think there were $236 million. I think that was the number last quarter. It was like $198 million or something like that. What were the key end markets and applications that drove the increase in bookings?
It's silicon carbide in China. That's the big driver.
The next question comes from Craig Ellis with B. Riley Securities.
I wanted to follow up on a messaging change that seems a little different than what we heard through last year. And it's regarding CS&I this year and next year. It seems like there's a more optimistic view about what upgrades and some other offerings that the company has developed can do for annual revenues in that area. Is that correct? And can you provide any color on what specifically you're doing that is driving the growth that you'd expect in '24 and '25?
Craig, it's Russell. So, as you're aware, our installed base has grown really rapidly, particularly in terms of Purion. And that has a very strong platform. So as we look at our CS&I aftermarket business, we're looking to focus on contracts, and we're looking to focus on high-value upgrades. So we are actively developing upgrades that add significant value for our customers so we could sell those. And at this point of the cycle, often what you see is the utilization starts to go back up, then you see customers buying upgrades that can support increased capacity, and then they start buying machines. So this is a perfect time to be working with our customers, qualifying these upgrades, working with these upgrades, and building that part of our business out stronger.
This concludes today's question-and-answer session and presentation. Thank you for your participation in today's conference. You may now disconnect. Have a great day.