[Non-English content]
Good afternoon everyone, and welcome to TSMC's fourth quarter 2020 earnings conference call. This is Jeff Su, TSMC's Director of Investor Relations and your host for today.
To prevent the spread of COVID-19, TSMC is hosting our earnings conference call via live audio webcast through the company's website at www.tsmc.com, where you can also download the earnings release materials. If you are joining us through the conference call, your dial-in lines are in listen-only mode. The format for today's event will be as follows. First, TSMC's Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the fourth quarter 2020, followed by our guidance for the first quarter 2021. Afterwards, Mr. Huang and TSMC's CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then, TSMC's Chairman, Dr. Mark Liu, will host a Q&A session where all three executives will entertain your questions.
As usual, I would like to remind everybody that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the Safe Harbor notice that appears in our press release. Now, I would like to turn the call over to TSMC's CFO, Mr. Wendell Huang, for the summary of operations and the current quarter guidance.
Thank you, Jeff. Happy New Year, everyone. Thank you for joining us today. My presentation will start with the financial highlights for the fourth quarter and a recap of full year 2020. After that, I will provide the guidance for the first quarter of 2021. Fourth quarter revenue increased 1.4% sequentially in TWD terms or 4.4% in U.S. dollar terms, as we saw strong demand for our 5 nm technology driven by 5G smartphone launches and HPC related applications. Gross margin increased 0.6 percentage points sequentially to 54%, mainly thanks to cost improvement, partially offset by the margin dilution from 5 nm ramp and an unfavorable exchange rate. Our utilization rate in the fourth quarter was at an extremely high level, partially due to more production output, of which some of the wafers will be shipped in the first quarter. Total sequentially for 43.5%.
Overall, our fourth quarter EPS was TWD 5.5. Revenue by technology. 5 nm process technology contributed 20% of wafer revenue in the fourth quarter, while 7 nm and 16 nm contributed 29% and 13% respectively. Advanced technologies, which are defined as 16 nm and below, accounted for 62% of wafer revenue. On a full year basis, 5 nm revenue contribution came in at 8% of 2020 wafer revenue, 7 nm was 33%, and 16 nm was 17%. Advanced technologies accounted for 58% of total wafer revenue, up from 50% in 2019. Moving on to the revenue contribution by platform. Smartphone increased 13% quarter-over-quarter to account for 51% of our fourth quarter revenue. HPC decreased 14% to account for 31%. IoT decreased 13% to account for 7%. Automotive increased 27% to account for 3%. Digital consumer electronics increased 29% to account for 4%.
On a full year basis, smartphone, HPC, and IoT saw strong growth of 23%, 39%, and 28% respectively. DCE also increased 2%, while auto decreased 7% in 2020. Overall, smartphone accounted for 48% of our 2020 revenue, HPC accounted for 33%, and IoT accounted for 8%. Moving on to the balance sheet. We ended the fourth quarter with cash and marketable securities of TWD 791 billion. On the liability side, current liabilities increased by TWD 29 billion, mainly due to the increase of TWD 57 billion in accounts payable and the increase of TWD 38 billion in accrual liabilities and others. Offset by the decrease of TWD 69 billion in short-term loan. Long-term interest-bearing debt increased by TWD 28 billion, mainly as we raised TWD 30.5 billion of corporate bonds during the quarter. On financial ratios, accounts receivable turnover days decreased one day to 39 days.
Days of inventory increased 15 days to 73 days, primarily due to the ramp of leading nodes. Let me make a few comments on cash flow and CapEx. During the fourth quarter, we generated about TWD 259 billion in cash from operations, spent TWD 89 billion in CapEx, and distributed TWD 65 billion for first quarter 2020 cash dividend. Short-term loans decreased by TWD 67 billion, while bonds payable increased by TWD 30.5 billion due to the bond issuances. Overall, our cash balance increased TWD 56 billion to TWD 660 billion at the end of the quarter. In U.S. dollar terms, our fourth quarter capital expenditures totaled $3.2 billion. Let's look at the recap of our performance in 2020. We saw a strong growth in 2020 as our technology leadership position enabled us to capture the industry megatrends of 5G and HPC.
Our revenue increased 31.4% in U.S. dollar terms and 25.2% in NT dollar terms to reach TWD 1.34 trillion. Gross margin increased 7.1 percentage points to 53.1%, primarily due to a high level of capacity utilization and cost improvement. Operating margin increased 7.5 percentage point to 42.3%. Overall, full-year EPS increased 50% to TWD 19.97. On cash flow, we spent TWD 507 billion in CapEx, while we generated TWD 823 billion in operating cash flow and TWD 315 billion in free cash flow. We also paid TWD 259 billion in cash dividends in 2020. I have finished my financial summary. Now, let's turn to our first quarter guidance. Based on the current business outlook, we expect our first quarter revenue to be between $12.7 billion and $13 billion, which represents a 1.3% sequential increase at the midpoint.
Based on the exchange rate assumption of $1 to TWD 27.95, gross margin is expected to be between 50.5% and 52.5%. Operating margin between 39.5% and 41.5%. The sequential decline in first quarter gross margin is mainly due to a slightly lower utilization rate in the first quarter, albeit it is still staying at the high level, as well as an unfavorable foreign exchange rate. I would like to talk about the tax rate. We expect our 2020 tax rate to be in the range of 10%-11%. This will be equally applied to all four quarters of the year. This concludes my financial presentation. I would like to start with the key messages for the quarter. I will start by making some comments on our capital budget in 2020 and 2021.
Every year, our CapEx is invested in anticipation of the growth that will follow in the next few years. Our capital investment decisions are based on four disciplines: technology leadership, flexible and responsive manufacturing, retaining customers' trust, and earning the proper return. In 2020, we spent $17.2 billion to capture the strong demand for our advanced technologies and support our customers' capacity needs. In order to meet the increasing demand for our advanced and specialty technologies and further support of customers' capacity needs, our 2021 capital budget is expected to be between $25 billion and $28 billion. Out of the $ 25 billion-$ 28 billion CapEx for 2021, about 80% of the capital budget will be allocated for advanced process technologies, including 3 nm, 5 nm, and 7 nm. About 10% will be spent for advanced packaging and mask making, and about 10% will be spent for specialty technologies.
Next, let me talk about our capital intensity outlook. As we have said previously, our long-term capital intensity is in the mid-30s percentage range. However, when we enter a period of higher growth, our CapEx needs to be spent ahead of the revenue growth that will follow, so our capital intensity will be higher. For example, during 2010 to 2014, our CapEx spending increased threefold as compared to the previous few years, and our capital intensity ranged between 38%-50%. Because of the increased investment, we were able to capture the growth opportunities and deliver about 15% growth CAGR from 2010- 2015. Today, as we enter another period of higher growth, we believe a higher level of capital intensity is appropriate to capture the future growth opportunities.
We now expect a higher growth CAGR in the next few years, driven by the industry megatrends of 5G and HPC-related applications, which C.C. will discuss in more detail. We also expect this higher level of capital investment to continue to drive our technology leadership, enable flexible and responsive manufacturing, and earn customers' trust. While our leading nodes capital costs continues to increase due to increasing process complexities, it is expected to be compensated by continuing to sell our value, which includes the value of our technology, service, quality, and capacity support, and diligently working on cost improvement. With this level of CapEx spending in 2021, we reiterate that TSMC remains committed to a sustainable cash dividend on both an annual and quarterly basis. Now let me turn the microphone over to C.C.
Thank you, Wendell. Hi, everyone. This is C.C. Wei. Good afternoon. We hope everybody is staying safe and healthy during this time. Let me start with our near-term demand and inventory. We concluded our fourth quarter with revenue of TWD 361.5 billion, or $12.7 billion, which was in line with our guidance, mainly due to strong demand for our 5 nm technology, driven by 5G smartphone launches and HPC-related applications. Concluding 2020, the semiconductor industry, excluding memory growth, was about 10%, while foundry industry increased about 20% year-over-year. TSMC's revenue grew 31.4% year-over-year in U.S. dollar terms. Moving into first quarter 2021, our business continues to be strong, supported by HPC-related demand, recovery in the automotive segment, and a milder smartphone seasonality than in recent years. On the inventory front, our fabless customers' overall inventory was digested throughout the fourth quarter.
We now expect it to approach the historical season exceeding 2020 better than our forecast three months ago. We observed that the supply chain are changing their approach to inventory management amidst the lingering macro uncertainties. Looking ahead, we expect the supply chain and our customer to prepare a higher level of inventory compared to the historical season level for a longer period of time, given the industry's continued need to ensure supply security. Next, let me talk about the automotive supply tightness. The automotive market has been soft since 2018. Entering 2020, COVID-19 further impacted the automotive market. The automotive supply chain was affected throughout the year, and our customers continued to decrease their demand in the third quarter. We only began to see certain recovery in the fourth quarter. However, the automotive supply chain is long and complex.
While many of our technology nodes have been tight throughout 2020 due to strong demand from our other customers. In the near term, as demand from the automotive supply chain is rebounding, the shortage in automotive supply has become more obvious. In TSMC, this is our top priority, and we are working closely with our automotive customer to resolve the capacity support issue. I will talk about our 2021 outlook. For the full year of 2021, we forecast the overall semiconductor market, excluding memory, to grow about 8%, while foundry industry growth is forecast to be about 10%. For TSMC, we are confident we can outperform the foundry revenue growth and grow by mid-teens percentage in 2021 in U.S. dollar term.
Our 2021 business will be supported by strong demand for our industry-leading advanced and specialty technologies, where we see strong interest from all four growth platforms, which are smartphone, HPC, automotive, and IoT. Let me talk about TSMC's long-term growth outlook. We are entering a period of higher growth as a multi-year mega trend of 5G and HPC related applications are expected to fuel strong demand for our advanced technologies in the next several years. We expect global smartphone units to grow 10% year-over-year in 2021. We forecast the penetration rate for 5G smartphone of the total smartphone market to rise from 18% in 2020 to more than 35% in 2021. We expect the silicon content of a 5G smartphone to continue to increase as compared to a 4G smartphone.
We continue to expect faster penetration of 5G smartphone as compared to 4G over the next several years, as 5G smartphone benefit from the significant performance, bandwidth, and latency improvement of 5G networks to drive more AI applications and more cloud services. We believe 5G is a multi-year mega trend that will enable a world where digital computation is increasingly ubiquitous, which will fuel the growth of all four of our growth platforms in the next several years. As we enter the 5G era, a smarter and more intelligent world will require massive increases in computation power and greater need for energy-efficient computing, and therefore, require leading-edge technologies. Thus, HPC is an increasingly important driver of TSMC's long-term growth and the largest contributor in terms of our incremental revenue growth. With our technology leadership, we are well-positioned to capture the growth from the foundry industry mega trend.
We now expect our long-term revenue growth to be 10%-15% CAGR from 2020 to 2025 in U.S. dollar terms. Now I will talk about N3 stages. N3 will be another full-node stride from our N5, with up to 70% logic density gain, up to 15% performance gain, and up to 30% power reduction as compared with 5 nm. Our N3 technology will use FinFET transistor structure to deliver the best technology maturity, performance, and cost for our customers. Our N3 technology development is on track with good progress. We are seeing a much higher level of customer engagement for both HPC and smartphone application at N3 as compared with N5 and N7 at a similar stage. Risk production is scheduled in 2021, and volume production is targeted in second half of 2022.
Our three nm technology will be the most advanced foundry technology in both PPA and transistor technology when it is introduced. We are confident our 3 nm will be another large and long-lasting node for TSMC. I will talk about TSMC 3DFabric. TSMC has developed an industry-leading and comprehensive wafer-level 3D IC technology roadmap to enhance system-level performance. Our differentiated chiplet and heterogeneous integration technology drive better power efficiency and smaller form factor benefit for our customer, while shortening their time to market. These technology, including chip stacking solution such as SoIC, as well as advanced packaging solutions such as InFO and CoWoS. We observe chiplets are becoming an industry trend. We are working with several customer on 3DFabric to enable chiplet architecture. SoIC small volume production is targeted in 2022.
SoIC is expected to be first adopted by HPC applications where bandwidth performance, power efficiency, and form factor are aggressively pursued. We expect revenue from our back-end services, which including both advanced packaging and testing to grow at a rate higher than corporate average in the next few years. This concluding our key message. Thank you for your attention.
Thank you, C.C. This concludes our prepared statements. Before we start the Q&A session, I would like to remind everybody to please limit your questions to two at a time to allow all the participants an opportunity to ask their questions. Should you wish to raise your question in Chinese, I will translate into English before our management answers your question. For those of you on the call, if you would like to ask a question, please press the zero, then one on your telephone keypad now. Questions will be taken in the order in which they are received. If at any time you would like to remove yourself from the questioning queue, please press zero too. Let's begin the Q&A session. Operator, please proceed with the first caller on the line.
Yes. The first one to ask question, Gokul Hariharan from JPMorgan.
Thank you for taking my question. Happy New Year and fantastic results and guidance. Let me ask a question first on 3 nm. Dr. Wei, how should we think about the size of 3 nm? What we have seen is over the past two years, 28 nm was a very big node. 7 nm came out to be roughly 17% bigger, if you think about peak revenue, compared to 28 nm when you had your application coming in. Given the big CapEx plan that you're also applying, should we think that 3 nm, once it ramps up fully, will be substantially bigger than 7 nm, in terms of peak revenues? Just wondering how we should think about the size of this process node. Could you also talk a little bit about the opportunities within HPC? Right now you are already engaged with multiple HPC customers.
Could you talk a little bit about CPU, x86 CPU, obviously, which is something on everybody's mind. Could you talk a little bit about how TSMC would be exposed to this market as well as we go into the three nm era?
Okay. Gokul, sorry, this is Jeff. Let me please summarize your questions, two questions. We'll take them one by one. Gokul's first question is with regards to 3 nm, and about the size of our 3 nm. He notes that in the past we have had very big nodes such as 28 nm and then 7 nm. Gokul wants to know, in terms of the peak revenue contribution, do we expect or should N3 be substantially bigger than N7? That's his first question, correct?
Yes. Especially since you bring the setup and CapEx as well. Thank you.
Well, Gokul, let me answer your question by saying that we do expect the 3 nm will be widely used in HPC related applications in addition to the smartphones. With this kind of engagement with our customer, we do expect our revenue will be bigger, certainly. There's no doubt about it. What is the next question?
Gokul, I think the second part of your question is looking at what are our opportunities in high performance computing. Gokul notes that we have multiple customers engaged, but in particular, he is asking about the progress or the status of CPU opportunity. What do we see as the drivers of HPC.
Gokul, we don't specifically name one of our HPC applications such as a CPU, to say that what is the growth rate. Let me tell you that CPU networking and AI accelerator will be the main growth area in the HPC applications. Did that answer your question?
Could you be a little bit more specific on x86? I think you already had good success in 7 nm in creating the x86 market. Should we think that the x86 market share continues to move up a lot as we get into 3 nm?
Okay. Gokul, I guess your question is really on the x86 and looking at seven nms done well. As we get into three nm, will our exposure to x86 continue to increase?
We don't specifically comment on a very specific area. We work with our customer continuously, and to supply the very good technology to support their business.
Okay?
Thank you. I will go back to the queue. Thank you.
All right. Thank you, Gokul. Operator, can we move on to the next person on the line, please?
Next one to ask question, Randy Abrams, Credit Suisse.
Okay, yes. Thank you. I have two questions to ask. First, you talked about the automotive, and I assume also your mature nodes are very tight. You traditionally haven't added that much capacity on mature nodes in 8 in. Could you discuss within that, because you have some mix of that, how you're seeing a strategy to add capacity for those nodes? Could you also look at auto has been only about 3% of revenue. Should we expect a meaningful pickup in this vertical, both from mature applications and also from new areas like EV and ADAS?
Okay. Randy, let me summarize your question. You're asking first, on the automotive side, he notes our comments that automotive supply is tight. Do we expect a pickup in the automotive vertical? Also in looking at the mature nodes. Will auto benefit our mature nodes? ADAS and other trends in automotive, how do we see?
Well, let me say that now we see the automotive industry need a lot of semiconductor component, and that including the leading-edge technology for the ADAS system, and also some of the mature technology for a lot of applications, like a sensor, like a power management IC. We do see right now, it's a little bit shortage on the automotive, the mature technology supply. We are working with customer to mitigate the shortage impact.
Randy is also asking second part. On our mature nodes, given the tightness, will we consider to add capacity for the mature nodes?
We always work with our customer to plan our technologies, capacity, all those kind of thing. For mature node, we used to convert some of the large capacity into specialties. Right now, the trend stays the same.
Okay, great. My second question is often two parts. Just want to ask on gross margin and inventory. The gross margins, you've improved four points year-over-year. Part of that utilization, depreciation also was up 45%, TWD against you six points. Could you discuss if you've had a breakthrough on the cost reduction side? If now, I think last quarter you said about 50%, given what you've seen on cost reduction and coming off 54%, if you could have better confidence on margin, could continue to do better. I just want to ask a quick on inventory. Was it up 15 days? Historically, you draw down WIP at the fourth quarter. Maybe the trend why inventory was rising into early in the year.
Okay, Randy, let me summarize your questions. Two parts. First is on the gross margin. He notes that our gross margin improves throughout the year. Randy wants to know if there is a breakthrough on the cost side, and therefore, the long-term outlook for our gross margin, is it still 50% or not?
Right. Randy, this is Wendell. You just mentioned that our depreciation increased 45% year-over-year. I think the number should be 15% year-over-year.
Okay. I was looking Q4 to Q4, I think just the fourth quarter over fourth quarter.
Right. Now, in terms of gross margin in the long term, we believe 50% gross margin is reasonable and achievable. There are six factors affecting our profitabilities. The ramp of leading-edge technology, price, cost, mix, utilization, and foreign exchange rate. Take foreign exchange rate, for example. In 2020, the average U.S. dollar against NT rate was TWD 29.43. It is now trading between TWD 27.90-TWD 28. That is already a 5% appreciation of NT. Every 1% of appreciation of NT will affect our gross margin by 40 basis point. The other thing is in the fourth quarter of last year, as we mentioned, the utilization rate was very high, extremely high. That's the abnormal level of high utilization rate cannot sustain. In this quarter, we believe the utilization rate will come down a little bit, albeit is it still at a very high level.
Every point of utilization rate change will impact the gross margin by 40 basis points. A third example will be the ramp in our leading-edge technologies. We mentioned last time that we expect N5 ramp in 2021 to affect our margins by 2-3 percentage points, and we still think that will be the case. If you take all of those into considerations, we believe 50% gross margin is reasonable and achievable in the long term.
Randy had also asked about our days of inventory.
Right.
Increasing in fourth quarter.
Right. That's partially because as we have a very high utilization in fourth quarter, but some of the wafers will be shipped in the first quarter as opposed to shipped in the fourth quarter.
Okay. Thank you. Thank you, Randy. Operator, can we move on to the next caller, please?
Next one, we have Sebastian Hou from CLSA.
Hi. Thanks, gentlemen, for taking my questions. Happy New Year. First question is I want to follow on gross margin side. If I look back in the past two quarters, your gross margin actually turned out to be either at the high end or the surprise to the upside to your original guidance. While revenue is much on the high end of the guidance, while the TWD has continued to appreciate second half of last year. Which means that the margin turns out to be better than what you originally guided for two quarters consecutively. My question is whether or not the 1Q outlook margin is too conservative again. Second to that is whether our structural profitability will need to revise up just as our five-year revenue growth CAGR has just been revised up officially. Thank you.
All right, Sebastian. Let me summarize your first question, your observation that in the past two quarters, our gross margin has come in at the high end or slightly above the high end of our guidance. Revenue at the high end and the currency appreciation is there. Sebastian, his question is first, is the first quarter gross margin guidance too conservative? What about the outlook for our longer-term structural profitability? Does it need to be revised up?
Okay, Sebastian, if we compared fourth quarter to first quarter, 54% in fourth quarter and the midterm of our guidance for first quarter is 51.5%. The 2.5 percentage point difference actually mainly come from the utilization as well as the unfavorable foreign exchange rates. So at this moment, we're still sticking to this guidance, although obviously, we will work hard to continue to improve the gross margins. As for the long-term gross margin, as I just reported earlier, that we are maintaining the 50% gross margin to be reasonably achievable based on the elements, the six factors that I just talked about. Each of those factors will affect our profitability in long term.
Okay. Sebastian, do you have a second question?
Yes, I do. Thanks, Jeff, and thanks, Wendell. My second question is on your CapEx outlook. Apparently, at least that's a little bit upside surprise to me and I think also to the consensus estimates. The last time, I think when company raised the CapEx from TWD 10 billion-TWD 12 billion level to the TWD 15 billion-TWD 17 billion level, that result in the 30% revenue growth in 2020. My question is that, I think that CapEx invest for the future growth. Whether or not this another step of the CapEx to TWD 25 billion-TWD 30 billion this year will represent a re-acceleration of the growth in 2022 or 2023. Thank you.
Sebastian's question is, looking at our CapEx guidance for this year, TWD 25 billion-TWD 28 billion. It is above his expectations. He's looking at the last time we have an increase in acceleration to CapEx from TWD 10 billion- TWD 12 billion to TWD 15 billion- TWD 17 billion, resulted in us growing 31% this year. What is the outlook for our growth in 2022 or the future years?
Okay, Sebastian, it's too early to talk specifically about 2022. As CC mentioned, in the next five years, our target CAGR is between 10%-15%. That's already higher than the original target of 5%-10% CAGR that we used to have before the last conference call. That's also because of the higher capital investment that we are ready to make to capture the higher growth opportunities underpinned by the multi-year mega trends in the industry.
Well, let me add something. This is C.C. Wei. This is a 10%-15% CAGR is based on a very high number for 2020. We still forecast a 10%-15% CAGR. That will tell you how much of capacity we need to invest.
Okay. Thank you. Thank you, Sebastian.
Thank you.
Operator, can we move on to the next caller, please?
Next one, we have Bruce Lu from Goldman Sachs.
Hi. Thank you for taking my question. Great result and great guidance. I think the big difference is this time is that you raised the long-term revenue CAGR from 5%-10% to 10%-15%. Can you tell us that in terms of this kind of incremental changes, how much of the growth is coming from HPC and what are the other drivers for that? In terms of smartphone growth, I mean, the 5G transition is already 30% something in 2021. Moving forward, how much growth for you is coming from the dollar content growth or the shipment growth, or can you provide more color on the growth?
Okay, Bruce. Your question is really about our long-term growth outlook with our growth target CAGR of 10%-15%. Your question basically is by the different platforms such as HPC, what is the growth contribution? In looking at smartphone, how much is dollar content, how much is unit contribution?
Well, let me answer the question by, actually, the growth rate from the HPC application is higher than the corporate average. Smartphone is very close to the corporate, automotive is higher than the corporate average. IoT is close to that corporate average. Did that answer your question?
Yes. Thank you. Okay. My next question is, I want to ask more about structural profitability. I understand that all these six factors for the profitability, that's based on the assumption that structural profitability remain unchanged. Do we consider to move up the structural profitability because of the current structural growth for the company or the structural tightness for, especially with legacy technology node?
Okay, Bruce, your second question is on the structural profitability, given the higher growth outlook, and also the tightness in supply at legacy nodes or legacy technologies. Would we consider to move up the structural profitability target?
Yes.
Bruce, as I just mentioned, we are maintaining the financial objective, i.e., the structural profitability goal of 50% gross margin. Of those six factors, every one of them can affect the profitability. For example, I just use an example in foreign exchange rate, utilization, and also the ramp of leading-edge nodes. For example, the leading-edge technologies, the complexities increases, the CapEx per K is more expensive than before. We are working very hard with the customer to sell our value, the service value, the technology value, and also the capacity value, and firm up the wafer pricing. At the same time, we also work very closely with our suppliers to continue improve our cost, so that altogether we can maintain and earn a proper return in the leading nodes compared to those of the previous few nodes.
As a result, we are maintaining our structural profitability goal as 50% of gross margin. Okay?
Understand, let me clarify that whatever you gain in terms of your cost saving, you will still return it to your customer and maintain your 50% profitability target.
There are six factors. You add all of them together.
Understand.
Yeah.
Understand. Thank you.
Thank you.
Okay. Thank you, Bruce. Operator, can we move on to the next caller on the line, please? Thank you.
Next one to ask question, Charlie Chan, Morgan Stanley. You are on now.
Thanks for taking my question. Happy New Year. First question is also about CapEx. In the past, for you to spend huge CapEx on leading edge is usually for the smartphone application, given that the key user is Apple. This time, you almost double your CapEx level. Does it mean that there's a significant upside of the Intel CPU sourcing? This is the first question. Thanks.
Okay, Charlie. Your question is on our CapEx. Basically, Charlie notes that in the past, our large CapEx on leading edge historically has been for smartphone platform. This year, of course, our CapEx number is much higher. Therefore, he is wondering whether it's intended for a particular customer on the CPU side.
Well, Charlie, let me answer the question. In fact, we don't comment on specific customer or specific area. Our CapEx guidance is based on the current long-term demand profile, underpinned by the industry's mega trend.
Okay, Charlie?
Yeah.
Do you have a second question?
Yes, I do. Just some feedback to C.C. I think we all understand the mega trend 5G and HPC. The last question was just to understand whether there is additional kind of growth driver, for example, IDM outsourcing on top of the organic growth. My next question, I think it should be more related to your strategy, because, I think your existing customer, Intel, two days ago, they also commented, "Don't rule out the possibility of a licensed foundry process." Actually, 20 years ago, back in 2000, I think you also licensed the latest semi process to National Semiconductor. I'm not sure if TSMC, after 20 years, do you still consider this kind of option? You know what I mean?
License your foundry process to your IDM customer or even consider some option like a joint venture for this fab operation with your IDM customer. Thanks.
Well, again, we don't comment on the specific topics or specific customer, but let me tell you that we are working with our customer continuously, and to expand TSMC's business and to support our customers' demand.
Okay. Got you. I will be back to the queue. I have some follow-up. Thanks.
Thanks, Charlie. All right, operator, let's move on to the next person on the line, please.
Next to ask question, Brett Simpson from Arete Research.
Yeah, thanks very much. Questions maybe first for Wendell. On the revenue guide, I guess you're starting the year with a far better than seasonal Q1. I just wondered, how do you see the year playing out? Should we expect in the second half typical seasonality this year? In terms of the CapEx guide for this year. Obviously, there's a big step up, and spending this year is normally a reflection of how you think about future capacity growth beyond 2021. Can we assume from the big increase in CapEx this year that your implied revenue growth in 2022 will be higher than 2021? Thank you.
Brett has two questions. One on the revenue guidance. We guided for mid-teens for the full year growth for 2021. He wants to know how does it play out throughout the year. Will we see the typical seasonality, first half, second half split? That's his first question.
Yeah. From what we can see now, second half is still higher than the first half.
The second part is also CapEx and growth. Looking at the increase in our CapEx investment in 2021, noting that we typically spend CapEx in advance of the growth that will follow. Brett wants to know should we expect a big year or a large growth year in 2022? Sorry.
Brett, as I said, it's a bit too early to discuss 2022 in detail. C.C. just mentioned over the next five years, we're looking at a higher range of CAGR. The CapEx spent this year means future opportunity and growth not just for the next year but also the years after that. We're looking at multiple years of growth opportunities.
Okay.
Maybe just one for C.C. Wei on N3. You mentioned N3 would have the best PPA. We're seeing a lot of transistor innovation at Intel and Samsung in the next couple years. You're planning to stick with FinFET at 3 nm. I'm just wondering how you see the transistor density at 3 nm. I think at N5, you've talked about 175 million transistors per millimeter squared , is the potential of N5. How should we think about N3 in that regard and relative to some of the transistor innovation we're seeing at Intel and Samsung? Are you happy with the FinFET roadmap? Thank you.
Okay. Brett, your second question is regards to our N3, and our decision to continue to use FinFET transistor structure at 3 nm. You note that at 5 nm, we can deliver about 175 million transistors per millimeter squared. You want to know how this falls out at N3 or maybe in terms of our three nm in comparison to Samsung or others. How does it compare?
Well, as I said in my statement that N3 still provide 70% of the logical density gain in addition to all the performance gain and the power reductions. Whether that's at a N5, you got 175 million transistor per millimeter square. That one depends on what the number in N3, I think that will depend on customers of design. We continue to say that we offer the FinFET because of the technology maturity, the performance, and the cost are the best combination for TSMC to serve our customer.
Okay. Thank you, Brett.
Great. Thank you.
Thanks. Operator, can we move on to the next caller, please?
Next one, we have Roland Shu from Citigroup.
Hi, good afternoon. Congrats on a very good result. My first question is also for the CapEx spending, and there are two parts of my question. With this sharply increased CapEx spending, are you considering to sign long-term contracts with customers, especially to those customers who are new to adopt your most leading-edge technology to ensure a proper return of your investment? Second part of the question is, it seems that you have spent ahead in CapEx in EUV because the lower productivity for EUV when you first ran EUV. I would like to know how much CapEx downside you expect after you have improved EUV productivity to the optimized level. Thank you.
Okay, Roland. We'll take your questions one by one. Both of them relate to CapEx. First one is that with the higher level of CapEx that we have in 2021, Roland wants to know that would we consider signing long-term contracts with customers, especially with customers that are new to TSMC to ensure that we are making a proper return.
Roland, signing a contract to guarantee the loading in the future is not our common practice. We always work with our customer and continuously work with customer to serve their demand. We also put our CapEx or expanding our capacity according to our current long-term demand forecast. All right. Did that answer your question?
Okay. Yeah. I take it.
Okay, Roland. Your second question is also related to CapEx. Roland, let me summarize. I think you are saying that in our CapEx guidance, your assumption that the lower productivity of EUV is leading to a higher CapEx level for TSMC. Your question is that if the productivity of EUV improves, then how much reduction in CapEx could we see? Is that your question? Am I summarizing that correctly?
Yes, exactly. Thank you.
Well, let me answer that. We continue to improve the EUV's productivity because we are working closely with suppliers. So far, the improvement is obvious but still not up to our expectation yet. As for the CapEx, will it decrease because of improved productivity? This is in our CapEx plan already.
Okay?
It means for going forward, even if you have a higher EUV productivity, the CapEx intensity probably will be still high next year or maybe in the near future.
Okay. Roland, his question is that even with EUV productivity and factoring into our CapEx, that our capital intensity could remain high even into next year.
Well, the CapEx remained high or the CapEx intensity remained high is because of technology complexity. It's actually that N5 is much more complicated than N7. N3 much more complicated than N5. Most of the CapEx intensity coming from this technology advancement. Of course, EUV is a part of it, but it's not the only one reason.
Okay, thanks.
Okay.
Okay.
Thank you.
My second question.
Roland, I think that's two questions already. Sorry. We still have several people in the queue. I would kindly ask you to get back into the queue, so we can allow everyone a chance.
Okay. Thank you. Yeah.
Thank you. All right, operator, let's move on to the next caller on the line, please.
Yeah. The next one we have Sunny Lin from UBS.
Hi, good afternoon. Thank you for taking my question. My first question is that I want to follow up on 3 nm. I think, just want to get a bit of color on your current visibility for the customer adoption into second half of next year. How does it compare with the historical ramp of 5 nm and 7 nm, and also the cost per transistor for 3 nm versus 5 nm? Thank you.
Okay, Sunny. Your first question is on 3 nm. You want to know the visibility into customer adoption of 3 nm into second half 2022, and how does it compare to 5 nm or prior nodes? Also, the cost per transistor at 3 nm, is it still declining?
Let me answer that. The cost per transistor actually is continuing to decrease. For your question about engagement with the customer, we see a lot of customers, especially from the HPC field, they are engaged with their activity with TSMC.
Okay. Sunny, do you have a second question?
Right. Just a very quick follow-up to my first question. Wonder if C.C. will be able to provide any color regarding the ramp for N3 for second half of next year. Thank you very much.
It's early adoption from our customer. It's both in smartphone and HPC related applications. That's all I can say.
Got it. Thank you.
You're welcome.
My second question is for your 2021 gross margin. With PPAC going up significantly, how should we think about your depreciation growth for this year, and also the impact on gross margin? Thank you.
Sunny's second question is on the 2021 overall gross margin with a higher level of CapEx spending. She wants to know what will be the year-on-year increase in depreciation, and what's the impact to the overall 2021 gross margin.
Sunny, the depreciation in 2021 is expected to be between mid to high 20s% higher than 2020. The impact to gross margins, well, it's too early to talk about the remaining quarters of 2021. As a general feeling, you look at the capacity utilization that I just mentioned for exchange rate unfavorable, and also the N5 ramp negative impact on our profitability. Those are the factors that may affect our whole year 2021 gross margins. As I said, it's too early to talk about details on the remaining quarters. Okay?
Got it. Thank you very much. Very helpful.
Sure. Thank you, Sunny. All right, operator, let's move on to the next caller, please.
Right now we're having Laura Chen from KGI. Go ahead, please.
Hi. Thank you for taking my question, congratulations for the good result and outlook. I also have the question about the CapEx and the gross margin trend. I think given your strong position in the most advanced technology node and the extremely high CapEx in recent years, I believe there must be some strong conviction on the outlook with your major clients. Can you share with us your view that for the N3 first-year contribution will be similar to N5, that will have probably more than 10% revenue for the first year mass production? Can we expect that to happen? Also on the gross margin side, given there might be some swing factor of your major IDM clients for outsourcing opportunity, how would you manage the utilization rate, which may impact your gross margin substantially? That's my first question. Thanks.
Okay, Laura, I think that's two questions. Your first question is on the N3, noting our strong position in the advanced nodes and also the higher CapEx as an indication of the strong conviction on major clients. Laura wants to know will the revenue contribution of 3 nm in its first year be similar to, or how does it compare to 5 nm in the first year?
Okay. Laura, it's really too early to talk about that at this moment. As C.C. said, we believe N3, when it's out, it's going to be another large and lasting node for TSMC.
Okay, got it. Thanks.
And then-
And also-
Sorry, go ahead.
Yeah, probably the swing factor of the utilization rate that may impact the gross margin potentially, and particularly for advanced node, how should we look at the trend? How you manage that?
Okay. Laura, second question is looking at our gross margin and then also looking at opportunities, for example, in a particular IDM, if there's swings in utilization, how would we manage that, and how would that impact the gross margin? Is that correct, Laura?
Yes. Thank you.
Okay.
Laura, we don't comment on specific customers or business outlook. What we can say is we continue to work with our customers closely and to ensure that we provide this proper capacity to them, and we also maintain a good utilization out of it.
As Wendell. Yeah.
Laura, let me add some colors. I think our business has been driven in the past few years by smartphones. Starting from this year on, the HPC also jump on the wagons. Therefore, we forward-looking, we see the traditional seasonality can be moderated with multiple big customer in multiple market segments. That's our confidence. The other confidence is our CapEx includes 3 nm, also 5 nm. Our 5 nm is also very strong, stronger than we expected three months ago. Those two combine to give us the confidence to increase our CapEx.
Okay, Laura?
That's very helpful. Yeah. Thank you very much. That's very helpful.
Great. Thank you, Laura. Operator, can we move on to the next caller, please?
Next one we have Robert Sanders from Deutsche Bank.
Yeah, hi. I've just got one question, actually. Just could you please comment more on the wafer shortage situation and how severe it is at present? At which node do you see the shortage most acute? Is it 65 nm, 90 nm, 0.13 nm, whatever it is? How far out are you essentially booked out at some of these nodes? Do you think there's upside to wafer pricing at these nodes? Thank you.
Robert, your question is on the tightness or shortage in the wafer. He is asking, is it at particular nodes such as 65 nm, 90 nm, 0.13 nm? How short it is, and how long it will last?
Robert, most of the shortage actually is in the mature node. It is not in the five or seven nm per se, but in all the mature node, especially in 0.18 µm, in 14 nm, and 55 nm, in those area.
Okay.
Can I just start with one follow-up, which is just you haven't traditionally built capacity there, but they could become path dependencies for the industry if they are continuing to be short. Would you actually consider building greenfields to help the industry, or you think that other foundries will handle that?
Robert, your follow-up question is, given the shortage or tightness on some of these mature nodes, will we consider to expand, build new capacity at these mature nodes to alleviate any potential bottleneck risk?
Well, actually, we are working with customer closely and moving some of their mature node to more advanced node where we have a better capacity to support them. In addition to that, we also try to manage this shortage condition, try to mitigate the impact from this shortage.
Okay.
Thank you.
Thank you. Operator, let's move on to the next caller, please.
Next one, we have Rick Hsu from Daiwa Securities.
Yeah. Hi. Happy New Year, guys. This is Rick. My first question is, I guess you guys mentioned that now your customers are happy living with a higher inventory than the historical pattern because of the macro uncertainties, such as COVID-19. I wonder if your customer would still be happy living with a higher inventory than the normal historical pattern if COVID-19 is contained. This is my first question.
Okay. Thank you, Rick. Your question is, are the higher level of inventory that we're seeing partly is attributable to COVID-19? What if COVID-19 is no longer, everyone has vaccine, then it's no longer an issue. Will this continue?
Well-
Yes.
Yeah. First, let's say that we really hope that the vaccine will work, but even if it is working, it takes time. Also our customers still, at least today, they still have a different approach for the inventory management, as we said, because of the security of the supply is more important than anything else in today's situation. We don't think it's really to revert back to the historical level of the inventory.
Okay. Thank you. That's helpful. My second question is also regarding your CapEx, because number this year is really high. About 80% of your high CapEx this year is going to be spent for leading edge. I wonder how much of that portion is actually for preparation of the capacity build for 2022 and beyond, not for this year. Can you share your idea with us?
Okay. Rick, your question is on our CapEx. About 80% is for the advanced nodes. He wants to know how much of this spending for the advanced nodes is in preparation for capacity for 2022.
Rick, we invest this year, actually for future year, primarily. It's not only be for 2022. It may also be for the years following that. I think that's something that I'd like to share with you. Okay?
Okay. That's helpful. Thank you.
Yeah.
Yeah. Thank you so much.
No problem. Thank you, Rick. Okay, operator, let's move on to the next caller.
Next one, we have Andrew Lu from Sinolink Securities.
Morning, good afternoon. Thank you for taking my question. Can you hear me?
Yes, we can hear you.
Okay. My first question is, if your customer has its own design rule nodes with a different metal and poly pitch spec from TSMC's one, can this customer use in-house manufacturing and TSMC foundry based on the same design, or it needs to redesign the chip based on TSMC five nm, three nm design rule?
Okay. Andrew, let me try to summarize your question. Your question is about customers' design rules. If the customer has their own design rules, but with different metal and different poly pitch from TSMC's, could this customer use TSMC foundry or use their in-house manufacturing, or do they need to use TSMC's design rules, basically?
Andrew, we always-
That's correct.
work closely with our customer to support their design into TSMC's process technologies. We can manufacture inside TSMC.
Customer doesn't need to change its own design?
Okay. I cannot answer this question because of its two parties cooperation. As I said, we work closely with them to support their design.
Okay.
Understood.
Yeah.
My second question is, since our 3 nm, 4 nm nodes will be ramping out next year, what about second half this year? Will we have something like a 5 nm plus or revision 5 nm process node for second half this year? Thank you.
Andrew's second question is looking at second half of this year, noting that next year we'll have, for example, N3 and N4, then second half of this year, do we have any new node or continuous enhancement?
Andrew, we always continue to improve the technologies. Last year, we introduced our 5 nm to the market. This year, we continue to improve it, and next year we'll improve further. We never stop.
Okay.
Something like a 5 nm plus?
If that's what you are naming.
Yes.
Okay. Thank you.
Okay. Thank you, Andrew. Let's move on to the next caller, please.
Next one, we have Mehdi Hosseini from SIG.
Yes. Thanks for taking my question. First question has to do with the revenue mix forecast for Q1 by technology and platform. It would be great if you could provide some color, and I have a follow-up.
Okay. Mehdi wants to know for the first quarter, revenue by technology and revenue by platform.
Okay. Mehdi, in the first quarter, HPC, automotive, and IoT will increase sequentially. While smartphone will experience a milder seasonal decline compared to its recent seasonalities.
We do not provide a breakdown guidance of revenue by technology, Mehdi. Okay, do you have a second question?
Yes. Just a quick follow-up on CapEx. Does your $25 billion-$28 billion CapEx guide include investment for infrastructure in the U.S.?
Mehdi's question is, does our CapEx guidance this year include any investment for the U.S. fab infrastructure?
Yes, it does. The U.S. fab starts construction this year.
Okay.
Can you elaborate how much of the CapEx is for the U.S.?
Not at this point .
Right.
Okay. Thank you, Mehdi.
Thank you.
Thanks. Operator, let's move on to the next caller.
Next one, Krish Sankar from Cowen and Company.
Yeah, hi. Thanks for taking my question. I also had two on CapEx. Number one, pretty nice step up in CapEx this year from last year. Is it fair to assume your investment in EUV is also up this year relative to last year? I had a follow-up.
Okay. Krish's first question is that with our increase in CapEx that we guided for in 2021 versus 2020 being an increase, does that also mean an increase in the CapEx we spend on EUV?
No, we do not disclose that details.
Got it. As a follow-up, C.C., you mentioned that how capital intensity is going to be high all the way to 3 nm. You also said long-term capital intensity should be in the mid-30s. I'm just trying to square that by, what do you mean by long-term? It looks like the 3 nm is still going to be high for the next few years. Capital intensity might be higher than mid-30s. At what point should we expect it to get to mid-30s?
Krish's second question is in terms of capital intensity, with the capital intensity or CapEx per K at 3 nm being higher, and then we're having a long-term capital intensity returning to mid-30s. He wants to know when will we return to mid-30s capital intensity level. Is that correct, Krish?
Yes.
Yeah.
Thank you. Yes.
Yeah. We mean long-term meaning three to five years. I think 2010- 2014 can be an example. During that period of time, the capital intensity rose from 38%- 50%, maintained at high 40s for a couple of years, and came down afterwards. Something like that should be a reference.
Okay?
Thank you.
All right. Thanks, Krish.
Thank you.
Operator, let's move on to the next caller, please.
Next one, Gokul Hariharan, JPMorgan.
Hi. Thanks for taking my follow-up question. One question on CapEx and depreciation. Are we having to spend CapEx a little bit ahead of what we used to spend in the past in the EUV era? Is that a function of having to spend maybe six to nine months ahead compared to, let's say, in the immersion era? That's one. How should we think about depreciation with this jump in CapEx? Wendell, could you give us a little bit of guidance in terms of how we should think about depreciation for this year and going ahead as well, given the higher level of CapEx?
Okay. Gokul, let me summarize. Your first question is in terms of the CapEx. He wants to know that with CapEx, are we having to spend CapEx earlier now? Is this because of EUV that we need to spend more CapEx earlier?
Well, let me answer the question. The answer is yes, because there is a long lead time for the EUV tools. The tools are very complicated, and the supply chain for the EUV takes a long time to prepare for it. As a result, TSMC also had to plan in advance. That's longer than the normal tools that we used to have.
Okay. Gokul's second question is looking at with the higher CapEx, the depreciation outlook.
Right. For this year, Gokul, we expect the depreciation to increase by mid-20% to high 20% for 2021 over 2020.
Okay, Gokul?
Okay. Even with that, we are comfortable with the 50% structural gross margin.
Even with the higher growth in depreciation, Gokul is asking, are we still comfortable with a 50% gross margin?
Yeah, 50% gross margin as a long-term target, we think it's reasonable and achievable.
Thank you.
Okay. Operator, in the interest of time, I think we'll take the last two callers. Can we proceed with the next caller on the line?
Okay. The next caller is Randy Abrams, Credit Suisse.
Okay. Yeah, thank you. My first follow-up on U.S. and China, your overseas sites. For the U.S. site, you bought 1,100 acres. Do you have plans to build out a mega fab or potential to build out multi-phase of 20K wafers? For the China business, post Huawei, where it's down to single digits, how's your outlook for the China and also expansion of the China from 20K?
Randy, your first question is regards to capacity and fab expansion overseas. Randy is asking in the U.S., in Arizona, we target 20K. Will we continue to build it out into a mega fab type of site? He also wants to know in China, and I guess you're referring to Nanjing, do we have plans to further expand the capacity in Nanjing? Is that your question correct, Randy?
Yeah, that's the question. Just the outlook to rebound China just post HiSilicon, where it's down to mid-single digit contribution.
Yeah, this is Mark. Let me take your question. Yeah, we recently acquired a big piece of land in Phoenix, 1,100 acres. Definitely, that was the long-term plan to have a mega scale production site. Currently, our plan is, only work on the phase I production and target in 2024 with 20,000 wafer per month. Going forward, we'll see, according to the market condition and the cost economics, and provided by the government support to mend the cost differences to decide the next steps. On China, yes, we do have plan to continue expand in China. Of course, the business in China, of the leading edge, does have a reset. We do expect the demand in China will continue and we will gradually, accordingly, increase our capacity in Nanjing.
Okay. Great. My second question, if you could give, I think you gave first quarter, but the full year, growth for each of the platforms and also for the back end where you're doubling CapEx. What's leading that investment between the InFO, CoWoS, SoIC, in growth outlook for back end?
Okay. Randy is asking about 2021 growth. First, growth outlook by platform, then growth outlook by the back end. Between the back end InFO, CoWoS by segment.
Okay. Randy, for 2021 by platform. We think HPC and automotive growth will be higher than the corporate average growth. Smartphone and IoT will be similar to the corporate average growth in U.S. dollar terms. In terms of our back end business, we expect it to grow slightly higher than the corporate, in 2021. We do not disclose details within the back end business. Okay?
Okay. Thanks a lot.
All right. Thanks, Randy.
Thank you.
Okay, operator, can we move on to, in the interest of time, the last caller then?
Okay. next one we have Sebastian Hou from CLSA.
Yeah. Thank you. I'm pretty lucky to be the last one and ask again. Thank you. Two follow up. The first follow up is to follow on Mark comments that, I think that Mark would say that the company has noted its 5 nm demand also stronger than you thought three months ago. Just curious if you can give us more details about which applications are you seeing a stronger than expected demand?
High- Performance Computing.
For High-Performance Computing, is the typical those customer electronics or is more typical HPC or blockchain related?
Sorry, we didn't hear the last part. Sebastian.
Yeah, sorry. I'm saying that for the HPC part, is it more related to your existing customers or more related to the blockchain related product?
Let me just add a little bit color on this. High performance computing, as Wendell just said, will be the major growth driver of our business. This field is currently under exciting changes. The High-Performance Computing architectures, as you know, from different customers, everybody is striving to get the best performance with different architectures. Many more players getting into this field. We see a stronger innovation is coming our way, on N3 as well as on N5, cryptocurrency. It's not on cryptocurrency, Sebastian. We don't count on that, we support that.
Okay. Yeah, that's fair. The second follow-up is follow to Wendell's comments on that. I think this year's based on the guidance that we will see the CapEx intensity to go up to 50%. If we calculate based on the revenue guidance, if we do some calculations, which means the free cash flow for this year, the growth will likely be pretty slow or even flat, depends on how things go. Definitely not as strong as the past few years. My question is the company still sticking to the dividend policy that is 70% of free cash flow?
Sebastian, your question is, in looking at the CapEx, looking at our revenue guidance, the capital intensity this year being about around 50%, then the free cash flow growth may slow this year. What is the outlook for the dividend? Do we still use 70% of free cash flow as the cash dividend for rule?
Right. Sebastian, our dividend policy has two parts, 70% of free cash flow, but not to be lower than the previous periods. We remain committed to a sustainable and steadily increasing cash dividend. During the periods of higher investment, the focus will be more on sustainable. As we harvest the growth, the focus will be on steadily increasing.
Okay. Thanks, Wendell. Given that you're paying the investors getting the dividend, in this quarter, and which is the earnings you made like three quarters earlier. If we do the calculation simulation, which means that in the next 24 months, the investor will probably still getting TWD 2.5 and TWD 0.90 per quarter. Is that a fair calculation assumption?
At least.
Okay?
Okay.
All right. Thanks, Sebastian.
Okay. Thank you.
Thank you, everyone. This concludes our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within four hours from now. The transcript will become available 24 hours from now, both of which will be available through TSMC's website at www.tsmc.com. Thank you for joining us today. We hope everyone continues to stay healthy and safe, and we hope you join us again next quarter. Goodbye and have a great day.