Hello. I am Ken Hsiang, the Head of Investor Relations for ASE Technology Holdings. Welcome to our Second Quarter 2021 Earnings Release. Thank you for attending our earnings presentation 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 disconnect 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, our dollar figures are generally stated in New Taiwan Dollars, unless otherwise indicated. As a Taiwan-based company, our financials are 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 in Chinese GAAP. I am joined today by Dr. Tien Wu, our COO, and Joseph Tung, our CFO. For today's call, Dr. Tien will first give a mid-year update. I will then be going over our financial results. Joseph and Tien will then be available to answer questions during the Q&A session. I would like to now hand the floor over to Dr. Tien Wu.

Thank you, Ken. Hello, everyone. Thank you for joining our conference call. 2021 has been an exciting year for all of us. I wish everyone safety and health. For today's conference call, I would like to give you a brief report on two items. First, I would like to give you a business update, including the second quarter and first half achievements, then the third quarter and the second half outlook, which will touch on business sentiment towards 2022. Second, business outlook, short term and long term. In the past weeks, there have been reports from several key customers and partners with some conflicting signals, causing some speculation about the current state of business and the landscape. I would like to give you ASE's perspective, so you have another angle in solving the puzzle. To begin with, I would like to give you our business update. Our second quarter '21 and first half '21 revenue and margin are both on track. Second quarter '21 ATM revenue grew 8% quarter-over-quarter in US dollar terms. First half '21 ATM revenue grew 20% year-over-year. If you exclude the EAR affected revenue, it will be 48% year-over-year. First half '21 test revenue recovered ahead of schedule. First half '21 revenue grew 54% year-over-year, if we exclude the EAR affected business. So starting from the third quarter, we will see assembly and test business both grow. Second quarter '21 HoldCo revenue grew 77% quarter-over-quarter. First half '21 HoldCo revenue grew 28% year-over-year. First half '21 HoldCo operating margin improved 2.7 percentage points year-over-year. Let me talk about the second half '21. We expect third quarter and fourth quarter, quarter-to-quarter revenue and margin improvement, as we have previously indicated. We're seeing very strong ATM demand compared to our previous target. Our last guidance estimated semiconductor will grow 10%, and we will do better than twice that. Right now, our sentiment is better than our previous guidance. The momentum will last into 2022. First half '21, ATM gross margin is at our full-year target, up 25%. In other words, we have achieved our full-year target in the first half. Therefore, we expect further gross margin expansion in the second half, Q3 and Q4. HoldCo 2021 operating margin target should exceed or be at the higher end target of 2.5% to 3%, as we previously guided. Let me turn to the business outlook. I will first talk about the short-term demands indicating a strong 2022 with another better than seasonal first quarter. As you know, the first quarter of 2021 has been stronger than all our expectations. We're looking forward to another strong Q1 in 2022. Many customers are extending their long-term service agreements beyond 2022 into 2023. Let me comment on the expansion, which has been a question from many people. Capacity expansion needs to consider holistic and balanced supply continuity across the complete material, equipment, and process ecosystem. Our estimate is that the earliest the balance of demand and supply will occur is sometime in 2023. In other words, in 2022, we still need to be very smart and efficient in managing the bottlenecks. Next, people ask about double booking and inventory control which may exist. However, that should be localized and temporal with the overall demand profile, with very little impact on the overall business momentum, at least from ASE's OSAT perspective. Next, I would like to comment on the long-term business outlook. What I'm trying to do is to share with you ASE's perspective, and maybe ASE's OSAT perspective on the longer-term outlook. On this page, I have a diagram of a pyramid. What I'm trying to illustrate is a conceptual framework about the current state of the semiconductor business. As you know, the semiconductor business is mainly driven by innovation. Just imagine that innovation is driven by technology, located at the tip of the pyramid. As innovation becomes more pervasive, the pyramid grows taller; to support a larger pyramid, the length, width, and height must all increase proportionally. This is not an exact description of our ecosystem; however, conceptually, it may help you understand. Today, we have two driving forces. The first one is what the industry has recognized for the longest time, including 5G, AI, EV, IoT, smart manufacturing, etc. For this type of innovation to become widespread in scale, a new infrastructure must be developed, which will stimulate new demand for the system and therefore demand for all semiconductor devices. However, in the last two years, unexpectedly, we have faced the impact of COVID-19. The COVID-19 pandemic has put a significant demand on existing systems without the need for new infrastructure. The industry is accustomed to building up capacity at a slower pace while developing the necessary infrastructure. However, COVID-19 leveraged existing infrastructure and created a sudden demand for a large quantity of new systems. Thus, the industry has been caught off guard. The impact of COVID-19 may last from two to five years; we truly do not know how long it will last. What we do know is that we are in the short term. Consequently, the industry is responding accordingly by building up wafer capacity and assembly and test capacity. The entire supply chain is ramping up its capacity accordingly. This has provided a significant incentive for the industry to start developing a manufacturing infrastructure. Even if the impact of COVID-19 dissipates in two to five years, there will be a new wave of innovation—characterized by 5G, AI, EV, IoT, and smart manufacturing—in which demand for IoT devices, electric vehicles, and autonomous driving will require a new infrastructure and entirely new systems. Thus, our perspective is that the semiconductor industry is healthy. In the short term, we have a strong incentive to build our capacity to accommodate the system requirements driven by COVID-19 while also preparing for future demand increases driven by a new paradigm shift. Moving to the next page, let me discuss three tailwinds from our perspective. The first is consolidation. The supply chain constraints have pressured the industry, forcing our customers and their buyers to adopt more standard, flexible, and secure supply alternatives. This benefits open platform service providers like Foundry and OSAT. Long term, this thesis suggests that OSAT and Foundry will gain more market share as they consolidate over proprietary suppliers. The second tailwind is the Taiwan cluster. The efficiency, economies of scale, and supply chain flexibility of the Taiwan cluster have long been recognized. Over the last few years, the Taiwan cluster has made heavy investments in CapEx, including ASE. The ASE/SPIL merger and the resulting synergies exemplify the efficiency of the Taiwan cluster. With such efficiency and additional CapEx investment, more customers are selecting the Taiwan sector as their preferred choice, thereby creating a positive or virtuous cycle. The final tailwind is ASE HoldCo. Today, ASE Holding Co. has demonstrated clear leadership in terms of scale, market share, margin, and efficiency. We have a solid understanding of how the new wave of 5G, autonomous driving, and smart manufacturing will drive heterogeneous integration, including silicon with non-silicon sensors. Our understanding of future AI, big data-driven manufacturing, and automation is clear. We are currently the de facto choice, and an indispensable manufacturing partner for the semiconductor ecosystem. With that, I thank you for listening. I will turn the floor back to Ken. Thank you.

Thank you, Dr. Wu. I will now go more in depth into our financial results. First off, I would like to clean up an order of business that needs a bit of explanation for the sake of reporting transparency. As you all know, our subsidiary USI completed its acquisition of Asteelflash in 2020. Given the complexities of the purchase price allocation process or PPA, IFRS generally allows companies up to a year to complete this valuation process. After the valuation is completed, a retroactive adjustment is usually made. Asteelflash's purchase price allocation was completed during the second quarter. Accordingly, we have retroactively adjusted our balance sheet by NT$0.4 billion, representing 0.1% of our total assets as of the first quarter. On our P&L, the purchase price allocation results in incremental expenses booked into the first quarter totaling NT$88.5 million or NT$0.02 per share. First quarter consolidated holding company reported gross margin has been reduced by 0.1 percentage points, while operating margin has been reduced by 0.2 percentage points. For the second and future quarters, the PPA impact on net income will be approximately NT$37 million per quarter. Impacts on future growth in operating margin will depend on future revenues, but in the current period, such impact is considered negligible at less than 0.1%. This will be added to our quarterly PPA adjustment. Please turn to Page 7 where you will find our second quarter consolidated results. Intercompany transactions between our ATM and EMS businesses have been eliminated during consolidation. For the second quarter, we recorded fully diluted EPS of NT$2.30 and basic EPS of NT$2.40. Consolidated net revenue increased by 6% quarter-over-quarter and by 18% year-over-year. This sequential increase was primarily driven by our ATM business. We had a gross profit of NT$24.8 billion with a gross margin of 19.5%. Our gross margin improved by 1.2 percentage points sequentially and two percentage points year-over-year. Both margin improvements are principally the result of a higher ATM business mix, offset partly by NT dollar appreciation. NT dollar appreciation had a negative 0.3 percentage point impact on sequential gross margin and a negative 2.1 percentage point impact on year-over-year gross margin. Our operating expenses increased by NT$0.6 billion to NT$11.6 billion sequentially. Our operating expense percentage was flat at 9.2% sequentially, with a decline of 0.5 percentage points year-over-year. For the year, we're now expecting improvement rather than targeting maintenance at last year's 9% level. Operating margin increased by 1.3 percentage points sequentially and 2.6 percentage points year-over-year to 10.4%. During the quarter, we recorded a net non-operating gain of NT$0.2 billion. This primarily consists of gains related to our foreign exchange hedging activities, investments, and asset sales offset partly by net interest expense of NT$0.6 billion. Tax expense for the quarter was NT$2.6 billion, with an effective tax rate of 20%. For the third quarter, we expect to record our annual undistributed earnings tax. For modeling purposes, please use an effective tax rate of 21% for the third quarter to account for such tax impact. Net income for the quarter was NT$10.3 billion, representing an increase of NT$1.9 billion sequentially and an improvement of NT$3.4 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 NT$25.7 billion with a 20.3% gross margin. Operating profit would be NT$14.4 billion, with an operating margin of 11.3%. Net profit would be NT$11.5 billion with a net margin of 9.1%. Basic EPS excluding PPA expenses would be NT$2.67. On Page 8 is our ATM P&L. It is worth noting here that the ATM revenue reported here contains revenue eliminated at the holding company level related to intercompany transactions between our ATM and EMS businesses. As Dr. Wu indicated, our ATM business looks very healthy for this year and going into 2022. For the second quarter of 2021, revenues for our ATM business were NT$79 billion, up NT$5.2 billion from the previous quarter and up NT$9.5 billion from the same period last year. This represents a 7% increase sequentially and a 14% increase year-over-year. Our ATM revenues came in ahead of our expectations. On a US dollar basis, our ATM revenues grew by 8% sequentially. Gross profit margin for our ATM business was 25.6%, up 1.2 percentage points sequentially, and 3.9 percentage points year-over-year. The sequential gross margin improvement was primarily due to higher loading. The year-over-year gross margin improvement was primarily the result of higher loading, improved efficiency, product mix, and a more favorable ASP environment. ATM gross margin improvement was achieved despite NT dollar appreciation having a negative 0.5 percentage point impact quarter-over-quarter and a three percentage point impact year-over-year. We expect to deliver quarter-on-quarter improvements in ATM gross margins during the last half of the year even with ATM gross margin for the first half of the year already reaching our 25% full-year target. During the second quarter, operating expenses were NT$8.4 billion, up NT$0.3 billion sequentially, and up NT$0.5 billion year-over-year. The sequential and annual operating expense increase was primarily driven by increased employee bonus accruals, based on a profit-sharing model. Our operating expense percentage was 10.6%, down 0.4 percentage points sequentially and down 0.7 percentage points year-over-year. Operating margin was 15%, improving 1.6 percentage points sequentially and 4.6 percentage points year-over-year. The strengthening NT dollar had a negative 0.5 percentage point impact quarter-over-quarter and three percentage points impact year-over-year on our operating margins. Without the impact of PPA-related depreciation and amortization, ATM gross profit margin would be 26.7%, and operating profit margin would be 16.4%. On Page 9, you'll find a graphical representation of our ATM P&L. When we see our ATM gross margins here almost linearizing and hitting historical highs, I think it's fair to mention that we do not believe that our business is immune to future cyclicality inherent to electronics. However, we believe that having the scale synergies and the benefits of the four tailwinds mentioned by Dr. Wu will allow us to achieve margins with gradually higher peaks and shallower troughs. On Page 10, you will see our ATM revenue by market segment. You can see here a decline in our communications market segment, with shares picked up by automotive, consumer, and other products. Meanwhile, our computing segment has held roughly steady since 2020. From what we can see, our near-term performance has not been driven by communications-related devices. More importantly, with such a decline in our communication segment, the speculated widespread overproduction of communications-related components seems less likely. Our near-term performance has been driven primarily by growing consumer demand and general semiconductor expansion. This supports Dr. Wu's earlier statement that new technology and products create an expansion of supporting devices. On Page 11, you will find our ATM revenue by service type. There is generally too much noise in trying to understand each quarter's individual movement here. However, when the chart is taken as a whole, it tells a more complete story. You can see here the gradual improvement and underlying strength of our wire bond-related business. Meanwhile, services for advanced products have seen a gradual decline, some of which is due to the impact of the US EAR. However, we believe that our advanced services will rebound in the back half of this year. On Page 12, you can see the results of our EMS business and a graphical representation of our EMS revenue by application. The information provided on our EMS business may differ materially from the information directly provided by our subsidiary, as they report independently using Chinese GAAP. As mentioned earlier, the figures from the first quarter have been retroactively adjusted for PPA costs. Our second-quarter revenue usually represents the end of our seasonal trough. However, what is more unusual this year is that many of our customers are experiencing the impact of component shortages in the second quarter. This is the main reason why we saw our EMS revenues fall slightly short of our initial expectations. However, we believe that most of this revenue shortfall will get pushed out into the third quarter. The second quarter EMS expenses tend to be characterized by training investment and preparation, readying our factory lines for the third and fourth quarters when things ramp up to full mass production speed. This is especially true this year with two new factory locations ramping up during COVID. As such, we incurred extra operating costs related to R&D, logistics, and factory start-up costs in the second quarter to set the stage for growth in the second half. During the second quarter, EMS revenues increased by 3% sequentially and 24% year-over-year. Our EMS gross profit was NT$4.5 billion, increasing NT$0.5 billion sequentially and NT$0.8 billion year-over-year. The higher sequential and year-over-year EMS gross profit was the result of product mix. Gross profit margin for EMS came in at 9.1%, an improvement of 0.7 percentage points sequentially and a decline of 0.3 percentage points year-over-year. The sequential improvement is primarily a result of cost differences from different product mixes. The annual decline in gross margin is primarily due to higher operating overhead. Our EMS second-quarter operating expenses were NT$3.2 billion, increasing NT$0.4 billion sequentially and NT$0.7 billion year-over-year. The sequential operating expenses were primarily up due to increased R&D and factory start-up costs. Annual operating expenses increased primarily due to a larger operating base. Our operating expense percentage increased by 0.6 percentage points sequentially to 6.5% while increasing by 0.2 percentage points year-over-year. We expect our operating expense percentage to decrease during the back half of the year as our mass production revenues ramp up during our typical seasonal upcycle. Our EMS business has had a more challenging start this year due to worsening COVID operating conditions and component shortages. The underlying conditions have changed, making it more difficult and expensive to run than expected. We do not see component shortages or extra costs subsiding in the near term; therefore, our target of a 4% operating margin for our EMS business has become more challenging. On the bottom half of the page, you will find a graphical representation of our EMS revenue by application. The second quarter change here with consumer products declining 5% is seasonally driven, while the increase in the industrial segment is largely due to industrial products recovering after a year of COVID softness. On Page 13, you will find key line items from our balance sheet. The only additional item is that our total unused credit lines amounted to NT$276.4 billion and our net debt to equity ratio dropped to 60%. On Page 14, you will find our equipment capital expenditures. Amounts on this slide are denoted in US dollars. Machinery and equipment capital expenditures for the second quarter totaled US$611 million, of which US$50 million were used in packaging, US$116 million in testing, US$39 million in EMS operations, and US$6 million in interconnect materials and others. At the end of the second quarter, we are still operating in a capacity-constrained environment. Currently, we see our capital expenditures up from 10% to 15% from last year, though it leans more towards the higher end of that range. However, the timing of this year's capital expenditure may be more volatile than in previous years. The timing of equipment may differ or accelerate. With that, we would like to provide our third quarter business outlook as follows. In US dollar terms, ATM third quarter 2021 volume is expected to increase by 12%, with ASE Holding status stable compared to second quarter 2021 levels. We expect a similar sequential improvement in ATM gross margins for the third quarter of 2021 as in the second quarter. For our EMS business, in US dollar terms, EMS third quarter 2021 business levels should be slightly higher than the average levels in the third and fourth quarters of 2020. The EMS third quarter 2021 operating margin should be around our targeted full-year operating margin for 2021. With that, I'd like to open the floor for questions. We're doing it slightly differently this time around. We have people scattered throughout different rooms and such. When we get the question, I will repeat it. And then I would direct it to Joseph and Tien. So please go ahead with your questions.

Our first question is from Mr. Gokul Hariharan of JPMorgan? Gokul?

Yeah, thanks for taking my question. My first question is on what Dr. Wu mentioned in terms of demand/supply balance coming in 2023. Could you talk a little bit about what you expect pricing trends and margin trends to be next year? It looks like you're still going to be in a similar situation as 2021. So how should we think about this? Secondly, could you also talk a little bit about how book-to-bill is looking as we head into early 2022? You talked about Q1 being better than seasonal, and you did allude to double booking, which is a very big topic for investor focus. How do you estimate double booking within your order book? And how does that affect your book-to-bill when you think about next year's demand/supply? That's my first question.

Okay, Gokul. I have here that you were asking about pricing contracts for next year, and also an outlook for 2022. And then somehow circle back over to the concept of double booking. Generally, we try to limit the number of questions to two per round. So let's go with your first two: the pricing and the outlook for 2022.

Alright, so our first priority is to ensure customer delivery and fulfillment. We have efficiently handled all of the customer requirements throughout 2020 and 2021. Occasionally, we may need to make pricing adjustments, either for higher material costs or expedite fees. I believe for the second half of 2021, as well as for the full year of 2022, we will continue with the current trend. It is very difficult to provide a quantitative estimate for pricing since this is dynamic. Our primary focus is making expedited deliveries. However, the pricing environment remains favorable. Regarding the overbooking, we've seen some customers making adjustments. Some equipment deliveries made a timing adjustment, which may not be completely due to the business slowdown. I mentioned about supply chain continuity: sometimes there might be a BGA substrate shortage, or lead frame shortage; it makes no sense to have overcapacity on one equipment or process while lacking materials. The comment I made about 2023 is that a more holistic and balanced capacity supply/demand balance will not be present in 2022, but hopefully in 2023, we will have an easier time executing customer delivery.

Thank you. Got it. Just one clarification. So on the guidance for this year, I think it sounded like you are looking at 20 plus percent growth for IC ATM on a USD basis. Just wanted to clarify if I got that right?

That's correct.

Okay, thank you, I'll queue back up. Thank you.

But I do want to mention that this is on top of the recovery of our affected business via US EAR. So as Dr. Wu mentioned in his presentation, if we exclude that part of the business from the equation, our actual first half overall ATM growth was 48%. Additionally, in terms of tests, not only are we ahead of our schedule in making a full recovery in the first half, but again, if we exclude the EAR affected business, the actual growth is about 54%. So we're seeing very strong growth momentum in terms of our ATM business at this point. Yes, and it seems that it's also leading into 2022.

Next question.

Next question is from Mr. Randy Abrams of Crédit Suisse. Randy?

Okay. Yes, and thank you. Okay, I'm on the phone line, hopefully, you can hear me. The first question, back to the comment about the tightness until 2023. How do you see the supply side? We've seen pretty heavy industry bookings for equipment and lead times are stretched out, but I assume those equipment would get delivered in the next year. So I'm curious, on the supply side, how you're viewing it and if you think that bottleneck on the back end equipment gets resolved moving through next year? And then from the demand side, how are you factoring in parts of the environment? There's fear about some of the COVID-related consumer PC and home electronics coming off the high base. I'm just curious what you're reflecting for next year, if those factors either on the supply or demand side give you confidence for tightness continuing into 2023?

So Randy, you're looking for a question regarding the situation or the...

Yeah, equipment.

Relating to back-end equipment tightness, and also looking for an impact on next year's overall market demand and whether we see changes in the overall market demand structure.

Well, let me talk about machine delivery. Currently, the lead time is as stretched as it was during our last conference call. It has not improved; the lead time is extending. I believe the lead time will start to improve not this year but sometime in 2022. However, once the lead time improves, we will still need to balance capacity expansion. As I discussed regarding materials, processes, and everything surrounding that. I believe reasonable capacity build-up will occur sometime between this year and next year, and I think in 2023, we will see a more balanced profile. Regarding demand, we have seen some adjustments. For instance, some customer sectors are pushing out deliveries. However, due to the overwhelming demand situation, it's very easy for ASE on the assembly and test side to redirect necessary equipment into other applications with strong demand. Our wafer bank width is still very high, so right now we are not overly concerned about inventory corrections due to various reasons. Nevertheless, I believe local adjustments are somewhat welcome as it highlights the tight delivery situation, which has become mentally unhealthy for everyone. So I think demand will remain strong in the second half and is expected to be robust for 2022. I discussed the pyramid model earlier. I have struggled to articulate what is going on right now. COVID-19 has caused an unforeseen acceleration in system demand without asking for additional infrastructure. Therefore, all devices are in short supply, which leads to the challenges we currently face. When people discuss the potential for COVID-19 impacts to fade, we remain cautious. We understand this situation will dissipate eventually; however, we are facing multiple waves of demand driven by AI, IoT, and smart manufacturing, for which we are aggressively building the supporting infrastructure. This will in turn generate substantial new demand for semiconductor devices at all levels. Hence, we see the semiconductor industry will remain short on supply. This sentiment is echoed by our Foundry partners who indicate that 29 new fabs will be deployed in 2021 and 2022. While the industry lacks the incentive to build manufacturing infrastructure ahead of the curve, the COVID-19 situation has provided solid short-term motivation even though we cannot ascertain how long its impact will last—whether that is three years, four years, or two. We believe that all new capacity will be beneficial and necessary for the industry.

Next question, please.

Thanks, great. And the second question - okay, yeah, in the second question and one clarification in the first too. The local adjustments, if you think those are all driven by the constraints up and down the chain or are you seeing any pockets of application weakness? That's just kind of a clarification. And then my second question just on the guidance, I know you mentioned that first quarter above seasonal. For the fourth quarter, if you're coming up with above seasonal third quarter, do you expect to grow in the IC ATM in the fourth quarter? And on pricing being stable, I know you talked about there being expedites and a friendly environment. So I'm curious, given we're in peak season, what's keeping prices stable, or why you're not seeing a little bit of sequential improvement on pricing?

Randy, so you're looking for a fourth quarter outlook, as well as a pricing environment commentary for the rest of the year.

Okay, I want to start by answering your first point. Yes, we are seeing local adjustments, but we do not know the reasons behind them. They could be business-related or component shortage-related, but these adjustments remain localized and time-sensitive. Overall, they don’t seem to affect our momentum. Regarding Q3 to Q4, yes, we expect growth from Q3 to Q4, mirroring last year’s performance. I am also hopeful to see a record-breaking Q1 of 2022, better than Q4 of the previous year. Though I won’t guarantee that right now, it is my hope. If we experience a strong Q1 of 2022, I believe that momentum will carry forward throughout the year.

Additionally, regarding our margins, we expect to see sequential growth on a quarterly basis for the second half of the year, as we continue to experience volume expansion and efficiency improvements, including automation. This will notably enhance our bottom line, and for next year, there remains room for further margin improvement as we are observing very healthy overall financial development moving forward.

Let me add that I won't discuss the overall pricing comparison directly. However, pricing is dictated by market forces. It is not defined by any individual supplier. Under constant pricing conditions, if we make that assumption, we will analyze our quarter-to-quarter margin improvements closely. We are conducting a detailed analysis spanning the last eight quarters, scrutinizing how much of the enhancements derive from synergies and automation. All these statistics contribute to the confidence in our projected performance for 2022 and beyond. Ken Hsiang previously mentioned that we are seeking a more solid baseline moving forward.

Thank you.

Alright, the next question.

Our next question is from Mr. Bruce Lu of Goldman Sachs. Bruce?

Hello, can you hear me?

Yes.

Yes.

Okay. Thank you for the great result. Oh, my question is regarding your long-term contract agreements. I know that your business model is quite complex, with various types of wire bonders. Can you help me understand how this works for your long-term contracts? How much of your capacity is secured by these long-term contracts? How do I ask this?

So your first question relates to the character and nature of our long-term contracts?

Yes. Help me to understand.

Well, I can tell you that a large majority of our customers is covered by service contracts. I don’t think I can provide anything more specific; it represents a large percentage.

Yeah. This contract has actually secured most of the capacity or is it only for the incremental capacity?

It's for all of the capacity, not just incremental. Also, when customers secure assembly capacity, there is an inherent expectation that assembly capacity will have the necessary lead frames, molding compounds, substrates, and all of the processing materials to accompany it. Today, this assumption is significant. Therefore, long-term service contracts not only ensure assembly capacity but also support the customer and ASE, along with our supplier partners, to secure the necessary overall balanced supply continuity beyond 2021, into 2022, and 2023. This reflects our overall efficiency and flexibility management. The competition in manufacturing involves understanding the overall supply chain from wafers to materials and beyond, which encompasses how factors affect the industry dynamics.

I see. Do you believe that your competitors also have secured a reasonable amount of long-term contracts, or is it primarily suppliers who can guarantee this kind of long-term contract?

I don’t know the answer, but as a good competitor, I would assume they do.

Or on the other hand, how much of your customers' demand is fully secured by these long-term contracts?

Customers' motivation can best be illustrated by the long-term service agreements and future technology development. If you are the de facto choice for the customer, you will enjoy both short and long-term loading, along with future pipeline confidence. When we mention de facto status with our key customers in automotive or IC systems, we refer to existing loading and future pipeline potential.

I see. Okay, so my next question is for Joseph. I mean, for the gross margin for ATM, again, thanks, congrats on the great result, but I'm a little greedy; a lot of the semiconductor companies have reported historical high gross margins. When do you expect to exceed your historical high gross margin as a comparative basis?

Well, we can't be.

That is motivation.

And we've surpassed our historical peak.

Well, no, if you use a pro forma basis, if you add SPIL's gross margin in aggregate, not yet?

That's what I'm saying; even including SPIL's on a combined basis, we have surpassed it. If we, for clarity, if I don't account for FX or PPA, I'm still impending upon the goal of achieving the historical peak on an annual basis. In 2014, our highest gross margin was about 27%. In the second half of this year, I believe this will be passed, and I am highly confident that we will surpass that historical record on an annualized basis next year.

Okay, thank you.

Next question, please.

Our next question is from Mr. Rick Hsu of Daiwa Securities. Rick?

Now, can you guys hear me?

Yes. We can hear you. Go ahead.

Yes.

Okay, great. Yeah, just two little questions. The first one is the housekeeping question for Joseph. So what's your utilization rate across the board for wire bond, for assembly and testing in Q2, and what's the outlook in Q3?

In Q2, for assembly, we're about 85% and for test, it's close to 80%. In Q3, I believe assembly wise we will be over 85% for assembly and over 80% for tests; we are running at nearly full capacity now.

Okay, great. And the second question is about your pricing. When you asked about your friendly pricing for Q3 and presumably for Q4, is it pretty much across the board or specifically for wire bonding?

Rick, you're asking whether we've raised prices or plan to raise prices across the board or just on wire bonders? Is that the question?

Yeah, that's right.

Our utilization will be full for Q3. We are already full in Q2. In Q3 and Q4, we will start ramping up the SiP product, so our fabs will be quite busy. The pricing adjustment is a fluid concept. We need to follow the business dynamics based on both needs and requirements. I, unfortunately, cannot commit, and I cannot state that we want to raise prices across our product line; however, there is a possibility that we might.

Okay, yeah. That's good enough. Thank you so much. Thank you for your time.

Next question, please.

Our next question is from Mr. Szeho Ng of China Renaissance. Szeho?

Hi. Good afternoon, gentlemen. I have two questions. The first one is that before the merger, the company would take out their substrate self-sufficiency ratio. I'm not sure if you have the number ready for the latest quarter.

Can you repeat that question one more time?

Oh, yeah. The substrate self-sufficiency, in the past you provided that before the merger with SPIL. I'm not sure if you have the same percentage on hand that you could share with all of us?

Well, it was hovering around 22% to 25%.

Okay, do you think

I think it is still around that. I think the current tightness of the materials or substrates in a sense actually gives us an additional edge over our competitors, due to our stronger buying power and also our in-house capability.

Oh, that's my second question. Yeah, you already answered. Okay. And also the other question, could you provide an update on the wire bonder delivery schedule? Last time, you mentioned that the company is planning to add roughly 3000 to 4000 wire bonders this year. Just wondering if there will be upside to that number?

Well, in the second quarter, we added close to 1500 bonders, 1482 to be exact. We also added 135 testers. I think that delivery is still ongoing, and we are keeping our target for the year. Hopefully, in the second half, we will have a full delivery.

Okay, great. And delivery, I can remember last time you mentioned is roughly October timeframe, right? The full delivery, I mean?

Yes, but as Ken mentioned, things are very dynamic at this point. So that's the target, but we will see how it goes.

Okay, all right. Okay, great. Well, thank you very much and congratulations.

Thank you.

Next question.

If you have a question, please raise your hand now. We have a question from Mr. Gokul Hariharan of JP Morgan. Gokul?

Go ahead, Gokul.

Yeah. Thank you for taking my question again. Could you talk a little bit about how you think about capital spending? It looks like this year is going to be at the high end of the 10% to 15% range, or closer to the $2 billion mark or even higher. Do you feel next year also CapEx is going to remain in this high range, given supply is still going to be quite tight and customers are willing to sign up for longer-term agreements? That's my first question.

So you're looking for an outlook on our capital expenditure plans into 2022?

Yeah, just wanted to know, yeah.

Okay, well, for this year, we are still maintaining our previous plan of possibly raising our CapEx to the high end of the range, closer to 15%. And by the way it's going, I don't preclude the possibility of raising our CapEx again for this year. I think it's a little too premature to discuss 2022 CapEx since it largely hinges on the market situation overall. However, we can anticipate additional demand next year.

I would like to provide additional insight. Yes, we have talked about high CapEx for ASE and OSAT during 2020 and 2021. However, it will also be very interesting to look at the CapEx from all of the IDMs for the backend. I believe you will see a very different dynamic. Keep in mind, in 2022 and 2023, newly built fabs that have started since 2020 will result in incoming wafers. This raises the question of who will act as the backend service provider for these wafers if there is a system demand for infrastructure. If the IDMs are investing less for the backend, then our consolidation thesis will guide it. The OSATs would need to double down to compensate for the delta. Therefore, even though it may be early to say, in 2022, depending on business dynamics throughout 2021 and the early parts of 2022, we will have a much clearer view of overall system demand and the backend demand, as well as the demand trends among substrate and lead frame suppliers.

Understood, maybe one follow up on that front. Could you talk a little bit about how you envision returns on the CapEx, now that your gross margins have clearly increased? How are these LDAs being structured, and how are you thinking about CapEx in the future? You alluded to the fact there is a lot of capacity covered by LDAs. So how should we think about returns? You discussed the bulk of your capacity being accounted for in LDAs, but based on historical trends, OSAT is perceived as more cyclical in terms of margins and returns. Is there something we can discuss regarding this through-the-cycle performance like how you expect your returns to be higher given the pricing dynamics and customers' willingness to commit to longer-term contracts?

Gokul, you're asking about how we evaluate CapEx, especially in the context of the current semiconductor supply environment, right?

Could you give ROIC or ROE kind of, what are the thresholds or hurdle rates that you use when you consider CapEx and investment in general?

The returns are evidently improving as we are seeing margin expansion. However, it's important to note that FX does present a negative impact on the overall return situation. The target ROE we are targeting for our investments is approximately 20% to 25%.

25% for new investment, right?

That's correct.

Next question.

Our next question is from Mr. Roland Shu of Citigroup. Roland?

Hi, can you hear me?

Yes, we can hear you, go ahead.

Okay, sorry, I dialed in late. So excuse me if my questions are a little behind. The first question from me is, TSMC has several plans to build new fabs overseas. Are you considering to increase your global footprint as well to catch the business opportunity associated with the newly built wafer fabs ROI? And if you want to do so, how will it impact your CapEx spending plans in the mid- to long-term?

So your question relates to our longer-term thought process regarding our global footprint expansion, right?

Yes.

To answer that question, ASE is, by far, the most diversified manufacturing company in the OSAT world. We have factories across three continents including Japan, Korea, Singapore, Malaysia, the US, and China, as well as Taiwan. We currently maintain a globally diversified footprint. The next question involves how we will respond involving geopolitical sensitivity, US incentives, and initiatives by China as well as the plans by TSMC, Samsung, and Intel to build new capacity across the world. The short answer is that it will heavily depend on business dynamics. To be more specific, we observe the overall semiconductor demand, which aligns with the pyramid analysis I presented earlier. If you dissect this pyramid into segments, you ask which parts of it are cost-sensitive. Then, we need to develop manufacturing at a significant scale in order to deliver the most cost-effective, flexible solutions. After that, you will have a smaller piece focusing on national security, technology-sensitive aspects. You will study the end-user and identify the service provider. For the assembly and test components, we need to consider who can contribute to that value add; if there is no need for ASE, the question then becomes—what volume is required in comparison to any other alternative route? When all fails, we will adjust our investments accordingly based on business needs. As we stand, we already have a globally diversified footprint. Adjustments to this footprint will depend on our business dynamics; however, the business need is not yet clear for assembly and test.

Understood, thanks. My second question concerns your testing business. In the past two to three years, you have aimed to grow your testing business. However, last year was challenging due to EAR issues in the supply chain. How do you currently view your testing business, and how do you envision growth in the percentage contribution of testing to total IC ATM revenue moving forward?

Yes, in the first half of the year, we were quite busy realigning our test capacity to serve other customers. We have done this successfully. As previously stated, we have fully recovered our test business in the first half ahead of schedule. Moving forward, we aim to return our test business to growth mode, and we will make the necessary investments for further expansion.

Do you have a specific revenue percentage target from testing?

At our peak, we were around 10%, and that's the first figure we aim to recover.

Next question.

Our next question is from Mr. Randy Abrams of Crédit Suisse. Randy?

Okay, thanks for the follow-up. Yes, my follow-up was on the SiP business. If you could give an update on its contributions now for IC ATM and USI, what percent is it? Also, could you provide an insight on the pipeline outlook for expansion next year regarding your primary customer as well as diversification into additional customers?

So you're looking for an overview on SiP, its contribution, from our EMS and ATM entity, and also a potential view on the pipeline?

As for our overall SiP business, we experienced a 50% growth last year, and this year we are expecting a reduction in growth rate. However, I believe the momentum is still present, and we will make necessary investments to continue growing our SiP business. In the second quarter, overall SiP business represented about 17% of our total business from the holding standpoint, about 40% for EMS, and around 4% for assembly and testing. All these ratios are expected to grow in the second half of the year.

Okay, and for the EMS business, where things appear to have improved significantly for semiconductors relative to EMS? Is that 4% still, and does that seem challenging for you to be able to get back to your typical margins, or do you have a program in place to improve that or additional pricing to counteract the higher costs to get back to that margin target?

The ATM segment will continue to grow in terms of its share of overall business during the third and fourth quarters, and for the entire year, it will grow at a higher pace than what we observe in the second quarter.

EMS is currently operating at 4%?

Yes, we're still looking at 4%.

Okay, I get that. Okay.

It's becoming more challenging due to rising logistics costs and overall operating uncertainties. However, we are still maintaining that target.

Okay, and lastly, on the end packaging for things like SoIC you mentioned, and TSMC-SoIC as well as Intel's Foveros. With advancements to 3D stacking, do you see the OSAT or CAC potentially participating in that market, or would the shift to full 3D IC mean lesser involvement for you? Or is there still some opportunity for final assembly to substrates? I am curious about the opportunities available in that area.

The short answer is that we have always developed key customer relationships along the lines of 2.5D. We were the first to launch a 2.5D solution with our partner in Texas. The development of 3D-ICs has been ongoing. However, it's important to understand the service versus business model. If we have a company providing proprietary services to a unique client, that is not an OSAT-driven business. The definition of OSAT business includes having at least two alternative service providers and two customers. For developments in 2.5D, 3D-IC, or other chiplet architectures to be categorized as OSAT, there must be open platforms with multiple foundries competing against similar architectures at comparable costs and performance metrics. In such cases, ASE will certainly participate. It’s anticipated that this trend of open platform versus proprietary services will continue to shape the industry and drive competition. The good news is that packaging is becoming a more critical component of the overall semiconductor ecosystem—a very welcome shift.

Great, thank you.

Next question.

Our next question is from Mr. Bruce Lu of Goldman Sachs. Bruce?

Okay, I want to have a quick follow-up on the CapEx. What's the CapEx allocation across testing, bonding, and wire-bonding for this year?

The CapEx allocation or distribution.

Allocation?

For this year, the likely allocation will be around 65% for assembly, roughly 23% to 25% for tests, a little bit for materials, and around 9% to 10% for EMS.

Okay. So if we assume your equipment lead time is currently over a year, you should have a clear idea of your CapEx for next year?

Well, when discussing CapEx, we are referring to required CapEx, not just the testing allocations.

Oh, I see. Okay. Oh, then the next question is, can you give us the revenue contribution from the automotive sector or from the IDM in your ATM business?

Bruce, so you're looking for how much revenue the automotive sector represents?

Yes, in ATM.

Oh, roughly, the second quarter is around 6%, 5% to 6%.

Do you see a clear uptrend?

Yes. We are being aggressive in programming our automotive business.

Do you expect it to be more than 10% in 2022?

We will look at it. But we anticipate significant growth of over 50% this year.

Thank you.

There are no further questions.

Okay. I'll turn it over to Dr. Tien Wu to wrap up the call.

Well, thank you very much for your patience and support for ASE. 2021 has been a challenging but extremely exciting year for us. Much of the EAR impact, which plagued us last year, has been resolved. I would like to thank all of you for your support, and I look forward to a successful 2021. We will talk to you next quarter. In the meantime, please stay safe and healthy. Thank you.

Thank you very much.