管理層發言
Good day, ladies and gentlemen, and welcome to the Axcelis Technologies call to discuss the company's results for the second quarter of 2026. My name is Grace, and I will be your coordinator for today. I would now like to turn the presentation over to your host for today's call, David Ryzhik, Senior Vice President and Interim Chief Financial Officer. Please proceed.
Thank you, operator. This is David Ryzhik, Senior Vice President and Interim Chief Financial Officer. And with me today is Russell Low, President and CEO. If you have not seen a copy of our press release issued earlier today, it is available on our website. In addition, we have prepared slides accompanying today's call, and you can find those on our website as well. 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's 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 annual report on Form 10-K 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. Given the pending merger with Veeco, we will not be addressing questions related to the transaction. During this call, we will be discussing various non-GAAP financial measures. Unless otherwise noted, all income statement-related financial measures will be non-GAAP other than revenue and other income. Please refer to our press release and accompanying materials for information regarding our non-GAAP financial results and a reconciliation to our GAAP measures. Now I'll turn the call over to President and CEO, Russell Low.
Thank you, David. Good morning, everyone, and thank you for joining us on our second quarter 2026 earnings call. In the second quarter, we delivered revenue of $215 million and earnings per diluted share of $1.06, both above our expectations. Our results reflect strong operational execution as we capitalize on favorable demand trends in several of our key markets. In the quarter, sequential growth in systems revenue was driven by improvement in our power and general mature markets, partially offset by the expected moderation in memory due to timing of available fab space. Importantly, customer investment plans in memory remain robust, and we continue to make progress executing our strategy to expand our position within this market. CS&I delivered a strong quarter and continues to be an important driver of our overall performance. Revenue growth in the business has been supported by a growing installed base, increased customer utilization and a continued expansion of our aftermarket products and service offerings. As we shared, CS&I has been a deliberate multiyear strategic focus for us to drive growth and stability through market cycles. We are pleased to see these efforts gain traction, and we look forward to continuing to build this momentum. Bookings in the quarter grew slightly, driven by general mature and power, and book-to-bill has neared 1 over the past three quarters, suggesting greater stability in the end markets we serve. Turning to Slide 5. Sales to mature node applications accounted for approximately 84% of system shipments with memory and advanced logic making up the balance of our sales. Now on Slide 6, let me review our trends by end market. Within our power market, revenue for silicon carbide applications declined sequentially, consistent with our expectations as quarterly revenue can fluctuate based on shipment timings. However, bookings improved on a sequential basis and through the first half of the year exceeded the average levels we experienced over the past two years. In addition, we continue to expand our customer base, securing orders for two new customers in China during the quarter. We are also seeing next-generation technology development create opportunities for our high-energy implant capabilities. During the quarter, we secured orders from multiple customers for high-energy channeling applications using advanced superjunction architectures, further validating the value of our differentiated implant technology. From an end market perspective, long-term demand fundamentals for silicon carbide remain highly attractive. We continue to expect increasing penetration in electric vehicles, broader adoption in AI data center power infrastructure and expanding use across a wide range of commercial industrial applications that require greater power efficiency, particularly at higher voltages. In our other power market, second quarter sales grew sequentially, and we continue to view silicon power as a foundational part of the broader power semiconductor market, serving applications across automotive, industrial, commercial and data center end markets. During the quarter, we completed a successful evaluation of our Purion XEmax at a leading foundry for use in power management IC production. The evaluation demonstrated the system's ability to address increasing customer requirements for high-energy implant applications, leveraging its dual LINAC architecture and patented Boost Technology to deliver implant energies up to 15 MeV with industry-leading beam purity. In general mature, sales improved sequentially during the quarter. While we have not yet seen a pickup in order rates, we are encouraged by improving end market trends. Following a period of capacity digestion that began in 2024, customers in China continue to add capacity, and we are beginning to see signs of improving activity outside of China as well, supporting higher tool utilization rates. Our customers are benefiting from demand for AI-related data center applications manufactured on 28 nanometers and above process technologies, including optical connectivity, microcontrollers and analog ICs. General mature remains an important market for Axcelis given its high implant intensity and our broad portfolio spanning high-energy, high-current and medium-current systems. We're also seeing growing customer interest in our recently introduced Purion H6 High Current platform across general mature applications. Turning to advanced logic on Slide 7. As we noted on our last call, we shipped a system early in the second quarter for materials-modification applications supporting 2-nanometer production, and we also shipped a follow-on system in the third quarter for this application. We continue to work closely with this customer in support of its next-generation technology road map. In memory, despite the anticipated sequential decline from a strong first quarter, customer engagement remains robust. As a reminder, memory sales can be lumpy from quarter-to-quarter depending on customer fab space availability. We continue to anticipate strong year-over-year growth in 2026 with momentum extending into 2027 as customers accelerate cleanroom investments to support growing demand for DRAM and high-bandwidth memory applications driven by AI. We're also seeing our memory customer portfolio expand, highlighted by a recent order for multiple high-current systems in the current quarter, reinforcing the strength and competitiveness of our offerings. Reflecting this momentum, we continue to make progress with a leading North American memory manufacturer we referenced last quarter. Following the successful completion of our system evaluation last quarter, we received additional orders during the period to support new fab investments. On Slide 8, let me wrap up my thoughts and provide our perspective on the second half of 2026. I am pleased with the momentum we are seeing so far in 2026. Our team has executed well, delivering solid results while capitalizing on the attractive secular growth opportunities across our end markets. In addition to a strong memory outlook for 2026, we are seeing improved demand in our power market. We're also encouraged by improving customer engagement and utilization trends in our general mature market, while our CS&I revenue continues to build a growing base of revenue and profitability. As a result, we now expect second half 2026 revenue to be stronger than our initial expectations and expect to deliver year-over-year revenue growth in 2026 compared to our prior outlook for revenue to be relatively flat with 2025. Looking ahead, we believe the favorable demand trends we are seeing today are likely to continue into 2027 with memory investments expected to remain strong as customers expand fab capacity, continued improvement in our silicon carbide market and encouraging underlying trends across general mature applications. We anticipate another year of revenue growth in 2027. Before I turn the call over to David, I'd like to provide a brief update on the pending merger with Veeco. We continue to make progress on the remaining requirements for the Veeco merger, including with the State Administration for Market Regulation in China. We continue to expect the transaction to close in the second half of 2026. We remain very excited about the pending combination and the opportunity to build on the momentum and create a stronger company with enhanced capabilities, broader growth opportunities and meaningful long-term value creation potential. I want to thank our customers, employees, partners and shareholders for their continued support in Axcelis. With that, let me turn the call over to David for a closer look at our results and outlook.
Thank you, Russell, and good morning, everyone. I'll first start with the financial details of the second quarter before turning to our outlook for the third quarter. Starting on Slide 9. Second quarter revenue was $215 million, consisting of system revenue of $132 million and CS&I revenue of $83 million, both exceeded our forecast. By geography, revenue in China increased sequentially to 46%, up from 40% in the prior quarter. Korea was our second largest revenue-generating region and 26% of our total revenues. In our other regions, Europe was 11%, the United States, 6%; Taiwan was 2% and Japan was 1%. The remaining 8% of revenue came from the rest of the world. Bookings were $131 million, slightly higher sequentially, continuing the trend of improving order activity with a book-to-bill ratio of approximately 1x. We exited the quarter with total backlog of $452 million. Turning to Slide 10. I'd like to share some additional detail on our results. Gross margin was 42.7%, slightly below our outlook of 43%, primarily due to mix within our CS&I business as well as higher-than-anticipated services costs, which can fluctuate from period to period. Second quarter operating expenses were $60 million, slightly above our outlook of $59 million, primarily due to higher variable compensation associated with stronger performance and, to a lesser extent, higher fringe costs. Tying it all together, our operating margin was 14.7%. Second quarter adjusted EBITDA was $36 million, and adjusted EBITDA margin was 16.7%. Other income was $5 million, higher on a sequential basis due primarily to foreign exchange gains. Our tax rate was 11%, below our forecast of 15% due to the windfall benefit associated with our equity compensation. And finally, second quarter earnings per diluted share was $1.06. Turning to Slide 11. Free cash flow for the second quarter was $15 million. This includes approximately $6 million of cash transaction expenses associated with the pending Veeco merger. We exited the second quarter with a strong balance sheet, consisting of $577 million of cash, cash equivalents and marketable securities on hand. This includes $175 million of long-term securities. With that, let me discuss our third quarter outlook on Slide 12. We expect revenue of approximately $230 million. Revenue is expected to benefit from a higher contribution from power and memory, partially offset by lower revenue from the general mature market. We expect gross margins of approximately 43%. We expect operating expenses of approximately $62 million. Adjusted EBITDA is expected to be approximately $41 million. We anticipate a tax rate of approximately 15%. And finally, we estimate net earnings per diluted share of approximately $1.11. Looking beyond the third quarter, we currently expect revenue to increase sequentially in the fourth quarter, supported by the business trends we are seeing across our markets that Russell touched on earlier. As a result, we now anticipate full year 2026 revenue growth of approximately mid-single digits year-over-year compared to our prior expectation of flat revenue. We also anticipate gross margin to improve slightly in the fourth quarter relative to third quarter levels. At the same time, we remain committed to investing in the business, particularly in technology innovation and other long-term growth initiatives. As a result, we expect fourth quarter operating expenses to be slightly higher than third quarter levels. In summary, we're executing on our strategy and remain focused on disciplined cost management while continuing to make targeted investments to capture attractive growth opportunities. We're encouraged by the trends we're seeing across the business and remain focused on delivering strong results and value creation for our shareholders. With that, operator, we're ready to take your questions.
分析師問答
Our first question comes from the line of Craig Ellis with B. Riley Securities.
Congratulations on the nice execution, team. Russell, I wanted to start off with a question on memory, understanding how the tenor of interaction with your more established customers and your newer customers has changed over the last three months. And as we look at near-term dynamics, which I think were indicated with memory up in the third quarter, are we at a point where we should expect memory system sales to grow sequentially? Or are we still in a period where there can be two steps forward and one step back?
Craig, thanks for the question. So I think we're going to see memory being slightly lumpy this year. Again, until the new cleanroom space comes online, I think the customers are mostly focused on solving bottleneck issues in their existing fabs. So it's a little bit lumpy. One thing I would say, though, is 2026 is a significant improvement over 2025; although it's off a low baseline, we are seeing significant DRAM memory revenue this year. And then once those cleanrooms start to come online, we expect to see the momentum continue into 2027.
Yes, Craig, just to add, if you think about the second half for memory, at this point it probably looks similar to the first half. And then on a full-year basis, we expect a strong growth rate into 2027, although at a lower percentage growth rate than in 2026 because we're coming off a very low base. But we definitely see that momentum in memory.
Yes. And next year, we get NAND capacity help, not just DRAM. And then the follow-up question is on the CS&I business. So congratulations on the real nice quarter there. My question is really what drove the magnitude of sequential strength? Is it really just in this environment, customers are looking at CS&I as one of the quickest paths to incremental capacity where they need it? Or is it really just the efficacy of better attach rates on Purion tools that are out there in the installed base? Just help us understand what drove the upside and what it might mean for the back half of the year.
So there's a couple of things, Craig. Clearly, we are seeing a pickup in utilization rates. We're seeing memory has really high utilization rates. We're seeing silicon carbide and silicon power picking up in utilization rates. And we're starting to see the embers of a recovery in general mature. All those things happen before you start getting orders; typically, you start to see spares and consumables going up. So you definitely get a component of utilization driving CS&I. The second thing is we continue to add installed base as well. But I think the really big part is that as companies are looking to ramp, they're looking to use what they have. We've developed a lot of valuable upgrades and a lot of those upgrades are now selling, and they're helping customers who, in some cases, are four-wall constrained or that give them a bit more capacity such that they can continue to ramp their business.
Our next question comes from the line of Charles Shi with Needham & Company.
Congrats on the pretty steady execution and exceeding the guidance and good guidance for the third quarter as well. So maybe the first question, the power pickup — the order rate you mentioned has been exceeding the average level you've seen over the last two years. Can you elaborate where the power strength really is from? Because historically, your power is closely tied to, let's say, the automotive industry, but is there something different this time? Do you feel like there is more of a data center component? I know that capacity is probably agnostic for many of your customers, but are you able to tell what is driving the power order rate pick-up?
Yes. When I think about where we are with power, there are two components to it: silicon carbide and silicon; both are doing better. Our power business was in digestion for a while, but now it's in recovery, and we expect the second half to be stronger than the first half for power. We're seeing that through orders placed. What's driving that? Electric vehicles remain the number one driver of silicon carbide. There's more electric vehicles and greater penetration of silicon carbide, particularly 800-volt systems. When you get to 800-volt systems, silicon carbide is essentially the choice. There are also onboard chargers and DC inverters and other components. Electric vehicles are still doing really well. There are a couple of other growth opportunities as well. Data centers moving to 800-volt architectures will also drive silicon carbide adoption as the material of choice, and that will sit between the grid at kilovolts down to the rack at around 800 volts. While it's still a small area, it is growing quickly. As the cost of silicon carbide devices continues to come down and device packaging improves, you see more value at the subsystem level, opening up more applications. We shouldn't forget industrial applications like transformers, motors and solar, which take up a lot of silicon carbide. Those are secular drivers. And it's more than just China as well; it's beyond China that people are adding capacity in power. In some regions, customers are still working on next-generation technology; they are ramping and learning. Once those device technologies reach volume and yield, you'll start to see customers ramp more, including transitions from planar to trench and even to superjunction. As these devices become more complex, the density of implant steps goes up, and it tilts toward high energy, which is another tailwind for us.
We understand the data center part is probably still very small but faster growing. Based on what you see today, are you able to tell how much of your installed base or the silicon carbide capacity in general is going into the data center application versus electric vehicles? We just want to get a sense on how to gauge the growth frontier, especially from the data center side.
We don't know necessarily what products our customers are creating with the tools — that is their business. But you can see from many of the customers' public statements: when a power company announces a portfolio that goes after data centers, that gives an indication of the device mix and how compelling the business is. So while we can't directly quantify installed-base allocations by end application, customer disclosures and product road maps provide useful signals of evolving markets and demand.
Our next question comes from the line of Jed Dorsheimer with William Blair.
Congrats on a solid quarter. Russell, just more of a technical question for you and kind of a market one as well. Indium phosphide looks like implants are being used for surface passivation, for etch repair as well as electrical isolation. I'm curious if you might talk about that — that's a pretty exciting area with significant undercapacity at the moment. What discussions you might be having in that area, and would that be for continuous wave lasers used for optical and electrical-to-optical applications in data center racks?
Thanks, Jed. Indium phosphide is an interesting laser material and data centers moving into optical communications is notable. For us, it's a relatively small amount today. We'll continue to monitor the market, but the implant opportunities for indium phosphide are relatively small compared with other technologies. There are much bigger opportunities in other technologies such as MOCVD.
This concludes our question-and-answer session. I would now like to turn the call over. Hold on just one minute. We had two questions come through. Our next question comes from the line of Duksan Jang of Bank of America Securities.
One question on memory. If I'm not mistaken, I think you said second half memory is going to be similar to first half. So I think that indicates about $120 million of total annual run rate. As we look into the industry, obviously a lot of people are expecting more cleanroom space to come through really starting next year and into 2028. I'm curious what kind of upside you expect in this market? I assume it's going to be a strong acceleration as well.
A little too early to size memory for next year. But our indications at this point are that next year would be another year of growth. Our assumption is that next year's growth rate percentage-wise would be lower than what we're seeing in 2026 because we're coming off a pretty low base in 2025. Engagement with customers is strong; they have ambitious capacity plans. I'd also note that we're making progress expanding beyond our traditional strong positions into other areas within the memory market and with other customers. Implant is a critical step there, and we're pretty excited about the opportunity.
Got it. And then are you seeing anything on the NAND side where I think industry trends have clearly improved and we might see some capacity additions? As memory grows as a bigger percentage of your sales, it's historically been a bit more dilutive to margins. Anything on the margin front would be helpful.
I'll take the NAND part. When Dave talked about strength into 2027 and cleanrooms coming on, he was specifically talking about DRAM. NAND has a little activity but I wouldn't call it a trend yet. Anybody with space is using it for higher-value DRAM. I do see a need for more NAND coming. When people vertically scale NAND, that doesn't necessarily increase implant density and therefore doesn't support us; but when people add wafer starts, that's when NAND benefits our business by increasing implant demand. Historically, NAND has followed DRAM, and we expect that to occur, but there's no clear trend yet.
And on gross margins, as you pointed out, our memory business typically comes at a lower-than-corporate average gross margin on the system side. But when we place those systems, that drives significant CS&I aftermarket, which is higher margin than the corporate average. So over the life of the tool, long term, it's still a pretty attractive business for us.
Our next question comes from the line of Craig Ellis with B. Riley Securities.
I wanted to start with a higher-level question, just understanding where the company's capacity positioning was. We're annualizing at about $920 million revenues as we head into the back half of the year. I would expect, given where we've been over the last five years, our capacity would be at least 40% higher than that. Can you help calibrate that? Secondly, related to memory, a number of years ago, the team did a great job qualifying the Purion Dragon at one memory customer. I would expect that would be part of the solution set that's going out the door. Can you talk about the demand that you're seeing across different products and where Purion Dragon fits into that mix?
On capacity, you're right. We do have capacity to meet higher revenue levels than we are today. As we absorb more and our markets recover, we would expect some benefit to our gross margin as a result of absorption. Clearly, mix will play an important role, but we do get some absorption benefit with higher volume. On the mix of systems, we're not going to get into specific systems that we're shipping to customers, but high-current is an important part of the equation for memory. We have a strong and competitive technology there. We're strong in high-energy, although high-energy content is lower in memory, and we also have competitive medium-current solutions. So we feel pretty well positioned for memory.
Did that answer your question, Craig?
Yes.
This concludes the question-and-answer session. I would now like to turn it back to David Ryzhik for closing remarks.
Thank you, operator. Thank you, everyone, for joining the call and your interest in Axcelis.
Thank you for your participation in today's conference. This does conclude the program. You may now disconnect.