Asmpt Ltd

Good morning and good evening, ladies and gentlemen. This is Romil here from the Investor Relations team, and I will be the moderator for today's call. On behalf of ASNBT Limited, let me welcome all of you to the group's investor conference call for the first quarter of 2024. And we would like to thank you for your interest and your continued support in the company. Please note that all participants will be will be on listen-only mode when the management is presenting. We will start the Q&A only after the management has gone through the entire presentation. During the Q&A session, priority will be given to the covering analysts. As part of our standard disclaimer, please do note that during this conference call, there may be forward-looking statements with respect to the company's business and financial conditions. Such forward-looking statements could involve known and unknown uncertainties and risks that could cause actual results, performance, and 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, Robin, and the Group Chief Financial Officer, Katie. Robin will cover the group's key highlights, outlook, and the guidance for second quarter, while Katie will provide details on the financial performance. And then we will open the floor for Q&A. So with that, let me hand the time over to Robin now.

Thank you, Rom. Good morning and good evening to everyone today. It is a pleasure to have you all on our earnings conference call for the first quarter of 2024. Before we begin, let me take this opportunity to share some thoughts on the overall macro environment and some key highlights of our business. The macroeconomic environment continues to present challenges. While the slow recovery of the Chinese economy continues, there have been dynamic and often rapidly evolving geopolitical conflicts that have roared markets. Coupled with stubborn inflationary pressure and the more recent fear of interest rate hikes, overall consumer sentiment remained weak, trickling down and causing tepid electronics demand. Against this backdrop, our unique and broad-based portfolio shielded us to a certain extent as our two segments followed different cycles. For the first quarter, ASMPT delivered revenue at a midpoint of guidance with a higher revenue proportion from our SMT business. Our block-based portfolio serves diverse end markets, and this continued to be an advantage for us because weakness in some end markets can at times be compensated by strength in others. For this quarter, the group's automotive applications continued to form the highest proportion of our overall group revenue. Revenue from SMT's automotive application also grew compared with the previous order, despite softness in the overall automotive market. As for SEMI, its automotive solution benefited owing to exposure to certain specialized technology areas of the supply chain. such as power and silicon carbide modules and smart headlamps of high-end vehicles. An important development for the group in the first quarter was the book-to-bill ratio that moved above one after seven quarters. Bookings for the group grew in the first quarter as both business segments had higher bookings when compared to the previous quarter. Semi's booking growth was also due to a low base effect. SMT bookings grew quarter on quarter and began stabilizing after softening in the second half of 2023, mainly due to automotive and industrial air markets. The group bookings growth in the first quarter was mostly propelled by strong demand for its advanced packaging or AP solutions, with both SEMI and SMT contributing strongly to group AP's booking. The group's interconnect solutions, including thermal compression bonding, hybrid bonding, and flip chip high-precision dye bonding, were mainly driven by AI. Together with SMT's system-in-package, or SIP, tools, these contributed strongly to group's AP bookings in the first quarter. Let me share more about the progress of our AP solutions, which have the highest growth potential under our portfolio. Undoubtedly, advanced packaging continues to be a bright spot for the group. We strongly believe that the group has the industry's most comprehensive suite of AP solutions that serve a diverse range of applications. Moreover, we are deeply embedded in the supply chain of major generative AI and high-performance computing customers, and hence are in a commanding position to capitalize on the growing demand from an increasing range of such applications. Let me shift focus to our most popular AP solution, thermal compression bonding, or TCP. For logic applications, we continue to win TCP orders in the first quarter from IDM and OSAP customers. For a leading Foundry customer, we have started to deliver TCP tools in the first quarter for chip-to-substrate application as part of the meaningful orders that we won from this customer in the second half of last year. To support this customer's growing AP demand propelled by generative AI, We expect more TCB orders for chip to substrate application from this leading foundry and the supply chain partners in the second quarter and beyond. In addition, we also recently delivered our next generation ultra-fine pitch TCB2 for chip to wafer application to this foundry customer for joint development. And we are confident of winning cheap to wafer TCB orders in the coming quarters. Next, let me talk about high bandwidth memory or HBM potential. The group's TCB tools are already in production at a leading HBM player for 12 high stacking. We have shipped out a demo tool to another HBM customer and have more tools in the pipeline. I'm sure most of you are aware of the news circulating recently since last month on the relaxation of the package thickness on the next generation of HPM. We have experienced heightened engagement with multiple HPM players as TCB is emerging as the preferred solution for stacking requirements of 12 high, 16 high and above. This is due to TCB's continued advantage in terms of cost over performance or total cost of ownership criteria. We remain confident of the strength of our TCP solution in terms of accuracy, thin or large time handling capabilities, and our deep process knowledge gathered over more than a decade. With a competitive advantage and based on the above-mentioned customer engagement, ASMPT is in a strong position to benefit as TCB adoption accelerates for both logic and HBM application. Let me talk a little bit about hybrid bonding. Last quarter, we indicated that we expected more orders. I'm pleased to highlight that in the first quarter, we want orders for two more hybrid bonding tools for logic applications. and we remain confident of winning more hybrid bonding orders in the coming quarters. In addition to our semi-AP tools, booking for SMT AP tools also grew quarter and quarter, mainly due to SIP tools demand mostly coming from RF modules for high-end smartphones and wearables, from leading global players, and also from PC and server-related applications. With those highlights, let me now pass the time over to Katie, who will talk about group and settlement performance.

Thank you, Robin. Good morning and good evening, everyone. Let me take you through the group financials. This slide covers the group's key financial metrics for the first quarter of 2024. The group delivered revenue at midpoint of guidance. Revenue for Q1 was down quarter on quarter, mainly due to the prolonged semiconductor down cycle. that impacted our semi-business. Though SMT revenue also declined, it was a much smaller rate and SMT contributed a high proportion of group revenue. This highlights the advantage of our broad-based portfolio as semi and SMT segments follow different business cycles and provide some stability at the group level. Group bookings had growth quarter on quarter in Q1 as both segments grew. and the growth was mostly powered by advanced packaging solutions. Our backlog remained stable at about 849 million US dollars at the end of Q1. Group gross margin remained at a relatively high level in the first quarter, and it was down slightly quarter on quarter. For Q1, Group's adjusted net profit was 177.5 million Hong Kong dollars, an increase of 132.1% quarter on quarter. and adjusted earnings per share was 0.43 Hong Kong dollar, an increase of 138.9% quarter on quarter. The group continued to have a robust balance sheet with cash in the bank deposits at 5.25 billion Hong Kong dollars, and net cash at 2.75 billion Hong Kong dollars at the end of Q1. In the first quarter, group revenue was 401.4 million US dollars, Revenue was down 7.8% quarter-on-quarter due to both Semi and SMT, but Semi's revenue decline was steeper due to the prolonged semiconductor down cycle. Group bookings in the first quarter grew 17% quarter-on-quarter to $409.3 million, as both segments registered bookings growth. As Robin highlighted earlier, this booking growth was supported by robust demand for advanced packaging solutions on both SEMI and SMT. Group gross margin declined slightly by 40 basis points quarter-on-quarter to 41.9%, this due at a higher level compared with previous quarters. The slight decline was mainly due to SMT's lower gross margin, but partially offset by SEMI. Group operating margin improved quarter-on-quarter by 218 basis points to 7.6% in the first quarter. This was mainly due to lower operating expenses from ongoing cost measures and seasonality effects. Last quarter, we announced that we'll invest in R&D and infrastructure with an additional HK$250 million for 2024. The projects are on track and will intensify in the remainder of the year. CEMI delivered a revenue of 175.8 million US dollars in the first quarter, a decline of 13.7% quarter on quarter. The IC discrete business unit had quarter on quarter declining revenue, mainly due to industry weakness for the group's deposition tools. However, there was some sporadic demand for consumer related mainstream tools in the quarter. Optoelectronics business unit's revenue declined quarter on quarter. the business unit's revenue was mainly driven by high-end automotive headlamps and photonics-related applications. Revenue for CIS business unit grew quarter-on-quarter, a low base. Growth was mainly due to a high-end smartphone market. Semi-spookings grew 25.1% quarter-on-quarter from a low base to 199 million US dollars, mainly due to consumer and computers and market applications. CEMI's advanced packaging solutions also registered quarter-on-quarter growth. It is worth noting that CEMI's booking turned to growth on a year-on-year basis in Q4 last year, and it continued this trend in Q1 2024. Despite the lower volume, segment gross margin improved by 86 basis points quarter-on-quarter to 44.6%, mainly due to a one-off benefit from the sale of previously provisioned inventory. For Q1 2024, our SMT segment continued to deliver higher revenue than SEMI for a seventh consecutive quarter, and it contributed about 56% of group revenue. SMT registered revenue of $225.5 million, a marginal decline of 2.6% quarter-on-quarter. SMT revenue performance continued to be dominant by automotive and industrial and market applications, and mostly in Europe. Segment bookings grew 10.1% quarter on quarter to 210.3 million US dollars in the first quarter. Mainly driven by growth from advanced packaging and automotive applications. SMT gross margin was at a healthy level of 39.7% in the first quarter. There was a decline of 123 basis points quarter on quarter due to product mix. Let me now pass the time back to Robin for next quarter's revenue guidance.

Thank you, Kelly. The group expects second quarter revenue to be between $380 million to $440 million. At midpoint of $410 million, this represents a decline of 17.6% year-on-year and an increase of 2.2% quarter-on-quarter. The slight quarter-on-quarter increase is due to higher revenue from semi- partially offset by lower revenue on SMT. Based on a unique and broad-based portfolio, we remain optimistic about the group's prospects and the potential for growth over the long term. This confidence is further supported by long-term structural trends of automotive electrification, smart factories, green infrastructure, 5G, IoT, and AI growth across cloud, data center, and AI-ish devices. On a broader level, these structural trends are also moving in tandem with increased capacity spent from nations securing their supply chains via more onshoring and organizations preparing themselves to deal with more dynamic global supply chains. This concludes our presentation for the first quarter of 2024. Thank you. And we are now ready for Q&A.

Let me pass the time to Romil to facilitate it. Thank you, Robin. For asking questions, please either use the raise hand function or type your questions in the chat to ASMPT Q&A. Please ask your questions one by one and limit them to two questions at each turn. With that, can I first request Gokul to unmute yourself and ask your questions?

Yeah, hi. Good morning, and thanks for taking my questions. First question is on TCB. Looks like the progress is quite good on your Hungary and OSAT customers. Robin, does it now feel that you're – kind of PCB order momentum and shipment momentum is materially better than what you had indicated three to four months back where you talked about equaling the shipments that you did in the past 10 years until 2021? Or is it still pretty much in that same range? Secondly, on HPM, you were a little bit more cautious on the HPM adoption of PCB previously. Are you starting to see meaningful progress on 12 high and 16 high with many customers? I'm asking this because one of your competitors still insists that most of HBM will be sticking with mass reflow. Are your PCB tools ready to be used on mass reflow or you basically require the customer to move to local reflow to start using your high precision PCB tools in HPM? That's my first question.

Thank you. Okay. Maybe I'll break it down into two segments and Robin will answer these questions for you. First one is on the progress that's happening on the logic side, especially with the Foundry and OSAT customers. So I will request Robin to comment whether the order momentum for the order flow and traction with this customer better than it was three to four months back. Thank you, Ram.

Now, for your first question, I think you were trying to gauge a sense of you know, whether there's any difference, you know, three months ago, three, four months ago compared to now? I think yes, in a way, whatever that we, you know, were telling the street came to realization, right? So we want the We want the orders for the chip substrate application in this quarter to be shipped to the leading supply foundry. And we have already started to ship a couple of tools already. In addition, as you can see in our MD&A, in our announcement, we have also shipped a next generation of ultra-fine pitch tool. also to this leading foundry for joint development. And this is for the chip to wafer applications. So you recall, TCP has been started to use for chip to substrate at the logic side. And now, as we have alluded before, in time to come, TCP will also be used for chip to wafer applications. So I think that has come to, in a way, what we expected. Now, we are definitely on track. We said that TCP in terms of dollar revenue for 2012 to 2021, compared with 2022 to 2024, the amount will be the same. So in other words, in short, there is an accelerated adoption of GCP in the last couple of years going into 2024.

So for the second part of the question, it's more on the HBM potential for TCB. Gokul noted that we were a bit more cautious previously. And now whether the progress is more meaningful, especially in relation to 12 high, 16 height kind of HBM stacking. And how does the TCB compare in terms of tool-wise versus MR-wise? So maybe Robin can give an overview on all these engagements and his views on the HBM potential for TCD. Sure.

Now, Kuku, there's a distinction. You know, when you talk about MR, it's mass reflow tool. So these are basically for what we call a flip chip kind of mass reflow. So it's different from TCB. TCB is a local reflow process. So don't mix up MR with local reflow process like the TCB tool. So there are different set of tools altogether. Now, yes, we started more optimistic for HBM, because as we have alluded before to the street, that we see that increasingly for HBM going from 12 high to 16 high and beyond, This is our view that the MR has certain limitations in terms of accuracy and in terms of handling increasingly higher density memory die. And also as you continue, we believe that as the industry continues to step higher and higher, the memory die has to be thinner and thinner. So in terms of accuracy, in terms of packaging more IO in the memory die, and also in terms of thin die, it's opinion that TCP has better handling capability in all these metrics, comparing to a mesh reflow tool. So that's why we have been saying that, and also let me also highlight that in the recent development, there was some discussion that you know, the HPM high, you know, can be relaxed from now 720 micron to 775 micron. And this bodes well for a TCB too, right? Because previously at some point when there's a high restriction, there is some concern that TCB may not be able to, you know, use for a multiple high HPM because it will exceed the limit. But with this relaxation, there is certainly more room and more potential for TCP application for 12 high HBN, 16 high and beyond. Next question, please.

Okay, that's clear. Thanks. Robin. My second question, just want to follow up a bit more here. So if you look at these three distinct buckets of TCB demand, Logic, Foundry, OSAT, and Memory or HBM, let's say in three to four years, Robin, which is, can you rank these in terms of the size of the addressable market you'll see for TCB? Because today I think Logic is still the biggest and then foundry outside is catching up, HBM is kind of a long-term potential, but let's say three years, four years later, do you have any view on what is the kind of rank order of these three distinct markets for TCB? And lastly, I think if you have, could you share some rough thoughts on what percentage of your semi-solution business is going to support AI applications right now? Is it like mid to high single digit already, or is it lower than that? Just wanted to, given that a lot of other companies, your peers are also talking about their AI exposure in terms of total revenues. Thank you.

Okay, so first part of the question is basically on the demand. Can Robin break down the demand in terms of, say, logic IDMs, logic foundry and OSATs and HBM going forward?

Now, our internal view about these devices, HBM versus Logic versus the HPC, I think HBM in our view is that because as HBM continues to increase in hype going forward, From currently 8 high to 12 high to 16 high and beyond, I think it's intuitive that more volume will be skewed towards HBM in time to come compared to the logic side. Now, it's difficult to clearly, on the other hand, it's difficult to clearly distinguish or try to differentiate between the logic and the HPC side. To a certain extent, they are kind of interrelated, so it's difficult to really draw a clear line. So I think it's easier to try to distinguish between HPC project on one cam and HPM. So for that, We believe HBM potential for TCB is larger in time to come, because 12 high at this point in time is still at low volume. HBM players are now mostly coming out with 8 high, but we believe 12 high will take on the more higher proportion in time to come, maybe starting from 2025 onwards. But this is our view so far about this kind of applications. Thanks, Robin.

Gokul, I want to clarify your next part of the question. So you want to know on AP as percentage of the semi-business, do you want to know more from a revenue perspective or do you want to know more from investment perspective? And AP, are you specifically targeting on AI?

Yeah, I'm just asking of semi-solutions, what percentage of your revenue is coming from AI applications? You probably have a rough idea where these tools are eventually going.

Okay. So we can't Google, how are we understanding? We can't be too definitive over here because of competitive reason. But I can safely say that... you know, TCB application at this point in time, uh, of course split between HPC and, uh, uh, uh, also the AI, but it's sometimes difficult to clear, as I said earlier, it's difficult to clearly mark, you know, what is HPC, what is logic because they, they can be kind of overlapping. So, so from that perspective, uh, uh, we, you know, HBM, you know, is, um, Yeah, from that perspective, I think TCB is looking at a volume in terms of suite of AP solution is the highest contributor, right? So, and also, so that's mostly at this point in time for AI, I would say. And besides TCB, there are also other tools, you know, like our photonics tools, They are also AI-related. In fact, also our SMT tools as well, because SMT, we believe our customers are using our SMT tools to place what you call passive devices on 2.5D packaging as well. So we are also facilitating AI application also from our SMT side. And also not forgetting our next deposition tools. They are laying, you know, interconnect layers on organic substrates. So from that perspective, we are quite well represented in the generative AI space.

Okay, thank you.

Thanks, Google. Next, can I request Donny to unmute yourself and ask your questions?

Yes, thank you, Robin and management team for taking my question. My first question is housekeeping questions. So could you kindly give us some colors on the booking training to the second quarter this year and when exactly you expect the conventional packaging order to start to recover more meaningfully in the rest of the year? And second question is regarding to the chip to wafer TCP progress. So you mentioned about you have jointly develop the next-gen chip-to-wafer TCBs with leading foundries. So does that mean we are providing new demo tools to the customer? And where exactly we can see repeating orders or, you know, the orders for mass production in the next couple of quarters? Thank you.

Thanks. So, Donny, first part of your question is more colour for bookings for second quarter and specifically demand for the mainstream or conventional tools. So, we'll request Robin to answer that first.

Thank you, Donny. So, as you know, we don't give concrete guidance for a booking. We only give for revenue. But certainly, as usual, we can give you some colour how we see due to booking for this year trending. We expect bookings for Q2 to be kind of flattish on a Q2 basis for both segments, SEMI as well as SMT. On a year-on-year basis, bookings will be higher, and that's a good sign on a year-on-year basis because on a year-on-year basis, it really cuts out the seasonality factor. So we expect the increase to come mainly from the semi-segment, whereas for the SMT segment, probably a little bit down or kind of flattish compared to year-on-year. The reason is very simple because SMT last year, first half, they were still on a high in terms of booking, whereas Assembly has already in a way corrected for some time already since the beginning of last year. Now, we expect this year-on-year increase in booking to actually continue into the remaining part of 2024. I think that is probably a a positive sign that we are hoping to realize in time to come, right? At this point, this is our expectation, and you know, your PCT should increase going forward. I can give you a little bit more color as you asked about, you know, Main Street, but we look at bookings for Q2 in three areas, right? One area is, let me start off with advanced packaging booking first. We expect advanced packaging bookings to remain robust. In other words, the momentum for advanced packaging, in particular for TCP application, will continue to be robust. And AP bookings, in our opinion, at this point will continue to increase Q1Q, that means Q2 versus Q1, and also year-on-year Q2, this year versus last year Q2. The increase, I think we expect the increase mainly will come from the semi side rather than the S&T side. Now for SMT, the booking seems to be stabilizing and for number of quarters already. And we believe the bookings for SMT, the downside is limited. So it's kind of bottoming out also for SMT already. Now for semi-mainstream, we see bookings in Q2 continue to remain at a low level, mainstream, and mainly because at this point in time, we also do not have a lot of visibility in the near term. As you recall, for semi-mainstream tools, typically our lead times is only about three months, right? So in this current environment, our customers on the mainstream side tend to be a little bit more cautious. So they only place order when they are sure of the timing of recovery. So because of that, we lack also the visibility in the near term for semi-mainstream science. I hope I answered the first part of your question.

Yes. So the second question is more on the progress of chip-to-wafer application for our TCB. And we have demo tool for joint development already at the leading foundry. So Donnie wants to know more on the potential of this and when will this translate into water flow? Certainly.

I think one of the reasons we have been saying why the logic side for chip-to-wafer will have to move to TCP eventually because there is this increasing need for customers to pack more and more chips together at the chip-to-wafer level. And also, because of that, the pitch requirement and the accuracy requirement will continue to increase. And from that perspective, the TCP, in our opinion, is superior to the mass reflow tool. Now, so as these trends continue, you know, we are optimistic that, you know, our TCB2, you know, will gain more traction in terms of G2 Wafer application in time to come. In terms of timing, we, you know, we just shipped that ultra-fine pitch tool for G2 Wafer to the Li Ming Foundry, so it takes some time for them to test it out. So we are hopeful that maybe in the coming couple of quarters, we could see some concrete orders which would be delivered, you know, either, you know, the later part of this year or maybe the beginning of 2025. This is how we see the Chitra Wafer application going forward for T3.

Next question, TC. Next, can I request, Randy, can you unmute yourself and ask your questions?

Yes, hi, thank you. Actually, I want to ask maybe two follow-ups on those comments on TCV, just given the interest. First, for high bandwidth memory, where we're starting to get the 12-py stacks coming into production later this year, are you seeing potential for a steeper ramp-up? You mentioned HVM. could be the larger opportunity over time. So do you see the potential 25 could be the end of collection for your business? Or more reasonably, is it more tied to the ramp up of HBM4? So it could take a little bit longer to get qualified in, or at least for your tools advantage to really be recognized. I'm just curious when that cuts in for the volume ramp. And then I had a follow-up on the chip-to-wafer versus chip-to-substrate question. Do you view that opportunity larger as you come into it relative to chip to substrate? Could that ultimately be more meaningful? Is it a higher value tool that you'd get an ASP lift as well?

All right. Thanks, Randy. Okay. So for your first question, more towards TCB for HBM, acknowledging that some tools are already into 12 high production. So maybe I'll request Robin to comment on the demand, especially how it's going to take off, whether the demand and ramp-up is going to be steeper soon in 2025, and 2025 might be an inflection point, or... Will it be timing the HBM4 potential whenever that happens?

Thanks, Randy, for the question. Now, a little bit of a background first. I want to emphasize that for 12 high HBM, we already have a few tools already in production. So we believe for 12 height TCP, we are probably the first mover in terms of such applications. So we have, together with that leading customer, we have already generated a lot of data, you know, a lot of process knowledge in that area. So if 12 height takes off in a meaningful way, we believe we are in a good position, you know, to capitalize on the potential for TCB application for HBM 12 height and beyond. Now, yes, I think in terms of 12 height, as I said, it's still high volume, but in small volume compared to, say, AI at this point in time. So, Randy, we tend to believe, also sharing the same sentiment like you, that probably 12 height is more of a 20, 25 story, you know, and probably an inflection point for HP and application for TCP.

Okay, then next question is the potential of chip-to-wafer for TCB. Can this potential be bigger than the chip-to-substrate applications? And then in terms of value and ASP of the tool, is chip-to-wafer higher than chip-to-substrate?

Maybe first part of the question, technically and potentially yes for chip-to-wafer because for chip-to-substrate, you're talking about packaging a a large compound dye into a substrate. So it's basically one dye. Whereas at the chip-to-wafer level, they are using TCB or MR, for that matter, to package, to assemble more dyes. So from that perspective alone, in terms of the number of dyes that need to be packaged at chip-to-wafer level, compared to one dye at the substrate level, potentially, uh there'll be more tc tools you know required for chip to wafer level and not just in terms of number of type but it's also in terms of technical uh requirement because uh as we continue as i mentioned as we continue to add more and more chiplets i would say at the chip to wafer level um the the uh the requirement for accuracy becomes much higher, and that's why TCB is the choice of tool, the way we see it at this point in time. And coming back to your second part of this question, in terms of ASP, probably you can guess by now that because the higher accuracy is needed for jitter wafer tools, so from that perspective, jitter wafer tools potentially has a higher ASP compared to chip to substrate application because of the higher technical requirement in terms of pitch, in terms of placement accuracy.

Okay, great. And I wanted to ask a follow-up, just going into the hybrid binding, we have a couple orders for the tool. Could you discuss where you're seeing the traction, like the application and how you feel now you're positioned versus the, I think you're the competitor kind of early in the market, just how you see your position and also what you're expecting like.

So for hybrid bonding, we are, you know, not the first mover, right, for sure. And our G1 so far, we have managed to secure four tools already, two last year, two this year. So, as we said earlier, G1 is more like a tool for us to really understand a little bit more about the intricacy of the hybrid bonding. At the same time, in parallel, we are working on a G2 version. Now, for hybrid bonding, I think we continue to maintain a stance that at this point, we still see TCB at a sweet spot in terms of weather packaging for logic weather packaging for uh hbm uh because of really the cost of performance or total cost of ownership kind of perspective tcp sits right in the middle between mr on one side and hybrid bonding on the other side And also with the recent relaxation of HBM high, I think it plays into the potential for TCP to have a higher potential in terms of also TCP for HBM rather than HP. So that's how we see it. And secondly, On the TCP side, in terms of progress, in terms of technical capability, together we are customers. We continue to push the boundary in terms of pitch accuracy, in terms of placement accuracy. you know, towards more and more closer and closer to the realm of hybrid bonding. Of course, you know, hybrid bonding is still much more accurate compared to TCP, but, you know, because of this push of the TCP capability, you know, more and more towards HP, I think TCP potential from that perspective can continue to increase vis-a-vis HP. Just a few at this point.

Okay. Yes. No, thank you. Appreciate that. If I can fit one last question, just back to the bookings outlook. Second quarter, normally, you're still at a low base, and it traditionally used to have a seasonal pickup. Is the issue certain application or certain segments like IDM, auto, industrial issues? or from OSAT. I'm just curious where the softness versus history, and when you mentioned second half up year over year, if you could clarify, do you expect sequential that we get off the flat second quarter, we could get sequential, at least at this stage, your view for sequential growth in Q3 and Q4?

Do I need Let me request Robin to comment on the bookings potential for 2Q because historically, seasonality-wise, 2Q can be a bit higher than Q1. So why is it not going up? Is there any particular applications or reason which Robin can share? And we did mention that bookings on a year-on-year basis are sort of increasing. So will that trend continue for the remaining part of the year?

Randy, Justin, when I answered the question on Q2 booking kind of color, I did mention that for the semi-mainstream, we see that continue to be a low level. So if you look at our semi-business right now, it's running on two tracks. One, on the advanced packaging side, the momentum is robust. we see that momentum will continue, you know, going forward in the near future for sure. But on the mainstream side, like for example, we are talking about wire bond, for example, or, you know, the basic time bonder, not the advanced one, but more the basic one. These tools are more closely tied to a number of things. One is really there are more consumer related kind of applications. So if the consumer demand for electronic goods are still kind of muted, we don't expect those tools to be in high demand. So the consumer related and market application, White goods, for example, you know, that's one good example. Notebooks is another example. Smartphones is another example. So all these more of the consumer-facing kind of end market-related demands have to come back in a more meaningful way before we can see booking for mainstream semi-site, you know, to show an update. But having said that, we see some... a little bit more encouraging side, if I may put it that way, compared to last few quarters. Some of these customers, especially customers in China, they are a little bit more, I would say, more active in inquiring. Like, for example, they say, hey, guys, you know, if I want to order these tools, you know, what is the lead time? So there's some interest, I would say. But if you ask me whether they are willing to place, you know, an order today, you know, they will probably take will take a more wait-and-see approach to time their order so that they can catch the upside, rather than buying too much in advance and waiting for the upside to come. So from that perspective, we see that it's more sporadic, from 8, 12, 16 and beyond. That's why we have been saying we continue to see TCB application become more and more relevant. Now, of course, TCB technology will have to progress at the same time, right? Because when you continue to increase the density of the HBM, you know, to sort of match the computing power at So as you continue to increase the computing power and the logics, the industry also have to increase the density and the memory of the HBM. So they have to increase the IO per die, and with that, we need more and more accurate TCB tools. So we have already developed a next generation of ultra fine-pitched tool that can go below 10 micron kind of pitch, you know, one micron kind of accuracy. So that's put us in a good position to have this tool to serve both the logic side at the chip to wafer, as well as for the memory HVN side as the IO density per memory die also increases. So you need more and more accurate TCP tool going forward. So we already have developed an ultra-fine pitch tool that is ready for that kind of applications.

Okay. The next part of the question is, if HBM players switch to certain technologies, say MRMUF or some other technologies, what does it mean for us? And are our TCBs still relevant or how do we play in it? Yeah.

Our view is that increasingly as we move from 8 high to 12 and beyond, For that reason, I just talked about a couple of minutes back. We believe there's really limitation in terms of MR application. First, MR tools are not as accurate as TCP tools. And two, MR tools, we believe, has also certain limitation in handling Tindai. So from that tool perspective, in terms of accuracy, In terms of the capability to handle thin dye, especially for HBEM, TCB stands out as the preferred solution.

Kevin, do you have a follow-up?

All right. Yes. Just a quick follow-up on TCP for chip to wafer. I think previous discussion, we're talking about increasing, potential increasing in ASP due to the high precision. I was just wondering, what about in terms of number of tools needed? Would that be also requiring more? And what would you say...

let's say a total addressable market for chip to wafer compared to chip to substrate would it be like a two times three times addressable market let me answer for i think i have a sort of talk about it just now but let me sort of repeat again between chip to substrate and chip to wafer because the number of chips at the sub at the wafer level increase, right? Because the industry, they are packing more and more chips at the wafer level because of higher, you know, computing power requirement. So from that perspective, as you compare to chip to substrate, you're talking about one compound dye essentially, right? So remember, at the chip to substrate level, you're talking about one compound dye. At the chip-to-wafer level, you're talking about HITO reuse integration of chipless. And the number of chips that we believe the trend is will continue to grow. The chipless will get smaller and the requirement to pack more and more chipless into a fixed kind of real estate. So the requirement for accuracy in terms of packaging will continue to increase. So from that perspective, if you compare just chip-to-wafer and chip-to-subtract, the number of TCT tools, in our opinion, will have to be higher at a chip-to-wafer level compared to chip-to-subtract.

Thanks. Can I request a lapping? Can you unmute yourself and ask your questions?

Yes, thank you for taking my question. So the first question is, can you share some color about your booking momentum by regions, especially in China? So we saw the front-end equipment demand was very strong in China in the last, I think, one year or one and a half years. But do you think the back-end equipment will pick up sometime this year or next year? Thank you.

At this moment, I mentioned earlier, we don't see it in a very broad-based manner. Sporadic demand, yes. I mentioned our customer base in China tend to be a little bit more active compared to the last couple of quarters. So they make more inquiries, they're a little bit more anxious if the ramp counts would they be able to you know get their foods in time you know uh so so we see sporadic demand also coming from certain areas like uh for the more consumer related uh application but But as I said, in general, the semi-mainstream level of business demand continues to be low as in the previous order. So sporadic demand, yes. A little bit more enquiry, yes. Their utilisation, we check, we do channel checking, right? Customers utilisation have increased. Have they reached a level whereby they will buy in terms of volume? Maybe not at this point in time. So we probably have to wait and see for a while in terms of semi-mainstream. Zabing, do you have a follow-up question?

Yes, yes. So the second question, so I think the market has quite a large expectation on the growth of the Chinese memory makers. From your side, do you see any order from the Chinese memory maker or the logic maker on the advanced packaging, especially on HPM, similar to HPM-like products? Thank you.

Yeah, in general, we do see. We do see, you know, for advanced packaging solutions, no specific tools, but across the whole range of AP tools that we have, we see our customer base in China definitely getting more active over the last four quarters. That's a trend we're seeing.

Okay, so do you think that Chinese has their own HBM-like products or not? Yeah.

Um, I can't comment specifically, but, uh, I think in general, that's the trend going to HBM, you know, is a trend to, to match the computing power at a logic set.

Yeah. Okay. Thank you. Simon, can you unmute yourself and ask your questions?

Yes. Uh, can you hear me well?

Yep.

Yeah. Great. Yeah. So just, uh, thanks to Robin and team, uh, very high level question for us. Uh, Would you recap the overall trend of the advanced packaging sales portion? What happened to 2023? And then the Q1, I remember advanced packaging revenue portion out of a total semi is about 20%. So would you recap the trend? Thank you.

We don't report this matrix on the Q1 process, right? On a quarterly basis, only for half-year and full-year? Only for full-year, but I can give you some color. I think the AP trend is increasing versus the mainstream side lately. But you have to wait. We're under Q2 when we report, we'll give a little bit more colour as to the composition of the AP proportion versus the mainstream. But at this moment, I can give you indication that the AP momentum is robust. compared to the mainstream side.

So, Simon, just to add on, if you see our yearly announcements, and if you see the breakdown of revenue from APS percentage of the entire group, that has slowly been increasing. And for 2023, what we reported was about 22% of revenue.

Yeah, that's great. And then, again, I know you guys are discussing a lot about TCD, which is very important for long-term growth for ASMPT. So maybe let's think about your customers. If they have to buy new TCV for every different generation, for 8 high, 12 high, 16 high, is there any TCV which can cover broadly 12 high, 16 high, even 24? Or each customer must purchase new TCV to follow the higher, higher DLM die stacking? That's the question.

Okay, so... Simon is saying that, you know, for different stacking requirements, 8 high, 12 high, 16 high, do the customers need to buy different equipment or different TCBs to handle different stacking? Or is there a TCB tool which can handle multiple stacking requirements or multiple height-related stacking requirements?

Okay.

um yeah uh basically our tcb tools are quite fungible right so of course but customers because i we said before for ap especially for ap um there's no standard way of packaging right it differs from customer to customer but i think within customer itself the TCB2, I think the only distinction is whether you use the ultra-fine page or the current generation of TCB2. So there's already a distinction because ultra-fine pages, as the word implies, is really precise. So that would be a difference between the TCB2. within the customer itself, there was probably stick to one type of TCP tool, because they want consistency in terms of, But if you talk about between customer to customer, the configuration can be different because there is no standard way of packaging.

How about this way then, for example, ATVM3E for new GPU is coming more meaningfully massively. Currently the A-high, but the 12-high also coming. So to have the HBM-3e with a 12 high versus the HBM-2e or 3e, your recommendation to the customer, you know, the new TCV is needed, or old TCV can be upgraded?

What could be your view, sir? Between the current generation TCV and the outer 5-pitch, there are differences in terms of modules. So, not likely can be upgraded in the future, so it has to be purchased brand new for ultra-fine pH on the TCV.

Yeah, that's very great. So, very lastly, so when the customer installs the new TCV for the more advanced HVM, you know, obviously the temperature should be higher and the compression level also higher, right? That's the continuously, your TCB spec with the temperature level and the compression level must be increased, right?

In a way, I mean, it's a bit technical, but in a way, yes, it depends on the process. As I said, the process refers from one customer to another customer, yeah.

Okay, yeah, good.

Thank you very much, Robin. Thanks, Simon. Arthur, can you ask your questions?

Hi, thank you, Robin and Katie, taking my question. Arthur Lai from Macquarie. First, I want to congratulate the SMPT team has delivered such a leading advanced packaging. And I think many people asked about TCP, so I will not ask more. I want to ask another question. So recall 2018 company acquired Nexus and also Amica. And we actually believe that market is very important for the TSV or for the other processes. So can Robin walk us through how you think the total adjustable market and the synergy to your existing equipment and any good news you can share with us? Thank you.

Okay, thanks Arthur. A lot of synergy. Justin also mentioned in terms of Our solution for AI packaging 2.5D panel package, we are planning on multiple fronts. So TCB for a chip to wafer, TCB for chip substrate. Our next two, the depositions are also involved in laying interconnect layers on the organic substrate, right, in a panel form. So T-Connect is also contributing an area. And For Amicra, which is a photonics tool, we are also contributing to AI in terms of the transverse market. So we mentioned for the last two quarters that we see heightened interest in photonics tool because Again, as the computing power increases on the logic side, as well as the HBM, the communication capability, bandwidth, transfer rate, all these have to increase in a sense to match with the computing power. So there's this heightened interest in photonics application, which we are also a leader for a whole range of tools that is required from 100G, to 800G and beyond. So we have a whole comprehensive of tools to cater to that kind of photonics transceiver packaging application. So I hope I answered your question on that as well as Amiga.

Yes, yes. And one quick follow-up. How about the TAM?

We don't break down TAM by individual tools. We have the TAM for AP, right?

So what's the TAM for AP? Okay, our time, Arthur, is going to increase up to 3.3B in 2028 with a CAGR between 2024 to 2028 of about 18%. And this is for our entire AP addressable market. So we have a sort of a cadence of giving this kind of number once a year. And we just updated this number end of last year, which was announced towards end of February.

Yes, and Robin, just one qualitative comment. You mentioned that this market actually is getting very robust. So did you sense that a lot of, you know, resource need from your team? Like they ask you to add more headcount on the nesters or ask you more, you know, do more corporate, you know, allocation?

Hi, Arthur, this is Katie. Very nice to hear from you. So in terms of the investment, right, so at the end of February when we did the annual announcement, we actually mentioned that due to all the activities and just the high potential of APs, especially, right? So we have announced that we will, in addition to the current investment level, we'll be investing 250 million Hong Kong dollars for R&D, especially on AP side and also infrastructure of the company. And as I mentioned in the opening remarks, those investment projects are all on track and we do expect that the investment expenses as we go through this year will intensify.

Thank you, Katie. No more questions.

All right. With that, we'll conclude our Q&A session. Let me request Robin to say a couple of words and conclude this call. Thank you, Rob.

So let me quickly highlight some key takeaways from today's call. Our book-to-bill ratio recovered above one of the seven quarters. This was mainly due to growth in our group bookings fueled by strong demand for our AP solutions. The AP solutions contributed strongly to bookings across both SEMI as well as S&T. AP remains a bright spot for us and some encouraging points on its potential based on the first quarter would include the following. TCB having accelerated adoption and also our commanding position in the space. We see heightened levels of engagement with HBM players in addition to the auto momentum in logic for TCB. We see traction also in hybrid bonding. And last but not least, our market relationship in SIP or system and package tools under our SMT business segment. So this concludes our call and I'll see you all in the next quarter. Thank you and take care. Bye-bye.