Good morning, and good evening, ladies and gentlemen. This is Justin Tham, and I will be moderating today's call. On behalf of ASMPT Limited, welcome to our First Quarter 2025 Investor Conference Call. Thank you for your interest and continued support. Please note that all participants will be in listen-only mode during the presentation [Operator Instructions] Before we start, let me go through our disclaimer. Please note that there may be forward-looking statements about the company's business and finances during this call. 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, Mr. Robin Ng; and 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. Now I will hand over to our Group Chief Executive Officer, Robin. Robin?

Thank you, Justin. 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 2025. Now let's start with the key highlights of the first quarter. We achieved group revenue of USD 401.5 million, which met the midpoint of revenue guidance. The group's Advanced Packaging Solutions continued to be a major beneficiary of AI adoption. AP Solutions continued to deliver strong performance led by our Thermal Compression Bonding or TCB tools. The group completed the delivery of the bulk TCB order to the leading memory maker and these solutions are mainly used for HBM3E 12-high high-volume manufacturing. This quarter, the group expanded its TCB HBM customer base. We won initial orders from another global HBM customer, which has been followed by further orders in April 2025. There were also continued bookings for chip-to-substrate tools serving the logic market, where our tool is a process of record. And lastly, within TCB, our chip-to-wafer TCB tools enabled with Active Oxide Removal fluxless capability have progressed from qualification to pilot production and a leading foundry. This encouraging development demonstrates our technological advantage in fluxless tools. The strong progress made this quarter in TCB further solidifies our leadership in the market. And our focus for 2025 is on securing additional orders from both HBM and logic customers. Moving on to group gross margin. This quarter, we saw a rebound in group gross margin, which exceeded 40% and was driven mainly by better product mix in both segments. To finish this overview, the group's mainstream business continued to be affected by soft demand from automotive and industrial end markets. While the growth trajectory of the mainstream business is difficult to forecast given the current environment, we are fully prepared to seize opportunities when the market recovers. With that, 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's key financial metrics for the first quarter of 2025. Group revenue met the midpoint of guidance, totaling USD 401.5 million. Group bookings delivered USD 431.2 million, which was better than expected, showing a 2.9% growth quarter-on-quarter and 4.8% growth year-on-year. The quarter-on-quarter increase was mainly due to higher SMT bookings, partially offset by lower SEMI bookings from a high base effect in Q4 2024. The year-on-year increase was driven by SEMI, which has shown year-on-year quarterly bookings growth over the past 6 quarters. In the first quarter, the group's gross margin was up 371 basis points quarter-on-quarter, but down 97 basis points year-on-year. The quarter-on-quarter improvement was due to both SEMI and SMT. Disciplined cost control measures and seasonality reduced the group's operating expenditure by 11.3% quarter-on-quarter. However, year-on-year operating expenses were up 4.1% due to the investments in strategic infrastructure and R&D to drive growth in our AP business. Thus, the adjusted net profit was HKD 83.2 million, up 1.6% quarter-on-quarter, but down 53.1% year-on-year. The year-on-year decline was due to a combination of previously mentioned a slight reduction in gross margin and OpEx increase for strategic investments as well as ForEx effects. For the first quarter of 2025, SEMI revenue grew to USD 255.6 million, up 0.6% quarter-on-quarter and 44.7% year-on-year. SEMI contributed about 64% of the group's revenue. In Q1, SEMI recognized revenue for the bulk order of TCB tools delivered to a leading HBM customer. SEMI bookings were USD 222.9 million, down 19.5% quarter-on-quarter, but up 11.4% year-on-year. In Q1, there were new TCV bookings, which included initial orders from another global HBM customer with further orders placed in April 2025. There were also continued bookings for chip-to-substrate tools serving the logic market, where the tool is the process of record. In addition, there were mainstream wins for high-end smartphones and automotive applications. The quarter-on-quarter bookings drop was mainly due to a high base effect from the bulk TCB order in Q4 2024, while year-on-year increase was mainly due to TCB orders. SEMI's gross margin of 46.3% for Q1 2025 was up 368 basis points quarter-on-quarter, mainly driven by higher AP mix and benefit from one-off items that impacted Q4 margin. Gross margin was up by 167 basis points year-on-year. In addition, as mentioned earlier, OpEx reduced quarter-on-quarter due to disciplined cost control measures and seasonality. Lastly, SEMI's profit was HKD 235.9 million in Q1 2025, an increase of 215.9% quarter-on-quarter. Moving on to the SMT business. SMT delivered revenue of USD 145.9 million in the first quarter of 2025, a decline of 20.3% quarter-on-quarter and 35.6% year-on-year, in line with ongoing softness in its overall market. SMT bookings of USD 208.4 million were up strongly, 46.5% quarter-on-quarter, driven by strong seasonal system-in-package or SiP bookings. The automotive and industrial end markets appear to have stabilized but remained soft. SMT's gross margin of 31.5% was up 180 basis points quarter-on-quarter, but down 827 basis points year-on-year. Quarter-on-quarter improvement was due to favorable product mix, while the year-on-year drop was mainly due to lower sales volume. Segment loss was HKD 5.3 million in Q1 2025 due to lower sales volume. I will now pass the time back to Robin.

Thank you, Katie. In Q2, the group expects revenue to be between USD 410 million to USD 470 million, up 3% year-on-year and 9.6% quarter-on-quarter at midpoint. We remain confident of sustaining AP revenue and expect the mainstream business to improve due to both seasonality and better-than-expected Q1 bookings. Looking ahead, the mainstream growth trajectory is difficult to forecast given uncertainties from the indirect impact of tariff. However, the group remains confident in the demand for Advanced Packaging and our TCB solutions for AI and high-performance computing applications. In addition, our global manufacturing footprint provides flexibility to navigate the potential impact of the tariffs. The group will continue to monitor the situation closely and adapt as needed. This concludes our first quarter 2025 presentation. Thank you, and we are now ready for Q&A. Let me pass the time back to Justin to facilitate. Justin?

Thank you, Robin. Let us now proceed with the Q&A session. [Operator Instructions] May I ask Donnie Teng of Nomura to unmute yourself.

First question, as usual, housekeeping question. So, yes, after some slight rebound in terms of the quarter-on-quarter booking momentum, would you be able to comment on booking direction in the second quarter? And this time, I would specifically like to check the Advanced Packaging booking momentum in second quarter and wondering if there is any preliminary impact from the potential oversupply to drive some customers start to hold the orders? And secondly is regarding to the SMT business. So because of this tariff drama, I think potentially some customers may want to see some capacity expansion in different regions, for example, like Mexico or even in the future in the United States. For SMT, I think that's an important equipment, which could help customers to avoid sort of the uncertainty in terms of the tariff impact if they build the capacity in Mexico or U.S. directly. So would you able to comment on some of the dynamics you are seeing, particularly on SMT business after this tariff drama?

Donnie, I see that you have 2 questions. Let me start with your first question is on the Q2 bookings. Maybe you start with Robin. That's the first question, right? The second question is also the AP booking in particular on the preliminary impact on the [trade war] potential for supply that was mentioned. And maybe I'll start with Robin on the bookings question first, right?

Yes. Thanks, Donnie, for your questions. Now on the booking, as you know, we don't guide bookings actually, but we always try to provide as much color as possible to you guys, right? So I think for Q2 booking this time around, we remain confident that Q2 bookings will be within a similar range compared to the last few quarters. Of course, this assumes there is really no unexpected impact from the tariff landscape at this point in time. You talk about -- you alluded to a bit about AP as well. So maybe at the same time, I can provide some additional color how we see the Q2 bookings may pan out. For SEMI, AP momentum, we are confident that the AP momentum would be intact in Q2. In particular, TCB, we expect orders to continue to come in from both the memory and the logic side from our global customer base. SEMI mainstream, the way we look at it at this point in time is that it remains stable despite the macro environment. However, relatively speaking, compared to historical is still at a soft level for SEMI mainstream. For SMT, as you can notice, we came back quite strongly in terms of booking on the back of a very low quarter in Q4 2024. Similar kind of picture with SEMI. In terms of mainstream SMT is stabilized. And in fact, in the recent time, we have seen opportunities in AI server and also automotive market in China. So I think these are some kind of color we can give to Q2 booking for both segments.

I think Donnie do have a follow-up question regarding -- he was focused on the tariff impact. The question is if the SMT side, customers will see any capacity expansion. Maybe Robin, do you want to add some commentary on.

So maybe I comment in general first because -- I mean, the tariff is -- the situation is at the top of everybody's mind during this recent period. In terms of direct impact, we don't see any significant or material direct impact on our operations. From the perspective of our overall business landscape, we are confident that AP will continue to grow, benefiting from AI adoption and also our superior technology leadership in this particular space. I talked about Q2 mainstream as well. The mainstream business is stable, but probably on the softer side for both SMT and SEMI as well. But we are confident that the market is still growing. However, the indirect impact of the tariff makes the growth trajectory difficult to forecast. Now coming from the customer side, we see both positive and negative effects on customer decision-making in light of the tariff environment. While we haven't seen any order cancellations, some customers are, of course, evaluating -- taking some time to say, evaluating their investment timing and location going forward. However, on the positive side, there appear to be 2 kinds of effects playing out. Let me go one by one. First, some customers show a willingness to engage in spot buys or delivery tool-ins to take advantage of the tariff process. You recall we have -- I mean, the U.S. has imposed a tariff pause since April 9 for 90 days. Second, potential opportunities coming from incremental investments as customers diversify their manufacturing location to take advantage of the difference in tariff rates in various countries. So I think that Donnie sort of answered your SMT tariff impact as well. Let me end by saying something on the tariff. Our global manufacturing footprint provides flexibility in our opinion to navigate the potential impact of the tariff. We will continue to monitor the situation closely and adapt as needed. But having said all this, I think it's still unclear how the tariff situation will unfold and what is the medium-term macroeconomic impact will be. So I hope I answered your questions, Donnie.

Just one quick follow-up is that you mentioned about the AP momentum may be still there into the second quarter. May I just simply ask whether it's mainly driven by memory, HBM or logic?

Donnie, I would say both. As I said, even for Q2, we are seeing orders coming from both memory as well as on the logic side.

Yes. I'm just wondering which one would be stronger.

I mean trending-wise, I think lately with the HBM orders that we have won from the leading memory maker and in Q1 and Q2, we also won further orders from another memory maker. From the perspective, the HBM momentum, at least for SMT is picking up nicely.

Understood. Understood. So there has been keeping seeing those repeat orders, right, since fourth quarter last year.

What was the question again, Donnie.

Yes. So I just want to make sure that -- so you are still seeing repeat orders since fourth quarter last year from the first HBM customer?

No, we are working on it because the first order, as probably you understand is a sizable order. So it takes time for the customer to digest the capacity. But surely, we are working hard to continue to engage not just this memory maker, but the other big ones as well to garner more orders as we progress. On that note, I must add that we are doing well on the HBM side. We have proven ourselves that our tools in the first memory maker are already in high-volume production for HBM3E 12-high. And also, we have been saying that for HBM4, we believe that we are the first to package the very complicated architecture for HBM4 chip. So we have proven ourselves that we are able to do it. And in fact, the way we look at it is that the sooner HBM4 comes into fold, it plays to our advantage because HBM4 packaging is a step higher compared -- requirement is a step higher compared to HBM3E. And that's where I think we could differentiate ourselves quite nicely from the rest of the pack because we believe our technology is able to handle the very complicated AI architecture of the HBM4 in terms of bonding accuracy, in terms of coplanarity, control, in terms of die handling, ensuring die crack and so forth. So we are confident that we will continue to win more orders in the HBM space.

Thank you, Donnie. May I ask Sunny to unmute yourself.

So my first question, I also want to follow up on TCB. And so you mentioned that you are getting orders from the second HBM customer. Could you help us understand the magnitude of the orders that you're getting in Q2? And then for your first HBM customer, since the customer should start ramping HBM4 soon, maybe in second half of the year. And so based on your current engagement, would you expect to get the second bulk orders? And then that will be for fluxless TCBs maybe into Q3 or Q4?

I think this question is for Robin.

Yes, for the second HBM customer, that's a question, Sunny. Yes, the orders are relatively smaller compared to the first one, but meaningful, I would say. And we have received already 2 orders; one in Q1 and one in April 2025. Now in terms of HBM4, as I mentioned earlier, we are confident of our capability in this space. We believe we have very good technology. Now we are definitely planning for a market share for HBM4. I think the industry has probably recognize that ASMPT has successfully used our tools to package HBM4 already for a number of customers. So I think that put us in a good position to garner orders for HBM4. Now in terms of timing, we do believe that HBM4 should start some kind of ordering in the second half. Now whether it is in Q3, Q4, that's really not up to us. It's really up to our customer. So we don't have a lot of visibility at this point in time, but can happen in second half of 2025. Depending on the type of HBM4, we believe fluxless for HBM4 maybe needed for [MUF] application, sorry. For NCF, we cannot use fluxless. The flux tool will still have to be used for NCF HBM. So it depends on the customer. So if they are going for NCF, it will not be fluxless. But if they are going for MUF, chances are for HBM4, it has to move towards fluxless solution.

So a small follow-up on the technical details. Robin, would you mind sharing more thoughts on why NCF would not need fluxless TCBs?

Because of inherent difference in the die between NCF structure and MUF. So NCF, you do need fluxless. Basically, fluxless is good enough.

Got it. My second question is on the leading edge logic. And so for chip-on-wafer, which I think most people are paying attention to. You mentioned leading edge foundry has moved forward from qualifications to trial production for these tools. But I guess the question is, has this customer decided if they will use chip-on-wafer for -- to use faster TCBs for chip-on-wafer anytime soon or they are still taking some time. And I wonder what are some of the key factors that hold them back from a faster migration to faster TCBs?

I think for the last bit of your question, Sunny, you probably have to ask them what is really in a way, holding them back what you have described. But in any case, as far as we are concerned, I think we have made significant progress from qualification to now using our tool for pilot production. From our engagement with the leading foundry, we are in a good position. We are in a good position. We have proven our capability to be able to handle complex packaging requirement or chip-to-wafer application using our fluxless tool. So we have recently shipped another tool to the leading foundry for the evaluation. We believe, as I said, we have made very good progress. So we believe that we are in a good position to win the chip-to-wafer orders once the leading foundry make up their mind.

Got it. Is there any time line when this customer may be able to place some orders?

We hope second half of 2025, maybe, yes. So -- but having said that, I think I have been repeatedly telling you that for chip-to-wafer, even it happens this year, it's not going to be significant. Chip-to-wafer application, the demand, the high-volume demand will only come in 2026.

Thank you, Sunny. Now may I ask Gokul to unmute yourself.

First of all, on HBM, could you talk a little bit about your experience with this leading HBM customer? There's been quite a bit of competitive noise about this customer that ASMPT tools have had some issues with this customer. Could you talk a little bit about where you are in terms of mass production for 12-high 3E, which is what I think your tools were targeted for? And secondly, I think when you talk about ASMPT being in the pole position for HBM4, is it mainly for this customer, given that you mentioned that fluxless will be required for MUF application for sure. And yes, maybe start there.

Thanks, Gokul. Let me answer your question on HBM. I had sort of alluded earlier on the HBM question, how do we differentiate ourselves from a competitor. We differentiate ourselves by the quality of a bond in terms of accuracy, in terms of voltage control, in terms of coplanarity. Also not to mention, I think customers value the scalability of our solution. We can move from a flux-based tool to a fluxless tool seamless, right, by adding a module, AOR module, Active Oxide Removal module on site. So I think that's an important consideration also for our customer. And not forgetting also in terms of HBM stacking, I think the ability to be able to stack chip as close as possible in terms of height, in terms of vertical stacking, what we call the chip gap is also important, right? So it has to stay within a certain height level. So I think all this added up together put us in a good position in terms of technology vis-a-vis the other players in the space. Now I think your second question is...

Maybe, Robin, Gokul wanted to touch on the progress of HBM4 progress with the leading customer or even customers in the HBM space. Maybe you could give him some update.

Again, I think it all boils down to those advantages that I have elaborated quite a few times already this morning in terms of our HBM capability. So I think that put us, again, in a very good position to make a progress -- continue to make progress in the HBM space for 4 because as I said earlier, HBM4, the AI architecture is challenging compared to HBM3. So you need certain capability to handle HBM4. Now Gokul, I think you also mentioned there have been some talk that our tools may be facing issues with the leading player. Let's put it that way, HBM stacking using TCB tool, sorry, is I would say it's first of a kind, not just for ourselves, but also for the industry as a whole. Now typically, in our equipment space, when we roll out first-of-a-kind tools, we're bound to have issues on site because there's a lot of process requirements are needed to harden the tool. So I would say this is part and parcel of rolling out first-of-a-kind tool for HBM stacking. So it's nothing unusual. The fact that we have our tools have now been in high-volume production using high-volume production for HBM3E 12-high and we have demonstrated our capability to do HBM4, I think that put to abate whatever doubts the industry had of our capability for TCB stacking tools for HBM.

Got it. Just a follow-up on that, Robin. What is your confidence on follow-on orders from this main customer? I think you had a pretty big bulk order that you've largely fulfilled in Q1. How is that follow-on order looking like? And maybe now that you have the second customer also, for HBM this year, who's going to be bigger? Is it going to be the second customer or the original lead customer in terms of revenue contribution?

Hard to forecast. Right now we are deeply engaging all HBM players, including the third one, okay, we have also successfully -- we have successfully bonded HBM stack also for the third memory maker. We are getting into deeper engagement with the third one as well in terms of joint development kind of arrangement with the third memory maker. Now I think that sort of put us in a good position to garner more shares from the HBM space.

Okay. So just to clarify, so we should expect follow-on orders from the lead customer pretty soon.

We are hoping for that for sure, and we're working hard to get it happen.

Okay. Got it. My second question, I think last earnings call, I think you had outlined your HBM addressable market growth. And you also talked about 35% to 40% market share is your target for HBM. Given all the developments, especially on the HBM side you're seeing, what is the kind of rough market share you think you can secure, especially as we move to HBM4, which feels like where your tools seem to have an edge compared to competition. Do we get to that 35% to 40% market share for HBM4 or even higher? Any thoughts on that front?

Yes. Gokul definitely, we're working towards getting more market share constantly and relentlessly. Yes, you are right. We came up with a TAM and we did share. Our ambition is really to garner 35% to 40% market share in the...

For the TCB space.

For the TCB space.

It's not just for HBM.

It's a TCB space overall.

So in HBM, do you think it is going to be higher or lower than the 35% to 40%?

Let's put it that way. I think there is definitely more competition in the HBM space compared to the logic space.

Gokul, just to clarify a little bit, I think when we put the $1 billion market share that TAM out there last quarter by 2027, we had a little bit more color to it. I think the logic side, we actually started many years ago, right? We have a very strong foothold. HBM really, we are -- we were basically the new player to the market. So Robin is right, we aspire to get to that the market share range. But we actually -- you have to bear in mind that we actually came into this game at 0 market share. So it's actually a really huge step-up already.

Got it. One small follow-up to Sunny's question on substrate. Previously, I think the understanding was that ASMPT was kind of behind your competitor in terms of qualification for this large foundry for chip-on-substrate. Do you think the position has changed now? You're largely on par with your competition or still slightly behind competition?

I think, Gokul, you're probably referring to chip-on-wafer rather than chip-on-substrate...

Chip-on-wafer. Yes.

Yes. Chip-on-substrate, we are still the sole supplier. Now on chip-on-wafer, yes, because we have really made significant progress from competition to now using our tool for pilot production, I think we are well positioned in that space considering the progress that we have made.

Thank you, Gokul. Now may I ask Hiu King to unmute yourself.

I have two questions. The first question in your presentation, I think it's the Page 8. I see 8% improvement on the segment margin this quarter, although the revenue does not change too much. So what are the major change you did to improve the operating efficiency and how we should model the profitability of the SEMI solution or the OpEx level in 2025?

Are you referring to SEMI?

Not SEMI. The Semiconductor Solutions segment probably.

Yes, SEMI. And you're talking about...

Segment margin improved quite a lot this quarter, right?

Correct. Q-on-Q.

Yes, Q-on-Q. So although the revenue does not change too much, right? Yes, it's basically flat. So what are the major change and how we should model this OpEx level in the Semiconductor Solutions.

Okay. I just trying to understand. You're talking about -- sorry, I'm a little bit confused here. You're talking about gross margin, talking about OpEx.

Yes. In other words, why the segment's margin improved so much in this quarter? Is it mainly due to the OpEx change or is it mainly due to the margin expansion as well?

I got you. I'm sorry. Yes. Actually, the answer is both. So if you look at the gross margin level, definitely, as we mentioned in the opening remarks, the product mix for the segment definitely was very healthy because we're just spending some time talking about the TCB side of the momentum, that's for sure. And also, we also mentioned that in Q4, there was a negative one-off margin hit, right? So Q-on-Q, basically, they helped with a relatively lower comp. So on -- that's on gross margin side. Now on the OpEx side, as we mentioned in the opening remarks as well, there's continued cost control measures in this relatively low environment. If you recall, we did announce restructuring programs in Q4. So we are benefiting from some of these savings from those measures. At the same time, Q1, usually, if you look at our historical numbers, there is a seasonality that Q1 relatively is lower for OpEx. That's because of the incentive share provision schedule only kicks in the last 6 days of the quarter, while all the other quarters, we actually provide for the entire quarter, right? So there's a seasonality to it. So net-net, that's what really supported the SEMI profit improvement. Now I think you also have a question sort of looking out, right? We don't -- yes, we don't actually provide the outlook for the future profit and the margin. However, just direct, I would say directionally, at the group level, right, you see that this quarter -- I'm talking about the group level now, this quarter, we rebounded back to 40% and then also, as Robin mentioned, I think within the product mix, we do feel very confident with the momentum of AP. Therefore, we expect that the margin going forward -- gross margin at group level going forward will be in the range of 40%. And I hope that helps with some of the color to the future profitability.

Okay. So what will be the OpEx level when you try to keep maintaining in terms of revenue scale? It seems that your OpEx level is much higher than quite a lot of other semi company.

Yes. So let me kind of -- on the OpEx, so similar to what we communicated, we try to be quite consistent actually. Every quarter, if you look at the OpEx level, it's probably HKD 1.1 billion, HKD 1.2 billion a quarter. And at the beginning of the year and very similar to last year, we did announce that with the growth opportunity we have, especially in AP, we have committed HKD 350 million of incremental investment, especially again for those R&D programs in TCB in hybrid bond, et cetera. So we do expect the OpEx number to be relatively steady with some marginal increase against a very, very surgical, very specific programs that we're running, especially for the R&D side.

Okay. It's very clear. So the second question is I attended SEMICON China this March and listened quite impressive presentation by your representative about the Advanced Packaging. So my question is, what's your view on the AP demand in China? And what are the major best-selling AP products in China?

I think a little bit -- Robert.

We cannot comment too specifically. But in general, our Advanced Packaging Solutions serve a global customer base.

Okay. So what will be the, let's say, percentage of your AP within your China business now?

Sorry, we don't break this down for competitive reason.

All right. May I ask Kevin, Citi to unmute yourself.

I would like to check on the mainstream side of our business. So I think last time we talked, you mentioned that we're expecting somewhat of a bottoming out or a recovery in second half of this year. So given like the recent impact, especially from tariff, do we still think that's the case? And how much confidence do we still have compared to a couple of months ago? And especially, I think we saw that SMT booking has shown quite a rebound in the past quarter. I think this is due to in part to SiP seasonal recovery. But how do you think this is going to trend towards later part of this year?

Kevin, thanks for the question. I did mention earlier, but I can sort of repeat again. Now the way we see mainstream business in general for both SEMI and SMT is that it has stabilized the mainstream business. And although it's still on the soft side because of the certain segments are still slow or sluggish, for example, the automotive and the industrial end market, especially in the European and the American region, that's slow. So obviously, that also impacted, in particular, our SMT business. But in general, the good sign is that the mainstream business has now stabilized. And looking at the -- because of the tariff, we are confident that the mainstream business will continue to grow, but the indirect impact of the tariff makes the growth trajectory difficult to forecast. So I can't give you a very definitive answer whether the growth will come in the second half or later because of the tariff situation.

Right. Just a quick follow-up on the SMT part. Do you think the recovery in first quarter is going to continue?

Looking at the color that I've given you earlier, we expect to be in the same range in the last couple of quarters. Yes, I think the short answer is yes. We will probably not fall to the level in Q4 for SMT booking.

Right. Got it. My second question is, I would like to get some update on your hybrid bonding. I think I guess right now, the industry consensus is still that the industry might need hybrid funding when we move to 12-high. I guess the timing is probably around like '27, '28 or possibly earlier. But just wondering how ready are we in terms of our second-generation tool? And do we have any following subsequent orders coming in this year?

Yes. Kevin, I believe it's probably you should be saying 20-high rather than 12-high. The 12-high, definitely we don't need hybrid bonding, right? So TCB is a choice of true. Our strong belief is that even for 16-high, our true TCB is good enough for 16-high. Now there's been a lot of talk about HP. Too early to tell. Of course, HP is maturing as we speak. But we still believe the HP because of the total cost of ownership is higher as everyone is known by now, the industry will continue to use more cost-effective tool like TCB for both logic and also HBM as far as possible, yes.

I will have the last question from Gokul. Can you unmute yourself and ask your question.

First question, given I think we've seen a pretty sharp share price decline, you're still sitting on pretty healthy cash balance. Just wanted to ask a question to Katie. Are we thinking about anything other than the regular cash dividend in terms of shareholder returns, any kind of share buyback policy or anything like that, that you've been thinking about with the Board to kind of shore up the share price and help some of the investors? First question.

Gokul, thanks for the question. And your observation is right, the company actually had some very solid cash performance in the recent years. However, due to the macro and the tariff uncertainties, we do not think this is the right time to launch a share buyback. We acknowledge our current share price level and we're always evaluating different options to return capital to shareholders. So we'll continue to monitor the market situation and we'll look into buybacks when it's appropriate.

Got it. And the second question I had, I think just following up on OpEx overall. I think if I look at the overall OpEx, I think last year was about HKD 5 billion. I think it used to be like HKD 4 billion or so in the last upcycle, but the revenues obviously are much smaller than what they were. So when you think about managing OpEx over the next maybe 3, 4 years through the cycle, do you feel this is the OpEx level that the company should maintain? Or is there a point at which you think we should think about some meaningful like OpEx control measures or something like that, given the OpEx ratio is quite high. I think some of your competitors have actually managed this OpEx a little bit more flexibly over this last long down cycle that we have had.

Gokul, thanks for the question on OpEx. As you probably recall, we -- in this relatively long down cycle, we have conducted 2 or 3 restructuring programs already and we'll continue to put cost measures on the various functions. So we are mindful of the cost level that we have. Having said that, though, we do want to emphasize that as a technology company, we want to protect -- create the future of the company, which is really on the R&D programs for future products. That's why in the last 2 years, we actually proactively came out and announced the incremental investments. So we'll continue to do that. So I think it's -- really it's a fine balance between the two, right; investment for the future and sensible cost control in this, again, relatively prolonged down cycle. So we'll continue to make sure that we strike a good balance between the two.

Got it. Maybe one small follow-up on the chip-on-substrate side, how do we think about visibility of orders? I think this year, obviously, is going to be a pretty strong year. Do we think that we should continue to grow next year for chip-on-substrate related orders for this large foundry and some of the other OSAT players who are kind of having a similar supply chain?

We do, we do, definitely, not just in terms of the current architecture, but going forward, Gokul, you can expect the compound die or the compound die is such that it will get bigger and bigger because more capacity needed, more capability needed for the AI chip. So that means customers who have continued to buy new tools in order to support the program to [borne] a larger die onto the substrate. So this trend will continue in time to come for sure.

Okay. With this, this concludes our Q&A session. And let me request Robin to say a few words. Robin?

Thank you. Thank you for all your questions, and allow me to quickly highlight some key takeaways for our call today. So the group delivered a solid performance and gross margin rebounded to above 40% in Q1 2025. We expanded our TCB customer base, won new orders and our tools are being used in production for both logic and memory applications. This further solidify our leadership in the TCB market. While the growth trajectory of the mainstream business is difficult to forecast, given the current environment, we are fully prepared to seize opportunities when the market recovers. The group remains confident in the demand for AP and its TCB solutions for AI and HPC applications. In addition, our global manufacturing footprint provides flexibility to navigate the potential impact of the tariff. As I said earlier, the group will continue to monitor the situation closely and adapt as needed. This concludes our call, and I'll see you in the next quarter. Thank you very much.