[Foreign language] Good afternoon everyone and welcome to TSMC's third quarter 2021 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 third quarter 2021, followed by our guidance for the fourth quarter 2021. Afterwards, Mr. Huang and TSMC's CEO Dr. C.C. Wei will jointly provide the company's key messages. We will open the line for Q&A. 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 on our press release. Now, I would like to turn the call over to TSMC CFO Mr. Wendell Huang for the summary of operations and the current quarter guidance.
Thank you, Jeff. Third quarter revenue increased 11.4% sequentially in NT terms, or 12% in dollar terms. Our third quarter business was driven by strong demand across all four growth platforms, which are smartphone, HPC, IoT and automotive related applications. Gross margin increased 1.3 percentage points sequentially to 51.3%, mainly due to the improvement in back-end profitability and a more favorable technology mix. Operating margin increased 2.1 percentage points sequentially to 41.2%, mainly due to better operating leverage. Overall, our third quarter EPS was NT 6.03 and ROE was 30.7%. Now let's move on to the revenue by technology. 5nm process technology contributed 18% of wafer revenue in the third quarter, while 7 nm accounted for 34%. Advanced technologies, which are defined as 7 nm and below, accounted for 52% of wafer revenue.
Now moving on to revenue contribution by platform. Smartphone increased 15% quarter-over-quarter to account for 44% of our third quarter revenue. HPC increased 9% to account for 37%. IoT increased 23% to account for 9%. Automotive increased 5% to account for 4%, and DCE decreased 2% to account for 3%. Moving on to the balance sheet. We ended the third quarter with cash and marketable securities of NT 976 billion or equivalent $35 billion U.S. dollars. On the liability side, current liabilities increased NT 8 billion, mainly due to the increase of NT 24 billion in accounts payables and the increase of NT 6 billion in dividend payable, partially offset by the decrease of NT 21 billion in short-term loans.
Long-term interest-bearing debt increased by NT 50 billion, mainly as we raised NT 49 billion corporate bonds during the quarter. On financial ratios, accounts receivable turnover days decreased 2 days to 40 days, while days of inventory remained at 85 days. Now let me make a few comments on cash flow and CapEx. During the third quarter, we generated about NT 319 billion in cash from operations, including some customer prepayments, spent NT 189 billion in CapEx and distributed NT 65 billion for fourth quarter 2020 cash dividend.
Short-term loans decreased NT 18 billion, while bonds payable increased by NT 49 billion. Overall, our cash balance increased NT 106 billion to NT 854 billion at the end of the quarter. In U.S. dollar terms, our third quarter capital expenditures totaled $6.77 billion U.S. dollars. I have finished my financial summary. Now let's turn on to our fourth quarter guidance. Based on the current business outlook, we expect our fourth quarter revenue to be between $15.4 billion and $15.7 billion U.S. dollars, which represents a 4.5% sequential increase at the midpoint. Based on the exchange rate assumption of 1 U.S. dollar to NT 28, gross margin is expected to be between 51% and 53%, operating margin between 39% and 41%.
On July 12th, we announced we have completed the purchase of five million doses of vaccine as part of our efforts to help fight against COVID-19 pandemic in Taiwan. We recognized a small portion of the vaccine donation expense in the third quarter. The majority of it will be recognized in the fourth quarter, which will have around one percentage point impact on our operating margin. This concludes my financial presentation. Let me turn to our key messages. I will start by making some comments on our 2021 capital budget. Every year, our CapEx is spent in anticipation of the growth that will follow in future years. We are witnessing a structural increase in underlying semiconductor demand, underpinned by the industry megatrends of 5G related and HPC applications.
In order to support our customers' growth and meet the increasing demand for our advanced and specialty technologies in the next several years, we have budgeted our full year 2021 CapEx to be around $30 billion. Next, let me talk about our profitability. Our third quarter gross margin increased 1.3 percentage points sequentially to 51.3%, mainly due to better back-end profitability and technology mix. Based on the exchange rate assumption of $1 to 28 NT, we have just guided fourth quarter gross margin to be 52% at the midpoint. The midpoint of our fourth quarter gross margin guidance also implies that our full year 2021 gross margin is expected to be higher than 50%, despite the rapidly rising depreciation cost, the dilution from N5 ramp, and the unfavorable foreign exchange rate in 2021 as compared to 2020.
As we have discussed before, six factors determine TSMC's profitability. Leadership, technology development and ramp-up, pricing, cost, capacity utilization, technology mix, and foreign exchange rate, which is not controllable. Taking all these factors into consideration, we believe a long-term gross margin of 50% and higher is achievable. Now let me turn the microphone over to CC.
Thank you, Wendell. We hope everybody is staying safe and healthy during this time. First, let me start with our near-term demand and inventory. We concluded our third quarter with revenue of NT414.7 billion, or $14.9 billion, driven by strong demand across all four growth platforms, which are smartphone, HPC, IoT, and Automotive-Related Applications. Moving into fourth quarter 2021, we expect our sequential growth to be supported by strong demand for our industry-leading 5 nm technology. Based on the midpoint of our fourth quarter revenue guidance, our full year 2021 revenue is expected to grow about 24% year-over-year in U.S. dollar term. On the inventory front, we continue to expect our customers and the supply chain to gradually prepare higher level of inventory in the second half of this year as compared to the historical seasonal level.
Given the industry continued need to ensure supply security, we expect the supply chain to maintain a higher level of inventory for a longer period of time. In the near term, we continue to observe short-term imbalances due to interruptions in the supply chain brought about by COVID-19. We also continue to observe the structural increase in demand underpinned by the industry megatrends of 5G and HPC related applications and the higher silicon content in many end devices, including automotive, PCs, servers, networking, and smartphones. While the short-term imbalances may or may not persist, we believe our technology leadership will enable TSMC to capture the strong demand for our advanced and specialty technologies, and we expect our capacity to remain tight in 2021 and throughout 2022. Let me talk about TSMC's long-term growth driver and return. We are entering a period of higher structural growth.
The multi-year megatrend of 5G and HPC related applications are expected to fuel massive requirement for computation power and propel greater need for energy-efficient computing, which demand the use of leading-edge technologies. These megatrends will not only spur unit growth, but also by increasing semiconductor content in HPC, mobile phone, automotive, and IoT applications. COVID-19 also fundamentally accelerated the digital transformation, making semiconductors more pervasive and essential in people's lives. With our technology leadership, manufacturing excellence, and customers' trust, TSMC is better positioned to capture the growth from the favorable industry megatrend with our differentiated technologies. To address the structural increase in the long-term market demand profile, TSMC is working closely with our customers to plan our capacity and investing in leading-edge and specialty technologies to support their demand.
Our capital investment decisions are based on four disciplines: technology leadership, flexible and responsive manufacturing, retaining customers' trust, and earning the proper return. At the same time, we face manufacturing cost challenges due to increasing process complexity at leading node, new investment in mature nodes, expansion of our global manufacturing footprint, and rising material and basic commodity costs. As we continue to work closely with our customers to support their growth, our pricing strategy will remain strategic, not opportunistic, to reflect our value creation. We will also continue to work diligently with our supplier to deliver on cost improvement. Even as we shoulder a greater burden of investment for the industry, by taking such actions, we believe we can achieve a proper return that enable us to invest to support our customers' growth and deliver long-term profitable growth with 50% and higher gross margin for our shareholders.
Let me talk about our Japan fab plan. We are expanding our manufacturing footprint to sustain and enhance our competitive advantage in providing industry-leading technologies, the world's largest logic capacity, efficient and cost-effective manufacturing, and to better serve our customer. Our global manufacturing expansion strategy is based on customers' need, business opportunities, operating efficiency, and cost-economic considerations. After conducting due diligence, we announced our intention to build a specialty technology fab in Japan, subject to our board of directors' approval. We have received a strong commitment to support this project from both our customers and the Japanese government. This fab will utilize 22/28 nm technology for semiconductor wafer fabrication. Fab construction is scheduled to begin in 2022, and production is targeted to begin in late 2024. Further details will be provided subject to the board approval.
We believe the expansion of our global manufacturing footprint will enable us to better serve our customers' need and to reach global talent, while earning the proper return from our investments and deliver long-term profitable growth for our shareholders. I will talk about the N3 and N3E status. 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. We have developed complete platform support for both HPC and smartphone applications. N3 risk production is scheduled in 2021, and production will start in second half of 2022. We continue to see a high level of customer engagement at N3 and expect more new tape out for N3 for the first year as compared with N5. We also introduce N3E as an extension of our N3 family.
N3E will feature improved manufacturing process window with better performance, power, and yield. Volume production of N3E is scheduled for one year after N3. Our 3 nm technology will be the most advanced foundry technology in both PPA and transistor technology when it is introduced. With our technology leadership and strong customer demand, we are confident that N3 family will be large and long-lasting node for TSMC. This concludes our key message. Thank you for your attention.
Thank you, CC. This concludes our prepared statements. Before we begin 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 questions. Should you wish to raise your question in Chinese, I will translate it to 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 the one key on your telephone keypad now. Questions will be taken in the order in which they were received. If at any time you would like to remove yourself from the questioning queue, please press zero two. Let's begin the Q&A session. Operator, please proceed with the first caller on the line.
Yes. The first one to ask questions, Gokul Hariharan, JP Morgan.
Good afternoon. Congrats on the good results, and thanks for taking my question. My first question is on the long-term roadmap. Intel has now unveiled their long-term roadmap until 2025, with four process nodes, looking to catch up with TSMC and potentially even overtake. Could TSMC talk a little bit more about its own longer-term roadmap, timing of adoption of some of the new technologies like gate-all-around, High-NA EUV, buried power rail, et cetera? Where does TSMC see itself from a process technology leadership perspective in the next three to five years? The entry message is definitely well-received, but maybe could we talk a little bit more longer term, given some of your competitors are kind of addressing that kind of timeframe as well? That's my first question.
Okay. Thank you, Gokul. Please let me summarize your first question. Gokul's first question is about regards to our long-term technology roadmap. He notes that in IDM, has outlined their long-term roadmap for the next 3-5 years, and talking about catching up and overtaking. Gokul wants to know what are our views or plans, or our roadmap, I guess, around the timing of new technologies, such as new transistor structure like gate-all-around, High NA, et cetera, and how do we see our technology leadership position in the next 3-5 years.
Okay. Gokul, I don't comment on my competitors' technology roadmap or their technology approaches. For TSMC, we are confident that we are very competitive, and we do have a very competitive schedule, actually, let me say that, in our 3 nm technology and the 2 nm technology. I can share with you that in our 2 nm technology, the density and performance will be the most competitive in 2025. Of course, I can also share with you that the gate-all-around structure is being considered, although I am not ready to release more information about it. That's Again, let me conclude in one sentence. We are become very competitive, and we are confident that our technology leadership will be maintained.
Okay. Thank you, C.C. Gokul, do you have a second question?
Yes. Thanks for the answer. Looking at CapEx, I think back in Q1 results, TSMC indicated spending TWD 100 billion-plus in CapEx over the next three years. Since then, you have talked about Japan capacity expansion. Looks like there are some capacity expansion plans for leading edge in Kaohsiung as well. Could we talk a little bit about, is TWD 100 billion going to be enough, or do you still see some upside to this TWD 100 billion budget on the CapEx over the next three to four years since the growth seems to be stronger? If we see that upside in CapEx, are we still looking at the high end of 10%-15% growth CAGR, or do we believe that there could be faster growth than this 10%-15% or high end of 10%-15% that we had talked about previously? Thank you.
Okay, Gokul. Let me see if I can catch your second question. It's around our CapEx and growth, longer term CapEx and growth outlook. Gokul is asking with sort of our plans in Japan and plans for expansion in Taiwan, will there be upside to this TWD 100 billion CapEx number that we have talked about for the next few years? Also, will there be a higher long-term growth CAGR target as a result as well?
Okay. Gokul, this is Wendell. Let me answer your question. We are not able to comment specifically on next few years' CapEx. Our capital investment decisions are based on ffour disciplines: technology leadership, flexible and responsive manufacturing, retaining customers' trust, and earning the proper return, as C.C. just mentioned. Every year, our CapEx is spent in anticipation of the growth that will follow in future years. As we said, we are witnessing a structural increase in underlying semiconductor demand, underpinned by the industry mega trends of 5G related and HPC applications and increasing silicon content. As long as our growth outlook looks good, there could be upside to our CapEx plans, and we will continue our disciplined investment approach to support our customers and capture the growth opportunities. Now, in terms of our revenue CAGR, we're not planning to make any changes at this moment.
We will provide you with more information in our January conference.
Okay.
Thank you. Thank you very much.
Okay, thank you, Gokul. Operator, can we move on to the next participant on the line, please?
Next one to ask question, Bruce Lu Goldman Sachs Go ahead, please.
Hi. Thank you for taking my question. I think my first question is that 80% of TSMC's CapEx is focused on advanced node capacity expansion. Do you see that the mature node becomes the bottleneck for your customers? How do you ensure your customer can have enough mature node chips?
Okay. Sorry, Bruce, let me repeat your question. Question is around the mature nodes. Typically, majority of our CapEx is for the leading nodes. How can we ensure that our customers will not be bottlenecked or have enough on the mature nodes as well?
Okay. Bruce, let me answer that question. TSMC's strategy at mature nodes is to work closely with our customers to develop and invest in specialty technology solutions to meet customers' requirement and create differentiated and long-lasting value to customers. We take a holistic view and work with our customer to decide the optimal capacity to support their demand.
Okay.
Okay?
Okay. Does that answer your first question, Bruce?
Yes. Let me try to ask a different question. I think recently we have a lot of investors asking that there are a lot of noise from the end demand, such as from the TV or China smartphone. The inventory level is also at a higher level. The foundry, although we remained very positive, and almost everyone is raising the capacity and CapEx. Can you try to tell the investor what's the discrepancy and why the foundry can continue to see such a strong demand while the end demand is deteriorating?
Bruce, let me summarize your second question. Second question for Bruce is around looking at end demand and the foundry. Bruce notes that there's a lot of, I guess, noises about different types of end demand. However, the foundry outlook seems to be very positive. How do we explain this disconnect or discrepancy?
Okay, Bruce. Let me say that while we do not rule out the possibility of an inventory correction, but we expect TSMC's capacity remain very tight in 2021 and throughout 2022. This is because of our technology leadership position. Even there is a correction to occur, we believe it could be less volatile for TSMC than previous downturn as an underlying structural mega trend of 5G related and HPC applications. Actually, the increasing silicon content in addition to in the end devices will continue. Again, with our technology leadership, we are better positioned to capture the mid to long-term growth opportunities. I hope that answered your question. There is a discrepancy between the demand and why still very tight in capacity.
Okay, Bruce.
Understood. Thank you.
Thank you, Bruce. Operator, can we move on to the next participant, please?
Next one, Randy Abrams, Credit Suisse.
Hey. Yes. Thank you. Good afternoon. I wanted to ask, probably for Wendell, a few questions on the margins. You mentioned 50% and above. Just a couple of follow-ups on that. Should we think it's still within two points of 50%, or with efforts to firm up pricing, you could push it higher? Just reflecting to maintain return on capital, you will have a higher asset base. That's kind of first part of that question. Then within margin, if you could update us on the inefficiencies you had operating at high level, if you've worked that out, and also if you have an initial view on depreciation for 2022.
Okay, Randy, let me summarize your question. Your question is about our margin. Randy notes that we now say 50% and above. He wants to know, is this a couple points above? How high above? Will we be able to maintain our, I guess, ROIC or ROE as a result? Also, that last time we had talked about sort of running at a high level of utilization and certain inefficiencies. Has that now become more improved, or what is the outlook there? Also, the depreciation outlook for 2022.
Okay. Let me answer the last question first. Depreciation in 2022 will increase, but the magnitude, we are going to tell you next year in January. For the margin, how many percentage point over 50%? We don't want to disclose it right now, but we hope we can tell you more in the January investor conference. That will, of course, bring a better ROE than before due to the higher margin targets. You also asked about utilization
Yeah Randy also asking about inefficiencies. When we had talked previously about when we run at a high level of utilization less efficient cost improvement and things like that. Randy is wondering, is it continuing?
The utilization continues to be pretty high. At the same time, the cost improvement activities is ongoing. As a matter of fact, in the fourth quarter, we believe the margin will be better partially because of the cost improvement activities.
Okay, great. The second question I'll ask, and it's one quick follow-up, actually, related to utilization 50 and above. If you have a utilization view, is that a 90% or that full capacity? The second question I have is on the capital intensity. One of the equipment suppliers, Tokyo Electron, they put up a slide about a moderating increase in capital intensity. CapEx per K, they have it by 2 nm, just rising gradually to TWD 210 million per 1,000 wafers. Could you discuss, if you can, a CapEx per K, either absolute or how you see that trending? Do you see that continuing to accelerate up or actions you're taking to keep it more stable after the increase we've seen the past few years?
Okay. Randy's second question is about capital intensity. He notes that Tokyo Electron is showing that the capital intensity or the CapEx per K is moderating, the pace of increase, especially into 2 nm. He is wondering if we can just comment on our CapEx per K.
Well, Randy, what I can share with you is that CapEx per K for more and more advanced technology is normally higher. That's for sure. At the same time, through selling our values, and working with the customers and the suppliers, we believe we are able to still earn a proper return, which is, at this moment, a 50% and higher gross margin is achievable.
Great. Thank you.
Yep. Thank you, Randy.
Okay. Thank you.
Operator, can we move on to the next participant, please?
Next one to ask questions, Brett Simpson from Arete Research.
Yeah. Thanks very much. My question was on the N3 introduction next year. Can you talk a bit about the ramp-up of N3? Is it going to be a typical ramp, very similar to the last couple of node ramps, or do you see the timing of this being different? Also, just in terms of the cost, there's a lot of talk about costs rising above expectations for N3 as you add more EUV layers. Can you just clarify how you see costs at N3 and whether you can still achieve a 70% density gain at that node? Thank you.
Okay. Brett, your first question is around N3. Brett wants to know with N3 ramping in the second half of next year, what type of ramp do we expect versus the prior nodes? Will it be typical, or will the timing be different? Also, on the N3 cost, what does the N3 cost structure look like? Is that your question, Brett?
That's right. Thanks, Jeff.
Okay. Brett, this is C.C. Wei. Let me answer your second part of the question first. N3's cost definitely is higher than N5. That is because of technology complexity, and we have to use many new equipment, which cost higher. The ramp-up is very similar to the previous node with many customers' engagement. Actually, it's higher than what we observed in the previous node. Second half of 2022 will be our mass production, but you can expect the revenue will be seen in first quarter of the 2023 because it takes a cycle time to have all those wafer out.
Oh, okay. Basically on 3 nm, typically you'd see your first revenue Q2 or Q3. It's going to be later next year. Is that right?
That's right.
Yeah. Brett, I think we have been consistently saying that N3 will begin the production in second half 2022. That has been a consistent message since we first introduced N3 in 2019.
Okay, great. Maybe just a follow-up on 28 nm. You just talked about a new fab that's coming on stream in 2024 in Japan. Can you maybe just clarify the latest thinking in terms of Europe? Is that another region that we may see new fab expansion for TSMC? Then in terms of looking at the 28 nm node, there's a lot of capacity being expanded at the moment. Can you talk about what's driving this and why you think this will not lead to an oversupply situation in time? Thanks.
Okay. Thank you, Brett. Brett's second question is around 28 nm. Two parts. First, of course, that C.C. just announced our intention to build 28 nm in Japan. Brett wants to know, do we have plans in Europe? The second part is that, with 28 nm, what is driving the longer-term structural demand for 28 nm? Is there a risk of oversupply of 28 nm?
Okay. We don't rule out the possibility of building a fab in other areas including Europe. We do emphasize when we build up a new capacity for 28 nm, is almost all to serve the specialty technologies. For some of the specialty technology that is not offered by our competitor, TSMC is working with our customer to meet their demand. Is there any possibility of oversupply? Not for TSMC. Okay? That's all I can let you know.
That's great. Thank you, C.C.
Okay. Thank you, Brett. Operator, can we move on to the next participant, please?
Next one to ask questions, Roland Shu, Citigroup.
Hi, good afternoon. Thanks for taking my questions. My first question is, you have a global manufacturing expansion strategy to build more fabs overseas going forward. Like you said, you also don't exclude the possibility to build a fab in Europe. My question is a joint venture with a local government or key customers an option for you to build this new fab overseas, or you prefer to build a fab and 100% own, like what you did for those fab in China or U.S. you are building now?
Okay, Roland. Your first question is about our overseas fabs. Roland wants to know, as we expand overseas, will we consider joint ventures with local governments or our key customers, or will it be 100% owned, like what we have done in China and in the U.S.?
Okay, Roland. Let me answer this question first. Normally, as you mentioned, our overseas fabs, we normally own 100%. We do not consider a JV with government. However, JV with other companies or key customers can be considered on a case-by-case basis.
Okay, thank you. My second question is that, now you set a short-term goal of a zero-emission growth by 2025, but you have to continue to invest in 5 nm, 3 nm, or even 2 nm before 2025. How are you going to achieve this zero-emission growth target and also in the meantime, keep up with your expansion plan? Will these aggressive zero- emission growth plan to decelerate your investment plans to meet your target? Thank you.
Okay, Roland. Roland's second question is asking that our commitment recently announced to zero emissions growth by 2025, but as we continue to invest and expand on N5 and N3, how will we be able to achieve this target?
Okay Roland, it's actually a net zero in 2050, not zero emission in 2025. We are going to do this first by working ourselves to become more energy efficient, because a lot of the carbon emission comes from the electricity. Our production, we can try to minimize the carbon emission. Secondly, we are going to use more green energy, which will emit the most part of the carbon. For whatever is left, it will depend on carbon trading, the carbon rights in the future. That's the basic framework of achieving this net zero in 2050.
No, actually, I'm talking about zero emission growth. You have this near-term target, zero emission growth by 2025. I know this is different from this net zero in 2050.
Right.
Yeah.
The way to achieve those are pretty much the same.
I think, sorry, Roland, your question is about net zero emissions growth by 2025.
It is.
We have net zero emissions by 2050, right? I think what Wendell is saying is that we will continue to invest in technology, but we are also, as Wendell just said, our own internal efforts, our use of renewable energies, carbon credits, and also working with our suppliers and our supply chain on green manufacturing to achieve and deliver on these targets.
Okay. Understood. Thanks.
Okay. Thank you, Roland. Operator, can we move on to the next caller, please?
Next one to ask question, Charlie Chan from Morgan Stanley.
Thanks. Good afternoon, gentlemen. My first question is about the chip shortage situation. I think chairman took an interview by Time, and his view is that there should be more than sufficient finished chip in the supply chain. Can you help us or global investors to understand when do you think that chip shortage, especially for the automotive, can be fixed? Your advice to no matter governments or car makers, besides asking you to provide customer data, what would be the better way to manage the shortage issue going forward? Thank you.
Okay. Thank you, Charlie. Charlie's first question is around the chip shortage with several aspects to it. Charlie wants to know with the chip shortage and also sort of observations of customers stockpiling chips. He wants to know that how do we see the situation, and when can this be fixed, particularly for the automotive segment? Yeah. Let me stop there first.
Yeah. For the automotive, let me specifically point it out. The automotive supply chain actually is quite long and complex. Is more complicated than we initially thought. Let me say that TSMC's participation in the global automotive IC market is only about 15%, and we are doing our part to support our automotive customer with what they need. However, we cannot solve the entire industry's supply challenge. Recent factors such as pandemic in Southeast Asia also affecting the auto IC supply. Again, we are actively taking the steps throughout the first half of this year to address the chip supply challenges for our automotive customer. We also believe the wafers supply shortage is greatly reduced for our automotive customer, starting probably in third quarter. The end of OEM, probably we will wait for a couple of quarter to see it. That's our estimate.
Thank you very much. It's super helpful. Another question is about, again, the price hike, right? I think news were reporting you decide to hike the price by 5%-20%. May we know how to determine the different range applying to different customers? What's the kind of strategic reason behind for different range of a price hike? I know you don't really want to give the next year guidance, right? Based on that, 5%-20% price hike, in terms of a percentage of our gross margin improvements, can your Wendell comment on the margin improvement? Thank you.
Okay. Charlie's second question is asking about pricing. He is asking, that recently there's lots of news that we have increased our price by anywhere from 5% to 20%. He wants to know how we decide how much to increase for what types of nodes or customers, and then also, what will be the impact to 2022 gross margin. Is that correct, Charlie?
Yes. Thank you.
Hi, Charlie. In fact, we do not comment on our pricing. This is a very private discussion between TSMC and our customer. Let me say that we continue to work closely with our customer to support their growth. That one need TSMC to expand the capacity to support their growth. It's for both leading-edge technologies and specialty technologies. Our wafers pricing strategy continues to be strategic, not opportunistic or short-term, and so that we can be better prepared to support the capacity expansion. As for the return, let me emphasize it again, our gross margin will be 50% and higher. TSMC need to earn a proper return that can enable us to invest for the future expansion to support our customers' growth.
Okay. I guess my question is that whether your desired ROI or desired margin change, right? Meaning you, for example, you hike the price by certain percentage points, but besides the passing through the cost you just mentioned, whether that would lead to further margin expansion. I think that should be core of my question. Thank you.
Right. Charlie, I think we just said that, in the past, we always say above 50% gross margin. But now, we're saying that 50% and higher gross margins is achievable.
Okay. I see.
Thank you, Charlie.
Okay, understood.
Thank you.
Thank you.
Operator, can we move on to the next participant, please?
Right now, we have Nicolas Gaudois from UBS.
Yes. Good afternoon. Thanks for taking my question. Just going back to the confirmation, you just did on investing in Japan, should we understand that the portion of CapEx, in 2022, 2023, is incorporated in your overall guidance of $100 billion, or would that come on top, and could you specify it, if you can, at all? A related question to that would be, what kind of capacity are we talking about for 22 nm and 28 nm? Thank you.
Okay. Nick's first question is about our fab plans in Japan. He wants to know that, with today's announcement, is the CapEx for the Japan fab already incorporated in this TWD 100 billion target that we have talked about previously. Can we disclose the capacity for Japan?
Okay, Nick. The CapEx for this project, as we said last time in last quarterly release, was not included in the TWD 100 billion budget, as you mentioned. It will be incremental. Other than this, we really are not able to comment on the investment amount and other details until after our board's review and approval.
Right. Okay. Fair enough. Understood. Going back to N3 and N3E, you talked about an improved process window for N3E. Is that the only main difference, or is there a difference in performance as well between the two? Thank you.
Okay. Nick's second question is on N3E. He notes that we have talked about the improved manufacturing process window. He wonders if there's any other improvements in things like performance and et cetera.
Just a difference. As we said, N3E is an improvement. Improvement in the manufacturing window. The majority in the design rule or something is similar. We use the N3E to enhance the manufacturing window with a better performance.
Got it. Thank you very much.
Thank you, Nick. Operator, can we move on to the next participant, please?
Next one, we have Laura Chen, Citi.
Hi. Thank you for taking my question. I think we are talking about that seeing the solid demand across the board, thanks to TSMC's strong position and technology. On the other hand, on the demand side, we are also seeing that the smartphone growth is slowing down, particularly in China. I think back in earlier this year, we mentioned about a 5G smartphone shipment. We estimate that will be 500 to 550 million units. Just wondering, do you still keep that target? Do you have any idea or preliminary projection for the 5G smartphone into next year? What if the smartphone, if 5G moving toward more like a lower end or mainstream, that kind of segment, what's the implication to TSMC? That's my first question. Thank you.
Okay, Laura. Laura's question is focusing on the smartphone. She notes that recently it seems the smartphone momentum in markets like China are slower. She's wondering about what our forecast for the smartphone market this year is , as well as how do we see the 5G penetration this year. Also, the trend for the next few years. Is that correct, Laura?
Yes. Thank you.
Okay, Laura, let me answer this question. We see the proliferation of the 5G smartphone is still higher than the 4G at the same period of time before. Also, we're looking at about probably slightly over 500 million units of 5G smartphone for this year.
Right. Do you have any preliminary thought about the next year growth? Would that mainly driven by the lower end segment? If that's the case, what's the implication to our outlook? Yeah.
We'll update you about that information in January.
Okay. Thank you. My second question is also regarding our CapEx intensity. We already talked about the three-year horizontal, just wondering that, do we still expect the CapEx intensity to maintain high beyond 2023 since we are launching a gate-all-around or 2 nm in 2025. Can we expect TSMC will start to bear fruit like our previous CapEx intensity hike back in 2011, and thus we will maintain the high CAGR growth going forward? Thanks.
Okay. Laura Chen's second question is on CapEx intensity. She is asking what is the outlook for our capital intensity beyond 2023? Will we still have a very high level of capital intensity? She notes back in the 2010, 2011 period, of course, our capital intensity was higher, but then we were able to harvest the growth and capture the growth. How do we see the next few years playing out?
Okay, Laura. In 2020, the capital intensity was 38%. 2021 this year is going to be over 50%. As we said earlier, CC mentioned this earlier, our CapEx spend every year in anticipation of the growth in the future years. If we think the future growth outlook is good, then there's a possibility of higher CapEx. We're entering into a higher growth period because of the industry mega trends of 5G and HPC applications, plus the silicon content increase. The higher capital investment in the next few years is appropriate. As a result, we expect the capital intensity to be relatively higher than previous year, like in 2020, for the next two to three years, before gradually coming down maybe to mid to high 30s level from what I can see at this moment. Your observation on the previous investment cycle in 2011-2014 will be a good one.
Thank you. Thank you very much.
Okay. Thank you, Laura. Operator, can we move on to the next participant, please?
Next one to ask question, Sebastian Hou, Neuberger Berman. Go ahead, please.
Hey, thank you for taking my questions. I only have one. It's on pricing. I think last quarter, the company talked about firming up pricing to reflect the cost. Based on the higher long-term gross margin guidance that the CFO gave this time, a 50%+, I'm curious if this round of pricing adjustment is enough to absorb the higher CapEx intensity only for this year or next multiple years. I have a follow-up to this question, so I will stop here.
Okay, Sebastian. His first question is on pricing and that we have talked about firming our pricing, actually also talking about firming our value in the past. He is wondering now that we say 50% and higher gross margin, does that mean it is enough to cover the cost?
Okay. Sebastian, let me answer it this way. Well, first of all, we're not able to comment on detailed pricing discussion with the customers. We work closely with the customer to provide our value. After providing our value, we're now expecting that a long-term gross margin of 50% and higher is achievable as compared to above 50% gross margin previously.
Got it. My follow-up is that, given that the next two years CapEx plan is still fluid, and I think CFO also mentioned there could be upside to our CapEx plan because of the Japan or any other reasons. Does that imply this will be a continuous adjustment? Meaning that it won't be just one shot, but that we will evaluate the future pricing and what kind of value we can offer to customers based on the CapEx and also to balance the structural profitability. That means that we may continue to see potential upside in the pricing in coming years.
Okay.
Go ahead, Sebastian.
Sebastian, your follow-up is, again, let me summarize it. I think Sebastian is asking, our pricing, is it sort of a one time, or will this be sort of an ongoing thing?
Sebastian, this is C.C. Wei. Certainly, I will not be able to comment on the pricing discussion with our customers. We work with them, and we continue to plan our capacity and share our value. The capacity is one of the very important values of TSMC to support customers' growth. Our pricing is accordingly with our value. We prepare for that. This is a one time, or this is not, it's not a question. We do it strategically and not opportunistic and continue to work with our customer.
Got it. Thank you, CC and Wendell. At least I think we can make a fair conclusion that the higher margin guidance outlook this time is a strong reflection or evidence of that a customer is willing to pay higher because we offer value add service. Is that right, fair to interpret as it?
Yes. Simple.
Got it. Thanks.
Okay.
That's all from me.
Okay. Thank you, Sebastian. Operator, can we move on to the next participant, please?
The next one to ask questions, Mehdi Hosseini, Susquehanna International Group. Go ahead, please.
Yes. Thanks for taking my question. I want to go back to your comments on Korea and N3 and N3 plus. Can you tell me how I should think about EUV double patterning and how it will impact your cost structure? I have a follow-up.
Okay. Mehdi's first question is about on N3, and actually, Mehdi, it's N3E, not N3 plus. His question on N3 and N3E. He's wondering about the impact of things like EUV and double patterning. What impact does this have on the cost structure for N3 and N3E?
Well, let me answer that question. From N3 to N3E, we provide a better value on the transistor performance and have a better manufacturing window. As for the cost, they are similar. We think our customer will enjoy a better yield, better die density, and better transistor performance.
Okay. Thank you. My follow-up has to do with your earlier commentary on customer prepayment. In the past, you've had one or two largest customers that have provided prepayment. Should I assume that there is a diversification and larger number of customers that are providing these prepayments?
Okay Mehdi's second question is on customer prepayments. He observes that in the past we may have had one or two customers who do prepayments. He wants to know, are we seeing a diversification? Are we seeing a larger number of customers doing prepayments today?
Okay, Mehdi, let me answer these questions. Yes, in the past, there was only one or two customers providing the prepayments. As we've been talking now, we expect to invest a higher capacity, higher capital expenditures in the next few years to satisfy the strong demand. In order to secure our customer's commitment, we are able to secure the prepayments for some of those customers. The number of the customer, I cannot disclose, but it's more than before.
Thank you.
Okay. Thank you, Mehdi. Operator, can we move on to the next participant, please?
The next one will be Rick Hsu from Daiwa Capital Markets.
Yeah. Hi, this is Rick Hsu, thank you so much for taking my questions. The first question is about, it's like a follow-up to Bruce Lu's question earlier about the disconnect between sell-in and sell-through demands. I think C.C. Wei mentioned that he doesn't rule out the possibility of the inventory correction. May I know if that happens, when do you expect that to happen? Also, if that happens, which area would feel the more impact in terms of technology node and in terms of the end applications? Thank you.
Okay. Rick's question is again going back to the disconnect of sell-in versus sell-through. Also, that we have said we do not rule out the possibility of an inventory correction. Rick wants to know, if one were to occur, when would it occur? What particular end segments or applications could be more impacted?
Hi, Rick. I said we do not rule out the possibility. It's just a possibility. All I say is that TSMC's capacity will remain very tight in 2021 and throughout 2022. Which market sector? So far, we observe a little bit soft in smartphone and PC market. If you ask me to predict, I cannot give you a very accurate prediction. We are the only one I can give you a hint as we continue to say, it's not for the semiconductor industry. The demand does not only come from the unit growth but also is increasing silicon content in end devices. Even you saw some smartphone unit become soft or even decrease, that doesn't mean that semiconductor or the business or the demand will drop. Does that answer your question?
Yeah, perfect. That's very good. Thank you so much. The second question is on the technology migration. I remember that at 7 nm, you defined a 7+ as a node for you guys to have a very good transition of EUV. I'm just wondering, are you going to do the same thing, to define a particular technology node for the GAA transition?
Rick's second question is about technology migration and transition. He notes that in N7, we had introduced also N7+ to transition and start to adopt EUV. He's asking if we will incorporate a similar transition as we move to a new transistor structure.
Well, I don't think we can have any more information to share with you as we move from N3 to go to the next more advanced node. Today, I only announced that the N3 to N3E that will have a better transistor performance and better manufacturing window. For N2 GAA, we will share with you when we are getting more ready.
Okay. Thank you. Thank you so much.
Okay. Thank you, Rick. In the interest of time, operator, let's take the last two participants, please.
Okay. Now, the one who is going to ask question is Krish Sankar, TD Cowen. Go ahead, please.
Yeah. Hi. Thanks for taking my question. I just wanted to follow up. One is on C.C. Wei's prepared comments. You said the industry is going to maintain a higher level of inventory. Can you tell us which specific end market and which specific technology nodes you're seeing those higher level of inventories? Your comment that you're seeing softness in smartphone and PCs, is that a function of end demand slowing or is it a function of not being able to get the components to make those products? I'm going to add a quick follow-up.
Okay, Krish. Krish's first question is on the higher level of inventory that we see preparing in the supply chain. He wants to know which end markets or applications specifically, or which technology nodes do we see this higher level of inventory. Also, the slower momentum in the sell-through of smartphone or PCs, is this related to component tightness or shortages?
Well, let me answer the question. The high level of inventory is actually caused by some of the necessity for that to be disruption in the supply chain. It's across the board. Actually, it's not any node or any product. It's across the board. We say it will be continued for a period of time. That is because of today, all those elements to drive the people to prepare more inventory still continue exist.
Did that answer your question?
Got it. Yes, it did. It did. Just the second part of the question, which is the softness in smartphone and PCs, is that end demand related or component tightness related? I'll ask a final question along with it. The gross margin upside you saw in Q3 from back-end, was it a one-time thing or is there more upside for that in the future? Thank you.
Krish's question also sort of the weakness that we see in areas like smartphone and PC. Is this related to end demand or is it related to component shortages?
Both, actually. Let me answer the question quickly. Actually, end market is a little bit soft. It's slow, but we think it's partly due to the component shortage.
The second part, or Krish's second question, I should say, is on the gross margin side also the improved back-end profitability.
Right. The back-end business is sort of seasonal. It has high season, low season during the years. Normally, second half is high season, especially third quarter. As a result, the profitability of back-end will be better in that quarter.
Thank you very much, gentlemen. Thank you.
Okay. Thank you. Then operator, we will take the last participant, please.
Yes, the last one to ask questions is Andrew Lu from Sinolink Securities. Go ahead please.
Thank you for taking my questions. C.C. Wei, I want to ask, this year, you just guided 24% year-over-year growth. I think this number is probably in line with the industry. It is clear we have a stronger growth in advanced technology but losing some share in the legacy. Earlier, Wendell Huang told mention you are build more mature technology based on customer's demand. If our CapEx unchanged, we adjust down our advanced CapEx, but increase more CapEx on mature technology.
Okay. Andrew's first question, he's looking at our growth in 2021 to be around 24%. He sees the strong leadership in the advanced nodes, but his note is that we're losing share in the mature nodes. Going forward, will there be any adjustment in our CapEx strategy, leading versus mature? Is that correct, Andrew?
Yes. Correct. Thank you, C.C. Yes.
Andrew, let me answer that. We did not change our strategy or philosophy in our CapEx plan. Certainly, the most important thing is that we are working with our customer to support their demand. This is very important. That including under specialty technologies. Actually, we share them to increase the mature nodes capacity. As we announced, the Japan Fab actually is a mature technology. It's a 22, 28 node.
Can we say that in the future, we should have a higher percentage CapEx, in terms of total CapEx, compared to the past?
Andrew really wants to know.
Yeah.
Andrew wants to know, will the CapEx spending proportion of the mature nodes versus leading edge, will we have a higher proportion for the mature nodes in the future years?
Andrew, not yet, because of we increase our CapEx, right? Even the same proportion, the mature nodes, actually, we spend a lot of money also.
Did that answer your question?
Thank you. Yes. My last question is, since we are adjusting our price based on the cost increase or whatever, how do we factor into our model for next year? What can blended ASP increase should we factor into our model? I have been observed average price on blended basis for the last three years, including this year. Our price for last three years, including this year, about 7%-9%. If next year we have additional adjustment on the apple-to-apple pricing level, should we say easy to have a 10% blended basis increase on ASP? Thank you.
Okay, Andrew's second question is on the blended ASP outlook. He wants to know, in essence, can he model a 10% or greater blended ASP increase for 2022?
Andrew, it's too early to comment on 2022. We will provide you more color in January. We don't really comment on ASP anyway.
Thank you.
Okay. Thank you. 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, and the transcript will become available 24 hours from now, and both of which are 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 will join us again next quarter in January. Goodbye, and have a good day.