ASMPT Limited (522) Earnings Call Transcript & Summary

Good morning, and good evening, ladies and gentlemen. This is Justin from ASMPT, and I am moderating today's call. Before we begin, let me pass the time to Robin. Robin?

Hello, everyone. I'd like to take this opportunity to acknowledge and honor our late Head of Investor Relations, Mr. Romil Singh, who passed away very suddenly about 2 weeks ago. Many of you have known Romil personally or been familiar with him hosting this Investor Relations call or during investor meetings and roadshows. The ASMPT team here and I are all shocked and saddened by his departure and miss him very much. I hope you'll join me in taking a moment to remember our colleague and friend. Thank you. Let me pass the time back to our moderator.

Thank you, Robin. On behalf of ASMPT Limited, welcome to our fourth quarter 2024 investor conference call. Thank you all for your interest and continued support. Please note that all participants will be in listen-only mode when the management is presenting. We will start the Q&A session after the presentation. During the Q&A session, priority will be given to covering analysts. Let me go through our disclaimer. Please do note that there will be -- there may be forward-looking statements about the company's performance, finances during this call. Such forward-looking statements could involve known and unknown uncertainties and risks that could cause actual results and performance events to differ materially from those expressed or implied during this conference call. For your reference, the Investor Relations presentation for our recent results is available on our website. On today's call, we have the Group Chief Executive Officer, Mr. Robin Ng; the Group Chief Financial Officer, Ms. Katie Xu. Robin will cover the group's highlights, outlook and next quarter's guidance, while Katie will provide details on the financial performance. With this, let me hand over the time to Robin. Robin?

Good morning, and good evening, everyone. It is a pleasure to have you all on our earnings conference call for the fourth quarter and the full year of 2024. I would also like to mention that ASMPT is celebrating its 50th anniversary this year. We have come a long way from our roots in 1975. Thank you all for your support as we continue enabling the digital world. Let me now proceed with the rest of our agenda this morning. Key highlights. In 2024, momentum for logic and memory packaging applications was driven by strong demand for generative AI and high-performance computing-related semiconductors. Our Advanced Packaging solution emerged as a key beneficiary of this accelerating AI adoption. AP solutions delivered a strong performance, increasing revenue by 23% year-on-year and demonstrating significant future potential by achieving key milestones. In particular, the group's Thermo-Compression Bonding, TCB solutions won further orders with multiple customers across logic and high-bandwidth memory, HBM, during the year and delivered its highest yearly revenue and bookings in 2024. The breakthrough in HBM led to a significant increase in overall bookings for TCB year-on-year. We remain optimistic that our business is well positioned to capitalize on the significant growth of the market, which we will talk about later on, along with our projections for the total addressable market size. The trend of generative AI will continue to bolster the demand for AP solutions and the group remains focused on growing its AP business as it enhances product offerings to support the technology road maps of major AI players. Lastly, we are fully prepared to seize opportunities in SEMI mainstream packaging and SMT when the market recovers. Advanced Packaging. With that overview, let me go into more detail about our Advanced Packaging solutions. As the most exciting part of our unique broad-based portfolio and the one with the greatest growth potential, our AP business increased its revenue contribution from 22% in 2023 to nearly 30% of the group's revenue in 2024. We are confident about the prospects of our AP business. We estimate the total addressable market for AP to grow at a compound annual growth rate of about 18%, reaching approximately USD 4 billion in 2029 from about USD 1.8 billion in 2024, driven by AI and HPC applications. Advanced packaging CoWoS opportunities. This is a new slide that sets up our portfolio of CoWoS packaging solutions. On the left is a graphic that sets up how memory and logic chips fit into the CoWoS structure. We are the clear leader for CoWoS packaging solutions, and I will spend some time on this slide to explain our solutions and unparalleled technology capabilities. Our portfolio includes a range of high-precision bonding solutions across TCB, hybrid bonding and mass reflow tools, all designed to meet the diverse interconnect requirements. Let's start with the top of the slide, our TCB solutions for HBM. The group secured substantial TCB orders from major HBM players, particularly in the second half of 2024. A key milestone of the bulk TCB order wins was achieved in Q4 2024 with a leading memory IDM to support its HBM3e 12 high HVM high-volume manufacturing demand ramp. Shipments under the bulk order commenced in Q4 2024 and are in line with the customers' ramp-up plan. I'm pleased to share that more recently, in this quarter, we secured an initial order of several tools from another global HBM player. We are also in advanced discussion for repeat orders with multiple customers, which strengthens the group's confidence to gain market share in this thriving HBM market. A key value proposition of the group's TCB tools is the ability to seamlessly upgrade to flux application for 125 and beyond, offering fungibility to handle different HBM packaging processes. These tools have best-in-class die placement accuracy and ultrafine pitch bonding capabilities and can also handle ultra-thin dies of below 30 micron at a cheap gap below 10 micron. For chip-to-substrate TCB towards the bottom of the slide, the group secured substantial TCB orders from its leading foundry customer and the customer's OSAT partner. serving as a sole supplier of TCB tools for chip-to-substrate application for these customers, the group delivered high-volume shipments of TCB tools in 2024 and expect a continuation of strong order momentum into 2025. At a couple of points in this slide, you will see a reference to TCB fluxless. This refers to our next-generation active oxide removal or AOR fluxless TCB. Our engagement with a leading foundry customer for ultra pitch chip-to-wafer logic applications is progressing well. The tool has demonstrated robust performance and is currently undergoing volume manufacturing qualification at the customer site. With the ability to achieve bond accuracy below 1 micron and ultrafine pitch bonding below 15 micron, the AOR fluxes process enhances package reliability. The process is completely residue-free, which is crucial since residue can degrade chip performance over time. In addition, the tools offer a total cost of ownership advantage by eliminating the need for downstream cleaning operations to remove residue, salts, forming acid and other corrosive elements. For mass reflow or MR, the group's high-precision flip chip bonding tools are utilized in the mass reflow process at a leading foundry and OSAT. The group is expecting further order flow for its tools in 2025 as MR is expected to remain the POR in the near term for these customers, while the AOR fluxless TCB tool awaits qualification. Featuring placement accuracy of 1.5 micron and bump pitch below 30 micron, these tools have been gaining traction in AI and HPC applications, which require varying degrees of accuracy across cloud and data centers as flip chip and TCB tools are expected to coexist together the group is positioned to capture opportunities in both technologies, engaging leading foundry, HBM and OSAT customers for both chip-to-wafer and chip-to-substrate applications. Finally, within hybrid bonding or HB, the group achieved a major milestone in Q3 2024 with the delivery of its first HB tool to a logic customer. During the year, the group also secured maiden orders for two next-generation HB tools for HBM applications. These tools are set for delivery by mid-2025 and have enhanced capabilities such as improved placement accuracy, reduced bonding pitch, significantly higher throughput and a more compact tool footprint. The group's HB order wins in 2024 demonstrate a strong recognition of its technology and competitiveness in this emerging solution. The group is confident in securing more orders in the coming quarters, positioning itself favorably to capitalize on the high-volume manufacturing demand ramp. So to summarize this slide, we have comprehensive solutions across MR, TCB and HB tools. We have deep involvement with multiple key customers across logic and HBM. Finally, and very importantly, we also have unparalleled technologies capability. Today, for the first time, we are presenting our estimated total addressable market of TCB, a market that has significant growth potential, estimated at USD 1 billion in 2027 and one that we continue to be excited about. This continued rapid expansion of the TCB market is supported by accelerated AI adoption. Our projection of USD 1 billion in 2027 at a CAGR of more than 45% since 2024 will be driven by both logic and memory applications. As we said earlier, we achieved record TCB revenue and bookings in 2024. That makes us the current leader in the overall TCB market. We are the market leader in logic applications, and we achieved a breakthrough into HBM market through a Q4 2024 order. The group is also currently in substantive engagement with multiple HBM players. Looking ahead, we are targeting market share in this expanded market of between 35% to 40%, leveraging our industry-leading technologies for AP interconnects, which set us apart from competitors and our significant entrenchment in the AI customer base. Let's move on to our Photonics and co-optic package or CPO solutions. As you can see on the left side of this slide, packaging is evolving rapidly to keep up with the ever-increasing bandwidth demands. Our solutions are key here, especially when it comes to accurate die placement and lens attachment. The rapid growth of AI is driving bandwidth needs in data center, which in turn is fueling demand for faster optical transceivers and CPOs. Our market-leading photonics solutions have the best-in-class placement accuracy below 3 micron and the highest throughput in the industry. We expect continued strong momentum because of our leadership position among all the major global transceivers market. Let's look at the CPO now. This is a game changer. Our silicon photonics solution have established a significant edge and CPO assembly through the leading high-precision bonding solutions. We have achieved an exceptional placement accuracy of 0.2 micron, which is essential for the precise integration of optical and electronic components. Plus, our system is highly capable of handling multiple bonding processes. While the CPO market is still in an early phase, the group's position as a photonics leader and its active collaboration with leading CPO players globally put us ahead of the competition. We are well positioned to deliver high-precision, reliable and efficient CPO assembly solutions, which sets us up perfectly to capture market share in the future. Not shown on the slide, but for completeness, we wanted to update you on our system in package or SiP business within AP. SMT won strong orders in the first half for its SiP solution, mainly from leading global high-end smartphone players for radio frequency modules and wearables. In addition, the group has been engaging multiple customers with its next-generation tools. These next-generation tools are gaining traction for AI and server-related applications with shipments to leading foundry and OSAT players. The group expects more orders in the coming quarters. Automotive. Moving to end markets. Our automotive end market application accounted for the largest proportion of the group's revenue in 2024 at about 20%. Even as the market softened, our comprehensive range of automotive solutions encompassing electrification sensor technology, displays and high-speed data transfer continue to contribute strongly with significant revenues coming from both SEMI and SMT. SEMI's contribution came from solutions serving various areas of the automotive supply chain, including technology packaging capabilities in power and electrification, specifically silicon carbide modules and high-end LED headlamps. SMT definitely navigated ongoing softness in the automotive market and maintained a strong position via continued engagement with a significant installed customer base. We have been actively engaging a growing number of leading automotive customers, including prominent EV players and leading subcontractors. These engagements across a significant proportion of the automotive ecosystem enable us to quickly scale when future demand ramps up, driven by the industry electrification trend. Looking ahead, we estimate the total addressable market for automotive end market applications to grow from approximately USD 1.3 billion in 2024 to USD 2.1 billion in 2029 at a CAGR of about 11%. With those highlights, let me now pass the time over to Katie, who will talk about our group and segment performance. Katie?

Thank you, Robin. Good morning, and good evening, everyone. This slide covers the group key financial metrics for the full year of 2024. In 2024, the group delivered revenue of USD 1.7 billion, a decline of 10% year-on-year, mainly impacted by SMT as its revenue declined 22.9% year-on-year, while SEMI registered 6.9% revenue growth year-on-year, contributing about 51% of group revenue. Group bookings were at USD 1.6 billion at the end of 2024, an increase of 4% year-on-year. SEMI registered 36.7% year-on-year growth in bookings, driven by AP, including strong contributions from TCB solutions, which offset a decline of 21.2% year-on-year in bookings from SMT. The group ended the year with a backlog of USD 779 million, a decline of 8.5% year-on-year. Group gross margin was up by 70 basis points to 40% year-on-year, mainly due to an increase of 418 basis points in SEMI's gross margin. This was partially offset by a decline of 346 basis points in SMT's gross margin. With lower revenue and a flat OpEx, the group's operating profit declined by 49.4% year-on-year to HKD 558.3 million. In line with reduced revenue and operating profit, the group adjusted net profit declined 42.8% year-on-year to HKD 426 million. Adjusted earnings per share was HKD 1.04, a decrease of 42.9% year-on-year. The group's total dividend payment for the year will be HKD 0.67 per share, which I will explain more later. We had a strong balance sheet at the end of 2024 with strong cash and bank deposit of HKD 5.1 billion compared to HKD 4.8 billion at the end of 2023. Our net cash was HKD 2.4 billion. Before we move on to Q4 financials, I would like to highlight that in 2024, we made incremental investment of HKD 180 million into infrastructure and AP solutions. As Robin has highlighted, we firmly believe that AP is a strategic growth area with significant upside potential, and we're prioritizing R&D resources and capacity investments to further strengthen our leading position. This year, in 2025, we plan to continue the investment to the tune of about HKD 350 million for strategic investments in AP R&D and infrastructure. With that, we expect OpEx to be marginally higher than previous years. In the fourth quarter of 2024, the group achieved a revenue of USD 437.6 million. The number grew 1.8% quarter-on-quarter and was flat year-on-year. Group bookings of USD 419.4 million were up 2.8% quarter-on-quarter and 19.2% year-on-year, mainly due to strong AP bookings for SEMI, but partially offset by the decline in SMT. As mentioned just now, the group ended the year with a backlog of USD 779 million, a decline of 3.3% quarter-on-quarter. Group gross margin of 37.2% was a decline of 379 basis points quarter-on-quarter and 508 basis points year-on-year. With reduced gross margin, the group's operating profit of HKD 5 million was down 97.1% quarter-on-quarter and 97.3% year-on-year. The group's adjusted net profit was HKD 82 million, an increase of 177.5% quarter-on-quarter and 7.2% year-on-year due to foreign exchange gain. Adjusted earnings per share was HKD 0.20, up by 150% quarter-on-quarter and 11.1% year-on-year. On this slide, let me give you more color on the key financials. In the fourth quarter of 2024, group revenue was above the midpoint of revenue guidance. As Robin has highlighted, the strong demand for AP solutions contributed to the improved SEMI performance with revenue growth of 24.1% year-on-year to USD 254.3 million, while the group's SMT business was impacted by continued softness in the overall market. For bookings, SEMI registered a 73.3% year-on-year growth, mainly driven by AP, whereas the SMT segment saw a decline of 25.9% year-on-year as automotive and industrial end markets remained weak. Group's gross margin year-on-year and quarter-on-quarter decline was due to both SEMI and SMT segments, which I will explain more in the next two slides. For the fourth quarter of 2024, SEMI recorded a 10.5% increase in revenue quarter-on-quarter, totaling USD 254.3 million, which represented a 24.1% increase year-on-year. The IC/Discrete business unit had a steady quarter-on-quarter growth with the highest revenue growth contribution from AP. The Optoelectronics business unit had a quarter-on-quarter decline, mainly due to weakness in advanced displays. The CIS business unit's revenue remained flat quarter-on-quarter at a low level with reduced revenue from high-end smartphone applications due to seasonality. SEMI bookings were up by 16.0% quarter-on-quarter to USD 277.1 million in Q4 2024, mainly driven by AP with bulk TCB orders from a major HBM maker. SEMI's book-to-bill ratio remained above 1 for full year 2024. Moreover, its quarterly bookings continue to show year-on-year improvement since Q4 2023, recording strong year-on-year growth of 73.3% for this quarter, driven mainly by AP. SEMI's gross margin of 42.6% for Q4 2024 was down 594 basis points quarter-on-quarter, mainly due to product mix, high base in Q3 and the sale of a first-of-a-kind deposition tool to break into the emerging glass substrate market with a major IDM customer. Gross margin was down by 115 basis points year-on-year. Lastly, SEMI's profit was HKD 74.7 million in the fourth quarter, a decline of 47.1% quarter-on-quarter. Moving on to the SMT business. SMT delivered revenue of USD 183.2 million in the fourth quarter of 2024, a decline of 8.4% quarter-on-quarter and 21.3% year-on-year, in line with ongoing softness in SMT's overall market. Segment bookings of USD 142.3 million followed a similar trend, down 15.8% quarter-on-quarter and 25.9% year-on-year. The automotive and industrial end markets continue to remain weak for SMT. SMT's gross margin of 29.7% for the fourth quarter was a decline of 260 basis points quarter-on-quarter and 1,130 basis points year-on-year. Margin was adversely impacted by lower sales volume and product mix. SMT segment profit was HKD 19.9 million in Q4 2024. This slide highlights the ASMPT's management's best estimates of revenue breakdown by end market applications for 2024 compared with 2023. These end markets highlight the extent of our broad-based portfolio and our exposure to diverse end market applications. Automotive continued to be the highest contributor to the group revenue at approximately 20% despite the overall softness in this market, particularly in the second half of the year. The group's comprehensive range of automotive solutions and a strong customer base contributed to this end market's performance. Consumer end market was the second highest contributing to group's revenue at about 16%, with year-on-year revenue growth coming mostly from semi midstream solutions, in line with higher revenue from China market. Communication end market contributed 15% to group revenue. Year-on-year revenue growth was boosted by demand in photonics and high-end smartphone-related applications. Computer end market maintained its year-on-year revenue contribution to the group's revenue at approximately 12%, with the highest contribution from TCB solutions. Industrial saw its revenue contribution decline to about 12%, in line with weak market conditions. Lastly, the other category includes revenue from spares and services and other applications accounting for 25%. This contribution has remained stable year-on-year. As you can see from this slide, we are a truly global business that partners with customers all over the world. China recorded year-on-year revenue growth and its share of group revenue increased from 31% to 38%. Revenues from both Europe and Americas declined year-on-year, mainly due to market softness in SMT. Europe's share of group revenue declined from 28% to 19% and Americas from 18% to 16%. Customer concentration risk continued to be low for the group as its top five customers accounted for approximately 14% of total revenue in 2024. We remain fully committed to enhancing shareholder value. We have an existing dividend policy of distributing about 50% of the annual profits as dividends. For 2024, the Board has recommended a final dividend of HKD 0.07 per share in line with this policy. In addition, the Board has also recommended a special dividend of HKD 0.25 per share to shareholders. Together with the interim dividend of HKD 0.35 per share paid in August 2024, the total dividend payment for the year 2024 will be HKD 0.67 per share. Let me now pass the time back to Robin for our outlook and the first quarter 2025 revenue guidance.

Thank you, Katie. As 2025 unfolds, strong momentum in the group's TCB solutions for AI and HPC applications is expected to continue to drive overall AP revenue growth. As such, we see our AP solutions constituting a greater proportion of our group revenue. The substantial progress the group's TCB solutions have made in logic and memory applications further cements our status as the TCB market leader. Looking at the near term, the group's AP revenue growth will be offset by ongoing weakness in mainstream market, particularly automotive and industrial. As such, we expect Q1 2025 revenue to be between USD 370 million to USD 430 million, flat year-on-year and down 9% Q-on-Q at midpoint. This concludes our fourth quarter and full year 2024 presentation. Thank you, and we are now ready for Q&A. Let me pass the time to Justin to facilitate it.

Thank you, Robin. Let us now proceed with the Q&A session. [Operator Instructions] I will start with Wern Juan.

So I'm extremely saddened by the passing of Romil. I just wanted to pay my respects once again briefly before I move on. Yes. So essentially, my question is regarding your TAM disclosure into 2026 and 2027. I noticed there's a big step-up into 2027 from 2026 for about USD 500 million to USD 1 billion. So my question is really where do you see that big step up? It seems to me that given the time lag, you're assuming TCB adoption in new end markets. Maybe it could be auto and logic or maybe it could be other edge AI devices. So my question to you is, is that so? And what has changed materially from the last quarter in terms of what you're seeing from a road map perspective in foundry?

Yes, this is Justin. I just want to -- your questions. You're asking about the TAM disclosure that we have put, and it is a big step up. I'll pass this question to Robin to address.

Maybe I'll just say a quick word on the way that we -- how we come up with the TAM estimate, and then Robin will put more color to it. So, Wern Juan, we basically undertook a systematic process, right, using the industry research and our own estimates to come up with the TAM. So the basis is really the CoWoS wafer number in the market, and that's market data. You guys probably have access to various versions of it. So that's the starting point. And then we use the certain market research and our knowledge to convert them into the number of interconnects as required with certain assumptions of [ SiPh ]. So with that, we also coupled that with the knowledge that we have with our customers or potential customers to estimate the TCB adoption rates across logic and HBM. Then after that, we factored the ASP evolution based on the progressive adoption of advanced solutions such as TCB AOR. So that's kind of the methodology that's behind those numbers. So as you were looking at the stepping up, it could be driven by either the number of wafers, the adoption rate, et cetera, right? So I'd now like to pass the time to Robin for more color.

Wern Juan, thanks for your question. In terms of application market or the end market, because our projection is up to 2027, the way we see is that the -- the TAM increase can be attributed twofold. One is the increasing adoption of TCB as the HBM memory market migrate from 12 high to 16 high and even eventually 20 high, which is that will be further down the road. And also, we see the increasing adoption of TCB for chip-to-wafer applications for logic. As we have been saying many times before, as the industry continue to pack more and more chiplets, more and more HBM onto the interposer, increasingly, TCB will be the tool of choice. And in particular, the fluxes operation will be the tool that will be needed to package such AI chip going forward. We hope we answered your question, Wern Juan.

Thanks, Robin. Thanks, Katie. Just a quick follow-up. Just so the additional USD 500 million, fair to say mainly driven by logic. I mean, obviously, there's some increase in high-bandwidth memory. Is that a fair statement?

I would say more coming from HBM than the logic side, yes.

Gokul, you may unmute yourself.

Just passing on our team's condolences as well on Romil's passing. Maybe first one on the TCB projection that you have provided, Robin and Katie, just one clarification. does it include the flip chip MR solutions that you're shipping for on-substrate or chip-on-substrate for the leading foundry? Is that part of the estimate? Or do you consider that more as a flip chip tool and not really a TCB tool?

Yes, you're right. That's not a TCB tool. That's more of a mass free flow. So that calculation is not included here, Gokul.

Okay. Understood. But that is still part of your AP revenue.

Yes, it's part of the AP revenue for sure. Yes.

Okay. So my question first to start with is on HBM. Could we talk a little bit about what is the progress with your first customer? I think you basically indicate that you are a POR or a process on record for this customer looks like in your slide deck. That means that our confidence on repeat order should be fairly high. How should we think about the repeat orders from this customer for HBM, given that you're probably going to be done shipping the first batch of tools by sometime in Q1 or Q2?

Yes. Yes, you're right, Gokul, we'll be done shipping by end of Q1. In fact, the progress is good. We have installed most of the tools by now at customer sites. Of course, customers are undergoing a qualification at this point in time. So yes, I think we are definitely pleased with this development. This entry into these important customers, I think we have established ourselves a good name in terms of HBM market. I think the industry is recognizing that if we are able to supply to this leading customer, we are good enough for most of the HBM players globally. I think in short, the development is good. The progress is good. Thank you.

And do you think -- what is your confidence on repeat order from this customer? Because there's a lot of noise in the market, including some of your competitors saying that this was a onetime order and you're not going to get a repeat order. I just wanted to address what is your confidence on the repeat order from this customer?

We are definitely confident. We are definitely confident of our own technology. As we said many times, we believe the reason we won this order, although we are not an incumbent, is customers like our technology. They know that we have a technology that is scalable. We have been saying also before that as the industry move from 12 to 16 and beyond, we strongly believe that the industry have to migrate to what we call a flux solution. And we strongly believe that in terms of flux solution, as we have also mentioned in our conference call, in our MD&A, we're in a good position to provide active oxide removal solution for -- as the industries continue to scale up in terms of accuracy, the demand for accuracy continue to scale up, the bonding pitch continue to tighten. We believe our AOR technology will be best suited for such development going forward.

Understood. Last question from me is on your logic fluxless or foundry fluxless PCB qualification. I think you classify that as undergoing qualification right now.

Yes.

There has been some kind of indications that one of your competitors has already got qualification for this chip-on-wafer portion. What is your expectation? Do you expect to ship tools in 2025? If the -- if your competitor is a POR for this tool, is it going to kind of preclude ASMPT from shipping into this customer? And also on COW versus COS, I think, Robin, you had said COW is probably a much bigger market than COS in terms of addressable market. Is that still your current evaluation?

Yes. Let me address your question by question first. In terms of our progress, we are progressing well. As far as we know, unless we don't know, but as far as we know, the customer has not made a decision which tool they will go for. So we are still in the running for sure. We are progressing well in terms of qualification at the customer side. Now yes, in terms of market size, down the road in the future, for sure, we see chip-on-wafer potentially could be a larger market size compared to chip to substrate. But in 2025, the way we see is that the volume, even if we get qualified, we've chosen as a POR, the volume for chip-to-wafer at the logic space in 2025 will not be significant. because the current POR process, which is MR is still the process of choice at this point in time. But increasingly, as I mentioned earlier, as the industry move towards more and more chiplets, more and more HBM stacks and also more and more IPD, I think it's inevitable at some point, maybe in 2026, we will see more adoption of chip-to-wafer fluxes tool at the logic side. So I think this is our assessment at this point in time. Do I answer all your question?

Yes, that is clear. Maybe one last thing is, so do you expect to ship to this foundry customer for fluxless for COW this year? Is that still the expectation?

Yes, yes, definitely, definitely. We are hopeful that definitely if we win the qualification for sure, the order should come in. But as I said, the order will not be substantial in 2025.

Donnie, you may unmute yourself.

And my first question is regarding to some housekeeping numbers. So I think could you -- or Robin, could you give us some color about the booking trend into the first quarter by different businesses? And also in terms of the gross margin, I think could you elaborate more on the gross margin in the fourth quarter is like because I think we have delivered some HBM TCB shipment. But inevitably, our semiconductor business gross margin still declined sequentially, while SMT looks like to be worse. So what could be the possible gross margin trend into the coming quarters as well? So this is the first one.

So Donnie, the Q1 booking highlight color will be addressed by Robin. The GM trend -- GM question will go to Katie. We'll start with Robin first.

Thank you, Donnie, for your question. Now bookings color, we think Q-on-Q for Q1 versus Q4 last year will be kind of flat sequentially. There are a couple of reasons here. You recall, we booked a significant bulk order of TCB tools in Q4 for HBM. Now I think you guys would appreciate that such orders cannot continue to repeat quarter after quarter. So for that reason, we see semi down, but slightly propped up by some improvement in SEMI mainstream bookings. I would say Q1 typically for mainstream SEMI tend to be higher than Q4. So it's some kind of seasonality. So in short, SEMI, we see kind of down because of AP, but propped up slightly by semi mainstream booking. However, for SMT, we do sort of guide in Q4 and Q3 earnings call that Q4 bookings for SMT will be at a low point. Indeed, that's true that we are confident that SMT bookings for Q1 will be up, led primarily by its AP bookings. We are still pretty strong in terms of SiP solutions, which are typically more front-end loaded. That's the kind of seasonality we expect.

Katie?

So, Donnie, before I start talking about the margin, I just want to make sure a note there, you mentioned about the bulk order for HBM player. In Q4, so as Robin mentioned earlier, in Q4 and Q1, we've been shipping the -- towards the customer. However, there's no revenue recognized for this bulk order in Q4, okay? I just want to put this clarification out there. Now coming to the gross margin in Q4 for SEMI, as I mentioned in the opening remarks, there are three drivers. First, if you recall, in Q3 last quarter, we mentioned that SEMI had a really strong margin in Q3 '24 due to the risk build for the same bulk order we just mentioned, manufacturing utilization was abnormally high. So that was one of the reasons that our Q3 margin was higher. So this created a high base for Q4 comparison. Secondly, product mix in Q4 was unfavorable due to consumer-related products. And then lastly, again, I mentioned in the opening remarks, we are developing a new deposition technology to break into the emerging glass substrate market with a major IDM customer. In Q4, we recognized revenue on the tool for this kind of -- we call it first-of-a-kind tool, but with no margin. So these are the three negative drivers that caused the Q4 semi margin to be on the lower side.

Understood. And just a quick follow-up on Katie, your explanation. So for the gross margin of like on-wafer TCB or HBM TCB, do you think that may carry higher than SEMI business average gross margin in the coming quarters? And my second question is regarding to your long-term TCB-10 estimate. So from 2025 to 2027, in terms of the dynamics, are you seeing stronger HBM TAM growth or logic market growth? I just wanted to have a sense about how the dynamics in between these 2 different markets from 2025 to 2027 and which one will be growing faster.

Katie will address your gross margin question. The subsequent question on TAM, Robin will address. Katie?

So on the question for gross margin for the HBM shipment, Donnie, as we communicated, and I think we've been consistent that for the advanced packaging TCB specific type of products, the margin is accretive to SEMI and group margin rate. And Robin, I'd like you take this.

Sure. So Donnie, on the TAM question as to what's the driver from the application. I actually answered earlier, but maybe it's worth repeating. The way we see in the near term from '25 to '27, the main driver in terms of the TAM growth for TCB, a few things. One is we see increasing adoption of TCB for HBM application. As the industry move, as I said, from 12 high to 16 high to 20 high, I think HDB solution -- sorry, TCB solution will be the tool of choice. So that's quite a clear trend. Now secondly, chip-to-wafer migration from current MR process to TCB, that is also a kind of trend that we see coming in the next couple of years. And of course, we continue also the continued adoption of TCB for chip-to-substrate application as AI acceleration continue. I hope that explains to you what -- how we see the TAM growth what are the drivers.

Donnie? Okay. Sunny, you up next. You may unmute and ask your question.

I'm very sorry about Romil's passing. So my question number one, I want to start from TCB as well. And so within the 30% sales for advanced packaging in 2024, could you let us know roughly how much is from TCB? And then you have this 35% to 40% market share target for TCB. I wonder what's your share assumption for HBM and logic, respectively? And how do you come up with that 35% to 40% market share?

Sunny, I just want to be clear with your first question on is it how much does TCB contribute to the 2024 AP.

You're talking about USD 4 billion, right, the 2029 number, is that what you...

2024.

Yes, for your own revenue, you quantified AP at 30% of total sales. So I want to know how much TCB contribute to the total sales?

Yes. So we cannot be too granular here, I hope, Sunny, because they'll be giving a lot of information away. But I can safely say that TCB is the main driver for the growth in terms of AP, yes.

Got it. And then the second part of the question is the market share target that you put at 35% to 40% for TCB. So I want to know what your logic behind?

Yes. I think there's certain logic, but as I said, we also cannot be too transparent here because of competition reasoning. But largely, our confidence is really grounded on the fact that we are the leading TCB for chip to substrate. Now that we have broken into HBM in a big way. And by the way, I don't know you guys have missed it, but we just -- we also disclosed that in the recent quarter in Q1, we have garnered an order of several tools from another global HBM player. So I think that speaks volume that the industry is really recognizing our TCB solution as probably the best-in-class for HBM. I think these are some of the assumptions that we have in mind when we came out with the number.

Got it. Thank you, Robin. So maybe switching gear to hybrid bonder. And so for this second-generation tool, I assume that should be able to achieve below 100-nanometer accuracy. And so are you still on track to ship? I think earlier, you said by mid of 2025 to HBM customers?

Let me confirm your question, right? You're trying to ascertain whether our hybrid bonder second-generation tool will achieve below 100 nano accuracy.

Yes, certainly. I think you all know we are not the first mover here. So I think there are -- we believe there are certain pain points that the industry are facing in terms of a new tool like hybrid bonding. So we are learning from it. And I think the good thing is that we are taking all these lessons on board. our R&D engineers are coming up with Gen 2, which we believe are highly competitive. Definitely, accuracy below 100-nanometer will be one of the criteria for the Gen 2 specs. So we are confident. We have won a couple of orders already for Gen 2. We are due to ship to our customers sometime in the middle of this year for Gen 2. And we are confident that we will continue to win more orders in the coming quarters.

Would you say for hybrid bonder, you may be making better progress for HBM in the near term versus in foundry?

Not really, Sunny, because HBM, the way we see HBM, TCB will still be the tool of choice. for some years to come. As long as you recall, as long as a couple of things as long as the high limit is not breached, I think TCB will be the choice to package HBM application because of a whole suite of reasoning. Cost is one reason and proven technology is another reason. So we believe HBM TCB will be a tool of choice for many years to come. For hybrid bonding, we believe it's more for logic, more of the logic space. Currently, it's being employed, we believe at the chiplet integration level, that's the top layer of the AI architecture. So I think that's where hybrid bonding will be most likely deployed in the years to come.

Got it. So my last question is on traditional packaging. And so recently, your key competitor seems to be suggesting there could be a possible recovery going to second half of 2025. Wonder if that also aligned with your expectation? Are you seeing any green shoots for recovery for this year?

Yes. The visibility is indeed very limited at this point in time. When we do channel checking with customers, I think their utilization rates are starting to creep up -- of course, we cannot generalize, but in general, it's starting to creep up. That's a good sign. So has it reached a level whereby they will start to order equipment, maybe not. That's why I say in the near term, in the near term we emphasize that the visibility for traditional mainstream packaging, including SMT are kind of limited at this point in time. However, having said that, we also have to take reference from industry experts, and they seem to say that second half recovery is on to cuts in 2025. So we certainly hope so. So as we continue to await the recovery, we get ourselves prepared. So when the recovery comes, we can capitalize on it.

Leping, you may ask your question.

The first question is about your SMT's margin. So do you think the Q4 last year was the bottom of the cycle? Or what should I -- what should we model the SMT margin looking forward? This is first question.

Yes. Look, so SMT gross margin is at a very low level, I can say. And like we mentioned in the opening remarks, volume is the main reason. And also the fact that Robin mentioned the automotive and industrial end market, right, where usually that's a relatively better margin could be gained is not in our favor, right? So with that kind of product mix, it's quite -- definitely is under a bit pressure. And also the other thing is that in this kind of slower market, unfortunately, our products are under pricing pressure as well, as you can imagine, right? All the competitors are trying to gain market share in this very tough market. So that's another headwind that we have.

So how we should model the SMT margin looking forward?

It really depends. I guess we need the best judgment for your model, but it really depends joking aside, right? It kind of see the -- if I kind of watch the mix, right, end market recovery, for example, right, that's really a key for SMT's margin rate.

Okay. The second question is about China. So where is the growth of China come from in 2024? Is it come from SMT or SEMI? And do you expect some HBM or CoWoS order from the Chinese customer in 2025 and beyond?

Yes. I think on the China side, probably, as I said, the utilization, just I mentioned about utilization, the utilization are creeping up. Sporadically, some Chinese customers have some subsidy program because they are into -- so they have subsidy program, they can -- they will start to invest, but these are sporadic investment. We see consumer-related applications getting a little bit more traction. So I think that's also a good sign. Now however, having said that, we keep coming back to our guidance is that having said all this, the near-term visibility for traditional mainstream for both SEMI and SMT are really limited at this point in time.

Robin, there's a second part of the question that is, that Leping is asking regarding CoWoS orders.

Yes, for sure, for sure. I think we serve a global market, right? So we don't limit ourselves to certain market segment. So definitely, especially for TCB, we will ship to customers where they are -- where they order from. So it's a global customer base for us for...

Simon, you may ask your question.

Yes. So yes, number one question is, yes, we do see the gross margin getting squeezed, but any chance to improve your net profit OP number, particularly OP margin these days only 1%, 2% range. So would you recap some -- your OpEx, R&D expenses after the gross margin, why the OP margin appears so low these days? That's the first question.

Yes. Simon, thank you for your question. Maybe let me address the question also from the OpEx because we talked about margin quite a bit gross margin already. So for OpEx, as you can see from our actual number, right, the last few years, OpEx has been pretty steady. A year ago this time, we actually announced that due to the potentiality of advanced packaging, we have decided or we have basically set aside investment for HKD 250 million last year in 2024. The actual came out to be HKD 180 million for 2024. So we largely have invested as we planned in R&D, especially for advanced packaging programs and some infrastructure as well. So that's the investment side. But with the investment, we managed to hold OpEx flat last year, fighting against certain merit increase, et cetera. So in Q4, we actually have done a restructuring program across the two businesses, given the slower recovery or the weak side of revenue, right? So that restructuring program actually impacted hundreds of people. And you can see probably in the financial reconciliation, right, there's restructuring cost about HKD 95 million and will bring savings more than that in an annualized way. So this is one of our efforts trying to be prudent in terms of cost measures. At the same time, we are committed to invest for future. So that's 2024. And if you think about going forward in 2025, as I mentioned in the opening remarks, we'll continue to do that, continue to invest in AP's R&D programs and also some infrastructure. And that's what we mentioned, right, will be at about HKD 350 million of investment. That strategy is not going to change. I hope that's enough.

Yes, very clear. But however, when we look at your income statement here, the R&D very consistent, about HKD 550 million a year ago, like Q4 2023 and then Q4 2024, also a similar amount, HKD 0.5 billion. I wonder why you have suddenly HKD 100 million of other gains versus other expenses is a little bit confused. But again, we don't see any meaningful change in your R&D expenses, Q4 2024 versus 2023 Q4, same amount. And also the SG&A, a little bit up, but pretty much close to HKD 300 million. I wonder why you are recognizing all your efforts for the AP extra expenses in non-R&D area.

So a couple of things. One is what you just talked about R&D being consistent is exactly the point I was trying to make that we're going to invest in selectively and strategically in the areas that we think that will bring future growth of the company. At the same time, we have implemented certain restructuring or cost measures to hold the R&D line relatively flat. You touched a little bit on SG&A. If you look at Q4 SG&A, it did go up a little bit. Think about all the opportunities that Robin just mentioned with multiple HBM players as an example, our people, not just R&D people, but the sales front, CRM people and all that, they've been really, really busy in Q4, actually chasing all these opportunities, traveling all over the place, working overtime, et cetera, right? So that did put some pressure on SG&A for Q4. I was not sure about you asking about some other, other profit, I'm not sure what exactly that was. But I hope -- what I just said is sufficient.

So computing is here, you are saying higher R&D expenses, some extra efforts for the Advanced Packaging related investment related. But according to your Q4 financial statement, I don't see any meaningful increase in your R&D Q4 2024 versus a year ago. That's the number one question. R&D expense is very flat year-on-year according to your financial statement. And then second question is, when you look at your consolidated statement here, doesn't we do see the other gains and loss and also some other expenses here, very unusually high number we do see. So I don't know how to understand this.

Let me chime in a little bit, Simon. Now the way -- I think the way we see it is that we will continue to optimize our cost structure regardless across functions, whether it's R&D, whether it's SG&A. However, at the same time, we are prioritizing certain investment into AP technology. So while we will be streamlining certain costs in R&D, there's not AP, but at the same time, we are piling in investment into AP. So we are really refocusing where we -- in terms of A, where we're going to focus on. Now coming back to your question on the overall OpEx, the way we see is that we will continue to optimize our cost structure where it makes sense, right? But at the same time, we need to have a certain size we need to keep a certain size and infrastructure to take advantage when the market, especially on the traditional and the mainstream market returns. So we cannot be too thin. But at the same time, we watch our cost while waiting for the market to recover so that we can capitalize when the volume returns. So I think this is our tactical plan going forward in the next couple of quarters.

Yes. One more question because everybody is now asking about TCB. Maybe one follow-up question is, sorry, my misunderstanding maybe you delivered TCB equipment for the customers, like memory makers, but the revenue recognition still in Q4. Any reason for that? And then the -- so far, up to 2024, only one memory maker ordered the equipment, but you are now seeing maybe multiple means, maybe three memory makers working with you to get the new TCB for the HBM?

Robin, I request that you answer that.

I'll answer the second part of the question, maybe Katie can chip in on the revenue recognition side now. Simon, no, we have more than 1 HBM players even in last year, right? So -- and continuing into 2025. We believe because of our statement wins in the large memory maker, I think the industry is now standing and say, okay, ASMPT is now is proven in terms of HBM solution. So we won other orders besides that global HBM player. And in addition, as I mentioned just now, we secured another order in Q1 from another global HBM player. So I think that is a strong testimony of our position, our strong position in terms of HBM. Now maybe back to Katie on the revenue recognition.

Simon, so for revenue recognition for those relatively complex tools, actually, the revenue recognition timing is different to the shipment timing. There is site installation and the customer certification involved. So overall, the revenue rec will be later than shipment timing. I hope that's clear.

So that means revenue recognition may happen in Q1 this quarter then or?

Yes. As customers' site is ready for us to install and as we go through certifications, we will be recognizing revenue in Q1 plus.

I see. So you are saying in 2024, particularly Q4, so two -- multiple customers already received some equipment. And then the new order took place this quarter. So maybe we can say two plus one customers means at least three customers working with you for the HBM memory PCB.

Yes. As Robin mentioned, we basically have multiple customers -- we have multiple customers, right? And in terms of revenue recognition, yes, starting from Q1, we start to recognize the revenue based on our shipment. I just want to make sure very, very clear, the new one that Robin mentioned with several tools, right, another global HBM player. Again, you need to keep that in mind that these tools have very long lead time. When we book the order from shipment to revenue rec, right, it's actually quite long cycle. I just want to make sure that you're very clear for Q1.

Yes. Yes. Okay. I'm not telling your presentation material, but today, conclusion is Advanced Packaging revenue portion much higher than historical average. But when we look at your revenue breakdown, the auto portion is #1, more than 20% of the total revenue, right, for 2024 but advanced packaging mostly for the computing or data center area. So I wonder where we can see your strongly growing TCB revenue contribution in your revenue breakdown by application because we don't see any meaningful revenue portion increase for some computing or some data center related to only the auto area. What do you think?

So largely, the TCB, Advanced Packaging solution, it goes into data center, it will be in computing. Some can go into communication, for example. Some can go into automotive. So we are not that granular, but I hope this is just a big picture where our solutions end up in terms of application.

Okay. [ Ethan Chan ], Morgan Stanley.

Sorry to hear about Romil passing. So just one quick question for me regarding recent OSAT restrictions on Chinese IC design firms. So as a result of these restrictions, some Chinese design firms might shift their orders from Chinese OSAT companies to non-Chinese OSAT companies. And my question is, do you see these restrictions impacting equipment sales to your Chinese OSAT customers?

Let me take that question. I think in general, Ethan, as an equipment player or maker, we are agnostic, we are agnostic who wins the order, who doesn't win as long as any customer in the world wants our equipment will ship -- so this may impact certainly some of our customers. We heard about it, okay? But it's too early to assess the implication of this impact at this point in time. Assuming that certain customers are affected, if the demand for such chips are still there, we are quite sure the system, the ecosystem will shift it out from one customer to the other customer. And since we serve a global customer base, I think the impact to us is really very limited, yes, at this point in time.

Okay. Thank you, Robin. And thank you to the team for all your questions. That concludes the Q&A question. But let me request Robin to say a few concluding words. Robin?

So thank you, guys, for all the questions, and this really concludes our Q&A. Now as we have said in the near term, we expect ongoing weakness in the mainstream market, but we are ready to capitalize on opportunities when this market eventually returns. Now looking at our AP and our flagship TCB solution, we have, for the first time, taken you through this TAM. This is an area where we see growth potential across both logic and memory applications. And we hope that so far, we have what we have shared, you share our segment as well. As the current TCB market leader, we are targeting market share in the expanded market between 35% to 40%. Now for the wider AP business, we are certainly confident and remain focused on growing our AP market share as we enhance our product offerings and to support the technology road maps of our major customer base. On this note, this concludes our call. I will see you in the next quarter. Thank you very much.