ASE Technology Aktie

Hello. I am Ken Hsiang, the Head of Investor Relations for ASE Technology Holdings. Welcome to our first quarter 2026 earnings release. I am joined today by Joseph Tung, our CFO. Thank you for attending our earnings release today. Please refer to our safe harbor notice on Page 2.

All participants consent to having their voices and questions broadcast via participation in this event. If participants do not consent, please do not ask questions or you may leave the session at this time.

I would like to remind everyone that the presentation that follows may contain forward-looking statements. These forward-looking statements are subject to a high degree of risk, and our actual results may differ materially. For the purposes of this presentation, dollar figures are generally stated in new Taiwan dollars, unless otherwise indicated.

As a Taiwan-based company, our financial information is presented in accordance with Taiwan IFRS. Results presented using Taiwan IFRS may differ materially from results using other accounting standards, including those presented by our subsidiaries.

For today's presentation, I will be going over the financial results, and Joseph will then provide the company's guidance. We will then be available to take your questions during the Q&A session that follows.

Our business, throughout the first quarter, remained resilient. Typically, we would expect to see manufacturing seasonality, as many consumer and corporate products wind down at the end of the calendar year. This year, we saw such seasonality in our EMS business, while demand for our ATM services did not slow at all. Even with less working days during the first quarter, our ATM revenues grew sequentially. We experienced continued strength in our LEAP services and traditional advanced packaging, while our wirebond also saw some pickup.

As our product mix shifts, typical seasonality may become more muted as AI-related products do not appear to follow the same seasonal patterns as typical consumer-driven devices. Our blended factory utilization rate was around 80% for the quarter. This percentage can be slightly misleading. Loading was actually a little better. However, we have been installing additional LEAP manufacturing capacities that are expected to start generating revenues weighted towards the fourth quarter.

We are also consolidating traditional capacities. These transitional activities would naturally bring down the blended utilization rates. As production resources get tighter, it needs to be emphasized that customers prefer manufacturing certainty. They like to know precisely the timing and pricing of wafers, substrates, packaging and testing services.

As uncertainties arise at key manufacturing points, customers perceive supply chain risk for their products. Our capacities along with those of our upstream foundry partners are finite with limited ability to be pulled forward at this point. With that said, we are seeing strength not only in our LEAP-related capacities, we are also seeing strong demand in wirebond and traditional advanced packaging.

From a financial perspective, first quarter ATM revenues came in slightly ahead of our expectations. We saw slight pickups throughout our customer base, especially as it relates to computing-related products.

We also saw incremental improvement in our profitability due to this pickup, with gross margin outpacing our original expectations. Our EMS business slowed slightly due to underlying product seasonality as per our original expectations.

Please turn to Page 3, where you will find our first quarter consolidated results. For the first quarter, we recorded fully diluted EPS of NTD 3.08 and basic EPS of NTD 3.24. Consolidated net revenues were NTD 173.7 billion, representing a decrease of 2% sequentially and an increase of 17% year-over-year.

On a U.S. dollar basis, our sales decreased by 4% sequentially and increased by 22% year-over-year. Our gross profit was NTD 34.8 billion with a gross margin of 20.1%. Our gross margin improved by 0.6 percentage points sequentially and by 3.3 percentage points year-over-year. The sequential improvement in margin is primarily due to NT dollar depreciation.

The annual improvement is primarily due to a higher mix of ATM revenue in addition to higher ATM factory utilization, offset in part by the annual appreciation of the NT dollar. We estimate that foreign exchange had a positive 0.6 percentage point impact to our gross margin sequentially and a negative 1.2 percentage point impact annually.

Our operating expenses increased by NTD 0.3 billion sequentially and NTD 2.1 billion annually to NTD 17.3 billion. The sequential increase in operating expenses is primarily due to higher R&D labor-related costs. Meanwhile, the annual increase is primarily due to higher R&D labor-related costs and to lesser extent, R&D supplies and other non-R&D labor-related expenses.

Our operating expense percentage eased 0.4 percentage points sequentially to 10% and declined 0.3 percentage points annually. Operating profit was NTD 17.5 billion, down NTD 0.2 billion sequentially and up NTD 7.9 billion year-over-year. Operating margin was 10.1%, up 0.2 percentage points sequentially and up 3.6 percentage points year-over-year.

During the quarter, we had a net nonoperating gain of NTD 0.7 billion. Our nonoperating gain for the quarter primarily consists of net foreign exchange hedging activities offset in part by net interest expense of NTD 1.6 billion. Tax expense for the quarter was NTD 3.6 billion. Our effective tax rate for the quarter was 20.0%. Net income for the quarter was NTD 14.1 billion, representing a 4% decline sequentially of NTD 0.6 billion and an 87% increase annually of NTD 6.6 billion.

On Page 4 is a graphical presentation of our consolidated quarterly financial performance. The chart effectively shows the impact of ATM business growth and its impact to our consolidated holding company-level results. For the first quarter this year, our ATM business represented 65% of our consolidated holding company revenue while representing 91% of our operating profit. This is compared to 58% of consolidated holding company revenue while representing 86% operating profit in the first quarter last year.

On Page 5 is our ATM P&L. The ATM revenue reported here contains revenues eliminated at the holding company level related to intercompany transactions between our ATM and EMS businesses. For the first quarter of 2026, we had record revenues for our ATM business of NTD 112.4 billion, up NTD 2.7 billion from the previous quarter and up NTD 25.8 billion from the same period last year. This represents an increase of 2% sequentially and 30% annually.

Our ATM business was able to avoid a seasonal decline in revenue during the first quarter upsiding our original expectations. We also delivered higher-than-expected margin performance via improved profitability from higher utilization of equipment. However, we were unable to avoid the higher running costs during the Lunar New Year holidays.

Gross profit for our ATM business was NTD 29.2 billion, up NTD 0.4 billion sequentially and up NTD 9.6 billion year-over-year. Gross profit margin for our ATM business was 26%, down 0.3 percentage points sequentially and up 3.4 percentage points year-over-year.

The sequential gross margin decline was primarily due to a higher rate of labor during the Lunar New Year holiday and also a higher percentage of depreciation from preparing equipment for further expansion. These negative impacts were mostly offset by a positive foreign exchange environment and efficiencies from higher loading. Meanwhile, the annual gross margin improvement was primarily due to higher factory utilization and a higher LEAP product mix, offset in part by negative foreign currency impact.

As a note, during this year, we see our depreciation rising faster than revenues as a result of our ongoing investments in LEAP. The granularity of our OS and full-process LEAP equipment differs from our more traditional advanced packaging lines.

Our LEAP lines must be installed at scale and together as a full set of differing machinery instead of small incremental units like wirebonders and testers. This results in our LEAP lines taking a greater amount of time to bring up and tune when compared with more traditional packaging capacities.

Currently, our full-process LEAP lines are in the midst of tuning and qualification. As a result, we will continue to see gradually increasing depreciation without significant amounts of associated revenue during the tuning and qualification period. Meanwhile, revenues associated with these full process lines will ramp mostly during the fourth quarter.

With that said, we continue to believe our margins will increase sequentially quarter-over-quarter and reach the higher end of our structural margins by the end of the year.

During the first quarter, operating expenses were NTD 13.3 billion, up NTD 0.6 billion sequentially and NTD 2 billion year-over-year. The sequential and annual increases in operating expenses are primarily related to higher overall R&D costs and labor expenses.

Our operating expense percentage for the quarter was 11.8%, increasing 0.2 percentage points sequentially and down 1.2 percentage points annually. The sequential increase was primarily due to relatively higher R&D labor costs. The annual decline was primarily due to higher revenues generating a higher operating leverage during the quarter.

During the first quarter, operating profit was NTD 15.9 billion representing a sequential 1% decline of NTD 0.2 billion and a 90% annual increase of NTD 7.5 billion. Operating margin was 14.1% down, 0.6 percentage points sequentially, while up 4.5 percentage points year-over-year.

On Page 6, you'll find a graphical representation of our ATM P&L. The chart highlights the improvement in our gross profit margin. It should be noted here that our second and third quarter 2025 margins were heavily impacted by NT dollar strengthening.

On Page 7 is our ATM revenue by the 3C market segments. LEAP services are primarily used within our Computing applications, with a lesser amount being used in the Communications applications for infrastructure hardware. As can be seen here, the Computing application percentage has been growing steadily with our Communications segment declining. Our Automotive, Consumer and Others application appears to be consistently growing in line with our overall ATM growth.

On Page 8, you will find our ATM revenue by service type. We did not see any meaningful changes here during the quarter. On Page 9, you can see the first quarter results of our EMS business. During the quarter, EMS revenues were down 10% sequentially and 1% annually to NTD 61.9 billion. Sequentially, our EMS business's gross margin increased by 0.5 percentage points to 9.5%. This change was principally the result of product mix.

Operating expenses within our EMS business decreased by NTD 0.3 billion sequentially and increased by NTD 0.1 billion annually. The sequential decline is primarily the result of lower profit sharing during the quarter. The slight increase annually is primarily attributable to higher R&D headcount.

Our first quarter EMS operating expense percentage of 6.4% was up 0.2 percentage points sequentially and 0.1 percentage point annually. The sequential and annual operating expense percentage increases are due to underlying product revenue seasonality relative to more stable operating expenses.

Operating margin for the quarter was 3.1%, up 0.3 percentage points sequentially and 0.5 percentage points year-over-year. The higher operating margins are primarily the result of product mix. Our EMS first quarter operating profit was NTD 1.9 billion, down NTD 0.1 billion sequentially and up NTD 0.3 billion annually.

On the bottom of the page, you will find a graphical representation of our EMS revenue by application. The Communications applications decline is primarily due to underlying product seasonality. The Computing applications increase is primarily due to a pickup in new AI accelerator products.

On Page 10, you will find key line items from our balance sheet. At the end of the year, we had cash, cash equivalents and current financial assets of NTD 114 billion. Our total interest-bearing debt decreased by NTD 7.6 billion to NTD 265.3 billion. Total unused credit lines amounted to NTD 419.4 billion. Our EBITDA for the quarter was NTD 38.2 billion. Our net debt to equity this quarter was 40%.

On Page 11, you will find our equipment capital expenditures relative to our EBITDA. Machinery and equipment capital expenditures for the first quarter in USD totaled $1 billion, of which $636 million was used in packaging operations, $327 million in testing operations, $40 million in EMS operations and $1 million in interconnect materials operations and others.

In addition to spending on machinery and equipment, during the quarter, we also spent $771 million on facilities.

With that, I'll hand the presentation over to Joseph to present the company's outlook.

Thank you, Ken. Before we get into the guidance for second quarter, I would like to give you a bit of a color for the full year.

Now first of all, we are upping our CapEx for the year, which includes additional [ NTD 0.9 billion ] for buildings and infrastructure, as reflected in our recent announcements and an incremental USD 0.6 billion in machinery, driven by stronger demand for LEAP services this year and next. The majority of this additional machinery CapEx will be allocated to LEAP, particularly wafer sort, and expect it to be deployed in the fourth quarter to support capacity ramp-up in 2027.

For ATM 2026 revenue, we now expect LEAP services revenue to be around 10% above our prior guidance, reaching over USD 3.5 billion, while the mainstream segment remains on track to grow at a similar rate with last year. And for 2027, we continue to see strong LEAP business momentum and expect even stronger incremental revenue growth than this year.

Lastly, on ATM profitability, our strong market position continues to support a favorable pricing environment throughout the year. We reported first quarter ATM gross margin of 26%, which is ahead of our original expectation of 24.5%. As we continue to expect sequential improvement for ATM margins with second half gross margin to reach the upper end of our structural gross margin range, second quarter gross margin improvement, however, will be partly offset by higher costs associated with early resource deployment for product transitions.

Now with that, let me give you the second quarter outlook. Based on our current business outlook and exchange rate assumption of USD 1 to NTD 31.8, management projects overall performance for the second quarter of 2026 to be as follows.

At the consolidated level, in NT dollar terms, our consolidated second quarter revenue should grow by 7% to 9% quarter-over-quarter. Our consolidated second quarter gross margin should increase by 20 to 100 basis points quarter-over-quarter. Our consolidated second quarter operating margin should increase by 50 to 120 basis points quarter-over-quarter.

For ATM, in NT dollar terms, our ATM second quarter revenue should grow by 9% to 11% quarter-over-quarter. And our second quarter gross margin should be between 26% to 27%. For EMS, in NT dollar terms, our EMS second quarter revenue should grow at least 10% year-over-year. Our EMS operating margin should be similar with second quarter 2025 levels.

With that, I'll open the floor for questions. Thank you.

[Operator Instructions]

We have our first question from Mr. Gokul Hariharan of JPMorgan.

Great results. Congratulations on that. Thanks for the opportunity, Joseph and Ken. First question is on LEAP. So the 10% upside in LEAP in 2026, could you talk a little bit about what is the reason for that? Is it mainly coming from on substrate and wafer sort or any composition?

And also just to clarify, '27, you said LEAP could grow at similar kind of stronger incremental revenue growth than this year. So this year is around NTD 2 billion incremental revenue based on the new guidance. So are we implying that next year, LEAP can grow at a NTD 2 billion plus kind of incremental revenue? Is that how we should read that statement? And any early read on what is the composition of LEAP next year?

Gokul is looking for clarification on our LEAP guidance for CapEx -- and actually, just primarily on revenue side, right?

Yes, revenue. Yes. This year and the 10% upside where it's coming from and '27, like the NTD 2 billion or more than NTD 2 billion incremental revenue. I just want to clarify if that's what you meant.

For this year, I think we are seeing a stronger than anticipated demand, particularly in the LEAP's part of the business. And as such, we need to increase our CapEx to support that expansion. And also a lot of the increase of CapEx is really to prepare ourselves for next year's ramp up as well.

In terms of momentum, I think both for assembly -- LEAP services, both in terms of assembly and tests, we are growing at pretty much the same pace. And in terms of combination, I think it's about 75% in assembly and 25% in tests. With tests, we have about 75% in wafer sort and 25% in final test.

Understood. And on '27, could we talk a little bit about, I think, Joseph, you mentioned incremental growth can be higher than this year. So does that mean that we can grow from the NTD 3.5 billion this year to at least NTD 5.5 billion next year, is that what you implied?

I'm only implying that we are continuing to see very, very strong momentum going forward into 2027. And the NTD 2 billion or NTD 1.9 billion is really just to serve as a baseline. In fact, we are -- we can safely expect even stronger momentum than that, although I'm not giving out any particular number at this point.

Understood. And just within that 2027 LEAP, any context on the full-process CoWoS like or full process for [ CoWoS ] kind of composition? And are you getting -- do you think that you will have some accelerator related -- AI accelerator-related full process revenue in '27? Or is it going to be mostly like the CPU and some of the other adjacent kind of full process revenue?

I think right now, the full process is going on track, and we still believe that we can reach this NTD 300 million revenue mark for this year. As we continue to expand that part of the capacity, I think, for next year, we can see quite substantial growth in that area as well.

And also not just on the revenue growth, we are also expanding our customer base in this particular area. And we believe that any other application that will be moving into CoWoS like kind of packages, I think we will be able to entertain those part of the request.

Understood. My second question is on gross margins, Joseph. You're already reaching probably the higher end of the structural range of gross margin, and you're already like at the midpoint in Q1 and Q2.

Why not raise the structural range of gross margin, given you have higher-margin products coming in next year as well with full process and more testing as well? Like is there anything that is holding you back from kind of raising the gross margin range, structural gross margin range?

So Gokul, you're looking for more clarification on the context of our structural margins?

Yes.

Well, I think things are moving very fast at this point, and it's very dynamic. And I think we will -- of course, we already said that even in first quarter and second quarter, we are ahead of our original expectation in terms of margin improvement. And we are even more confident that in the second half, we'll be reaching the upper end of the structural margin.

But I think at this point, because things are moving very fast, I think we will rather wait until things are more settled. Maybe next year, we will -- we view the situation and see if this is justifiable to revise up our structural margin.

Is there any risk factor that holds you back? Is it just the depreciation growth is very strong? Or is there anything else that is holding you back?

Well, I wouldn't call it holding us back. I would say that because as Ken is mentioning, right now, the investment or the CapEx that we put in are in pretty large chunks rather than -- for we can incrementally add capacity.

So -- and before the capacity is fully ramped up, it's kind of -- that normally creates a some of -- some pressure on the margin during the ramp-up stage. So it's -- we need to wait until a more steady state kind of situation before we can find the suitable range for margin.

Our next question is from Ms. Sunny Lin of UBS.

Congrats on the very strong results. So my first question is actually to get your thoughts on CPO. And so would you be able to share a bit more color on what will be the role that ASE can play for CPOs packaging and test? And how should we think about the revenue opportunity [ ASE ] from CPO? And when would you expect the sales to start to contribute to ASE from CPO?

So Sunny's first question refers to the context of CPO and the various financial aspects of that, right?

Yes.

We are working very closely with the foundry as well as our customer, the end customer in terms of the development of CPO. Right now, we don't have a number for it. But what I can say is once it gets into volume scale, we would definitely be a very, very critical partner within the overall effort. And I think things are moving on track, and we continue to make progress on this CPO development.

Got it. So may I follow up? And so one, what will be the timeline that you start to expect revenue contribution? And then secondly, in terms of packaging and test, what may be the role that [ ASE ] may play if you could share a bit more details?

No, I don't have that much of a detail, except we will start with the packaging part of it and eventually -- because the test is more complicated than the regular chips that we are testing now. So I think things -- we'll have to sync up our technology road map with our partners and also our customers to find the most suitable work combination between the three of us. And whatever happens, it will be a natural division of works for each of us to do what we do best.

And my second question is on CapEx. And so now with the higher CapEx for this year, capital intensity should be above 30%. And so how should we think about the capital intensity for the coming maybe 1 to 2 years? Should we expect given a very strong demand, you need to continue to accelerate capacity expansion and therefore, capital intensity will remain high at about 30%? Or should we think about maybe in 2027, given the revenue generation, maybe CapEx will start to slow in terms of intensity? How should we think about the growth between the sales and CapEx?

Sunny, you're looking for clarification on our CapEx trend going forward and the capital intensity involved, right?

Yes.

Well, we're still in this megatrend, and we're certainly not going to be shy about making the necessary investments, not only just on the capacity itself but also on the R&D spending that we need to put in. And from the momentum that we're seeing now, I would expect a pretty heavy CapEx for this and maybe even going into next year.

Although, as I said, things are moving very fast and there's a lot of variables in front of us, I cannot give you a number at this point. But I will say that we will be making the necessary investments, and we do have multiple cost-effective funding sources that we can leverage on to continue to support the CapEx requirement that we'll be facing for this year, also next year.

Thank you. So if I may have a follow-up. And so earlier, you mentioned lots of capacity expansions [ throughout ] this year for the output from Q4. So would it be fair to say, therefore, [ LEAP ], we should expect a pretty meaningful step up in terms of revenue going to Q4, and that should meaningfully improve your gross margin on a sequential basis?

So while the direction is upward in the coming few quarters, but in terms of magnitude, should we expect a pretty meaningful expansion from Q3 into Q4, given a lot stronger LEAP growth?

Sunny, first of all, I don't see how that's a follow-up. But I guess I'll let you ask it. For -- you're looking for basically what type of trend we're looking at in terms of LEAP, given that we're upping the outlook currently? Is that probably the right summary of that?

Yes, sure. Thank you, Ken.

Yes. I think from the get go, I was saying that we're still expecting a very strong year going into 2027 as well, particularly in the LEAP services, we see -- we do see the need to up our CapEx to support that momentum.

And I think I already answered Gokul's questions that we are -- right now, we're expecting even stronger incremental revenue growth for next year in terms of LEAP. And whatever that we're seeing this year, the incremental revenue will serve as a baseline for us. And we do see further potential for further upside.

Next, we have Ms. Laura Chen of Citi online.

My question is also regarding our LEAP advanced packaging progress. We know that [ NC ] also build up your own full process. So I'm just wondering that for our like NTD 3.2 billion and even stronger growth into next year, how should we think about that, the full process of our LEAP business? And what would be the gross margin profile comparing to the substrate part?

So Laura, you're looking for more expansion on our full process business opportunities. Is that correct?

Yes. Yes. And what's the percentage? And also what's the gross margin profile?

I think we -- like I said, we are on track at whatever capacity that we're putting in and the customer engagement that we are having at this point. We do see very good potential going forward as we have multiple customers requesting for this kind of capacity. But this is a very complex process that we need to go through, and we are in the midst of trying to tune up capacity as well as the yield that we can generate.

So we're going to take one step at a time at this point in time, although we do see a very good potential, but we want to be cautious in making the -- making this part of the business with a very good execution. And so at this point, I think it's not very easy for us to say, to give you a number for the business for next year. But all I can say is we do see very good potential and it's just a matter of how we can quickly execute the expansion plan.

Sure. That's fair. Very clear. And also my following question is just wondering that as we see that many of your clients or the AI chip makers, they are looking for various, different back-end support, including the panel base or a large -- like radical support for -- from the OS, not just at the foundry side. So I'm just wondering that how is the progress at ASE on the panel base? And how would you think about these business opportunities in your LEAP revenue?

Laura, you're looking for our opportunities in terms of supporting large [ reticle ] size manufacturing in [ panel ], right?

Yes, the panel based like the reticle that kind of packaging design.

Yes. On panel, I think we're also on track. And right now, we have already installed a fully automated pilot line for customer qualification. We do expect to have -- starting with small -- mass production volume starting from next year. And we'll see how it goes from there and that point on.

And I think in terms of the panel, right now, is still at its early stage. It is something that's going to happen. But it's -- we're still at the early stage of the overall development and the -- how the industry infrastructure will be built remains to be -- there's still some work to be done going forward. Yes, there is potential for that. And if anything else, we'll be the first to start the capacity ramp-up.

Charlie Chan of Morgan Stanley will be the next to ask questions.

Also congratulations for very strong results and guidance. My first question is actually the AI or so-called [ broader semi ]. So over the past 3 months, do you see smartphone consumer semi demand bottoming out or it continue to deteriorate? Because we are getting a bit confused why some mature node foundry are hiking wafer price. And [indiscernible] instrument also suggest that they want to hike the price, but overall end-market demand are so weak. So I just want to get some thoughts from management.

Charlie, you're looking for some characterization of our non-LEAP related revenues and how we view that going forward, right?

Yes, yes, exactly. And also, the tendency, whether it's getting better or getting worse and how ASE kind of benefit or suffer from those incremental trends?

I mentioned earlier on that we are maintaining -- I think the general market revenue growth, if we will maintaining the same guidance as we made last quarter, which is a repeat of last year's growth rate of about 13%. And yes, we do see that the PC and cellphone market softness continues, and it seems to be softening a bit more.

But on the other hand, I think a lot of the softness is being picked up by the more IC content in devices, different devices. We are seeing the AI peripheral chips emerging, and we are seeing good recoveries in terms of Automotive and Industrial segment. So put everything together, I think we're still very, very confident that we will repeat our last year's growth in terms of general market.

I see. So can I clarify it with Joseph, what did you mean by the AI peripheral chips? We do have some stance, but I just want to get some clarification from Joseph?

They are power-related connectivity, sensors, all different kinds of edge devices that we'll be using or the -- which we don't -- we were dividing this general market with LEAP. So it's not in industry segment per se, kind of a differentiation or segmentation. We're just only looking at the process itself to decide what's the general market for us.

I see. Thanks for the clarification. And my second question is actually about your pricing strategy. I mean, you're investing heavily. But there is also some cost increase from the recent geopolitical events, right? So I just wanted to get a sense, what's your prediction for your fab utilization in the coming quarters? And would you -- that circle [ tightenings ] into further pricing power to your customers?

So I believe, Charlie, you're asking what the impact of recent political volatility would be on our pricing and then also different components that go into what we do, what our pricing strategy is relative to our services that we provide. Is that correct?

Yes. So yes, thanks, Ken. Yes, so maybe it's more like from a strategic perspective. One is that you can just passively pass through those additional costs. Like you said, right, there's increase in energy costs or chemical costs. I'm not sure how much of a chemical cost you have an exposure to. Or I think it differently, right? Your fab in the coming quarters if the utilization continue to be getting higher and you feel like is sustainable. So why don't you further kind of hike the price and it's more like proactive price hike?

Well, I can never give you a direct answer on whether we're going to raise our prices or not. What I can say is we will continue to seek for the suitable pricing strategy, given the situation and also considering the customer relationships that we need to maintain. Also, we will set the pricing to reflect the margin requirement that we will have for ourselves.

I see.

Whatever the cost increase there is because of the market uncertainties, those can be fully passed through. And also, at the same time, we will continue to set the pricing to reach our margin requirement.

Next, we have Haas Liu of BofA.

Congrats on the solid results and guidance. I guess my first question is also a follow-up on the near-term guidance. Just want to get more clarity regarding your view on what is going to be the key driver for second quarter sequential growth for your IC ATM business? Would you be able to quantify it between LEAP versus traditional business on the growth momentum?

And then I think a quick follow-up also on the full-year outlook for your overall IC ATM business that you did provide more detail on the non-LEAP business and the reason why you are still holding that kind of growth guidance similar to last year. But would you be able to provide the breakdown on the unit and also pricing assumptions for that kind of growth outlook?

I was just wondering if in the past 3 months, if there's any like meaningful change on the mix of assumptions on the pricing that you may be able to improve your pricing further or you simply have not yet seen your customers have been cutting their orders at this stage. So I just want to get more clarity on the second quarter and also full-year IC ATM growth momentum between LEAP versus non-LEAP business.

Haas, I believe you're looking for clarity between -- on our guidance relative to LEAP and non-LEAP context. Is that correct?

Yes, on a quarterly basis as well as on the full-year outlook basis.

I think second quarter revenue growth is really broad-based. I think both LEAP as well as the general market, we're seeing similar kind of a growth to support the quarterly growth of revenue for us.

I think the -- as I said, the -- we are more than doubling our LEAP revenue this year. We just upped our LEAP revenue by another 10%, showing the strong momentum that we are seeing at this point. General market, we continue to see same kind of growth rate from last year.

And as I tried to explain, I think although we're seeing some softness in some certain segments, we do see that the so-called AI peripheral or AI adjacent type of chips that are being brought on stream, and then we are seeing the demand for those chips are more than offset what we're seeing the softness in these particular segments.

Okay. I think just a quick follow-up here is that to my first question is that so you are actually not changing your pricing strategy in the past 3 months or in the next few quarters to reflect the potentially higher cost on the materials or the labors. And you pretty much -- on an aggregate basis, you are not seeing much of the customers' demand dynamic change in the past 3 months. Should I just make that kind of assumption for your full-year outlook?

I'm not sure I quite get your question. Can you...

I was just trying to get more clarity that you simply are not seeing any of the order cuts from your customers, even though that the end-market demand, including like smartphone, PC at this stage. And also, the overall end-market demand is just tracking slower.

But you are seeing incrementally automotive industrial-related demand. And also, some AI peripheral chipsets are making up some of the weakness in the other parts of the market. And so net-net, you are not seeing much of the unit demand weakness versus 3 months ago for your full-year outlook. Is that correct?

Yes, yes.

Okay. Okay. Yes. And then my second question is on CapEx and also gross margins. Joseph, so you just mentioned that on the CapEx for this year, you raised it by like 20% versus your original guidance. And I understand there is a lead time between your [ pulling ] equipment and build clean [ room ] versus it starts to contribute to revenue.

But would you be able to help us to triangulate that growth rate or the incremental 20% hike versus the 10% hike in your LEAP business for this year? What is the main reason why that you raised your CapEx by 20%, but in the meantime, your LEAP business in general, only being raised by like 10%?

And on top of that, I was just wondering with the current supply tightness in the back-end equipment supply chain, do you think there is any further potential for you to raise CapEx throughout this year, if you would like to have more capacity being installed throughout the next few years? And I will have a follow-up on gross margins.

Haas, you're looking for commentary in terms of how our capital equipment expenditures are correlated to our leading-edge advanced packaging revenue outlook, correct?

That's right. Thanks, Ken.

First of all, 2/3 of the increase of our CapEx is for building and facilities. As we've been saying that the suitable new factory or facility is one of the one of the gating factors for us to expand. And in that area, we are making progresses, as shown in our past announcements that we've managed to start some of the building constructions and also to acquire some of the existing facilities that are suitable for our operation or for our LEAP services business operation.

So -- and for the machinery, it's largely for next year's ramp-up and the bulk of it or -- NTD 600 million worth of the CapEx is really for wafer sort. That will be -- the capacity will be ramped up in next year. As for this year, the -- whatever the capacity that we're building or the CapEx that we're spending for machinery should be sufficient to support another 10% growth in our LEAP service.

Got it. And then I think just a quick follow-up on this part of the...

I think you got your follow-up earlier.

Sorry about that.

Next, we have Mr. Bruce Lu of Goldman Sachs.

Let me ask a simple question to save job for Ken. Can you give us an update for the global capacity investment, global capacity expansion?

Global capacity expansion, you mean outside of Taiwan?

Yes. U.S., Southeast Asia, anywhere, any places.

So you're looking for a geographical differentiation between our capital expenditures? Is that what you're looking?

No, no, no. I'm looking for like is there any incremental update for your investment outside of Taiwan. Your current investment is mostly highly concentrated in Taiwan. You have limited capacity, you have limited clean room. Why not go outside? And I asked 2 quarters ago, and I'm going to ask again.

I think outside of Taiwan, I think the main investment that we're making today is really in Malaysia, where we announced that we've acquired a factory from ADI. And that will be used as our buffer capacity to serve the demand coming from outside of Taiwan.

And in other areas, I think the -- we are making mostly maintenance kind of CapEx to support the operation there. However, aside from Malaysia, we are also -- I think we are also ramping up quite a bit in Singapore for our test operation, which is mostly to serve the AI-related test requirement.

I see. I want to ask the next question, is that can you give us some update for the profitability for the -- I think you asked that before, but I want to ask a bit different. For the full process of your [ LEAP ], right, because the CapEx requirement is high, the utilization might be low, may be [ unknown ], right, at the current stage.

But assuming the blue sky scenario, once you eventually will deliver the good yield with reasonable [ cash ] rate, can that -- can the full process profitability be higher than [ LEAP ] services gross margin or higher than the corporate average?

Bruce, you're looking for our profitability potential of our full process leading edge advanced packaging lines going forward in the case that everything goes perfectly. Is that correct?

Yes, which eventually will deliver.

Well, we haven't reached the blue sky yet. So it's kind of difficult to answer that question. But in theory, it should be accretive to LEAP.

[ Accretive ] to the LEAP.

In theory, yes, given the complexity of it.

Next, we have Mr. Rick Hsu of Daiwa.

Okay. I just got one question here. I think, for your very strong demand for your IC ATM, the driver behind that, apart from AI and [ E&I ] -- sorry, Automotive, Industrial, do you guys see any order pull in from your customers across the board because of a concern on the supply chain uncertainties?

And if so, do you -- are you concerned about any demand rebalancing or order rebalance in the second half? Because apparently, like [ Charlie ] said earlier, right, there is a disconnect between the end demand, non-AI end demand and the wafer and chip loading. That is my only question.

So Rick, you're looking to see whether we see order pull-ins and then how we may respond to such situations. Is that correct?

Exactly, especially for second half.

Yes. As I said, in certain segments, we are seeing softness, but those are more than offset -- being offset by the AI peripheral chips demand that we are getting. Yes, I think in the first quarter, we are seeing stronger than normal seasonality demand may be coming from pull-in.

But having said that, we do -- we don't have -- really have the capacity because our capacity is really limited. So we don't have the capacity to entertain those pull-in demand, so to speak. So I think our revenue is very solid. It's not because of -- particularly in the first quarter growth. I don't think it's because of the customers' early pull-in that really bumped up our revenue for the quarter.

Next, we have Mr. Gokul Hariharan of JPMorgan coming back for the second round Q&A.

One question on CapEx, Joseph. Depending upon what you're hearing from your customers and your foundry partner, could you talk a little bit about -- do you feel that this year is a peak of CapEx? Or you think that you will have to keep spending on CapEx even into next year, given it feels like we're going to be a little bit free cash flow negative this year?

So I just wanted to understand, how you're thinking about CapEx into next year as well? And if CapEx is growing next year, any thoughts on how we are looking to kind of fund that CapEx into next year?

By the way it's going, I think next year could be another CapEx heavy year for us.

Okay. So we should assume that CapEx keeps growing next year then, yes?

It's more likely, yes.

Got it. And any thoughts on the funding, given the free cash flow is kind of probably kind of flattish or close to zero or negative this year given the heavy CapEx? Like is there any like new fundraising or something like that, that we need to think about?

Yes. Whatever the funding gap that we're going to have for this and next year, I'm not sure about next year, though. But for this year, I think the funding gap will mostly be funded by additional borrowing, which we have multiple sources for that.

Okay. Understood. Maybe one more question from my side is on the testing business. I think we did talk about final test as potentially something that we wanted to win with a large GPU customer. Is that still the objective right now? Or you are too busy with too many wafer sort kind of projects on hand to kind of really go and support that kind of demand with the heavy capacity on final test?

Yes. I think the -- primarily, I think we will be focusing our resources in wafer sort at this point. I think we are tight in both facility as well as in capacity. So for the time being, I think the main focus will still be on wafer sort.

Next, we have Haas Liu of BofA to ask questions.

I was not able to fully ask my questions in the first round. I think just on the back-end equipment, is -- the supply right now is getting tighter and the lead time for delivery is also getting longer. Do you think there's like further potential or possibility that you need to raise your CapEx again just to show your equipment supply chain more commitment or to ensure that you could expand the capacity as scheduled throughout this year, if needed, that you need to raise your CapEx again this year?

Haas, you're looking for characterization on the likelihood of our ability to spend more on CapEx. Is that...

Yes, that's right. This is the straight way to ask.

Haas is looking to understand whether we have any potential to increase our capital expenditures during this year.

I wouldn't say no. I think the -- like I said, the things are moving very fast, and there are a lot of -- our forecast is -- has been adjusted pretty rapidly at this point. And I wouldn't preclude that there will be a further need for us to increase our CapEx for the year.

Okay. Yes. And my second question before I jump back to the queue is that I was wondering if you could quantify the impact of the new technology ramp transition-related costs impacting your first quarter, second quarter or even like second half this year as you have more capacity gradually on the line, but not yet entering into mass production at this stage, contributing to your revenue.

Would you be able to quantify the potential negative impact to your gross margins? Because we already see some of the mildly negative impact in first quarter and also your second quarter outlook. And should we actually expect more of that negative dilution to happen throughout the second half of this year before your full process or even other parts of the LEAP business capacity ramp is going to contribute to your revenue growth in the following year? Is that kind of a factor -- if you have any way to quantify any of the potential negative dilution throughout the next few quarters, would be pretty helpful.

Haas is looking to understand whether we have previously included the incremental depreciation into our original gross margin outlook for the year. Is that correct?

Yes.

Yes. I think when we look at the margin trend, of course, we will consider all factors, including the CapEx that we're going to make, including the timing of the revenue that will be generated.

I think the margin profile that we presented in the beginning of the year remains the same that first half of the year -- all quarters will be within our structural margin range, except first quarter, we are already ahead of our original plan. And second quarter, we remain to be very confident that we will be reaching the upper end of the margin range. Yes, everything is included. And so far, we haven't changed our view on our margin profile for the year.

Got it. So is there any way you could quantify probably like 50 basis points or 100 basis point impact to your first quarter and second quarter margins? And then into second half, if there's similar like 50 to 100 basis point negative impact from that margin dilution for the capacity ramp. Is it like a reasonable assumption?

Haas, I think we haven't been able to necessarily release that as of yet, but there is expansion into the depreciation percentage. We have not been able to create an immediate dilution effect at this point. Maybe we can have that for you next quarter.

[Operator Instructions] Next, we have Ms. Laura Chen of Citigroup to ask questions.

Just a quick follow-up. I think just briefly talk about the CPU opportunities on the packaging and also the testing front. Although it's still in relatively early stage, so can you share with us what would you see that the most difficult part or potential technology bottleneck and what ASE now is working? And when would you see that will be kind of an indicator we can see better breakthrough?

Laura, you're looking for an understanding as to where we see a technology bottleneck going forward?

Yes. And what would be like a key [ watching ] point [Audio Gap] we can judge that when it will be becoming a stronger revenue catalyst at ASE?

Well, as I said, this is a team effort between us and our foundry partner as well as our customer. I think at the end of the day, each will be doing what they do best, and there will be a natural division of works. I can't tell you what is hard, what is not. It really depends on what our respective expertise is and then suitable works will be allocated to different partners.

So that's how the thing is being managed at this point. Like I think everybody understands that CPO is still -- it's a must-have going forward. But at this point, it's still at a very early stage. So I think it's a bit premature to say when and how much is going to be a major revenue or profit contributor to us.

Next, we have a question from Stefan Chen.

This is Stefan from Aletheia Capital. Congratulations for your good result and guidance. So two questions from me. The first is regarding the 2 fab acquisition for the [indiscernible] announced, one for steel and one for ASE. I'm just curious, in what time frame do you expect these 2 new acquired facility can start to contribute to the company revenue?

Stefan, welcome to the call. You're asking about our 2 recent acquisitions in regards to buildings and facilities. Is that correct?

Yes, that's correct. Thank you.

First of all, both buildings are going to be for LEAP services. And I think we will start ramping up progressively starting from early next year.

Okay. So a quick follow-up. So if early next year, and I think especially for the second fab, in the news release, you mentioned that is for the packaging. So does that mean the tool installation for these 2 facilities, the CapEx will be in this year or that will be in the next year?

You're looking for the timing of the facilitization and equipment of these 2 facilities?

Yes. And also the impact of the CapEx, whether it's this year or next year.

I think the -- at least for equipment, it will be for next year. And this factory will be -- how should I put this? Because the current facility needs to be vacated before we can start moving in and start installation and before that becomes available, there are also some construction works for facilities and some of the wiring, construction that needs to be done.

So if we're breaking down the CapEx between -- for building facilities or for equipment, I think the -- for the building part of it, it will start earlier. But for equipment, it will be starting from next year once we -- the space becomes available and we progressively put the machineries in.

And also one last quick follow-up is so we discussed about full service and even some -- the new technology panel. But I'm curious if you can give us any color, how do you see co-op right now?

Co-op?

Yes, chip PCB.

That's -- I think it's further away. So I'm not an expert in this. So I don't think I'm qualified to comment on that.

We don't see anyone raising their hands online now.

All right. Thank you very much. See you next quarter.

I think to sum up, we had a good quarter, and we'll continue to have a good year. And next year, we see a very, very strong momentum continuing. And we will do our best to execute our expansion plan to meet the demand from our customers. And we will continue to see our profit return to expand. And come next year when things are more settled, we will be reviewing our structural margin range to make it a more suitable range for us. Thank you.

Hello. I am Ken Hsiang , the Head of Investor Relations for ASE Technology Holdings. Welcome to our fourth quarter and full year 2025 earnings release. I'm joined today by Dr. Tien Wu, our COO; and Joseph Tung, our CFO.

Thank you for attending our earnings release today. Please refer to our safe harbor notice on Page 2. All participants consent to having their voices and questions broadcast via participation in this event. If participants do not consent, please do not ask questions or you may leave the session at this time.

I would like to remind everyone that the presentation that follows may contain forward-looking statements. These forward-looking statements are subject to a high degree of risk, and our actual results may differ materially. For purposes of this presentation, dollar figures are generally stated in new Taiwan dollars unless otherwise indicated. As a Taiwan-based company, our financial information is presented in accordance with Taiwan IFRS. Results presented using Taiwan IFRS may differ materially from results using other accounting standards, including those presented by our subsidiaries.

For today's presentation, Dr. Wu will be delivering the company's keynote. I will be going over the financial results, and Joseph will then provide the company's guidance. We will then be available to take your questions during the Q&A session that follows.

With that, let me hand the presentation over to Dr. Tien Wu.

Good afternoon. I would like to give you a 2, 3 years' outlook for the ASE business. This represents the best perspective that we have as of today from our partner as well as customer.

Let me talk about the megatrend and future opportunities. The AI server cycle continues, primarily led by hyperscaler and the data center development. There's a lot of activity in the physical layers via edge applications. For example, we're seeing more design perspective regarding to the robotics and the drone, also the equipment surrounding the automotive and the smart manufacturing. And I think the volume will gradually show up in the next 2 years. which is what we're looking for. We are seeing last year, the mainstream business recovered. We believe the mainstream, namely the IoT, the automotive, the general sector, the mainstream business will recover better this year comparing to last year.

The second category is what I call the ASE and the Taiwan cluster. I would like to give you 2 backgrounds. Many of you have raised the questions, there seems to be a very fast evolution in technology as was demand. So the question is, how will ASE and the Taiwan cluster react to the fast evolution of technology and manufacturing requirement? The second question is, there seems to be a lot of constraints in substrate and maybe memory. And how would that change our perspective regarding to the manufacturing for our partner?

I will try to answer that using a longer time frame. I think there's no question about the leadership in semiconductor manufacturing in Taiwan for this year as well as a few years down the road. I don't think there's any question about our position, Taiwan as was ASE. We also understand that what is driving the business in AI is mainly the system optimization, which includes chip level optimization, packaging level optimization as well as power delivery, silicon photonics, manufacturing efficiency as well as thermal.

I would like to remind you, Taiwan has manufacturing leadership in all sectors. In other words, not only we exemplify the strength in each sector, there's a lot of cross-collaboration, co-design, co-optimization and co-manufacturing in this era. This is particularly true when there is a supply constraint. The leadership will have a first-mover advantage when there is a supply constraint. In the fast evolution, amidst all uncertainty, customers tend to go for the leader to manufacturing the first product in order to maintain the leadership position. So ASE and Taiwan collectively will have competitive advantages over our competitor in this space.

Many of you ask if ASE is going to ramp up last year, this year, as was potentially 2027 and beyond, how can we manage all of the factory space, CapEx, resource management? It is a difficult question, but ASE is not doing this alone. In Taiwan, let me give you a few examples, there's a lot of technology collaboration with our partner, upstream and downstream. In terms of factory space, yes, we are building factory from scratch. We're also acquiring existing factory from our partner, even with clean rooms already installed. Resource planning, you have to look at the whole cluster. So here, I am particularly grateful to our customer as well as our partner for supporting us, helping us to ramp up.

Now once again, when there is a supply constraint, when there's a fast evolution, where everything is running against time, the Taiwan cluster has demonstrated the best efficiency and speed in terms of manufacturing ramp-up. This will set the stage for many of the conversations that we will have throughout this call.

Lastly, I want to talk about ASE's Taiwan Plus One. Many of you have asked, what is our strategy in terms of Taiwan Plus One? Our objective is to support all of our customers, satisfying their manufacturing requirement on the global footprint. Here, I want to give you a very simple classification from ASE's perspective. In the future 5 to 10 years, there will be wafers out of Taiwan. There will be wafers not coming from Taiwan. For wafers that are coming from Taiwan, ASE has a very good opportunity to do packaging and testing inside of Taiwan. Might not be all true, but that is the assumption.

Now for wafers that are not produced in Taiwan, they might also come to Taiwan, but they might also not come to Taiwan. So ASE is building footprint primarily in Penang, mainly for the automotive and the future potentially robotics, to capture customers and wafers that are not produced in Taiwan, but would like ASE to use our automation as well as all of our advanced technology to help them produce the system package and the system optimization. We're also building footprint in Korea and Philippines. But Penang will be the main sector that we'll be ramping up simply because the Penang cluster has been well established second to Taiwan.

With that, I would like to give you the 2025 recap. The consolidated revenue grew 12% at a holdco level with ATM revenue up 23% led by Leading Edge Advanced Packaging services and Testing business. The LEAP services reached $1.6 billion, accounting for 13% of ATM revenue, up from $0.6 billion in 2024 or 6% of ATM revenue. The general segment grew 13% year-on-year. Testing business grew 36% year-on-year in 2025, supported by expanding turnkey and leading-edge test. Machinery CapEx totaled $3.4 billion. Building facilities automation CapEx was $2.1 billion in 2025, mainly driven by LEAP services and testing investment.

Next page, 2026 outlook. Our CFO will give you more elaboration after this highlight. We expect revenue uptrend to continue 2026 and beyond, driven by leading-edge solutions and broad-based semiconductor demand related to AI proliferation and general market recovery. ATM business, leading-edge assembly packaging service to double from USD 1.6 billion to USD 3.2 billion, with roughly 75% from packaging and 25% from testing. General segment continues to grow at a similar pace as last year. Overall, ATM revenue to outperform the logic semiconductor market.

Stepping up CapEx expenditure that Joseph will talk about numbers with the investments in R&D, human capital, advanced capacity and smart factory infrastructure to support the multiyear growth. So our view is ASE is in a very good position together with our customer as well as Taiwan partners, and we understand the short-term need, we're trying to run against time to fulfill the supply. Long term, we are deploying our floor space outside of Taiwan to capture the next-generation opportunity. Thank you.

Thank you, Tien. For the fourth quarter from a financial perspective, our ATM factory loading was slightly better than originally anticipated. With higher loading, we were able to extract higher operating leverage. Our ATM factories in Taiwan ran at or near full capacity with LEAP and traditional advanced packaging utilization rates outpacing that of wirebond. Non-Taiwan utilization rates continued to show improvement. Our overall ATM utilization rate was around 80%. Our EMS business slowed slightly due to underlying product seasonality. Revenue and profitability was aligned to our initial outlooks.

Please turn to Page 6 where you will find our fourth quarter consolidated results. For the fourth quarter, we recorded fully diluted EPS of $3.24 and basic EPS of $3.37. Consolidated net revenues were $177.9 billion, representing an increase of 6% sequentially and 10% year-over-year. On a U.S. dollar basis, our sales increased by 2% sequentially and 14% year-over-year. Our gross profit was $34.7 billion with a gross margin of 19.5%. Our gross margin improved by 2.4 percentage points sequentially and by 3.1 percentage points year-over-year. This sequential improvement in margin is primarily due to higher loading in our ATM business and NT dollar depreciation. The annual improvement is primarily due to higher factory utilization offset in part by the annual appreciation of the NT dollar. We estimate that foreign exchange had a positive 1.1 percentage point impact to our gross margins sequentially, and while having a negative 1.2 percentage point impact annually.

Our operating expenses increased by $1.4 billion sequentially and $1.6 billion annually to $17 billion. The sequential and annual increases in operating expenses are primarily due to higher R&D labor-related costs. Our operating expense percentage increased 0.3 percentage points sequentially to 9.6% and edged up 0.1 percentage points annually. Operating profit was $17.7 billion, up $4.5 billion sequentially and $6.5 billion year-over-year. Operating margin was 9.9%, up 2.1 percentage points sequentially and up 3 percentage points year-over-year.

During the quarter, we had a net nonoperating gain of $0.6 billion. Our nonoperating gain for the quarter primarily consists of net foreign exchange hedging activities, offset in part by net interest expense of $1.7 billion. Tax expense for the quarter was $3.2 billion. Our effective tax rate for the quarter was 18%. Net income for the quarter was $14.7 billion, representing an increase of $3.8 billion sequentially and $5.4 billion annually. On the bottom of the page, we provide key P&L line items without the inclusion of PPA-related expenses. Excluding PPA expenses, gross margin would be 19.8%, operating margin would be 10.4% and net margin would be 8.7%. Basic EPS, excluding PPA expenses, would be $3.55.

Please refer to Page 7. Here, you will find our 2025 consolidated full year results versus 2024 full year results. Fully diluted EPS for the year was $8.89 while basic EPS was $9.37. For 2025, consolidated net revenues improved 8% as compared with 2024. Our ATM business improved by 20%, while our EMS business declined by 5% annually. Our ATM business was 60% of our consolidated net revenue, up from 54% in 2024. Gross profit for the year was $114.2 billion, improving $17.3 billion year-over-year or by 18%.

In 2025, our consolidated gross margin improved 1.4 percentage points to 17.7% principally as a result of higher ATM revenue mix and higher factory utilization of our ATM equipment, offset in part by appreciating NT dollar and higher utility costs. Operating expenses increased $5.7 billion for the year and came in at $63.4 billion. Our overall operating expense percentage edged up 0.1 percentage points to 9.8%.

As a general trend, we believe our spending in R&D on an absolute dollar level will continue to increase as the technological complexity of services we offer continues to progress. However, as our R&D investments start yielding associated incremental revenues, such as those in our LEAP business, we should see increasing operating leverage. Currently, we see our 2026 ATM operating expense percentage declining by near 100 basis points with our consolidated operating expense percentage dropping 80 basis points. Operating profit for the year was $50.8 billion, increasing $11.6 billion. Operating margin for the year was 7.9%, representing an improvement of 1.3 percentage points from 2024. Our ATM business accounted for 87% of our 2025 operating profit, up from 80% in 2024.

We recorded a net nonoperating gain of $0.5 billion for the year, including a net interest expense of $5.6 billion versus $4.9 billion in 2024. Most of the net nonoperating gain was associated with our foreign currency hedging activities. Total tax expense was $9.5 billion, the effective tax rate for 2025 was 18.4%. We expect our effective tax rate for 2026 to be about 18%.

Net income for the year increased by 25% to $40.7 billion. On a full year basis, we estimate that the appreciating NT dollar had a negative 0.9 percentage point impact to our consolidated gross and operating margins. Removing the effect of PPA depreciation, our gross margin would be 18%, our operating margin would be 8.4%, our basic EPS would be $10.07.

On Page 8 is a graphical presentation of our consolidated quarterly financial performance. Our ATM business driven by expanding LEAP services continues to outgrow our EMS business. Looking into 2026, we continue to expect our ATM business to outgrow our EMS business. As such, we believe that ATM revenues and profitability will continue to become a larger share of our consolidated total and continue to positively impact our consolidated margin structure.

On Page 9 is our ATM P&L. The ATM revenue reported here contains revenues eliminated at the holding company level related to intercompany transactions between our ATM and EMS businesses. For the fourth quarter of 2025, we had record revenues for our ATM business of $109.7 billion, up $9.4 billion from the previous quarter and up $21.3 billion from the same period last year. This represents an increase of 9% sequentially and 24% annually.

Our test business' growth as a whole continues to outpace that of our assembly business. Test revenues grew 13% sequentially and 33% annually. Gross profit for our ATM business was $28.8 billion, up $6.1 billion sequentially and up $8.2 billion year-over-year. Gross profit margin for our ATM business was 26.3%, up 3.7 percentage points sequentially and 3 percentage points year-over-year. The sequential gross margin increase was primarily due to higher equipment utilization, depreciation of the NT dollar and the end of higher summer utility rates.

Meanwhile, the annual gross margin improvement was primarily due to higher equipment utilization, offset in part by the depreciation of the NT dollar. During the fourth quarter, operating expenses were $12.7 billion, up $0.9 billion sequentially and $1.6 billion year-over-year. The sequential and annual increases in operating expenses are primarily related to higher R&D costs and labor expenses.

Our operating expense percentage for the quarter was 11.6%, decreasing 0.2 percentage points sequentially and down 1 percentage point annually. The decline was primarily the result of higher revenues during the quarter. During the fourth quarter, operating profit was $16.1 billion, representing a sequential increase of $5.2 billion and an annual increase of $6.6 billion. Operating margin was 14.7%, up 3.9 percentage points sequentially and up 4 percentage points year-over-year. Without the impact of PPA-related depreciation and amortization, ATM gross profit margin would be 26.7% and operating profit margin would be 15.3%.

On Page 10, we have our ATM full year P&L. During 2025, we continue to see impressive growth of our LEAP-related services, but we also started to see a strong recovery of more traditional services toward the middle of the year. Full year 2025 revenues for our ATM business improved by 19% and with our packaging business up 17% and our test business up nearly twice a packaging at 32%. Gross profit for the year improved 25% to $91.4 billion. Gross margin for the year was 23.5%, up 1 percentage point from 2024.

Margin improvement was the result of higher factory efficiency, offset in part by the impact of the appreciating NT dollar. We estimate that depreciating NT dollar had a negative 1.4 percentage point impact on margins here. Adding that back, gross margin for the year would be well within our structural gross margin range. Our operating expenses increased by $6.1 billion during the year, led primarily by higher labor-related expenses. However, our operating expense percentage decreased by 0.5 percentage points to 12.1%. Operating profit improved $12.1 billion to $44.1 billion while our operating margin improved 1.5 percentage points to 11.3%. Without the impact of PPA-related expenses, gross profit margin would be 24% and operating margin would be 12.1%.

On Page 11, you'll find a graphical representation of our ATM P&L. We believe we have had 2 main drivers for our improvement trend in gross margin, higher utilization of factory equipment and a higher mix of LEAP services and revenues. Looking forward, we expect to continue to see a rising mix of LEAP-related business.

On Page 12 is our ATM revenue by 3C market segments. There aren't many changes here.

On Page 13, you will find our ATM revenue by service type. Here, you can see the 2 service types which pertain to our LEAP services. Bump and flip chip and testing, both are becoming a larger component of our overall business. Traditional advanced packaging with LEAP now accounts for more than half of our overall ATM business. Wirebond now accounts for less than 1/4 of our overall ATM business. Meanwhile, our test business during the fourth quarter reached 19% of ATM.

On Page 14, you can see the fourth quarter results of our EMS business. During the quarter, EMS revenues were flat sequentially at $69 billion, while down 8% year-over-year. The annual decline was the result of differing underlying device seasonality. Sequentially, our EMS business' gross margin declined 0.2 percentage points to 9%. This change was principally the result of product mix. Operating expenses within our EMS business increased by $0.4 billion sequentially and $0.1 billion annually. The increases are primarily the result of a higher head count and fluctuations related to our profit-sharing program.

Our fourth quarter EMS operating expense percentage of 6.2% was up 0.6 percentage points sequentially and annually. The sequential operating expense percentage increase is primarily from increases in compensation due to head count and related bonuses and profit sharing. Operating margin for the fourth quarter was 2.8%, down 0.9 percentage points sequentially and up 0.1 percentage points year-over-year.

Our EMS fourth quarter operating profit was $2 billion, down $0.5 billion sequentially and flat annually. On a full year basis, our EMS operations revenues declined 5%, gross profits for the year declined 3% with gross margin improving 0.1 percentage points to 9.1%. Operating profit declined 5% with operating margin staying flat at 2.9%.

As the electronics industry pivots towards various applications of AI, so will the focus of our EMS business. For the coming year, we'll see our EMS business continued to extend its system capabilities further into AI and AI adjacent applications such as server, optical and power solutions. There are a number of EMS projects in various stages of development that will help position the business for growth this year and beyond.

On Page 15, you will find a graphical representation of our EMS revenue by application. There was a slight shift from consumer devices to computing, automotive and industrial devices. The shifts here are generally due to underlying product seasonality.

On Page 16, you will find key line items from our balance sheet. At the end of the year, we had cash, cash equivalents and current financial assets of $102 billion. Our total interest-bearing debt increased by $22.7 billion to $272.9 billion. Total unused credit lines amounted to $400.6 billion. Our EBITDA for the quarter was $38.3 billion. Our net debt to equity this quarter was 46%.

On Page 17, you will find our equipment capital expenditures relative to our EBITDA. Machinery and equipment capital expenditures for the fourth quarter in U.S. dollars totaled $733 million, of which $485 million was used in packaging operations, $218 million in testing operations, and $28 million in EMS operations and $1 million in interconnect material operations and others.

In addition to spending on machinery and equipment, during the quarter, we also spent $456 million on facilities, which includes land and buildings. For the year 2025, machinery and equipment capital expenditures in U.S. dollars totaled $3.4 billion, of which $2.1 billion was used in packaging operations, $1.1 billion in testing operations, $139 million in EMS operations, $13 million in interconnect materials and others. For the year 2025, we additionally spent $2.1 billion on facilities, which include buildings and land.

With that, I'll hand the presentation over to Joseph to present the company's outlook.

Thank you, Ken. Let me go over the first quarter guidance. For '26 first quarter, we will be seeing a much stronger than normal seasonality for both our ATM as well as EMS businesses. Based on our current business outlook and exchange rate assumption of USD 1 to NTD 31.4 versus last quarter it was about NTD 30.9. Management projects overall first quarter performance in NT dollar terms to be as follows. On a consolidated basis, first quarter revenue should decline only by 5% to 7% quarter-over-quarter. First quarter gross margin should decline by 50 basis points to 100 basis points quarter-over-quarter. Our first quarter operating margin should decline by 100 basis points to 150 basis points quarter-over-quarter.

For ATM business, our ATM first quarter revenue should decline only by low to mid-single-digit percentage quarter-over-quarter. Gross margin should stay in our structural margin range but fall between 24% to 25%. The sequential decline in both revenue and gross margin in first quarter is largely due to less working days and higher labor costs as a result of higher overtime during Lunar New Year holidays. For EMS business, our EMS first quarter revenue and operating margin should be similar with first quarter 2025 levels.

With that, let me give you some color for the full year. For ATM, as Tien mentioned, we expect 2026 leading-edge revenue to at least double compared with last year, while demand continues to significantly exceed supply. As for general market, last year's growth momentum will continue this year given AI proliferation and automotive and industrial sector recovery. On ATM profitability, we're expecting a favorable pricing environment for the year. And as operating leverage continues to improve, we expect ATM gross margins to stay within our structural margin range throughout the year and to improve every quarter, while second half gross margin to reach the upper end of the range.

With increasing mix of lead services and overall testing, expanding scale as well as automation, we are optimistic on our mid- to long-term profitability. Lastly, on CapEx, we will remain aggressive in CapEx spending to support the strong business prospects for 2026 and beyond and to further extend our lead over competition.

This year, we plan to add another USD 1.5 billion in machinery on top of last year's USD 3.4 billion, of which about 2/3 will be for leading-edge services. Also needed investment in buildings and facilities is expected to be at a similar level versus last year's USD 2.1 billion. With that, thank you.

Thank you, Joseph. During the Q&A session that follows, we would appreciate if your questions could be as clear and concise as possible and ask singularly. We will start by taking questions from live participants and then alternating questions from our online participants. I, as the moderator, will be receiving each question and repeating and directing each ask question. After participants' initial question, he or she may ask a follow-up question, clarifications of the earlier question or another question entirely. Then we will move to the next participant. Participants may return to the queue for any additional questions or clarifications. Thank you. Questions? Can we get the microphone over to Sunny here?

Thank you very much. Hopefully, not too late to say Happy New Year. So to kick off, so on your LEAP business, you just guided revenue to double to $3.2 billion for this year. So appreciate a bit more color on maybe breakdown by OS, outsourcing, full process and also test.

So you're looking for a breakdown of our LEAP incremental business for the year.

I think Tien just mentioned that we're at least going to double our LEAP revenue next year -- this year, I'm sorry. And the momentum continues to be strong. I think there is still further upside if we're not constrained by a lot of the capacity build that we're scrambling at today.

In terms of the breakdown, I think we'll predominantly still be in OS and also on test side will be on the wafer sort. And I think the full process, which is going on track, and we do have engagement with multiple customers, but we're going to start seeing the meaningful revenue contribution by later of the year. And we expect to triple our full process revenue this year to reach about 10% of the overall LEAP service revenue.

In terms of final tests, I think we will also be putting a lot of -- most of our focus right now on the wafer sort. In terms of final tests, I think we will start to have a meaningful revenue as well by the later part of the year, and we should have roughly 10% of the business coming from final tests of the wafer -- of the test business coming from final tests.

Very helpful. And so I think one highlight from earnings was very strong margin expansion in Q4, given improving utilization rate and also better product mix. And so how should we think about from here? You just guided the IC ATM gross margin to trend towards the high end, maybe towards like high 20% or even 30%. But how should we think about from there? Meaning LEAP gross margin should be higher than the structural gross margin trend. And I do think your LEAP gross margin will continue to improve, given better scale and also improvement mix. And so how should we think about maybe going to next 2 to 3 years, how high could the gross margin get to?

So Sunny, you're looking for a longer-term guidance on overall margins.

We would like to take 1 year at a time. So we'll talk about this next year. Thank you.

Thank you, Sunny. Let's give the microphone over to Haas there.

This is Bank of America, Haas. Congrats on the good results and also guidance. I think first one is just regarding your mainstream business outlook for this year. You mentioned you will grow at a similar pace compared with last year. And in the near term, you also have that part of the business above seasonal. So just wondering what are you seeing regarding shipment versus pricing benefiting on that above seasonal in the near term and also for full year, the breakdown between shipment and also pricing outlook for that part of the mainstream business?

And then, I think also related on IC ATM, it's just you mentioned that you will exceed at least USD 3.2 billion for the LEAP business for this year. I remember last year, you only mentioned you will reach $1.6 billion. So it doesn't sound like that you're already fully factoring in the potential upside in the LEAP business. Could you also comment on that?

So Haas, you've asked 2 questions here. Going against the rules, huh, very aggressive. So the first question is relating to mainstream business and what we see in the mainstream business in terms of growth. And I'll summarize your second question after this.

The mainstream business has two sources. The first source is from IoT, automotive, industrial. Those are the general sector that we are familiar with. Not surprisingly, part of the general loading does come from the AI data center. For example, the power management, the switch router. So as they're building the data center, we have a difficult time to track which part of the general sector loading comes from the AI data center, which part comes from the general, right? But in general, when we talk to our customers, our general sector loading has been quite decent.

The pricing environment, I would say, friendly. I will not comment on pricing increase or any customer-specific information. The only thing I might comment is, the gold price, the substrate price, whatever the price is going up, then it's up to us and the customer. We do have long-term service agreement. In general, it's a friendly environment.

Should I answer the LEAP? Well, the second question is the, last quarter, we talked about $1.6 billion, most likely will go to $2.6 billion. And this time, we revised to $3.2 billion. And you're asking what happened? Because 3 months later, we have a better visibility about our factory space. I have already said many times, actually, the demand is far beyond the capacity that we're capable of building.

And I remember last year, one of you asked us, if this is so good, why don't you just do it? We are. However, we have to manage the quality, the delivery, all of the resource planning, and these are complicated progress. Processes, equipment lead time and all of the management and also the engineering training, they're not easy. And we just try to do this very, very carefully. So right now, our CFO and myself, our best view is we believe we can achieve $3.2 billion comfortably, right? And over time, if we have any good news, you'll be the first one to know.

Let's see if I can go online.

We have Mr. Gokul Hariharan from JPMorgan.

Can you hear me?

Yes.

We can't hear him.

Can you hear me now?

Yes, you are on the line, but we had some technical issues here. The audience on the floor cannot hear you?

Let's ask Gokul to hold off for 1 second. We can pass -- as you guys fix the technical issue, let's transfer over to Charlie over there.

Ken, let me try to fill the void. And Dr. Wu, Joseph, good afternoon. So my first question actually is about your subsidiary, USI, they announced the acquisition of EugenLight for the CPO repeat component, the optical engine. So I'm wondering what does it mean to the ASE group overall for your CPO business? And going forward, how you're going to split the process between yourself and also USI?

Charlie, you're asking about our fiberoptic acquisition at EMS, EugenLight. Do we have a comment on that?

The optical business is an important direction for the industry. Everybody understands that. At ASE, we're working with our foundry partner as well as the end customer, trying to implement the silicon photonics part of it, which is a CPO. At a system level, it is not a CPO. However, the optical, it needs to go through from the chip to the packaging as well as to the system as well as to the optical transceiver as well as to the rack and beyond.

So I think the optical technology as well as business, it goes very pervasive and very, very long. I think what you mentioned is USI is trying to piece together early deployment in terms of the future optical road map, and this is part of it. And I announced the silicon photonics quite a few years ago. The whole industry is trying to do early deployment and development and positioning. And I think we are going to see some early volume on the silicon and the CPO part of it. And if that goes well, and I believe the volume will start catching up.

So the technology development, the manufacturing capacity development needs to happen well before that. So I think this is the USI and the ASE story. There is no conflict. However, it is important that if we can manage the silicon, the CPO, the packaging, we need to have a good know-how as well as inside knowledge on the system level such that we can support our customers on the overall system optimization if we need to switch hybrid or electrical all the way to optical, right? It's a very long question, I gave you a longer answer.

So it seems like ASE get involved deeply, right, in the CPO supply chain. And we do hear there are some kind of technical challenging, for example, put together, there's a major compute die with those FAU together on a substrate. So I'm wondering is ASE very critical to solve this problem? Or it's more like your foundry partner figure out how to work it and then outsource to you guys with the missed production.

Charlie, you're looking for clarification on the actual processes that we can be involved with on CPO?

It's a very difficult question because at the silicon level, the technology complexity is very, very different. So the foundry partner, they can be working on like a really, really complicated issues, right? And this, you're dealing with at the silicon level, how they convert the light out. So I will not comment on that. So I have partner addressing those. I also have partner addressing more traditional way. If you do the photo or if you do the electrical separately, and those are more existing. So at our level, we have to manage the really complex PIC, EIC together or separately and via different kind of stacking configuration. For example, chip-to-chip, chip-to-packaging, chip-to-rack.

So at the packaging level, I will not say it is easier or more difficult. I think packaging level is easier. However, you have to go much, much deeper because you have to deal with different alternative of the electrical source, the light source and different configuration, different memory and who connects to what, right? At a system level, I will not say it's easier, but you got to go deeper. So the beauty about Taiwan is, we have the chip, we have the CPO, substrate. We also have the system. So in terms of sharing co-optimization trade-off, I believe within this ecosystem, you have a much better chance to hit the first product.

Thanks for your clarification. So my next question is to Joseph.

I think you got your 2 questions there. Sorry. Let's try to go see if we can get to online again. Is it working?

Hello? Can you hear me now?

Yes, I can hear you. But the audience on the floor cannot hear you.

Can you put the microphone to your ear piece.

Can you try again, Gokul?

Is it better now?

Still doesn't work. Can you write down your questions in the Q&A session?

Yes, I'll do.

Do we have further questions over here? Let's go to Rick over here.

Dr. Wu, Joseph and Ken, I just got 1 question here. Regarding your EMS, it seems to me it has been quite muted over the past 2 years. Are you guys strategically downsizing your EMS business only focusing on what adds the most value? The reason why I'm asking is because I remember you used to fund a big chunk of your EMS from consumer, like a SIP, for example, wearable, et cetera. So can you give us more color about your strategy about your EMS going forward?

The EMS business has a few angles, all right? First is your competitive landscape. The competitive landscape inevitably have to deal with the geographical location. And then the second angle will be either the consumer AI or the futuristic sector. And then you will talk about the synergy between the ATM business and the USA business. And I'll try to answer this in a very, very brief way, okay? So what has transpired in the last few years is, we understand our consumer business has ramped to a level that we're comfortable with. But we are facing competition. Also, we're feeling constrained.

At the same time, this AI sector and also the AI emerging general sector, I really don't know how to describe it. In the AI data center, for example, you have the glass, as one example, which is really not a consumer, it's more related to the AI space. Then you have the optical, which optical has been there forever. But because of the data center, we're pushing the system-level optimization to a brand-new level. And I can give you a few examples. The optical is one example, power supply, absolutely is the second angle. You will have optical transformation or upgrade or paradigm shift. You have a power delivery paradigm shift. All of this are hinting us with AI, also the system level optimization.

There are good opportunities we can divert the resource and start more synergistically with the ATM part of the business, and that's called the realignment or recalibration. So what we have hinted, which we will report in the next few quarters is, in the AI space of consumer what are the activity that we're engaging. Also in the AI data center or AI enable or AI motivated system-level optimization what are the efforts that we're making. So I think in the next few years, you will see a much stronger vintage in that direction. So no, we're not trying to downsize, but the market is shifting the geographic positioning is changing, the requirement of customers changing. So we just have to change accordingly.

Just one quick follow-up. So how long would this transition take? So meaning that this year, your EMS will continue to undergrow your ATM? How long this transition would take before your EMS catch up the corporate average?

I think this year, we are growing. Yes, this year, we're -- well, okay. So the question is the, okay, your ATM is growing very fast. I view that as a good news because the sector change, you tend to go to leadership and ASE happened to be a member, a key member of this leadership cluster. And then I think you were seeing some significant input on the EMS part of it. For the last few years, there has been a lot of design in pipeline, and hopefully, we can give you a better view into the revenue contribution. But as soon as this year, it's not the -- yes.

So do we want to give a try?

I can read out your questions.

I think we're trying to use a speaker phone methodology to microphone, very, MacGyver like.

Okay. Third time lucky, I guess.

Perfect.

All right. You could finally hear me.

Great.

Thanks for going through all the trouble. First question, Dr. Wu, given this very strong CapEx continuing in '26, could we understand how much of this is going to full process packaging. And when we look forward, how big is your full process business likely to scale from the 10% of LEAP this year? Can it get to like 30%, 40% of the total revenue eventually in like a 2- to 3-year time frame?

And I also wanted to understand whether it aligns with your partners' future plans? Because right now, there is a bit of a division of labor between you and your partner, but full process kind of a little bit of competition with them, but just wanted to understand how it aligns with their future plans.

Gokul wants to know about our full process plans and such.

Well, as Joseph already talked about, out of the $3.2 billion, we're looking at about 10% full process revenue by the end of this year. There are a few clarifications. We're not competing. And this is part of the cluster ideas, right? Because the wafer level, the foundry partner they will have the first right, the first knowledge and the first need to come up with the different configuration, 2.5D or 3D packaging. The long-term prospect or motivation, how long or how big do they want to grow the business is up to the foundry partner. Our understanding is, if we're fully capable of executing customers as well as partner would like to have second source, right? Therefore, there will be technology sharing not on the OS side, but also on the full process.

Now the configuration is a big question, right? It's complicated. As the HBM and also the TPU size grows, the wafer level, the panel level, they have all different configurations of stacking big panel substrate. So the technology also evolved very, very quickly between the foundry process and know-how, between the customer cost, architectural and design requirements and the packaging. This is where the co-optimization and the collaboration will come in.

Now everybody has the question, why you're spending so much CapEx? What if the customer switch? First, you have to understand it's not a one customer thing. They are mainly many customers. Are they competing? In a way. But in a way, they're not competing. They're trying to fulfill the diversification of the AI space in all kinds of inference, all kinds of learning large language model. It's a very, very big market. This is the beginning. What we are trying to do is as fast as we can come out with alternative and the toolbox. So our partner, the designers will have the total freedom to pick and choose whichever way they want to put this together for whichever application. I'm telling you 90% of the application we couldn't see yet. This is why the AI is so exciting, right?

So with that, I think our CapEx is fine. The partnership collaboration, competition is healthy. All the iteration and the fast evolution of technology from our customer is not a penalty for us, it's actually very, very healthy for the industry. And being the first mover in the Taiwan, first-mover cluster, you will have some natural advantage. So I believe this is the good time for us to do this.

Thanks, Dr. Wu. That's very clear. Second question, we did mention we are expecting the mainstream demand to keep growing similar to 2025, 10-plus percent. Within that, obviously, PC and smartphone related, especially smartphone-related is a pretty big chunk, and some of your larger customers have been turning down on the smartphone demand for the last few weeks or so. So how does this kind of sit within this expectation of mainstream demand continuing to grow at a similar pace? I'm sure you've done your math and come up with this assumption, but can you explain a little bit.

So Gokul is looking for characterization of our outlook on our mainstream market.

The mainstream market, the ASE has a large market share on the communication or the cell phone. Again, it's a good news, right? And thanks to our customer and long-term support. That sector will continue. There could be some fluctuation that I will not go into detail. But please remember, we are fully loaded. The AI data center now has the FPGA, microcontroller, power management, router and all kinds of loading coming to ASE. And you understand that we have signed -- we bought 2 factories for Infineon. We have also announced we are buying another factory in Penang for ADI. We already announced that. That is most likely to close in Q2 this year.

With all of this, the Infineon loading, the ADI loading, they're also coming to us as part of the acquisition, also the co-optimization and the co-design for today's device and system as well as for the next-generation device and system. So when I talk about there's a general sector recovery, I truly mean it. I have seen industrial automotive recovery. On top of that, I have unknown demand, could be AI data center, could be not, from the same guys asking us to run up the loading. And of course, because our fully automated general process, we could be gaining market share. And that we have to look at ASE versus our competitor in all different spaces. I will not comment on that. But in general, we feel comfortable for 2026 and beyond for our general sector recovery.

Do we want to go to who, another online Laura?

Congrats for the great results. I also have questions on the LEAP business. Actually, we see that there's various different type of advanced packaging as Dr. Tien mentioned that. Definitely, we see the great chance as he has seen promising growth with the AI chip involving, but we see that the chips and die sizes are getting more compact, and there are various different type of, like a panel base, even some are talking about the chip-on-wafer-on-PCB. So can you share with us your plans on various packaging type and also your strategic focus? And how should ASE fit and leverage your technology to get into these different type of packaging going forward.

All right. So the question, again, it's a loaded question. Now because of the AI system demand, there are 2 family of thoughts. We believe that the chip will get bigger. Therefore, the wafer size will get bigger. Therefore, the chiplet size will get bigger. And therefore, the 3D, the 2.5D and the HBM will get bigger, which is why you need the silicon photonics, which is why you need a new conceptual power delivery method to provide the vertical power supply. That's one family thought.

The second family thought is because of the efficiency of design, everything will go smaller, okay? I will not get into that debate. All I can tell you is the last 4 years, things just got bigger, right? Therefore, ASE will prepare both. If you can handle this in the 300-millimeter wafer form, we got that already. But if the chiplet size, the requirement are going unreasonably large, we will have the 310x310 panel to give you a better relief.

Now will the 310x310 go to 620x620? Depending on the volume, also the technology requirement. Also, our process capability, can we handle a larger panel size with the kind of resolution and I/O count they require? But ASE's job is by the end of this year, we will have a fully automated 310x310 in production. Today, we already have manual. But in my view, the manual process doesn't count. I can only count in this space, it needs to be fully light out. By the end of this year, we'll have that. That will provide the first try in the throughput and cost improvement on all of the wafer level process that we do.

Now our foundry partner and our customer will continue to drive the different configuration. I cannot comment. Our job is to provide all the toolboxes. So pending on your resolution I/O and cost requirement, we will give you different options. And I'm comfortable to say that Taiwan cluster overall, we have the most toolbox and the best option with capacity in place. We will be the fastest to ramp up in the world. ASE being part of this ecosystem on the packaging world, we are responsible to provide more packaging level toolbox, including panel, including CPO, including the next-generation power delivery for the VRM though we have not talked much about it.

But the whole idea is, we will leverage this opportunity to strengthen our portfolio and the R&D in-depth understanding about the system requirement. It's not just CoWoS. It's not just the full process. It's after this AI boom, the following through system-level paradigm shift improvement, we need to put all of this in place. And this is part of the ASE's vision. We will co-work with foundry as well as our USI EMS arm as well as the other system assembler partner and trying to come up with a total solution from chip all the way to system long term.

Very comprehensive, very exciting. My second question is about the gross margin. As we see that actually these advanced packaging contribution continue going up. And overall, we see that ASP trend is more friendly, as you just mentioned. Can we kind of expect the overall ATM gross margin later this year we'll be able to reach 30% or potentially to be higher?

Laura has a question in terms of the ceiling of our structural margin.

As I explained earlier on, I think, first of all, this year -- starting from second half of last year, we already started to see the improvement in our operating leverage as we continue to improve our ramp-up. And I think for this year, we are confident to say that we will see throughout the whole year, every quarter, we will have our gross margin fall into the structural range. And so in the first quarter, we will see -- well, our margin will be between 24% to 25%. And then sequentially, every quarter, we will see continuous improvement in our margin.

And also by a second half of the year, we will see our gross margin closing in on the upper end of the structural margin. There's further room for improvement as we continue to reach an optimal ramp-up stage and we -- in local term, as we continue to expand our leading edge services, expand our testing business, also continue to reach a full ramp-up, I think there's further room for improvement. But as Tien mentioned, in terms of the margin movement, we will take 1 year at a time and see how the product mix will shift and how the utilization will improve to see whether we need to adjust our margin guidelines.

I'm sorry, I don't know your name. Name and company, please.

Hello Joseph and Dr. Wu. My name is Alan Patterson. I'm with EE Times. You mentioned that you're buying clean room space from foundries. And I would just like to know, is that a first? I've heard from people in the supply chain that this is something that -- advanced packaging is something that's been done primarily by TSMC. So I'm just wondering if this move into advanced packaging, like 2.5, 3D is this something new?

The gentleman would like some clarification on clean room specifications related to advanced packaging.

First of all, I said that we're buying factory with clean rooms already built from partners. I did not specify foundry partners. We have many partners. And also in terms of doing the full process, 2.5D, no, it is not new for ASE. We have been doing this for quite some time.

And then maybe a follow-up question. Is Photonics a new area for you? You had mentioned that this is something that you're moving into. I mean there are many different flavors of photonics. So I wonder if there's any one that you see with greater potential for your company.

You would like an outlook in terms of how we view the CPO market.

Well, Photonics is a new technology. It represents a paradigm shift. And I think there will be a first-mover coming in, and then we will see how the first-mover fares, right? I will not comment, there are different alternatives. And we always believe that different alternatives will serve different solutions for different architectural requirements. So we will not take size in terms of the -- there's a high end, midrange and also low end and different requirements, between the scaled up, the scaled out chip-to-chip level, chip-to-package level or the package-to-package level, they're all different, right?

So again, our job is to develop the toolboxes for the designers, so they can -- they have the freedom to choose based on whatever. But keep in mind, the electronic signal and also the optical signal it is continuous. You got to go through all the way. And our job is to make sure that whatever technology is not limited to any kind of segment. You just need to go all the way down at a different cost of the potential.

We're going to restart the queue here. So Haas, do you want to ask another question?

I think just a quick follow-up on your CapEx because you mentioned you have a lot of amazing demand opportunity, no matter it is in mainstream or in the AI space going forward. But just wondering if you could share with us your view just regarding the capital intensity targets because this year, you definitely grow your CapEx from the amount perspective. But your sales definitely seems to be outgrowing this year based on your guidance just now. So just wondering if you have a guidance or a view on your capital intensity in the longer term, how should we think about the pace that you are going to grow your capacity versus your customers' demand? How would you balance that?

Haas is looking for a financial balance in terms of how we view our capital intensity over time and how that moves and changes.

There is no guideline. CapEx, Joseph and I, we were very comfortable with $2 billion for the longest time. And then for some reason, we just showed up $4 billion. And then we showed up like $5 billion. And then Joseph talks about even bigger number. I'm not comfortable. I think he is not comfortable. We're all under pressure. Long term, there's a little half of me saying that, yes, there are more. Half of me saying, I'm sure we're doing the right thing. Listen, we're a human being. We struggle exactly the same way you struggle.

The only way we can do is, we look at the landscape. AI is brand-new. This is like the beginning of a boom. I mean I won't call Big Bang. This is too religious, right? It's a boom. And then we are in the first-mover leadership position. We have all of the support from everybody and customer. This is our time to shine. And as we move in, we are responsible for the CapEx discipline.

And then there's a financial -- I mean, Joseph is very clever, managing different instruments alone. We're learning all of this. And our guys, we have, I mean, 64,000 people in Taiwan, 100,000 worldwide. It's a good number. You want to stretch them, but you don't want to stretch it to a point that all went to hospital. So as you're doing this and the customer is giving you different voices. I mean, they're a good customer, they're not so good customer. So you're dealing with a lot of this kind of thing.

Can I give you a view about 5 years from now, what's our CapEx? No. One year at a time, we would deal with the margin, execution. If anything, I do not want to disappoint my partner and customer. I will rather deliver whatever we have promised and make everybody happy going forward until the next stop. I will not stretch. Are we aggressive? Yes and no. I don't think we're pushing ourselves to the limit nor should we believe we should, all right? I'm not answering your question because I don't have an answer for you.

No, no, that's great. That's a great comment. And I think just a follow-on question that your LEAP business definitely has been growing quite significantly in the past few years and also this year and probably in the coming years. And as a good leading indicator, CapEx, that you have also been spending quite meaningfully in the past couple of years and also this year, so just wondering if you could share with us your metric regarding the ROIC for that part of the CapEx, especially on the advanced nodes versus your traditional business?

The logic is very simple, right? The CapEx is a leading indicator on technology capacity and PO, also margin. If you spend CapEx, you should come up with -- your depreciation will go up. It should at least cover that part of depreciation. Otherwise, you shouldn't be doing this, right? So if with CapEx, the margin shows improvement, that means you're on the right track. And then you add the CapEx. And not the improvement, you're on the right track, which means that the market supports you and you have a better visibility.

Also, your team becomes more sophisticated, and they're ready to launch the next level of endeavor. We're engineers. We climb stories, floors, one step at a time. So you're asking 2026, we will spend more CapEx. What does that imply to 2027? I already gave you the answer. Hopefully, we can achieve that. But we will not know until third quarter, fourth quarter of 2026 that we can comfortably say, listen, this is what we have come up with. We're making calibration. As we move along, we will give you better visibility of how that calibration is.

I think to sum up your question, I think we're still in this megatrend, and we're certainly not going to be shy on making the necessary investments in terms of CapEx also as well as in R&D dollars that we're going to put in to keep our lead in the -- and we are the chosen partner around this whole ecosystem here. So it's not the time for us to be conservative, I think. So this year, of course, our CapEx is much higher than last year. Last year is much higher than the year before. And we're seeing this trend actually continuing, maybe not just for this year, next year also. We will continue to spend quite large in terms of CapEx and R&D dollars.

But having said that, I think we're still in a very healthy financial condition here, and we do have multiple cost-effective funding sources to support our growth and support our investments. And in terms of return, we are actually seeing that -- first of all, these leading-edge services is margin accretive and certainly return accretive as well. So we're seeing that happening already. And from last year to this, I think both on an ROE or ROIC standpoint, we are seeing improvement, although I'm not giving out any real numbers at this point yet, but we are seeing that our investments are paying off.

Just one more comment. The CapEx dollar and capacity are not the best entry barrier, but it is an entry barrier.

Charlie, do you want to follow up?

So my question to Joseph, was very similar to Laura's question on gross margin, but more focused on those structural factors. So I think in the past, how you talked about utilization, FX, I think it's more short-term cyclical and you talk about the product mix improvement, I think it's more structural. How about pricing? We are hearing that for your wirebond flip chip, et cetera, it seems like there is a price hike this year. Do you think it's more like a structural price hike? And is that ASE specific? Or do you think it's kind of overall ATM industries enjoying this kind of a structural price hike?

Charlie is looking for commentary on the ASP environment.

The price hike is not part of the structure margin. During the COVID days, we have gone through that. Technology is the product mix, the value you provide is. And when Joseph comment, we have not included the price hike, the price hike is a very opportunistic approach and depending on the management philosophy, also in relation with the customer, it will be exercised when needed. But in general, we do not comment on the price increase with customers.

I think our pricing, we will continue to seek the most suitable pricing strategy depending on the situation and also the requirement of -- and our return requirement.

And a follow-up question on the CapEx and clean room part. So I'm wondering whether ASE have some concrete plan to solve the clean room, right? I think TSMC announced they have a piece of new land here and there. So you have so-called circa visibility, right? And if not, whether there is going to be a gating factor for your future spending for business growth?

Charlie is asking about the physical factory progress and development that we are working on, right?

In your CapEx, if you can break down your infrastructure or clean room portion, I think that would be great.

Okay. With separation of machinery and equipment and facilities, CapEx?

Right. This year, I think for the building and facilities, we are still looking at about $2.1 billion, which is about the same level as of last year. Yes, finding new locations and new factories, it's a bit of a challenge. We're doing all we can to look at over the island to find a suitable location for our new buildings. As Tien mentioned, that includes green greenfield factory buildings as well as buying some existing from our partners. So we are going out to find the suitable location for that.

Do we have a follow-up question? I think we have time for 1 more question. If we want to ask for 1 more question. If not, we can wrap it up at this time.

Thank you very much for attending our full year fourth quarter 2025 earnings release. Hope to see you next time. Thank you.

Happy New Year.

Good morning, good evening. Thank you all for attending the conference call with ASE. I'm Sunny Lin covering Greater [ China ] semis at UBS. It's my great honor to host Mr. [ In ] Cheng, Executive VP of ASE for sales and marketing. He will be sharing with us how [ a ] packaging innovations are evolving to support [ call ] technologies. ASE's IR team, Ken Hsiang, Tien Wu, [ Li ] will also be on the line as well to take questions with Ying toward the end of the event. Through the session, if you have any questions, please feel free to e-mail me the questions at [email protected]. So with that, let me hand over to you, Ying, for the presentation.

Thank you, Sunny. Good morning, everyone. Thank you for this opportunity to share ASE's view on advanced packaging and how we are driving AI forward.

So next page. I think this is really given that AI is really here. AI application is changing how we look at health care, telecommunication, retail, financial services. And this will dramatically increase our AI economy from $189 billion in 2023 to over $4.8 trillion in 2033. This is dramatic 25-fold increases. And this is generated by all the data that we are -- as a consumer put together for AI to consume, learn and then inference for all our future AI application.

Next. If you look at how AI is spending, we all know that the AI spending is exploding. In Q2 of 2025, we hit a new high of $87 billion out of only eight major hyperscale builders from Alphabet to Oracle. The dramatic increase in amount of CapEx for revenues also broke 45% in Q2 2025. We expect this trend to continue, and this is great news for our ASE semiconductor exposure.

Next. So if you look at how the data center CapEx and AI semiconductor spend, majority of the spend are in compute, which is where we're going to focus a lot of our talk, followed by networking and memory and followed by power. We also will address a little bit about the power concerns in the AI futures.

Next. So this chart shows a very interesting thing. So if you look at 2020, '21, '22 and even '23, the amount of revenue per device kind of correlate very well with the volume. So as you sell more, the value or the revenue increases. But starting from 2024, that trend dramatically changed that even though we don't sell as much, but the value or the money people wanting to spend on the devices dramatically increases.

So even though the growth for the volume is modest, the amount of revenue we generated are tremendously higher, which means that the value add companies such as ASE can put into an AI system also are being monetized and being valued in the AI equipment market. So this shows a tremendous promise for us going into the future.

Next. So what are the key demand challenges in AI? Well, obviously, performance is the number one requirement. The dramatic increase in compute through the latest large language model push us to create higher network bandwidth in memory and the amount of capacity in HBM that we put on each particular GPUs or accelerated chips. And this creates a problem in area because the number of chip we put down, so if you put from 4 HBM to 8 HBM to 10 HBM to as much as 16 HBM, the area of that package dramatically increases.

So now we are looking at 100 by 100 millimeter square and how do we deal with those large areas. And the next is power. For power, we are now looking at how do we power so many chips within just one blade to now thinking 72 blades together or 72 chip together or up to 154 chip together in the latest NVIDIA requirements. And obviously, if you put in power, the thermal will be the next key consideration for the packaging challenges in the AI era.

Next, one more tab. So this kind of shows you why compute is being driven so quickly. And this kind of shows you the data set and the compute performance by various large language models. And it's growing at 3 to 4x per year. And at the current rate, we're looking at 1.7x in number of chip quantity just for the AI industry. And then we need to improve performance of each chip by almost 1.5x per year, which is significantly faster than Moore's Law, which is a challenge for the chip designer, the chip foundry and -- but this is where advanced packaging can really come into effect. How do we put everything together, achieving this performance requirement without all the benefit of Moore's Law. And this is the benefit of advanced packaging.

Next. So just look at the AI model for this. You kind of see just from April '23 to October '25, which is a little bit over 18 months, the performance has increased almost 50% in terms of how fast the model has improved from OpenAI, from X, AI, from [ alphabet ]. And this is what the compute consumptions are, and this is what the requirement is putting together to the silicon players and to ASE as a company to come up with solution that can provide the compute that's ever increasing and insatable demand of compute power by this open AI or AI model requirements.

Next. So what we talk about with the compute, if we're able to achieve the amount of compute that we see in the previous slides in the various models, what you really require is the amount of data you need to feed those AI accelerators. So what we kind of see, if you look at the HPC, the high-performance computing HBM road map, the number of HBM per each generation continue to grow.

So from the left-hand side, which is the MI300 and 350, we didn't put the 450 on there. But if you look at the medium road map, they will be reaching 16 in Rubin 300. And in the various other models, we're also looking at at least 12 HBM in the coming year. This is we drive the area portion of our challenges as advanced packaging. So if you look at the very bottom, on the NVIDIA Rubin Ultra, the substrate area is 153x77. The interposer size is 124x50. This is a tremendous advanced packaging challenge for us to put this many die in such a small space.

Next. So you look at the HBM integration trend, the reason we are driving to HBM4 is trying to leverage the faster and faster bandwidth and allow us to transmit as many data as possible up to 1.5 terabytes per second. And by putting 16 HBM into 3 chiplet, which is that cartoon I show, this creates a much larger interposer requirement and also require more RDL layers to connect all these chips together. And this is the advanced packaging.

And this actually creates a challenge but also opportunity for system architect or chip architect to create a unique chiplet and memory solution to support the next-generation AI compute requirements.

Next. So the key challenges for large-size modules, what the big challenge is, as it gets bigger and bigger, the number of chip per wafer for 300 wafer drops significantly. So if you go by 100 by 100, there's already only 7 modules or 7 die per wafer and keep dropping as the package gets bigger and bigger, interposer get bigger and bigger. So this is the challenge for us is how do we maintain the yield while reducing the number of chip or interposer per wafer. So we have a solution for that in the coming slides. Next. So with the compute solution, one of the challenges for us is how do we deliver power. Power is very key to the success of the AI chipset. So we all know that when you route copper through the substrate, you have routing losses. And the distance from the VRM to the chip is very important. So how do we reduce the distance and reduce the voltage losses is a key challenge for us.

So for us is how do we put this power solution as close to a silicon as possible. And what are the solutions that we can come up with that can put up a vertical voltage regulation onto the package itself. So this challenge is something that we need to figure out and deliver precise power into this complex chipset and HBM structure.

Next, so if you look at the overall AI compute rack power, what you will see is that if you look at -- in 2020, we only look at 10 kilowatt per rack. But by 2024, you look at Blackwell, it's already 120 kilowatt per rack. And this is really driven by a number of chips within that same rack. And then if you look at what the future holds, we're already seeing 600-kilowatt rack solution and a megawatt rack solution will not be far behind.

So one of the key things for us is, how do we deliver this solution not only to the chip itself, but also find out the power solution to the rack. And I think that with our experiences in some of the high-voltage applications in other industry suit us very well in trying to create a power solution for this higher and higher voltage requirement into the most complex rack AI that will be coming into the marketplace in a very short time.

Next. So with power thermal. So we are looking at how do we do thermal solution in the very increasing watts and voltage environment. So this chart is [indiscernible] of AMD. And what it shows is if you look at the red bar on the left, that is the CPU power. As we go in time, the power gets higher. And then if you do the green bar, that's a GPU power and then GPU power actually gets increasing into 1,500 watts.

But ironically, the higher the watt, the temperature need to be operated actually drops. So if it's lower voltage, you actually can run the chip hotter. But with higher voltage -- higher voltage, you actually need to run the chip cooler. So create even bigger problem for us in trying to run the chip at optimum temperature with an increased power consumption. So this is something that the industry needs to work through and ASE is going to participate in how do we also work on the overall package thermal consumption and thermal requirements.

Okay. Next. So one of the things that we will do for the compute problem that we discussed earlier is we leverage the VIPack that we announced back in 2022. VIPack is a collection of advanced packaging technology from focus POP, package on package to 3D ICs to FOCoS-Bridge to Co-Packaged Optics and to FOCoS SiP and just focus, which is the fan-out chip on substrate.

Next. And what we're really focused right now is two type of VIPack. One is the focus, which is fan-out chip on substrate. This is what's very common today in a lot of the AI solution. And you kind of see some of the cross-section that ASE has done to create the latest chiplet architecture or heterogeneous integration solution that combine whether it's I/O buffer die with graphic accelerators or graphic accelerator with HBM.

And then to the right, FOCoS-Bridge is our next solution or the solution of choice for some of the higher density solution where we are using a silicon bridge between HBM and a graphic accelerator to maximize the I/O count between the connection while minimize the RDL that is needed to route between those 2 die or collection of those die. Maybe there will be 3 GPU with 16HBM. And kind of show you the cross-section between the bridge and a C4 pump, and that show you that our pitch can be as low as 130 microns.

Next. So this just show you the FOCoS Extension Platform, specifically talking about Bridge and they kind of show you the package size that go to 100 by 100 millimeters, and they kind of show you various constructions or various opportunity for the chip architect to create. It kind of show you that we can have maybe a GPU accelerated die with some memory controllers with memory itself or just parallel GPUs with memory, like a device or a collection of chiplets with I/O, SRAM GPU, neural network chips, all connected through FOCoS Bridge solution where the bridge are connected -- are connecting the chip to each other and to the next die next to it. So these are the opportunity that we see with FOCoS Bridge that creates a tremendous amount of creativity for the AI chip of the future.

Next. So one of the challenges for that is as the package gets bigger, the utilization on a 300-millimeter wafer start to drop. And we mentioned earlier, if once you get to 5 or 6 [ variable ] size, the number of those chips per wafer is dropped down to 8, 7, possibly 6 and that's only 57% of the utilization. So we really need to figure out a way how do we increase our utilization. So ASE has been working on a panel solution.

We have demonstrated 300-millimeter panel and also 600-millimeter panel that shows that we can increase the overall utilization from 57% up to 87%. And this dramatically allow us to produce this complex solution in scale. And that is the key for ASE is how do we scale this into as high a volume as possible.

Next. So this shows the actual example of panel focused bridge. This shows basically two chiplets with HBM onto a large panel. So you have two SoC dies and SoC 1. So this creates -- and then we put 10 chiplet with 10 silicon bridge onto that one section. And the middle chart shows you the whole panel, how we put it together. And this is the fan-out construction and it's died by laser direct via solution. So it's -- this is an increase in number of unit per panel versus a wafer on the bridge construction.

Next. So as a panel road map for ASE, we're looking at 310x310 for the HPC and AI for fine pitch for 2 micron and 2 micron line space with bridge and IPT. And then we are looking at large panel for fan-out MCM or some people call it wafer MCM that allow us to do mobile application or edge AI application that doesn't require a fine line space and allow us to do the full fan-out RDL as a substrate. So you actually create a very thin multichip modules with complex RDL underneath. So you kind of show you the 600-millimeter fan-out MCM and also the chip glass and bridge solution that we mentioned earlier that give us the 310 and then moving to 600.

Next. So once we are able to put all the chip onto a panel or wafer or modules, the next thing we really want to look at is how do we put the power solution to it? How do we drive all the chip with the necessary power with the first and second stage regulator. So we have created power SIP that basically allow us to put all of the first and second stage regular directly underneath the substrate. So we are creating the minimum amount of distance between the power source and the silicon itself.

So instead of putting a side-by side, as we show on the right-hand side, where we take a 12 volt down to 1 or 0.8 volts or even just do first stage and second stage, now we are actually putting both of them underneath as a vertical regulated modules, what we call a power SiP that allow us to deliver the power at the most -- closest to the silicon and reduce the overall loss and then achieve a maximum efficiency of the power delivery.

Next. And then this kind of shows you the latest thing for the power solution on data center. So if you look at today on top, if you look at the alternating current from the grid to the data center. And basically, we're dropping it from plus or minus 400 down to maybe 480 or lower or plus or minus 220, so 440 solution onto the data center and driving it at 400-volt direct current.

And one of the things that will be more efficient is convert that directly from plus or minus 400 ASE current directly down to 800 direct current. and then using a solid-state conversion and then drive the whole backbone of data center using higher voltage. And this allow us to create more power efficiency throughout the grid and also creates growth in the overall infrastructure going to the future. And a simpler distribution system and with a fewer point of failure will create a more robust data center solution. And this is also aligned with overall thinking in terms of leveraging what we already learned in some of the other industry that also leverage the 800-volt systems. So ASE is in a prime position to work with customers in developing this 800-volt DC systems.

Next. So this kind of shows you one example of using gallium nitride to our silicon carbide as a primary chip solution and module solution to create a monolithic low-voltage conversion into the silicon itself, but also allow us to drive from 48 volt down to 12 or 6 or 0.71. And this actually creates a better solution, a more solid-state solution instead of going through ASE conversions. And for us is this gives ASE another opportunity to create more value within the data center ecosystems.

Next. So driving power itself through electrons is one challenge. But another way to solve that same challenges is trying to convert the electron into photons. So ASE has put in a tremendous effort in trying to work on full-package optic or basically trying to communicate data transfer through photons. And we believe that is the future of data center, the combination of electron and photons.

Next. So this kind of shows you all the various toolbox that ASE has demonstrated through the passive alignment for fiber attach to creating a cavity through laser direct edge. To chip-on-wafer trying to put electric IC on top of photonic IC through a fan-out PoP solution. You kind of show the various cross-section, and that creates a silicon photonic engine that can be used in part of the CPO solution that will show a little later.

And then give you the various way to put in the laser diodes that provide the laser source for the photonic. And obviously, the submicron accuracy is important for all of the die attach or the chip-on-wafer integration. So these are the various tools that ASE already developed that can help our customers to create the next-generation optical solution for the AI hypercenter.

Next. So there are three key challenges in trying to create optical engine onto a CPO. So one is really to deal with warpage. There's warpage on the optical engine. There's a wage on the substrate, which is organic substrate. And then all these things have an impact the way we do fiber attach, whether we do active alignment or passive alignment, these are the key challenges for us to put together. So you kind of see the large different ring on top of this CPO demonstration that we did for the customer. And these are demonstrating that we could show on the next page.

Next page. So this is the thing that we did for CPO test vehicle, where you kind of see the network IC in the middle with different optical engine, and that allow us to connect eight different optical fiber onto this switch solution. And you kind of see what we did with the [ Ser ] rings to maintain the work on not only the optical engine, but on the substrate itself. And this is really a large package of 75x75 package size, and this is all joined by copper pillars with the Fan-Out PoP solution.

Next. And this kind of shows you just kind of the wealth of solution that we are offering in terms of toolbox, whether it's ASE on chip on substrates, whether it's optical engine, and it kind of show you on the same thing how each of the optical engine and the substrate are connected through copper pillars and copper bumps. And so those are demonstration that ASE can execute a large panel network, basically CPO solution in the next-generation AI solutions.

Next. So why do we want to do that? Because we believe that the high-density RDL packaging such as VIPack is only one solution that we can do, but that's not everything we can do. So we need to put in the photonic system. By putting the photonic system, we can dramatically increase the overall compute performances, while reducing the power because we don't have the same losses through photons as we do electron through copper wires. So with the combination of the high-density RDL such as VIPack and the CPO that we showed you earlier that this can dramatically increase the overall compute to meet the latest LLM compute requirements.

Next. So lastly, let's talk about thermal. So the thermal is that we use the same chart with GPU power over 1,500 with CPU power over 600. There are many ways that we are looking at it. So today, ASE is really looking at system solution, which is the standard cold plate that sits on TIM 1 that sits on top of the TIM 2 sits on top of the heat sink that sits on top of the die itself.

So right now, we are looking at various material that can improve the overall heat conductivity between the silicon and the cold plate itself. But we are also examining the potential of the silicon solution where the coolants actually are directly in touch of the silicon itself. So instead of have two, three or four different thermal interfaces, we are able to bring the coolant directly to silicon to dissipate the heat and allow the chip to run at its optimum temperature. And this is the next generation where you kind of see the solution is migrating from system to chip, which offer [ AAC ] another opportunity to develop silicon level solution and able to produce the next-generation compute power that's needed for the future AI requirements.

Next. So for the thermal TIM solution that we have talked about, today, we're looking at standard dispense method. We have developed the graphic method. We have done the solar -- that you kind of see the ability for us to increase the overall thermal conductivity from below 10 to right around 86. So we are looking at various level trying to improve the thermal conductivity between various interfaces. But as I mentioned earlier, the potential is not just improving the thermal coefficient in the interfaces, but also bring the coolant directly onto the silicon itself, and that will be the next-generation development.

Next. So we look at the overall packaging innovation and the packaging architecture and technology, if we look at performance, we already see the latest AI model push our overall performance by 2 to 7x. And this is what is required for the AI chip to meet. And then for that, we need to drive the memory. And for the memory to increase, then we need to drive the area. And once you put the compute chip together, then we need to figure out how to deliver the power or the precise power directly onto this array of silicon on top of the modules.

And then once you put in the power, obviously, the thermal will be the next consideration. And so with the packaging itself, we kind of show you how we are able to use the herogene integration that combine various functions, whether it's CPU, GPU, XPU or various I/O or memory chip solution along with HBM put together in whether it's in chiplet or in a fan-out solution with bridge or using 2.5D silicon interposer. And if the package get too big, then we need to look at 300-millimeter panels or 600-millimeter panel to leverage the overall efficiency and maximize the yield and also the scaling of the overall solution.

Then with power, we are demonstrating the vertical voltage regulators and that is the [indiscernible] and try to put in the power regulator as close to silicon as possible and trying to create the backside power that's needed. And if I then convert electrons into photons like CPO, that also reduce the overall power consumption in a given compute. Obviously, when we give more compute, then the CPO requirement will also increase.

And last thing is the thermal. So thermal is something that ASE is looking into in terms of how do we figure out the next-generation cooling structure beyond the thermal interfaces and the thermal interface material that we already are producing through silicon microchanneling or possibly even new material set. And this kind of -- give you kind of a summary of what we are looking at in terms of innovation to fulfill the compute needs for overall industry.

Next. So in summary, we kind of see AI and data continue to fuel the semiconductor innovation. The proliferation is given with the amount of money that the overall industry is asking us to produce the next breakthrough in terms of solutions, and we are accelerating through really the heterogeneous integration advancement. We're putting various type of function die together, various size of function die together. We're putting side-by-side. We're putting on the 3D format. And this type of heterogene integration is the innovation, I think, that fuels the AI data growth.

And last is, we truly believe that package creativity is the [ enablement ] for AI growth or AI path. So it really help us in terms of enhance the functionality and also improve the overall efficiency of any particular compute silicon solution.

With that, I thank you for your attention and time.

Sure. Thank you very much, [ Ian ], for your great presentation. So now let's move on to Q&A session. So once again, if you have any questions, please feel free to e-mail me at [email protected].

So let me kick off. So maybe, Ken, first question for you. Since we have you, lots of questions on how ASE and testing segment could scale going to 2026. In October, management did guide at over $1 billion sales upside going to 2026 on top of this year's USD 1.6 billion.

And so, could you perhaps share with us how the outlook has evolved going to 2026 after you reported? And then how we should think about the ramp for the business across maybe outsourcing, your full process CWoS and final test?

So the commentary for 2026 thus far, we have not given a tremendous amount of color as you've mentioned. The only real comment as of now that we've talked about is that leading-edge advanced packaging will be growing by more than $1 billion next year.

The components of that are -- haven't been particularly talked about, but I think it would be fair to say that they are being led by our traditional LEAP services, meaning on substrate and also, to a certain extent, the testing related to such devices. Towards the back half of the year, we should see a much more pronounced ramp-up in our full service type applications and services.

Also, in terms of the ramp of full service type of package going to second half of 2026, if you may, how should we think about the technology? Will it be driven by maybe more traditional focus or will it be a combination of focus and FOCoS-Bridge? And how should we think about your technology readiness for FOCoS-Bridge? Would you say now the yield has reached a good level and therefore, you are seeing increasing customer engagement?

FOCoS-Bridge, we do believe to be, as Ying mentioned in his presentation, quite an important part of the overall ramping in terms of AI system architecture at the chip level. It does provide incredible increased performance between the processing unit and the memory dies.

So this is something that is particularly important for our ongoing ramps going forward. We have not talked about, again, the makeup of 2026, but the full process, I think many sell side, including yourself, have written up articles on this. On this particular trend. But we do not -- again, we don't have any new information, but we do believe that this should be increasingly important.

In terms of yield, we are -- we have -- we don't generally disclose yield, but we do have full a process work that we are completing or providing right now during 2025. 2026, we should see maybe a different set of customer products, maybe ramp towards the back half of the year.

So on margin for [indiscernible] and testing, the company has got a higher margin and so accretive versus IC ATM. But IC ATM gross margin is at a low base for 2025. And therefore, when management talk about the segment being margin accretive, is it fair to say it's higher even compared with the high end of the range for IC ATM structural gross margin being about 30%?

So the segment gross margin should be over 30%. And so that's the first part of the question. And then the second part will be, how should we think about the margin outlook for the segment going to 2026? Should we expect maybe better margin given larger scale, maybe better yield and also ramping of full process?

So leading -edge advanced packaging is accretive towards our structural margins. So structural margins, we generally talk about in terms of maybe a 70% overall utilization being tied to a 24%. The trough of the structural range and then full utilization at around 85% or so tied to a 30% ceiling margin in terms of the structural range.

But leading-edge advanced packaging in total, all the components do create incremental margin or are accretive to the overall structural mix, right? So that would mean that each of those components do -- are higher, so to say.

In terms of what we're looking at for next year, again, we're not commenting a lot on that. But in total, we do believe that given that the -- ideally, the FX headwinds are behind us, and we should see a much more friendly environment for our margin structure. So this year, I think we did show a decent amount of margin recovery, especially if you do take the FX component out or adjust for the FX component.

So next year, we should continue to see the overall margin environment improve. And then I think Joseph talked about next year having full year margins well within the structural context.

Got it. Thank you, Ken. So maybe moving on to a question on panel-level packaging for HPC application that you talked about. And so where is ASE in terms of the technology readiness for, let's say, 300 for HPC?

And based on your current technology development and also client engagement, when do you think we should see the first wave of product migration? Will it be maybe 2028 that people talk about? Or do you think it would take a bit longer?

I think panel is part of an overall set of delivery set of services and products that we offer. In terms of moving towards an overall panel service and full readiness, I think right now, we have equipment coming in during this year.

We have some level of qualification for next year and then a very -- maybe a minor level of revenue towards the end of next year. But as of this point, we haven't seen a mass migration yet, but that's not to say that this won't happen. I think there's a lot that has to do with the overall ecosystem being ready, meaning machinery, meaning things that may not be within our control.

But again, this is part of an overall view. I think as panel does become more ready, I think it provides an opportunity for leading-edge advanced packaging type services to permeate into different levels of products, not just within this the very peak of the pyramid, so to say, in terms of electronics technology or electronics usage.

So maybe we might see things kind of drift off towards maybe a mobile application or other applications. And I think that might be -- might allow for leading-edge advanced packaging to grow even faster.

Got it. So can you take a step back and look at CoWoS or focus it may be fair to say ASE ramp maybe a bit later than some of your peers. And so now with the potential migration to panel level for HPC, would you say ASE start working very hard to be able to address the first wave of [indiscernible] it becomes?

I think our position has always been fairly steady. We like to do things when -- we don't like to get very much ahead of the technology. I think those situations result in less than optimal returns for the overall company. I think we do like to see machinery or standards become fairly well developed before we really scale things up.

So from our perspective, we are where we like to be. We don't necessarily have a timeline in place. But I think if the market does call upon us to scale up, I think we can be ready along with our foundry partners in this particular area.

Got it. And then on this very interesting topic around [ HVC ] -- so maybe if you could share a bit more color on how OSATs or ASE could play a more important role. What's the content, let's say, between the current solution versus HVDC? Should we assume the packaging for HVDC to be more complicated and therefore, opportunity for you to expand value going forward?

I'm unaware of the abbreviation you used there, the -- what did you for that?

So direct current power delivery. So basically, I talked about in the presentation that potentially [indiscernible] center could migrate to 800 volt in the coming future for better power efficiency, even less [ coverage ].

I think from our -- from where we sit, just from a natural perspective, voltage conversion getting closer and closer to the die becomes ever more important. I think I made that. He highlighted a couple of key points on that.

And then as power efficiency becomes more and more important to not just from a cost savings perspective, from like maybe a global power what am I looking for here, kind of an eco-friendly type perspective in which AI is projected to consume nuclear reactors worth of power, I think this type of power efficiency delivery or the capability to deliver that becomes increasingly important.

I thinkASE's position in terms of where we sit just from a geometric scale perspective as being the bridge to these dies that are consuming a lot of power. I think being -- not being able to do the monolithic methodology of doing -- providing power into those dies makes ASE a very natural provider for such technology or electrical delivery methods or changes in those methods.

So I think this is a very high opportunity for us as these products become or develop further and further. So I don't have a lot of extra information for you here, but we are working on a number of key fronts in this area.

No problem. So maybe back to full process, given a lot of attention for your ramp. So would you be able to share maybe a bit more on what are the type of products or clients that you are ramping going to late 2026? Some of your competitors talk a lot on the expanding product base beyond like accelerators going to 2023.

So for your ramp on full process, would you be able to share a bit more on what are the type of products and clients that you're ramping going to the second half of next year?

Again, we're not -- we have not given a lot of color in terms of 2026 in terms of which products we are involved with. But we do believe our foundry partners are fairly busy overall. There are a lot of opportunities available to us given the lack of resource across the entire industry. So at this point, we don't -- again, we don't specify exactly which customers, but we do have a very wide breadth of exposure in this area.

Sure. No problem. And maybe on CPO. So presentation also showcased several packing opportunities along the process from EIC and PIC stacking, FAU assembly, the packaging overall for maybe on substrates. And so how should we think about business model going to CPO? Would you say it could be multiple type of business models, just like for CoWoS and focus, you could work with foundry, you could also try to ramp full process. Basically, how should we think about the [ opportunities ] going to CPO and your positioning?

I guess, I think from our perspective, silicon photonics is particularly, again, a very interesting area that has been highlighted. Being next to the die or interfacing to these processing units, it puts us in a very unique area in terms of being able to provide interface as part of the silicon photonics solutions.

There are a number of different standards and methodologies being talked about and developed at this point in time. We don't have -- again, in terms of these types of situations, we try to be fairly agnostic. We don't try to -- we don't try to push one or the other. We just want to be part of the endgame solution when there are volumes and when there are returns to be had.

At this point, the silicon photonics revenue levels for us are still relatively small. So we're not talking a lot in terms of the end solutions that we're seeing. But when things do start ramping up in a more major way, I think we can talk about that then. And we are not -- I don't think silicon photonics in terms of a revenue perspective is a major part of the '26 outlook at this point.

No problem. Maybe if I may switch gear a bit to test -- so Ken, lots of expectations on you ramping, if any, on final test, especially for A accelerators in the coming few years. We all understand some engagement may take time. But could you share with us what's the latest progress on your ramp for final test for A accelerators?

I think from our perspective, we are seeing progress in terms of being able to do more final tests within this space. However, given the time lines in terms of how our buildings and facilities are coming ready and then timelines in which other products are going to be coming in.

I think the focus right now that we have and what we're seeing is probably more geared or more focused towards wafer probe. We should see significant wafer probe expansion during '26 and then probably see a little bit more final test exposure on the AI front towards the back half of the year.

But these are all subject to timeline and customer products and such. But we are fairly excited in our overall growth opportunities within test. I think when we're -- right now, we're going to finish the year closer to an 18% to 19% range in terms of how test is part of the overall ATM revenue.

A more natural percentage is probably closer to maybe a 30% or 1/3, 2/3 type relationship between testing and assembly. So we do have quite a bit of growth opportunity if we just test the products that we package. So we will continue to push that forward. And again, our overall test story is about overall test, not necessarily just focused on leading edge or maybe whatever customer that investors may be particularly interested in at this time.

Sure. No problem. But I guess for WaferPro, even the foundry, the fab space is quite constrained and therefore, it's indeed releasing, increasing demand opportunities for OSAT that you have been benefiting, but I think some of your peers also seems to be benefiting from this year. So how should we think about from here?

Do you think the market is big enough to accommodate multiple suppliers for WaferPro and therefore, there should be no impact on the pace for your ramp? Or would you expect maybe at some point, there may be some competitive dynamics that we need to watch?

I think WaferPro being the largest packager out there, we are uniquely positioned to take on more wafer probe opportunities. We also believe that our overall labor-free or maybe lights out type solutions do help contribute to this type of the cost and performance of WaferPro for us.

So we continue to expect WaferPro to continue -- to keep expanding. And as part of an overall turnkey type solution again, ASE being the largest packager, we should see -- we should be uniquely positioned again to receive WaferPro along with the final test -- in terms of services overall for [indiscernible] test.

Got it. No problem. So I actually got a question from an investor. It may be a bit technical. So if Ken or [ Ian ], you could answer. So if power delivery. So you mentioned in your presentation that ASE is able to provide like a [ power ] for voltage regulators as a backside power delivery. So for that voltage regulators, would you be like sourcing from the IC manufacturers? Or would ASE be able to make in-house?

Why don't I pass that along to Ying. Ying, do you want to take a stab at that question?

Can you repeat the question one more time?

Yes. So for the power delivery that you talk about the power set, ASE is looking to offer there are regulators in the. So will ASE buy regulators from the others? Or would you make it in-house?

I think the module itself, we make it in-house. But for the PMIC, it's typically customer specified or customer custom silicon. So it's a combination, I guess. We make the module ourselves, but the chip itself typically has been signed or bought.

Okay. Sure. No problem. It's about time to wrap up. Ken, anything you want to highlight before we close?

I think probably the key point here in terms of -- in this presentation, I think we've talked a lot about the technical aspects of what we encounter.

But I think the key point that people should remember is that as monolithic manufacturing becomes less and less capable in terms of providing these end solutions, a lot of the value of what used to be created on a single die are now spreading out into multiple die, thus having us provide more value in this particular space. So we are excited about the various opportunities that will be presented to ASE via this type of technology propagation.

Thank you very much. Looking forward to 2026.

All right. Thank you very much.

Thank you.

Bye-bye.

All right. Bye

Hello. I am Ken Hsiang, the Head of Investor Relations for ASE Technology Holdings. Welcome to our third quarter 2025 earnings release. I am joined today by Joseph Tung, our CFO. Thank you for attending our earnings release today.

Please refer to the safe harbor notice on Page 2. All participants consent to having their voices and questions broadcast via participation in this event. If participants do not consent, please do not ask questions or you may leave the session at this time.

I would like to remind everyone that the presentation that follows may contain forward-looking statements. These forward-looking statements are subject to a high degree of risk, and our actual results may differ materially.

For the purposes of this presentation, dollar figures are generally stated in New Taiwan dollars unless otherwise indicated. As a Taiwan-based company, our financial information is presented in accordance with Taiwan IFRS. Results presented using Taiwan IFRS may differ materially from results using other accounting standards, including those presented by our subsidiary using Chinese GAAP.

For today's presentation, I will go over the financial results, and Joseph will give the company's guidance. Afterwards, we will be available to take your questions during the Q&A session. With that, let's get started.

During the third quarter, both our ATM and EMS businesses outperformed our original sales and profitability expectations. Packaging and testing utilization percentages were in the high 70s. Loading on LEAP and traditional advanced packaging lines were generally full. Our wire bond utilization also showed some improvement. Our test business continues to grow faster than our assembly business, with our chip probe testing leading the way.

From a profitability perspective, with our factory loading being better than anticipated, we were able to extract higher operating leverage. However, the company's performance was still impacted significantly by foreign exchange. Despite the NT dollar's near-term decline in value against the U.S. dollar, for much of the third quarter, the NT dollar traded at a relatively appreciated level when compared with the second quarter. During the quarter, the NT dollar moved from an average exchange rate of TWD 31.2 to TWD 29.7 per U.S. dollar, strengthening by 4.6%.

Simplistically, we estimate that for every percentage point appreciation of the NT dollar relative to the U.S. dollar, we see a corresponding 0.3 percentage point negative impact to margins at the holding company level and a 0.45 percentage point negative impact to margins at the ATM level. Using this simplified approach, foreign exchange had negative sequential impacts to our holding company and ATM margins of 1.4 and 2.1 percentage points, respectively. And annually, negative impacts to our holding company and ATM margins of 2.4 and 3.6 percentage points, respectively. Heading into the fourth quarter, we expect a more stable NT dollar environment with an average exchange rate of TWD 30.4 per U.S. dollar.

Please turn to Page 3, where you will find our third quarter consolidated results. For the third quarter, we recorded fully diluted EPS of TWD 2.41 and basic EPS of TWD 2.50. Consolidated net revenues were TWD 168.6 billion, representing an increase of 12% sequentially and 5% year-over-year. On a U.S. dollar basis, our sales increased by 17% sequentially and 14% year-over-year. We had a gross profit of TWD 28.9 billion, with a gross margin of 17.1%. Our gross margin improved by 0.1 percentage points sequentially and 0.6 percentage points year-over-year.

The sequential improvement in margin is primarily due to higher loading and our ATM business, offset in large part by foreign exchange. The annual improvement is primarily due to higher utilization and beneficial product mix, offset by foreign exchange. We estimate that foreign exchange had a negative 1.4 and 2.4 percentage point impact to our gross margins on a sequential and annual basis, respectively.

Our operating expenses increased by TWD 0.2 billion sequentially and TWD 0.7 billion annually to TWD 15.7 billion. The sequential and annual increases in operating expenses are primarily due to higher R&D costs. Our operating expense percentage declined 1 percentage point sequentially to 9.3% and was flat annually. Operating profit was TWD 13.2 billion, up TWD 3 billion sequentially and TWD 1.7 billion year-over-year. Operating margin was 7.8%, up 1 percentage point sequentially and up 0.6 percentage points year-over-year.

During the quarter, we had a net nonoperating gain of TWD 0.8 billion. Our nonoperating gain for the quarter primarily consists of net foreign exchange hedging activities, offset in part by net interest expense of TWD 1.4 billion. Tax expense for the quarter was TWD 2.6 billion. Our effective tax rate for the quarter was 19%. Net income for the quarter was TWD 10.9 billion, representing an increase of TWD 3.4 billion sequentially and TWD 1.2 billion annually.

On the bottom of the page, we provide key P&L line items without the inclusion of PPA-related expenses. Consolidated gross profit, excluding PPA expenses, would be TWD 29.4 billion, with a 17.4% gross margin. Operating profit would be TWD 14 billion, with an operating margin of 8.3%. Net profit would be TWD 11.6 billion, with a net margin of 6.9%. Basic EPS, excluding PPA expenses, would be TWD 2.68.

On Page 4 is a graphical presentation of our consolidated quarterly financial performance.

On Page 5 is our ATM P&L. The ATM revenue reported here contains revenues eliminated at the holding company level related to intercompany transactions between our ATM and EMS businesses. For the third quarter of 2025, we had record revenues for our ATM business of TWD 100.3 billion, up TWD 7.7 billion from the previous quarter and up TWD 14.5 billion from the same period last year. This represents an increase of 8% sequentially and a 17% increase annually.

On a U.S. dollar basis, our ATM revenues were up 13% sequentially and 27% annually. Our test businesses growth as a whole continues to outpace our assembly business as a whole, growing 11% sequentially and 30% annually. Gross profit for our ATM business was TWD 22.7 billion, up TWD 2.5 billion sequentially and up TWD 2.9 billion year-over-year. Gross profit margin for our ATM business was 22.6%, up 0.7 percentage points sequentially and down 0.5 percentage points year-over-year.

The sequential gross margin increase was due to equipment utilization rate improvement, offset in large part by NT dollar appreciation. The annual gross margin decline was primarily due to NT dollar appreciation, and to a much lesser extent, higher electricity rates, offset in large part by higher loading.

On a constant currency basis, relative to our first quarter, we estimate our gross margin would be roughly 4.2 percentage points higher during the quarter. This difference would have put our adjusted third quarter gross margin of 26.8% in the middle of our previously stated structural ATM gross margin range.

During the third quarter, operating expenses were TWD 11.8 billion, up TWD 0.4 billion sequentially and TWD 1.2 billion year-over-year. The sequential increase in operating expenses was primarily related to higher overall R&D costs, including labor, equipment and factory supplies. The annual increase is primarily the result of R&D ramp-up and labor-related expenses.

Our operating expense percentage for the quarter was 11.8%, decreasing 0.5 percentage points sequentially and down 0.5 percentage points annually. The decline was primarily the result of higher revenues during the quarter. As we previously have mentioned, we believe our spending in R&D, on an absolute dollar level, will continue to increase. But as the associated LEAP revenue syncs up with the R&D spending, our operating expense percentage should continue to moderate.

During the third quarter, operating profit was TWD 10.9 billion, representing a sequential increase of TWD 2.1 billion and an annual increase of TWD 1.7 billion. Operating margin was 10.8%, up 1.3 percentage points sequentially and up 0.1 percentage points year-over-year. Without the impact of PPA-related depreciation and amortization, ATM gross profit margin would be 23.1% and an operating profit margin would be 11.6%.

On Page 6, you'll find a graphical representation of our ATM P&L. Please note the generally upsloping revenue bars. Using the first quarter's foreign exchange rate, we estimate the gross margin percentages for the second and third quarters would be 24.1% and 26.8%.

On Page 7 is our ATM revenue by the 3C market segments. You can see here that the Computing segment continues to become a relatively larger component of our business. This was largely driven by a higher percentage of LEAP based revenues.

On Page 8, you will find our ATM revenue by service type. Here, you can see the 2 service types containing LEAP services, bump and flip-chip and testing. Both are becoming a larger component of our overall business. We expect continued momentum in these areas heading into 2026.

On Page 9, you can see the third quarter results of our EMS business. The annual seasonality of our EMS business has been inconsistent over the last few years due to differing device ramp-up schedules. As such, we believe the annual comparability of our quarterly results may be impacted.

During the quarter, EMS revenues were TWD 69 billion, increasing 17% sequentially, while down 8% year-over-year. The sequential increase in annual decline were both primarily the result of differing underlying device seasonality. Sequentially, our EMS business's gross margin declined 0.2 percentage points to 9.2%. This slight change was principally the result of product mix.

Operating expenses within our EMS business decreased by TWD 0.2 billion sequentially and declined TWD 0.5 billion annually. The sequential decline is primarily the result of lower compensation and professional fees. While on an annual basis, the decline is primarily related to lower compensation expenses. Our third quarter EMS operating expense percentage of 5.6% was down 1.3 percentage points sequentially, while annually, our EMS operating expense percentage declined slightly by 0.1 percentage points on lower spending and revenues.

Operating margin for the third quarter was 3.7%, up 1.1 percentage points sequentially and up 0.4 percentage points year-over-year. The improvements are primarily the results of higher loading rate and some one-time inventory-related adjustments. Our EMS third quarter operating profit was TWD 2.5 billion, up TWD 1 billion sequentially and TWD 0.1 billion annually.

On the bottom of the page, you will find a graphical representation of our EMS revenue by application. The third quarter application mix shows the seasonal ramp-up of our customers' consumer products, with our consumer segment growing while all other segments declining in application share.

We believe, at a strategic level, our EMS business faces similar technological manufacturing trends as our ATM business does. Trends such as power delivery and thermal control are core themes at the forefront in both our ATM and EMS businesses. Having the ability to address customer challenges at both the ATM and EMS level allows us to provide a broader set of technical solutions to our customers.

On Page 10, you will find key line items from our balance sheet. At the end of the year, we had cash, cash equivalents and current financial assets of TWD 83.4 billion. Our total interest-bearing debt increased by TWD 55.6 billion to TWD 295.7 billion. This increase was primarily due to the completion of a TWD 50 billion syndicated loan to fund our CapEx. Total unused credit lines amounted to TWD 344.7 billion. Our EBITDA for the quarter was TWD 32.6 billion. Our net debt to equity this quarter was 63%.

On Page 11, you will find our equipment capital expenditures relative to our EBITDA. Machinery and equipment capital expenditures for the third quarter in U.S. dollars totaled $779 million, of which $534 million was used in packaging operations, $199 million in testing operations, $40 million in EMS operations and $6 million in interconnect material operations and others. In addition to spending on machinery and equipment, during the quarter, we also spent $716 million on facilities, which includes land and buildings.

The overall environment appears to be strengthening. For us, the upward seasonality during the third quarter has been the strongest since the COVID timeframe. From a customer sentiment perspective, the pendulum appears to be swinging from booking capacity on an as-needed basis to prebooking capacities and making sure raw materials are available. As a whole, our customers are now looking for more assurance and security in their supply chains.

For the quarter, LEAP and test services continue to lead growth for the company. LEAP continues to be driven by AI. Although we are seeing more customers target their products for use within the AI super cycle, many new products are inferring AI capability or AI readiness. Products are expanding new and smart AI capabilities and features. Newer generations of products are becoming more robust electronically, while allowing streamlined access to certain aspects of GenAI capability, such as video and document creation.

The key is whether the end consumers are enticed to integrate new generations of products into their lives. And to that end, AI does appear to be upping the basic standards of quality in various contexts, not just limited to the school, office and social media. And there does appear to be the not so subtle ominous angle of you need AI to be competitive. This is bringing an intelligence and capabilities arms race to everyone's front door.

In such a context, understanding the seemingly insatiable need for more capable chips and hardware seems fairly straightforward. From the packaging and test perspective, the higher the AI computational capability, the stronger the chips packaging and testing needs are. Critical improvement paths in power delivery, processing bandwidth and thermal performance will continue to drive our LEAP services.

With that, I'll hand the presentation over to Joseph to present the company's outlook.

Thank you, Ken. Let me give you the fourth quarter guidance. Based on our current business outlook and the exchange rate assumption of USD 1 to TWD 30.4 versus in third quarter, we have TWD 29.7 exchange rate. Management projects overall performance for the fourth quarter of 2025 to be as follows.

On a consolidated level, in NT dollar terms, our consolidated fourth quarter revenue should grow by 1% to 2% quarter-over-quarter. Our consolidated fourth quarter gross margin should increase by 70 to 100 basis points quarter-over-quarter. Our consolidated fourth quarter operating margin should increase by 70 to 100 basis points quarter-over-quarter.

For ATM, in NT dollar terms, our ATM fourth quarter revenue should grow by 3% to 5% quarter-over-quarter. Our ATM fourth quarter gross margin should increase by 100 to 150 basis points quarter-over-quarter.

For EMS, in NT dollar terms, our EMS fourth quarter revenue should stay flat or decline slightly quarter-over-quarter. Our EMS fourth quarter operating margin should be similar to fourth quarter 2024 level.

With that, let me also give you some color for the full year. For ATM, we're seeing better-than-expected momentum of mainstream business, given the continuing recovery of the general market. While our leading-edge revenue, we are on track to reach the USD 1.6 billion mark as planned. Altogether, we expect ATM 2025 full year revenue to exceed our target and grow over 20% year-over-year in U.S. dollar terms.

As for machinery CapEx, we expect to further increase our full year CapEx by another few hundred million U.S. dollars to meet customers' requests and to support continuing business momentum into 2026. The increase is largely for wafer probing for both AI and non-AI chips as well as for general capacity ramp and some new initiatives for year 2026.

With that, let's give it back to Ken to open the floor for questions.

Thank you, Joseph. During the Q&A session that follows, we would appreciate if questions can be kept concise and asked one at a time. I will be receiving each question and repeating the asked question to Joseph. Again, we'll be limiting the number of questions asked to 2 questions per turn, but asked one at a time.

The first question is from Gokul Hariharan of JPMorgan. Gokul?

First question, obviously, on LEAP, could you give us a little bit more color about how the progress has been on LEAP revenues this year? I think you had the TWD 1.6 billion guidance or additional TWD 1 billion guidance. What are we tracking to compare to that guidance now? And any indications for what it could do next year? I think based on our own math, it looks like it could easily double next year. And you're also raising capacity and CapEx pretty much every quarter. And also, on LEAP, what is the margin contribution from LEAP-related business? Is it already accretive or it will turn accretive once you reach a certain kind of revenue run rate, and any indications on that? That's my first question.

Gokul, you are looking for revenue progress and generally kind of what you're thinking for this year.

Yes.

Okay. Like I said, the -- we are on track in reaching our TWD 1.6 billion mark this year. Everything is progressing well. I think we have shown very strong momentum in the AI and HPC related part of the business.

In terms of the revenue mix, I think, because of the geopolitical uncertainties, in terms of packaging, we are a little bit short from our original target, but that was sufficiently replenished by our more than expected growth in our test business. So we are very, very confident that we will reach our TWD 1.6 billion mark for this year. And going forward into 2026, we see -- we continue to see very strong momentum. And we are very, very confident that we will gain another -- over TWD 1 billion kind of revenue increase for 2026 in this space.

CapEx-wise, we will continue to make heavy investments in our leading edge, I think, to support the strong momentum that we are seeing today. And I think AI or HPC is really -- the momentum is here to stay. We're not going to be shy on making the necessary investment to not just secure our dominant position in this space, also to expand that dominance against our competitors and to fully support our customers' needs.

In terms of margin and return, I think the -- as steady state, as we mentioned before, the LEAP would definitely be both margin as well as return accretive. And we are quickly reaching that point at this point.

Okay. That's very clear. Maybe one other question. Can you talk a little bit about pricing? I think, Ken mentioned in the opening remarks that you're pretty much running full on flip chip and bumping. You're pretty much running full on LEAP. I think last time around, I think Dr. Wu had discussed about potential price negotiations. Anything that you can report on what are we seeing on pricing for your overall offering? Should we expect that pricing should go up? I think OSAT pricing doesn't usually go up that much, but just wanted to understand how we should think about pricing going into next year.

Gokul, you're looking for commentary on overall just pricing environment for us for this year and next year.

Yes. And maybe also specifically on LEAP as well as your flip chip and bumping kind of advance -- the mainstream advanced packaging business as well because the customer set is slightly different. LEAP, you're kind of largely partnering with the large foundry.

Well, without giving -- without getting into specific, I think, in general, I think our pricing remains to be resilient. And I think it's very sensitive to talk about pricing. But as a whole, I think we will continue to set the -- our pricing, the most suitable pricing structure based on the current situation. I think there are a lot of moving parts, and there are a lot of uncertainties in front of us. But in general, I think we will continue to make our pricing a very, very resilient level.

Maybe if I kind of tweak it a little bit, Joseph, like what is customer feedback? I think I'm sure that everybody is talking about this. We hear that from your fabless customers as well. But I just wanted to understand like what is customer feedback to pricing even in -- I wanted to think about a little bit more on the mainstream stuff, like flip-chip CSP or flip-chip BGA, where there is no super cycle of growth. Even in those areas, are you able to have some like value app programs coming through?

Are you asking for expansion on the original pricing question there?

Yes, sir. Maybe talk a little bit more on the mainstream advanced packaging as well, yes.

For mainstream, I think we are seeing the continuing recovery of the general market. And therefore, I think pricing wise, I think, it's right now at a very stable level.

Next question is from Charlie Chan of Morgan Stanley. Charlie?

Yes. I just unmuted myself. First of all, congratulations for very good results and outlook. My first question is really on sort of supply chain related discussion. For example, what's the update plan for you to do the U.S. operation? Because your major customers -- major foundry partners are all very active in the U.S., and there seems to be -- your competitor, Amkor, in that presence. So one is that your updated plan for the U.S. operation to enjoy that ASME kind of growth. And also, we are very concerned about the sort of T-Glass shortage. I think a lot of customers are going through with your fab to see if they can secure more substrates, right? So I'm not sure if they would be kind of a gating factor for your next year's growth. So this is the first question.

That's -- Charlie, that sounds like 2 questions. So let's start with question number one, the U.S. building out perspective.

Okay. Thank you for your question, and thanks for coming to my concert.

Yes, it was a great one.

U.S., we don't have anything new to report except that -- let me reiterate what we mentioned last time that we were invited by our customers to look at the investment opportunities in the U.S. We are currently still engaging in discussion with our customers and we're evaluating different opportunities, but no decision is made at this point. But whatever decision we will eventually make, it will have to make economical sense for us.

In terms of the competition, I think Amkor has its own mind. So I think I'm not going to answer for Amkor. But overall, we will continue to be watchful on the overall competition landscape and see how we can better position ourselves in terms of meeting this competition.

So Charlie, do you want your second question to be about your previous question on T-Glass and such?

Yes, maybe we can save it for maybe second round. But my major second question is really the final test completion. So I know this one is a little bit controversial, but I wanted to get your updates or confidence level about your final test market share at major customers' next-generation GPU.

Yes, and by the way, congratulations for a very strong share price. So I think your efforts were recognized by foreign shareholders. Yes, so second question is really about your final test business updates.

So you're looking for a more comprehensive explanation or update on our final test market share gains.

Yes. Because your Taiwanese competitor seems to be very aggressive in the cashless purchase and capacity expansion as well. So I hope both can win. Yes, so I just wanted to get a little bit more color about your realistic assumption about your final test market share.

I think, as we mentioned, we have been aggressive, and we have been pretty successful in terms of expanding our test business. I think for this year, our test business growth is going to be twice the packaging revenue growth. And we will continue to make large investment into our test capacity. But our resources are also limited. We don't have unlimited resources to try to cover everything in the market. So right now, the main focus for our investment in test is really on the wafer probing. And I think we will continue to on this effort for the time to come.

And in terms of final test, I think we are making the investment -- necessary investment at this point to build up the capacity. And we're expecting to have meaningful revenue being generated in the later part of next year when we start serving the next-generation AI chips.

Next question is from Bruce Lu of Goldman Sachs.

Can you hear me?

Yes.

Okay. My question is regarding to your revenue split for your incremental TWD 1 billion revenue in 2026 for your AI-related revenue. We understand that the revenue contribution is more geared to testing for this year. Are we able to see incremental more revenue contribution from packaging? And to be more specific, can we get more like packaging-related business from both outsourcing as well as your own packaging or AI packaging business?

Bruce, you're asking for the incremental revenue for this year, right?

And next year, your -- because Joseph just said that we will see another additional TWD 1 billion revenue for next year, right?

He may have said that. So yes, okay.

For the TWD 1 billion increase of our leading-edge revenue, I think the breakdown is TWD 650 million from packaging and about TWD 350 million from test for this year. For next year, well, we'll see how things go. I think the -- we'll kind of give you a ballpark number saying that we will be having maybe at least TWD 1 billion revenue growth. But in terms of the exact composition, I think that remains to be seen, and we'll base on the current situation to allocate our resources and to grow both of the business, but without -- right now, we don't have a set mind on what kind of breakdown it will be. But what I can say is that test seems to be -- continue to have stronger momentum at this point.

I see. So the testing will grow faster than packaging next year within this TWD 1 billion?

It has been growing faster than the packaging. But come next year when the new generation product comes on stream, the competition may have some changes. But what I'm saying is that we are seeing -- we're continuing to see strong momentum in test at this point.

I see. Okay. The second question is for -- again, I want to drill down a little bit for the U.S. plan. I mean, TSMC has utilized asset plan to build some CoW process, and Amkor committed to build some substrate process. So it seems to me that they have -- your customer, your competitor seems to have at least one supply chain in U.S., which probably -- what's the strategy for ASE at the current stage? Obviously, you probably don't need a 2 supply chain in United States, right? So the potential -- losing some market share for TSMC Automotive business is definitely a threat for our future business, right? So can we elaborate more about like what's the strategy from ASE?

Bruce, you're looking for a reiteration on the U.S. plan on our behalf.

Yes. .

Well, we don't fight for market share just for market share's sake. We fight for the market share that makes sense or make profit for us. If we don't see return, if we don't see at least acceptable margin, then that's not the part of the business that we want to pursue. I think the -- like I said, regardless if it's U.S. or in any part of the world, for us to make an investment, it has to make economical sense. So that's -- if Amkor feels that with that kind of investment they can make a profit out of it, fine. But right now, we're not sure on that.

So there's no way to pass on the incremental cost to the customer in order to make the investment like profitable?

Well, it's not just about pricing, it's about the overall infrastructure, which -- that can support that kind of a business at a reasonable cost structure. And even with some premium pricing, whether that cost -- that can cover the costs associated with it remains to be seen. Right now, I think that's a very tall task actually.

Next question is from Laura Chen of Citigroup.

Can you hear me?

Yes.

I just want to consult, Joseph, your view on the gross margin outlook, and also, congratulations for the great result. Do you think we see quite full utilization rate like Ken just mentioned? At the same time, there is a stronger testing business. I recall, Joseph, you mentioned before that in the longer term, if the utilization rate get back to 80% plus, the gross margin could go back to high 20s. So just wondering how is the dynamic now. Are you also -- and also, you are increasing the CapEx for the future demand. So just wondering how should we think about the gross margin outlook into next year or longer term?

Laura, are you looking for commentary on the relationship between utilization and our margin structure?

Yes. And also -- yes, and at the same time, we are also increasing CapEx. I believe that there's also some increasing in depreciation costs. So just wondering the dynamic right now, how should we think about the gross margin outlook?

Well, if we exclude the foreign exchange impact, I think we have already come back to our structural margin. Like Ken mentioned in third quarter, if we were on the same ForEx level as quarter 1, our margin should be around 26.8%. And going into fourth quarter, there will be further margin improvement. And again, at the same currency level, we should be over 27%. So what we mentioned before, once our utilization reaches 70% and above, then we should go back to our structural margin range.

But unfortunately, the foreign exchange does have a pretty big impact on our overall margin. But having said that, I think we will continue to -- I think right now, the foreign exchange seems to be stabilizing now. We will start our margin effort from this level. And we are very confident that with the continuing expansion of our leading-edge business, we're confident that we will continue to see -- as the capacity being ramped up, we are confident that we will continue to see margin improvement. And right now, we are very confident that in 2026, for the whole year, we should be -- we should have a gross profit margin for ATM at the structural margin range.

We are looking for that. My second question is that -- about the leading-edge advanced packaging. ASE also developed your own focus technologies. Just wondering that how is the current progress in the customers' engagement. It's not just focused on the outflow opportunities on substrates. Also, how does ASE's -- your own focus progress?

Laura, you're looking for an update on our internal advanced packaging solutions, just as a focus.

Yes. Right.

Well, obviously, in terms of the overall capacity, I think for CoWoS or CoWoS like 2.5D, I think our foundry partner as well as ourselves is still scrambling to try to make the necessary investment for our capacity to catch up with the demand. And so given the tightness, I think, obviously, there will be customers -- other customers that would like to have other alternatives or solutions for -- to meet their demand, and that creates a very good business opportunity for us to try to sell our own solutions. And on that, we are making the necessary investment at this point. And we do have engagement with multiple customers. And -- but these things take time. I think the -- what we're expecting is that by latter part of next year, we will start to see meaningful full process revenue coming in serving multiple customers.

Okay. So does this also included in your at least TWD 1 billion revenue increase into next year?

Yes.

Next question is from Sunny of UBS.

Sunny, are you there?

Yes. Could you hear me okay?

Yes.

So congrats on the very good results and guidance. Glad to see LEAP business ramping up and gaining momentum going to 2026. So maybe a question on mainstream. Could you help us understand the recovery ahead? And so when you guide IC ATM sales to grow 3% to 5% sequentially, how is the growth by mainstream and LEAP? And how should we think about the cycle for mainstream going to 2026? Do you see the current utilization rate being a good base for critical recovery going to 2026?

Sunny, you're looking for basically our more trailing edge capacity or trailing edge plus traditional advanced packaging capacity.

So mostly on the mainstream, so wire bonding, die bonding?

Okay. You're looking for commentary on more traditional packaging and then -- for this year and into next year?

Yes. How should we think about the cycle from here?

Well, as I mentioned, the mainstream business is -- we're seeing better-than-expected performances. And I think that's a result of the general market recovery. And also, in some part of the -- in different sectors, we are also seeing ourselves gaining shares, particularly -- if we look at different sectors, I think communication and -- communications, and of course, PC or computing is recovering better than the other, like automotive and industrial.

But nonetheless, I think the recovery is very obvious at this point. Maybe in terms of automotive, it's kind of moving in a slower pace than the other 3 sectors. But on that, we actually posted a very, very good growth in our automotive business. I think for ATM this year, we're going to see over 20% growth in this part of the business. I think that's largely as a result of we continuing getting -- gaining market share in this space through our factory automation.

In general, I think, in the beginning of the year, we were saying that we will have our leading-edge giving us 10% growth and mid- to high single-digit growth coming from the mainstream. Obviously, as I mentioned in the beginning of the session, I told everybody that we're going to exceed our revenue growth target to over 20%. So that means the mainstream performance is much better than what we were expecting in the beginning of the year. And we're not seeing anything negative at this point in terms of mainstream business. So without giving you any further guidance for next year, but we do think that we are in a very healthy space at this point for both in the general market, and we're still seeing very strong momentum in the leading edge as well.

Maybe a very quick follow-up. So for Q4, is the utilization rate for mainstream continuing to recover a bit?

Yes. I think in the -- like what Ken just mentioned, I think our bumping and flip chip are pretty full. Wire bonding is improving, although it's not entirely full, but it is steadily improving.

Got it. My second question is on gross margin. So from here, one, with the improving measuring business, and then secondly, accelerating ramp probably for LEAP going to 2026 and a stabilizing FX, should we assume for IC ATM, the gross margin recovery should accelerate in the coming few quarters?

Sunny, you're looking for an update in terms of forward-looking commentary regarding our gross margin structure.

Yes, especially on the pace of the improvement.

Well, we're not in a perfect world, right? There's still a lot of moving parts and uncertainties in front of us, which includes foreign exchange movements. So yes, I think the general trend is very certain because as we continue to expand rapidly in our leading-edge, which is margin accretive, so that gives us a very good pace for our margin improvement going forward. But in terms of the pace, I think there's still -- I think right now, it's still too early to give you a clear path of what kind of pace we're going to have in terms of our margin expansion.

Got it. Also, on LEAP, is there a margin difference between outsourcing and full-process CoWoS, meaning if you start to ramp more full process from second half of next year, will that further boost your gross margin for IC ATM?

I think, in terms of full process, we're still at the early stage, so it's kind of difficult to make any meaningful comparison at this point. I think both needs to be at really a more stable level for us to make the comparison. I think theoretically, regardless it's our own full process or outsourced, leading edge does give us margin accretion.

Our next question is from Felix Pan of KGI.

Can you, guys, hear me okay?

Yes.

Yes. So my first question regarding to -- I have seen your foundry partner incremental CP test outsourcing demand. Just correct me if I was wrong, but I found it very difficult to quantify how big for the TAM is. Maybe for you, it's really hard to comment on the same, but maybe on the TAM side or even the percentage of the BOM, can you just give us some sense how can we quantify, how big for the CP test demand? Just any color will be grateful. That's my first question.

Felix, I think I'll take this one. In terms of the overall TAM, for something like that, I would say that that's not quantifiable, at least from our perspective. This is something that is probably known by our foundry partners. And you may want to address the amount of work that they want to outsource directly to them. We don't quantify that at this point.

Okay. Yes. So my second question is, I think during the TSMC's latest earnings call, I think C.C. emphasized customers of customers engagement. We do see the incremental engagement. From your perspective, do you see the similar pattern, the engagement from customer to customer as well? Or -- I think there's a lot of thing happening this month, so I just want to -- if any color you can share, is any business model change or you have seen incremental customers' customer engagement as a TSMC as well? Yes.

Felix, you're asking for how we look at our overall market and whether we actually look into our customers' customers? Similar to...

Actually, my question is there's any customer -- your customers' customer jump your customer to have the engagement with you guys, like to secure some critical capacity or something like that.

I don't know if we can talk about that. Joseph, if you want to take a step?

Yes. I think we have a very, very close communication with both our direct customers as well as our foundry partners. Those dialogues are being conducted on a routine basis so that we can better prepare ourselves in terms of our capacity and also our technology roadmap. So in this regard, we do talk to them. And I think our information source is not just coming from our customer, but our customers will definitely keep us informed of what they're expecting from their own customers and how the overall market will shape up. So it's a constant dialogue among the industry players to make sure that the demand is sufficiently being supported by the supply. That's an ongoing process that has been going on for them, maybe forever.

[Operator Instructions] Next question is from Gokul Hariharan of JPMorgan.

First one, could you help us understand what is the progress on the full stack focus or CoWoS like kind of processes going into next year? When do we expect this to start becoming more meaningful contributor to revenues, to the LEAP total revenues? And are the applications still similar in terms of like AI accelerator? Or are the applications becoming more diverse in terms of networking or server CPU and other kind of stuff as well?

Gokul, you're looking for an update on our -- more on our full process type services, is that correct?

I talked about this earlier. I think we are continuing our investment in full process, and we are currently engaging with multiple customers to plan for the capacity, and we expect that come later part of next year, we should start seeing some meaningful revenue coming from full process rather than just only from outsourced part of the business.

In terms of the application, I think there will be AI accelerators. There will be other adoptions in different chips requiring such capability. But at this point, I think it's a little bit too early to say the exact revenue -- scale of the revenue or the composition of that -- of such revenue. We just have to continue to work very closely with our customers, multiple customers to better understand what their demands will be, and we'll prep ourselves for the necessary capacity for them.

Got it. Maybe a slightly related question is on the CapEx. I think we are probably finishing this year above TWD 3 billion, well above TWD 3 billion in terms of machinery CapEx. How do we think about this investment cycle? Are we still going to be in this, like increased CapEx, likely to continue to increase CapEx over the next couple of years given the demand outlook that you're seeing from your customers and your customers' customers?

Gokul, you're looking for an update on our overall CapEx view. And also in the frame of the leading-edge advanced packaging, how it works?

That's right. Yes.

Like I said, we stay very close with our foundry partner, and they -- our foundry partner being the dominant player, they cover all the who's and who's in the -- whoever has any demand, they will be the one to supply. So they really have a very, very close connection with their customers and their customers' customers. Since we have a very close communication with them, so whatever information that they're gathering, we do have the benefit of sharing some of that information to better prepare ourselves for capacity expansion.

And as I said earlier, again, we're not going to be shy of making the necessary investment for -- particularly for the leading edge, so as to secure and also to expand our dominance in this space. And as such, we believe, at least for next year, we will continue to see pretty heavy investments in our capacity as well as technology in this -- in the leading edge.

So is it fair to say next year machinery CapEx is likely to be still higher than this year?

We will give you better guidance once we complete our budget cycle, which is starting now. And we will reserve this question to next quarter.

Next question is from Charlie Chan of Morgan Stanley.

Yes, that question is about T-Glass resulting on the shortage of substrates. I'm not sure if you've heard, there would be kind of a risk factor for ASE Group to grow your revenue next year because we start to hear some customers' hard time to get the substrate sourced. And how would the ASE to help our customers to get a more sufficient supply?

Your -- Charlie, your second -- or your third question is regarding overall T-Glass supply and how it impacts our -- whether it would impact our overall supply chain going forward?

Yes, yes. And how would the ASE manage or help your customers on this period of shortage?

Like I said, there's a lot of uncertainties that's ahead of us, so -- like running any other business, there still is going to be ups and downs, there's going to be changes. But right now, I think whatever we're seeing today, maybe some of the materials or -- don't ask me what T-Glass is, but some of the materials may have a longer lead time. But at this time, we haven't seen any real disruptions on our service to our customers at this point. I think if anything else, being the dominant player, if there's any problem, we're the ones that our customers come to, and we certainly have the best leverage in trying to secure the needed materials or variable components that will be needed for the -- for serving them.

Got you. So I would assume, for those materials or substrate, if there will be any cost or price increase, ASE would fully pass-through to customers? Is it right or you would charge some markup because those materials are getting harder to get?

We will find the most suitable pricing for current situation.

Next question is from Bruce Lu of Goldman Sachs.

I think I asked this question last quarter, but I want to ask it again. What is the CapEx to revenue nowadays? Or is there any changes in terms of like how long does it take to see the revenue after you invest your CapEx? The reason I ask this is that you invest for TWD 1.8 billion CapEx last year and 3-point something billion this year, right? But you generate additional TWD 1 billion of revenue this year, but you also can only generate additional TWD 1 billion next year. Theoretically, should be able to generate a bit more than $1 billion next year, right? Is there any changes in terms of CapEx to revenue or trying to generate revenue?

Bruce, you're looking for the magic solution in terms of CapEx to revenue, right?

Which Joseph used to give us.

Well, first of all, the TWD 3 million plus CapEx is not entirely for leading-edge. I think -- for this year, I think 55% of that is for leading edge. And bear in mind that that's just a number. We don't make capacity expansion overnight. Equipment needs to be delivered. You don't have this equipment all delivered at once, right? Things move progressively.

So just simple math, if it's TWD 1.8 billion worth of CapEx, that means now, on average, TWD 900 million worth of new capacity being put in. So this year, if it's a TWD 1 billion increase, that ratio seems to be still on track. Of course, the other half of the investment will start to generate revenue, but there is always a time gap between when the machineries -- or the CapEx being spent and when the revenue is being generated.

I'm not saying that. I'm not giving you -- I'm not saying that we can only generate TWD 1 billion worth of new revenue coming in. I'm just saying that at this point, we are very, very confident that we can have at least TWD 1 billion worth leading edge revenue -- new revenue coming in next year.

For the majority of the leading edge at this point, we're still in the earlier stage at this point, and we're still gathering data to come up with the more meaningful investment intensity on this kind of investment. But from the limited data that we've gathered so far, I think the traditional TWD 1 of investment creating TWD 1 of annual revenue seems still be the case for the main businesses that we are entering now, which is OS and test.

So one to one. That's the major number. It still works.

Still applies. But like I said, we are still in the process of gathering more data. And bear in mind that the -- our capacity is not in full ramp at this point. So it's going to take a little bit more time.

My plan is simple, right? 45% of your TWD 3-point-something billion CapEx is USD 2 billion, right? I mean, you just mentioned that TWD 3.8-something billion, 55% is for matured technology. Let's say, 45% -- let's say, 50% of your TWD 3-point-something billion CapEx this year, that's close to TWD 2 billion for next year in terms of incremental new revenue from AI. That's how the math works.

No, that's not how the math works. We don't live on math. We live in the real world.

Well, I only know math.

Well, if you're calling me conservative, well, call me conservative.

There's no question on the floor.

Okay. I guess, time has pretty much run out. I would like to thank everyone for participating in the call. I look forward to seeing you all, either during the quarter or at the next earnings release.

Okay. We are having a good run, and we'll continue to have a good run going into next year. And we're confident that we will continue to deliver good performances and good numbers for you. We'll see you next quarter. Thank you very much.

Hello. I am Ken Hsiang, the Head of Investor Relations for ASE Technology Holdings. Welcome to our second quarter 2025 earnings release. Thank you for attending our earnings release today. Please refer to our safe harbor notice on Page 2. All participants consent to having their voices and questions broadcast via participation in this event. If participants do not consent, please do not ask questions or you may leave the session at this time.

I would like to remind everyone that the presentation that follows may contain forward-looking statements. These forward-looking statements are subject to a high degree of risk, and our actual results may differ materially. For the purposes of this presentation, dollar figures are generally stated in new Taiwan dollars, unless otherwise indicated. As a Taiwan-based company, our financial information is presented in accordance with Taiwan IFRS. Results presented using Taiwan IFRS may differ materially from results using other accounting standards, including those presented by our subsidiary using Chinese GAAP.

I'm joined today by Dr. Tien Wu, our COO; and Joseph Tung, our CFO. For today's presentation, Dr. Wu will go over our midyear update. I then will go over the financial results, and Joseph will deliver the company's guidance and closing remarks. Afterwards, both Tien and Joseph will be available to take your questions during the Q&A session. I will be moderating the Q&A sessions that follow. We will start with some questions from the floor and then start alternating in some questions from virtual attendees.

With that, I'll hand the presentation over to Dr. Wu.

Good afternoon. We have not had a life one for quite a few years. Welcome back, and thank you for coming. To begin with, I would like to give you a recap for the first half of 2025. Everything will be in U.S. dollar terms. The unconsolidated revenue grew 9% year-on-year in first half of 2025, with ATM revenues up 18% year-on-year. Leading-edge advanced packaging and overall testing outpaced growth, while the general segment saw some recovery. The leading -edge advanced packaging and testing revenue were over 10% of ATM revenues in the first half of 2025 comparing to 6% for the full year of 2024.

Our testing business grew 31% year-on-year in the first half. The momentum will continue into second half on increased turnkey and expanding leading-edge test. Machinery CapEx was USD 1.9 billion, building factory facility automation was $0.9 billion in the first half of 2025, driven by advanced packaging and testing. So that is the first half recap. It has been quite busy for ASE team. The second half will be busier. Let me give you the outlook. For the ATM business, we expect momentum to carry into Q3. As of now, we also believe Q4 will have a quarter-to-quarter growth comparing to Q3.

The leading-edge advanced packaging and testing revenue, we target to increase by USD 1 billion. We made a statement at the early beginning of this year, contributing 10% of the whole year growth, while the general segment to grow by mid- to high single digit year-on-year in 2025. We maintain this view as of now. Even with all of the uncertainties that we have gone through, this view has not changed. We expect revenue uptrend continuing into 2026 and beyond, driven by leading-edge solutions and broad-based semiconductor demand related to AI proliferation, also general recovery that we believe will happen in 2026. Investment in R&D, human capital, advanced capacity and smart factory infrastructures are key to support the multiyear growth. This is putting a lot of pressure on ASE team. However, with detailed conversation discussion with our key partners and all of our key customers, we believe this is the best thing that we can do for ASE, for the Taiwan ecosystem as well as for the industry.

With that, let me go over to the third page. The third page, I only have bullet form. I would like to talk about 3 items: Market dynamics, operations and also the challenge. I might not go over things in the right order, but these are the bullet points I would like to cover. To begin with, I want to talk about how do we see the market heading and what is the current dynamics and also the future opportunities for ASE as was for many players in our industry. I want to touch on the technology focus because that's related to the market trend, also related to the ASE position. Then I'll cover the operation where the leading-edge capacity and all capacity in general in Taiwan is very full now. That's why we have to continue to expand into the second half. The overseas, we still have some idle capacity. Therefore, how do we manage the expansion in Taiwan as was outside of Taiwan and the resource optimization becomes key. And lastly, all of the challenges, I will touch base on that.

Okay. With that, let me just go over the megatrend. Everybody talks about AI, hyperscaler, data center. We're well into, I would say, the second of the three year of this trend. The recent announcement outside of U.S., you're seeing the announcement of mega data center worldwide. We believe there are 2 things that are happening. The first is the expansion of hyperscaler data centers worldwide. The second thing is in the middle of that expansion, the upgrade cycle are ongoing right now with the technology provider as well as the infrastructure provider. We have not touched based on the AI edge applications yet, but we believe there will be multiple waves in the next 10 years, starting with the hyperscaler data centers and then go through the inference and then go through the AI edge applications. What are important is in this AI paradigm shift, what has become clear to us by talking to our foundry partner and also talking to our key customers, some of the foundational technology requirements are identical.

I'm going to give you 4. The first one is integration. I think the 3D IC packaging, the density that is an example of the heterogenous integration that we've been working on this for a long time. The focus offered by ASE, the CoWoS offered by TSMC are examples of that integration. That trend will continue with the expansion as well with the upgrade cycle.

The second thing is the power management. We have not touch based on the power management. We believe power management is going to be a key hurdle that the industry needs to address. The third one is the silicon photonics. I think we've been talking about very vocal on the importance of silicon photonics. We have not seen the revenue uptick yet. But this is a foundational technology that will provide the bandwidth, the speed, latency and the efficiency. It needs to happen in order to trigger even more applications such as humanoid.

And lastly, will be the cost. In the cost, not only what you think about capacity, what do you think about material configuration, more importantly, we have to think about throughput and also the flexibility in designing the footprint and that's where the large panel coming in. If we have gone through all the foundational requirements, that we're going back to the ASE position. And there are 3 things I want to share with you how I view the ASE position, how I articulate ASE's position to our key partner and clients.

The first one is scale. I think from the news report, you can look at ASE scale, margin model as well as our cash flow and also the amount of CapEx investment we are making on behalf of the industry. The second item is speed. Because ASE is well positioned within the Taiwan ecosystems, that we will execute expansion in the second to none. In other words, it's very difficult to imagine in 2025 alone how much CapEx we have put in, how much more revenue we are going to introduce. The third one will be the synergy. So between scale, speed and synergy that pretty much outlined the ASE position in this new paradigm AI shift. We're at the beginning of the data center hyperscaler. In the future, there will be multiple cycles of expansion, upgrade, inference as well as the AI edge, that's where the real volume and the real application is going to emerge.

Because all of these future opportunities, we are seeing the leading edge capacities in Taiwan is very, very full right now. We do see the disparity between AI as well as the other general sectors, and that pretty much outlined the 2025 fist half scenario. In the second half, the disparity will improve. In 2026 and beyond, we believe the cycle will start showing less of a disparity. That is why it is putting a lot of pressure on ASE to accelerate on the capacity growth in Taiwan, especially in the leading-edge packaging and testing. At the same time, it prompt us to look at resource optimization between Taiwan and overseas. We have planned to looking at expansions in other countries of the world. with all the recent updates and changes, it prompt us to start looking at the business opportunities versus how do we deploy our capital and resources based on the new paradigm as well as all the new variables.

And lastly, foreign exchange. Foreign exchange, I think Ken Hsiang and Joseph will give you much, much more details on the impact of foreign exchange on ASE's performance for last quarter and maybe for Q3. And However, I want all of you to keep in mind, we are here for the long term. In the long term, we will have execution issues. We'll have regulatory control issues. We will have product mix issues. We have customers changing order issues. Of course, we will also have machine delivery issues and our own execution issues. When the management team is busy worrying about all the detailed operational issues, let's not forget the future opportunities here and the speed, execution, we would like to take all of the ASE partner and our customer to a much higher ground for this uptick, 10 years of AI cycles.

With that, thank you.

Thank you, Dr. Wu. Now I will go over our prepared remarks in regards to our financial results for the second quarter. We are trying to be more environmentally friendly. We are no longer providing printed copies of our slides. If you have not done so, we have a QR code, the QR code, please, here for the attendees to scan. After scanning, you'll be forwarded to our Investor Relations landing page where you can download materials related to the presentation today. The slide deck currently does not include our guidance section. After the presentation has concluded, the slide deck will be updated to include our guidance section.

During the quarter, we saw a material move in the NT dollar to U.S. dollar exchange rate. The NT dollar moved from an average exchange rate of NTD 32.8 to NTD 31.2 per U.S. dollar strengthening by 4.9%, with our revenues generally based in U.S. dollars, and a large percentage of our ATM expenses being NT dollar based. The foreign exchange fluctuation was detrimental to our financial performance. The impact of currency fluctuation will differ each quarter. However, on a simplified basis, we estimate that for every percentage point appreciation of the NT dollar relative to the U.S. dollar, we see a corresponding 0.3 percentage point negative impact to our gross and operating margins at the holding company level and a 0.45 percentage point negative impact to margins at the ATM level. Using this simplified approach, we can estimate that on a sequential basis, foreign exchange had impacts to our holding company and ATM margins of 1.5 and 2.2 percentage points, respectively. On an annual basis, we estimate impacts to our holding company and ATM margins of 1.0 and 1.5 percentage points, respectively.

As Dr. Wu stressed, our businesses are healthy and are generally on track to hitting most of our targets stated at the beginning of the year. However, foreign exchange movements have created a temporary misalignment between our costs and revenues. And as a result, we believe that our current financial results may not fully portray our underlying accomplishments. From a strategic perspective, we believe the current negative currency impact to be a near- to mid-term phenomenon financially. We fundamentally believe our businesses support a certain level of financial return. Such return expectations are intrinsic within our business evaluation and capital investment processes. The exchange rates we encounter are variables used within these calculations. As such, in time, we believe our margin structure can and will return to previously stated structural levels.

With that said, we are examining and considering the timing of a number of strategic initiatives. We are also reconsidering whether future business opportunities still align with our return goals. There is much to accomplish, but it does present an opportunity to reexamine our businesses in more detail. With that, let's go through the financial results.

Please turn to Page 7, where you will find our second quarter consolidated results. For the second quarter, we recorded fully diluted EPS of $1.70 and basic EPS of $1.74. Consolidated net revenues were $150.8 billion, representing an increase of 2% sequentially and 7% year-over-year. On a U.S. dollar basis, our sales increased by 7% sequentially and 11% year-over-year. We had a gross profit of $25.7 billion with a gross margin of 17%. Our gross margin improved by 0.2 percentage points sequentially and improved by 0.6 percentage points year-over-year. The sequential improvement in margin is primarily due to higher loading efficiency in our ATM business, offset in large part by foreign exchange. The annual improvement is primarily due to higher utilization and beneficial product mix offset by foreign exchange. We estimate that foreign exchange fluctuation had a negative 1.5 and 1.0 percentage point impact on gross margins on a sequential and annual basis, respectively.

Our operating expenses increased by $0.3 billion sequentially and $1.5 billion annually to $15.5 billion. The sequential increase in operating expenses is primarily due to higher consumption of factory supplies as our R&D activities ramp. The year-over-year increase in operating expenses is primarily attributable to increases in R&D staffing, factory supply consumption and other labor-related costs. Our operating expense percentage stayed flat sequentially at 10.3% and increased annually by 0.3 percentage points. Operating profit was $10.2 billion, up $0.5 billion sequentially and $1.2 billion year-over-year. Operating margin was 6.8%, up 0.3 percentage points sequentially and improved 0.4 percentage points year-over-year. During the quarter, we had a net nonoperating loss of $0.9 billion. Our nonoperating loss for the quarter primarily consists of net interest expense and net foreign exchange hedging activities, offset in part by profits from associates and other nonoperating income. Net interest expense for the quarter was $1.2 billion.

Tax expense for the quarter was $1.6 billion. Our effective tax rate for the quarter was 17%. Net income for the quarter was $7.5 billion, representing a decrease of $0.1 billion sequentially and a decrease of $0.3 billion year-over-year. On the bottom of the page, we provide key P&L line items without the inclusion of PPA-related expenses. Consolidated gross profit, excluding PPA expenses, would be $26.2 billion with a 17.4% gross margin. Operating profit would be $11 billion with an operating margin of 7.3%. Net profit would be $8.3 billion with a net margin of 5.5%. Basic EPS, excluding PPA expenses, would be $1.91.

On Page 8 is a graphical presentation of our consolidated quarterly financial performance. On Page 9 is our ATM P&L. The ATM revenue reported here contains revenues eliminated at the holding company level related to intercompany transactions between our ATM and EMS businesses. For the second quarter 2025, revenues for our ATM business were $92.6 billion, up $5.9 billion from the previous quarter and up $14.8 billion from the same period last year. This represents a 7% increase sequentially and a 19% increase annually. On a U.S. dollar basis, our ATM revenues were up 13% sequentially and 23% annually.

Gross profit for our ATM business was $20.2 billion, up $0.6 billion sequentially and up $3 billion year-over-year. Gross profit margin for our ATM business was 21.9%, down 0.7 percentage points sequentially and down 0.2 percentage points year-over-year. The sequential and annual margin declines were primarily due to NT dollar to U.S. dollar appreciation and, to a lesser extent, higher utility rates offset in part by efficiency from higher loading. On a constant currency assumption, we estimate our gross margin would be roughly 2.2 percentage points higher during the quarter, within our original margin expectations for the second quarter. During the second quarter, operating expenses were $11.4 billion, up $0.1 billion sequentially and $1.5 billion year-over-year. The sequential increase in operating expenses was related to slightly higher labor costs from workdays. The annual increase is primarily the result of R&D ramp-up and labor-related expenses.

Our operating expense percentage for the quarter was 12.3%, decreasing 0.7 percentage points sequentially and down 0.5 percentage points annually. The sequential decrease was primarily related to higher revenues on relatively stable operating expenses. We continue to target to lower our operating expense percentage. However, given the foreign exchange environment, the level of anticipated decline in percentage may be somewhat impacted. During the second quarter, operating profit was $8.8 billion, representing a sequential increase of $0.5 billion and an annual increase of $1.6 billion. Operating margin was 9.5%, down 0.1 percentage points sequentially, while up 0.2 percentage points year-over-year. Without the impact of PPA-related depreciation and amortization, ATM gross profit margin would be 22.4% and operating profit margin would be 10.3%.

On Page 10, you'll find a graphical representation of our ATM P&L. On Page 11 is our ATM revenue by 3C market segments. You can see here that the Computing segment continues to become a relatively larger component of our business. This was largely driven by a higher percentage of LEAP-based revenues. From a wider perspective, it is representative of AI's growing share of the electronics market.

On Page 12, you will find our ATM revenue by service type. Here, you can see the 2 service types containing LEAP services, bump and flip chip and testing. Both are becoming a larger component of our overall business. We continue to expect growth in these areas. It should be noted that we are starting to see a more visible pickup in our wirebond business. There are signs that this is related to a more general market recovery. On an absolute dollar basis, our wirebond business grew on a U.S. dollar basis but was outpaced by LEAP and testing.

On Page 13, you can see the second quarter results of our EMS business. The annual seasonality of our EMS business has been inconsistent over the last couple of years due to differing device ramp schedules. As such, we believe the annual comparability of our second quarter results may be impacted. During the quarter, EMS revenues were $58.8 billion, declining 6% sequentially and 7% year-over-year. The sequential decline was primarily the result of underlying device seasonality.

Sequentially, our EMS business' gross margin improved 0.5 percentage points to 9.4%. This change was principally the result of product mix. Operating expenses within our AMS business increased slightly by $0.1 billion sequentially and declined $0.1 billion annually. Our second quarter operating expense percentage of 6.9% was up 0.6 percentage points. Annually, our EMS operating expense percentage was up 0.4 percentage points on lower revenues. Operating margin for the second quarter was 2.6% flat sequentially and down 0.5 percentage points year-over-year. The annual decline was primarily due to lower revenues. Our EMS second quarter operating profit was $1.5 billion, down $0.1 billion sequentially and $0.4 billion annually. On the bottom of the page, you will find a graphical representation of our EMS revenue by application. As you can see, the second quarter mix of application revenue was relatively steady sequentially.

On Page 14, you will find key line items from our balance sheet. At the end of the year, we had cash, cash equivalents and current financial assets of $76.9 billion. Our total interest-bearing debt increased by $8.5 billion to $240.1 billion. We continue to anticipate increasing our debt outstanding throughout the year. Total unused credit lines amounted to $355.3 billion. Our EBITDA for the quarter was $27.4 billion. Our net debt to equity this quarter was 52%. As a reminder, we anticipate that our net debt to equity will be peaking this year during the third quarter.

On Page 15, you will find our equipment capital expenditures relative to our EBITDA. Machinery and equipment capital expenditures for the second quarter in U.S. dollars totaled $992 million, of which $690 million were used in packaging operations, $251 million in testing operations, $49 million in EMS operations and $2 million in interconnect material operations and others. In addition to spending on machinery and equipment, during the quarter, we also spent $531 million on facilities, which includes land and buildings. We continue to see the complexities of semiconductor design requiring step-ups in our LEAP offerings. Progressing device memory, thermal and power requirements in addition to traditional bandwidth expansion continue to necessitate advancements in our capabilities, equipment and facilities. Our packaging products are now more than ever on the critical path of chip design.

As we get closer to 2026, we are seeing a number of initiatives starting to activate. Aligning with customer requests, we are trying to be more aggressive with time lines and as a result, we are potentially seeing some of the capital expenditures slated for 2026 being accelerated into the fourth quarter of 2025. At this point, the delivery and installation schedules are still fairly dynamic. But we are potentially looking at a bump up in 2025 capital equipment expenditures by a few hundred million dollars.

With that, I'll hand the presentation over to Joseph to give the outlook for the coming quarter.

Thank you, Ken. Let me give you the guidance for the third quarter. Based on our current business outlook and the exchange rate assumption of USD 1 to TWD 29.2, the management projects overall performance for the third quarter of 2025 to be as follows. This time, the guidance will be given in both U.S. dollar terms as well as in NT. So it's a little bit more complicated, so please bear with me.

At the holdco consolidated level in U.S. dollar terms, consolidated third quarter revenue should grow by 12% to 14% quarter-over-quarter. Whereas in NT dollar terms, our consolidated third quarter revenue should grow by 6% to 8% quarter-over-quarter. Our consolidated third quarter 2025 gross margin should decrease by 1 to 1.2 percentage points quarter-over-quarter. Our consolidated third quarter 2025 operating margin should decrease by 0.1 to 0.3 percentage points quarter-over-quarter.

Now coming down to ATM. In U.S. dollar terms, our ATM third quarter 2025 revenue should grow by 9% to 11% quarter-over-quarter. While in NT dollar terms, our ATM third quarter revenue should grow by 3% to 5% quarter-over-quarter. Our ATM third quarter gross margin should decrease by 0.9 to 1.1 percentage points quarter-over-quarter.

EMS. In U.S. dollar terms, our EMS third quarter 2025 revenue should grow by 18% to 20% quarter-over-quarter. In NT dollar terms, our EMS third quarter revenue should grow by 12% to 14% quarter-over-quarter. EMS third quarter 2025 operating margin should increase by 0.3 to 0.5 percentage point quarter-over-quarter. That is the overall guidance for third quarter.

With that, I would like to also make very short comments on our margin. Well, on top of the overall higher cost environment that we're working in today, the NT dollar appreciation since early May put further pressure on our margin and will have a 5 percentage point negative impact on ATM third quarter '25 gross margin. If excluding such currency impact, our ATM gross margin should be around 26%. We're near the midpoint of our structural gross margin as originally targeted during our last earnings call.

Now looking forward, as we continue to improve our costs through efficiency improvement, leveraging our scale and capabilities to align our pricing and investment strategies with our value proposition and to aggressively expand our leading-edge packaging and testing business and start easing up on our early-stage ramp-up costs. We are very confident that we will get back to our structural margin range in 2026.

With that, thank you very much.

Thank you, Joseph. During the Q&A session that follows, we would appreciate if questions can be kept concise and asked one at a time. I will be receiving each question and repeating the asked question to Joseph and Tien. Again, we will be limiting the number of questions asked to 2 per turn, but ask one at a time. We will start off by taking questions from the attendees on the floor. After some of those, we will look to the virtual queue to see if we have some questions online.

So my first question is about Dr. Tien your comments about your fab is going to be very busy in second half. And actually, your ATM guidance is pretty good, right? How do we kind of reconcile with the sort of comments about PC, smartphone, automotive seems to be very slow. Is that because of ASE's share gain or any other factors that we don't consider.

Charlie, you're asking about the general dynamics of variable markets or...

Yes. So end market seems to be pretty slow. I mean, excluding the AI, right? But your guidance and also the comments about your fab utilization seems to be very busy. I just wanted to get a sense how do we reconcile your performance and also the end market weakness?

You want some consolidate -- some reconciliation between what people believe and what we're experiencing.

Well, we're in a much better position to look at our forecast with our customers. We're providing the guidance based on the customers' input. Those orders are firm and it covers the AI. It also covers other area, for example, the wireless, also the industrial and automotive. So I'm not going to go to the second half forecast on the segments, and I don't think we're in the position to do that. But right now, based on the forecast committed order, we do see a very strong -- I wouldn't use the word very strong. We do see the word strong -- or as strong as Q2 outlook.

Charlie, your second question?

Actually, if I can add, I think we are seeing a very strong demand for HPC and AI, of course. But on the general market, we're also seeing a healthy recovery in the second quarter. Actually, in all sectors, we're seeing actually double-digit kind of growth quarter-on-quarter on each different segment. So we are seeing a recovery in the general market as well along with the hyper growth in terms of the leading edge.

And my second question is more focused on your advanced packaging and testing business because foundry, TSMC revised the full year, and they kind of attribute that upward revision to the strength in AI and HPC. So why company doesn't revise up your kind of additional $1 billion revenue guidance?

Charlie, you're asking for -- actually, I don't understand what you're asking for. Can you rephrase this question?

Yes. So TSMC revised their full year, right, attribute to AI strength. But you maintained your kind of revenue increase from the advanced packaging remains to be at USD 1 billion. So is that because of any conservatism or why TSMC revised up and you don't?

Charlie is asking about whether we can comment on a leading-edge advanced packaging outlook for the year?

We're in a very interesting position because our capacities are full. The incremental capacities are new capacity that we put it in. I hope it answers your questions. We're capacity constrained right now.

So why don't you revise up the CapEx in?

There's a little thing called execution, operation, human talent, land, space, facility and machine delivery. We're doing the best we can. I think I already made a comment that the Q4 well into 2026, many of the customer pipeline require similar foundational technology capacity. So it's not a matter of we are reluctant to put in more CapEx. But we also be mindful like any machines, you don't want to overstress it on top of the -- we do have to put in some buffer for typhoon for all of this, right? So hopefully, we can do a better job, but we're very, very busy.

Next question, Bruce? Name and company, please.

Bruce Lu, Goldman Sachs. I think I want to -- the first question, I want to clarify something that Ken, when you did the prepared remarks, you were talking about certain like prioritized initiative talk with the customer. Last time when I remember when ASE was talking about this, a little about the SiP project, by them, try to prioritize those projects. Given the current currency environment, do we expect a new pricing strategy? Or do we expect to get a different -- set a different bar for the future project? For the future business discussion, does that only goes to like new SiP business or the existing business with the foundry partner or the assisting new business when we needed better pricing, otherwise? You might not want to do the business. So can we expect the margin upside moving forward?

So Bruce, you're asking for us to elaborate on our strategic initiatives.

Yes.

There are many aspects on the strategic initiative. Some of the items, I cannot go over in detail because once I said it, you understand exactly what I'm talking about. I think the SiP project as well as many that becomes a resource recalibration, we're looking at the floor space, we're looking at the cash flow, capacity, investment and most importantly, we're looking for the resource deployment. Now in the resource recalibration, obviously will come in the optimization. And the optimization by default will improve the end results, which will be a margin improvement. So that's one aspect.

The second thing, the pricing strategy that has always been on the table. It really depends on the customers, also the future product as well as the timing, for example, for the RFQ, new product, old product, investment requirement, the each one will be different. The third large item will be the overseas expansion. We were thinking about expanding in many strategic areas. But given the dynamics of the tariff as well as the exchange rate we have to rethink priority and also the dollar amount. So our approach is going to be instead of going to multiple places. We will focus on 1 or 2 overseas areas and make a much, much bigger investment just to make sure whatever we do counts, right?

So those are the recalibration that we're going through right now. Why all we're very, very busy investing in Taiwan and trying to build up with speed, scale as well synergy with the Taiwan ecosystems. But the important thing is whatever we put in, we understand the fluctuation of the market, but we tend to look at the basic characteristics. What are the foundational requirements we would like to put in capacity that we believe will run for 7 to 10 years, and that's what we're building on now.

The second question is for your AI-related business, which is highly skewed to testing business for this year. Do we expect packaging will account for a larger portion of the business in 2026 and onwards? And if that would be the case, can the margin maintained at a similar level with both packaging and testing for the AI business?

Bruce, you're asking about our LEAP services headed into 2026.

The concept is to grow our turnkey business that covers the AI-related leading-edge. If you only look at the growth rate, it tends to be very misleading because the testing business has a larger base. We talk about the 31% first half and will probably be a better number in the second half. That will be the overall testing growth rate. If you just look at the leading-edge packaging, then the growth rate is much higher. But the concept is to grow this in tandem. However, you cannot look at it because the base is different.

But just to answer you briefly, we do intend to make investment on packaging as well as testing because the requirement tend to go in tandem. For example, once you go to chiplet, silicon photonics or any kind of HPC or AI related, then the thermal fluctuation that you have to deal with a whole lot more variables for the testing arena. And the lead time tend to be multiple cycles, much longer. You also have to deal with the interim testing to guarantee yield. With all of this practical concern, I think ASE has a very good position to run the leading edge, assembly packaging as well as testing just the nature of logistics and also the technology and the cash flow.

Can I have a very quick follow-up because we were guiding for $1 billion additional business for this year. Given the huge CapEx we invest this year, the incremental revenue will be a lot bigger than $1 billion for next year. Is that right?

You're jumping way ahead of me.

Yes, I don't think we have any comments out on '26 yet.

But Joseph used the comment about like capital to CapEx per revenue, right? If you use the similar terminology or similar methodology, you can provide some color, again, nothing that had these things.

I'm not sure I understand what you're asking. I'm nodding my head just to show my appreciation to your question. Anyways I think we do have -- we have been pretty aggressive in terms of investing into test. And the growth rate that test is showing actually shows or demonstrates the progress that we're making. For this year, we're still expecting growth of test to be twice the pace of packaging. In terms of test overall to ATM, we will be approaching 20% of ATM revenue being test by fourth quarter this year.

And going forward, I think aside from the -- right now, the leading-edge part of the test is about -- over 20% of the overall leading-edge revenue. And I think that ratio will continue to grow because right now, we're expanding our test mostly for leading edge at the wafer sort level. And in the later part of the year, we will be starting getting into final test burning included. So I think there's still a lot of potential for us to grab going forward, particularly in the test which overall will bring up our margin as well. So it's a very accretive business for us to further penetrate.

Does that answer your question, Bruce?

I got to say yes, right?

If we could switch it up and maybe take a question from virtual?

We have a question from Mr. Gokul Hariharan.

First question is on gross margins. So you mentioned that gross margin can get back to the structural gross margin levels, mid- to high 20s next year. Is that assuming that currency stays at this level? So if so, what are the variables that are contributing to that? Do you assume that there is a big increase in non-leading-edge utilization? Or is it just a function of better mix of LEAP, better testing and potentially better pricing?

Gokul, you're asking what levers are available to get back to our structural margin levels. Is that correct?

Yes, especially for next year. I think Joseph mentioned, I think for next year, we hope to get back there, right?

Yes. I was talking about our margin for next year, and we are very confident about coming back to our structural margin range in 2026, I think, through several different directions. One is, of course, we will continue our effort in improving our efficiency. That includes further automation of our overall operation to bring down our costs. And I think a lot of the early stage ramp-up costs will start to see that easing up. And we're actually saying last time that our operating expense ratio will start to level off because initially, we're putting a lot of R&D dollars into our investment.

So as we continue to grow or expand our leading edge, I think we're going out of the ramp-up stage, I think that part of the expense can be more controlled. And of course, all the leading-edge businesses, including tests are margin accretive. So all these put together, I think it gives us a high confidence level that going into next year, we will be able to get back to our structural margin range. Of course, that's based on the assumption that the currency doesn't further appreciate from the TWD 29.2 or TWD 29 level at this point.

And just a follow-up, how much of a lever can be pricing given for LEAP, you are pretty much the only vendor out there spending so much money. Your competitor doesn't really seem to be spending much CapEx. Your foundry partner seems to be a lot more aggressive about selling their value as well. So I just wanted to understand how you were thinking about this, given your market position today seems to be a lot better than maybe 5 years back.

Gokul, you're asking about our -- what we perceive as our market position within LEAP. Is that correct?

And leverage on pricing as a result of that market position.

The pricing power is what every company wants to have. The pricing power includes your design, attachment rate, long-term loyalty, trust as well as your technology and overall solutions. I think I talk about the current scale and the speed and synergy of ASE being in Taiwan as well as being supported by many ecosystem players and their foundry partners. I think we need to clearly demonstrate our trust and also the overall solution support to our key customers. Pricing is always an option. However, we prefer to use technical strength, offer long-term growth prospect, which I think is much, much longer lasting. However, everything is being considering. And by recalibration and resource the optimization, that by itself is one pricing strategy.

Gokul claims that, that's a follow-up question. Do you want to let him on for his second question?

Second question is on the LEAP platform. Dr. Wu, can you talk a little bit about application of 2.5D, 3D packaging, whether it is CoWos like, PoCoS, FOCoS-Bridge, et cetera. For applications beyond AI accelerator, it definitely looks like '26, '27, you start to see a lot more of these applications coming to you for CPU or other HPC applications. Could you talk a little bit about what you are seeing and how important these are going to be in terms of the growth going in '26 and '27, given this year mostly, it seems for AI accelerators only?

Gokul, you're asking for Dr. Wu to expand upon our LEAP offerings?

Yes, beyond AI accelerator, yes.

Well, I talk about the first wave is the accelerator. I think the second wave, you will see more infrastructure people. You all see the ASIC people start coming in. But however, they all share the same characteristics. They would like to have very, very tight configuration and very, very large design footprint and also very efficient power delivery systems and also the transmission, high bandwidth, low latency. I do see this round of AI accelerator that were cascaded to the CPU, yes. And we'll go to the ASIC, yes. And in the future, for the AI edge applications, I believe most of the devices will share the similar multifunctional heterogeneous integration that will demand much of the characteristics I just talked about it. So I believe the volume, we obviously are optimistic that the volume will be increasing over time.

Question from the floor. Sunny?

Sunny Lin from UBS. So my first question is on advanced packaging, especially on the leading edge, just want to learn your strategy on how to scale the business into 2026 and 2027. I believe for 2025, most of your cell supply for packaging may be through the outsourcing from foundry. And so going into 2026, where are you in terms of the progress to ramp full process CoWos or FOCoS? And how should we think about the further upside from the foundry outsourcing.

Sunny wants us to describe in a little bit more detail what we're going to do with our LEAP.

Well, the foundry outsourcing will continue pending on the foundry partner as well as our end customers requirement overall. So that is ongoing. The second layer is going to be the ASIC players. The ASIC players will make their choice and they can either go through different routes. And also, there will be a peripheral packaging and testing requirements on the other devices that will also go into the same infrastructure. And then there will be the CPU guys. I don't think I can go into even more details, but if you really believe that the AI is a new paradigm shift, if you look at how many players are really embracing the AI and how many other people are jumping on it, you can pretty much figure out, we are still seeing the first wave. Even the CPU, the ASICs, whatever we can name today, it will be the first wave.

The real definition of the paradigm shift is, there will be others that we cannot see. That's the definition of paradigm shift. So I think for that, building an efficient, viable and flexible infrastructure today at a speed is extremely critical, which is why ASE management choose their hard route that we want to accelerate investment and then just take the heat. Even with the risk, we might miscue it. Also put a lot of stress on the supply chain. But we believe in this round, gaining customers' confidence will dramatically improve the long-term royalty as well as explaining to our customers, not just the first wave, the second wave and the third wave, you would need to have a footprint, flexibility. That's where the large panel, the 300 x 300 fully automated line will be put in by the end of this year.

The 600 x 600 fully automated line will be also delivered by Q4. And then silicon photonics, we have been extremely vocal. And also on the new power management, the VRM solutions, we're working with many of our key customers trying to define at IP level, the technology level, also at a capacity level to the next generation. So I think this campaign is not just CoWoS or FOCoS. It's really the overall the solution. And I think once in a lifetime, you will encounter something like this. So I'm personally very excited.

For sure. So maybe let me try to ask the question maybe in a different way as a follow-up, if I may. And so for us to think about 2026 for your advanced packaging cells, is it fair to assume that your full process, advanced packaging could account for a meaningful portion of the opportunity, given your current customer engagements?

That will be the aspiration. But of course, you have to understand the customer loyalty means it's a trusting partners, whatever people ask us do, we will try to fulfill it. But different customers will have different choices. And the market will take us to the rightful place. And we should not look at just 2026. You should really look at the multiple years. I think that statement is very true sometimes.

I think just to give you a short answer, I think we are in full production in terms of full process. We do have a couple of customers that are using our capacity. We are investing for making further investment into full process. We do expect to have some meaningful business coming in next year, mostly from ASIC customers.

My very quick second question is on testing for AI. Just want to learn more on how you think about the competition. Obviously, the key competitors continue to be very aggressive in the in-house solution, whereas you may lean more toward a third-party platform with Advantest. And so how should we think about the progress of you with the clients? I'm sure at some point ASE will be a very important supplier for final test for AI. But how should we think about the timing for you to gain meaningful market share?

Sunny, you're asking about our test position related to final test on AI.

I think we have been talking about this for a long time, and we are expecting to have some final test revenue coming in by maybe the fourth quarter. We'll continue to expand that part of the business going forward into next year. And I'm sure we will gain a meaningful market share in this aspect as well. I think everybody saw some short report saying whatever about our competitor was. So I'm not going to comment on our competitor. I'm just saying that what we're doing ourselves, we do have a goal, we do have a plan. We are making progress. And we believe that whatever we invested is going to be fruitful for us.

We have another question on the floor. Brad?

I'm Brad Lin from Bank of America. So I got two questions. One is that, well, Ken, you just mentioned heterogeneous integration. So basically, I think ASE has a very unique position. It has ASE, SPIL and also USI. So how will that help you maintain a competitive advantage across multiple potential application in the future in AI, including humanoid robots, HAI? And then if available, may you share any initial view on the time line for those to happen.

Brad is asking about the positioning of our brands, including our EMS side of the business in terms of whether this allows us to tackle incremental AI opportunities.

You probably know that SPIL and ASE ATM business, we're running independent. And one of the advantage of that is the we're free to choose our own customers. And what so happened is, right now, the AI initial set of customers are extremely happy because each one have a different focus group with a very independent firewall in between, if we may call that advantage. The second one is on the USI. I think with the SiP, we clearly demonstrated a synergy and also the vertical capability for our designing as well as manufacturing also logistics perspective.

For the AI, I think we're waiting for that opportunity. For example, power management is a key area because the power management, you go through 2,000 volts all the way to 0.1. And then the assembly people have their sweet spot. And then the EMS team will also have their sweet spot. If you take this kind of a hierarchical approach into the PCB, into a different kind of architecture, you can derive the similar thesis.

So over time, it is the value between ASE, SPIL and USI and painting on the market demand. I think we've demonstrated this a few times, and I do believe in the AI, you will see more optical, you will see more power management that will bring the synergy back again. On top of that, you have the China, you also have the other part of the world. You have the U.S., you have the China alliance. I think that presents another competitive advantage if we decided to go that route.

Just one comment on the testing portion. There's one thing I want to articulate is, testing is not about wafer sort and final test. If we only fixate on the wafer sort and the final test, I think we're missing the point. The point is, in the future architecture, the testing needs to be an integral part of the processes. And the question is, how do you do this, on a discrete format or an integrated in situ format? I am giving you a 10-year outlook on the overall manufacturing strategy for very, very, highly evolved, automated, complicated assembly and testing process.

Thank you for the explanation. So with that comment, may we assume that actually advanced testing will greatly outpace the so-called advanced packaging in that view because that will be integrated into, well, a lot of the new process?

Don't take my 10 years comment back to 1 year. I'm always struggling with people, right? The technology direction is correct. The testing has a larger revenue base. If you really want to compare initial, of course, you can make that statement. However, you have to look at the overall testing business percentage as well as the nature and how do you define the segments, then it is very difficult to make that comment. But the overall trend, yes, I mean, the heterogeneous integration, a very, very long cycle time on the FOCoS and the CoWoS process, and also the value of the components and also the yield impact at the module level. And these are the issues that defines and segregates the people who can play in this market versus people who cannot afford it. But this all becomes competitive advantages in any kind of business.

So my point is, when you evaluate the overall ecosystem at a worldwide level, at a geographical level, you have to put all of this into consideration. The AI has initiated a paradigm shift in terms of algorithm and application. Accordingly, each country, each geography, each company needs to evaluate their own paradigm shift in accordance to this macro paradigm shift. Maybe I'm over stretching, but I think this is the kind of thing I'm very excited to see in the next 10 years what will play out.

So my second question would be a little bit on the financials. So basically, we know we are putting a lot of the CapEx in the LEAP business. And I believe, well, that would not be a problem for gross margin because it's margin accretive. But well, if available, can you kind of share some of the initial thoughts about how the growth rate of the depreciation that we are looking at in this year and maybe next year?

Right, you're looking for some guidance on what depreciation looks like, right?

I think that for this year, I think given the heavy investment in our CapEx, I think the depreciation percentage is going to -- depreciation is going to grow about 14% from last year. At the holding -- I think at the ATM level, it will reach about TWD 69 billion for the year. That's at holding level, right? At the ATM level, it's about TWD 59 billion. I think financially, I think we will see this heavy investment for some time. And we are currently funding our investment through additional debt. As you can see in second quarter, already, we are seeing our net debt-to-equity ratio has come up quite a bit. We expect this to peak actually in the third quarter as we put in additional debt to our balance sheet.

And in the third quarter, we should see a ratio roughly below 70%. And then it will start to come down as our target level remains to be 60% to 65%, which is the level that we feel comfortable with. And we're seeing that peak at third quarter and start coming down on a quarterly basis back to our more comfortable level of around 60% to 65%.

The next question we are taking from the virtual queue.

We have an online question from Laura Chen of Citi Group.

I have questions on the broad-based recovery on the wirebonding. What would be the key driver behind? Just curious especially, we see some of the pooling happening in the first half already. So just wondering that you mentioned there are some power management IC demand. Is that also related to AI or do we see any market share gain or HAI device or just simply the demand gradual recovery? That's my first question.

Laura is looking for a rationale as to why we're seeing or potentially seeing a wirebond recovery within our businesses.

The wirebond capacity is tight in Taiwan mainly due to 2 markets. The first one, it could be related to AI. As AI are building AI systems, it does require some of the chips. They are using wirebond to support those AI systems. But we believe the larger percentage of the demand actually comes from the automotive based on customers that we are serving and maybe they are using the ASE capacity because the ASE wirebond are highly automated, and we have the 100% fully automated line. Therefore, it gives you a better throughput and quality that could be. Our overseas wirebonder remains to be somehow idled. So we do see a disparity between the Taiwan wirebond as well as the overseas wirebond.

Just wondering, is that also kind of because of China or non-China, that kind of recovering trends or the reason that customers prefer to place more orders in our Taiwan fab, along with the probably main chip for AI or something like that?

You're looking at the -- due to the BIS regulatory control, some of the loading are migrating from China to Taiwan. We see some of that, but that is not a major variable in our report.

I think also the -- if it's BIS, I think most of the packages are more advanced rather than wirebond.

Okay. Sure. Also, I think Ken and Joseph, you probably already talked a little bit about that, but as we see a lot of leading-edge advanced packaging happening for different type of design, like CoWoS or panel-based or CPO, what will be the key consideration on your capacity preparation? And particularly, everything seems to be quite tight and AI chips development is very fast. So just wondering how you would plan your capacity and CapEx parity. And also for the final testing starting from Q4 this year, I'm just wondering, is that also including the burning and also the system-level testing?

So Laura, you're asking about our LEAP services and how we prioritize funding those LEAP services, including to some extent, the final test services provided for leading edge.

The investment on factory facility, that is very flexible. I mean that's the nature of OSAT, and that's what we do. A large portion of the equipment, they do have some intrinsic fungibility. In other words, why we're going through different code names. And I think you know the code names better than I do. Intrinsically, we are doing this every day, not just this year. So even for the advanced packaging, there's still intrinsic fungibility that any company needs to master, although more complicated. So the investment that we put in, we want to make sure our capacity has the fungibility, the flexibility which you can guard at 5 to 10 years of utilization, right? So that's one question.

In terms of the testing, I'm pretty sure our testing revenue will include burn as well as system level test. I won't be able to give you a percentage of that in detail. So as customers are requiring more ATC as well as thermal variation, I think that becomes one area included in the power management as well as the testing solution. I think that's part of the technology development that we're going through. In the right time, we will come out and announce the overall power management solution, including assembly and testing as well some of the IP innovation. Hopefully, we can schedule this sometime next year.

We have more questions from the floor. We have one additional question from Laura. No. We have one additional question. Are there any more questions on the floor? No? Bruce has a question.

I just wanted to clarify a couple of things. Number one is that when we talk about the overseas investment, U.S. is not in the shortlist. Is that right?

I'm going to be very careful answering these questions. Please go read our announcement. We have made clear announcement that we are supporting one of the large customers going into United States. I'm not going to say anything other than that. Anything related to U.S. investment, please read our press release.

Okay. The second thing is that Ken, can I double check that what is the latest CapEx forecast for the full year in '25 for equipment and total CapEx.

You're looking for the full -- an update on the CapEx.

For this year?

Sure.

I think Ken just mentioned that we are upping -- because the capacity right now is very tight and the demand is very high, so we are upping our machinery CapEx a few hundred million dollars. By a few hundred million, I means $300 million to $400 million. And the bulk of the -- I would think almost 90% of this increase is really for leading edge as we see demand continues to be very strong, particularly going into 2026. So I think the increase of CapEx is really trickled into the fourth quarter.

It's more geared to packaging or testing?

Both.

Joseph and I had a long debate on this. As a CFO, he tends to be very concerned about the operation guys are spending money like crazy. My advice to him is, everybody wants to spend CapEx. The only people who has order can spend the CapEx. The question now is the -- that puts a lot of burden on execution. If the operation team cannot fulfill the things online, it becomes a problem. We understand that. So the CFO concern is legitimate and right on. The team collectively decided we're going to go through this CapEx ramp mainly because our partner as well as our key client for the next 10 years, they would like to have a trusted partner that can execute. The speed of that will define how are we going to invest further in Taiwan as well as outside of Taiwan. So this goes to a long way. Therefore, we decided to bite the bullet and just go through that.

I think to give you a little bit color. I think for this year's machinery CapEx, roughly 60% will be for assembly and 30% for test and 10% for others and material. That includes material and EMS, right? And also on this machinery CapEx, I think about 60% is for leading edge.

Don't look at the numbers too much because the machine delivery, it can change. The execution, including building new buildings, hiring new people, I mean there's a lot of variable. But right now, we do have a high aspiration to do this.

I believe we have one more question from virtual.

We have a question from Gokul Hariharan of JPMorgan.

So Tien, just wanted to zoom out and think about this. This is the first time I've seen ASE enter this like massive CapEx investment and hyper growth kind of stage. And I think as you discussed earlier, I think it's a very careful consideration that the management team has placed. Where do you want to be, let's say, in 3 to 5 years' time? I think ASE is -- I think if I just look at ATM, it's a $12 million revenue operation, maybe mid-30s market share in OSAT. What constitutes success of this plan? Let's say, if we roll out maybe 3 to 5 years, like what does a successful ASE look like at the kind of culmination of this hyper investment phase?

Okay, you're looking to see how we define our own successes within the company?

Yes, I'm sure you are spending billions of dollars, so you do have some targets in mind. So just wanted to understand what do you define success as at the culmination of this heavy investment phase?

I think this is the exercise we're going through. For example, if you look at the traditional OSAT, you can define the revenue, the margin and the CapEx you have some ratio. If you look at the key foundry supplier, their CapEx on their assembly and test will have a different ratio. So we're going through is to take this from the traditional OSAT ratio to the foundry ratio, and we're climbing stairs. Then the question is, why are we doing this? Well, because the foundry partner wants you to do this because their end customer wants both of us to do this.

The question now is, who has the capability to execute this to the scale that becomes a competitive advantages over time. This includes not only the key players, also includes the whole ecosystem supporters that will enable you to do this. In terms of the margin model, I think you can apply equally. How do we define success? Going through the execution without issue and earn the larger clientele base on the longer-term basis, that will be the success.

All right. Gokul, do you have any more questions there?

Yes. So since you referred your foundry partner, in that process, their own margins have gone up maybe 10% or more. Their growth rate has accelerated quite a bit. So how do we think about our margins? I think you've talked about structural margins heading to mid-20s to high 20s. Is that where we get to? And is there any expectations on growth? Like does $12 billion become $20 billion in 5 years, is that our definition of success in terms of ATM revenue growth?

Well, I'm sure everybody knows that there's still a lot of moving parts and uncertainties in front of us. The world is changing very fast. Right now, we're not trying to reinvent the wheel. I think from what we learned, from what we know our own business, we are keeping our structural margin range as it is for now. And we just will see how situation changes. And if we need to make a change, we'll make a change. But at this point, there's nothing structurally different from what we originally expected.

I think we've hit our hard cutoff time frame. It's 4:30 over here. Thank you very much for attending our earnings release. See you next time.

Thank you.