大家午安，我是台积电法人关系处的苏志凯，欢迎您参加台积公司2021年第四季的法人说明会。为了防范新冠肺炎疫情扩散，本次法人说明会仍采电话会议进行。由于本法说会是向全球投资人同时连线转播，所以我们会全程使用英文，请您见谅。
Good afternoon, everyone, and welcome to TSMC's fourth 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 fourth quarter 2021, followed by our guidance for the first quarter 2022. 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 the 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 on 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 financial highlights for the fourth quarter and a recap of full year 2020. After that, I will provide the guidance for the first quarter 2022. Fourth quarter revenue increased 5.7% sequentially in NT dollar or 5.8% in US dollar. As our fourth quarter business was supported by the strong demand for our industry-leading five-nanometer technology. Gross margin increased 1.4 percentage points sequentially to 52.7%, mainly due to the continuous cost improvement efforts. Operating margin increased 0.5 percentage points sequentially to 41.7%, slightly ahead of our guidance as we enjoyed higher operating leverage, and a portion of the vaccine donation expenses got pushed out to the first quarter. Overall, our fourth quarter EPS was TWD 6.41 and ROE was 31.3%. Now, let's move on to the revenue by technology. Five-nanometer process technology contributed 23% of wafer revenue in the fourth quarter, while seven-nanometer accounted for 27%. Advanced technology, which are defined as seven nanometer and below, accounted for 50% of wafer revenue. On a full year basis, five-nanometer revenue contribution came in at 19% of 2021 wafer revenue. Seven-nanometer was 31%. Advanced technologies accounted for 50% of total wafer revenue, up from 41% in 2020. Now, moving on to revenue contribution by platform. All platforms increased in the fourth quarter. Smartphone increased 7% quarter-over-quarter to account for 44% of our fourth quarter revenue. HPC increased 3% to account for 37%. IoT increased 3% to account for 9%. Automotive increased 10% to account for 4%. And digital consumer electronics increased 2% to account for 3%. On a full year basis, all six platforms experienced year-on-year growth. HPC, IoT and automotive saw strong growth of 34%, 21% and 51% respectively. Smartphone also increased 8% and DCE increased 2% in 2021. Overall, smartphone accounted for 44% of our 2021 revenue, HPC accounted for 37% and IoT accounted for 8%. Moving on to the balance sheet, we ended the fourth quarter with cash and marketable securities of TWD 1.2 trillion. On the liability side. Current liabilities increased by TWD 84 billion, mainly due to the increase of TWD 22 billion in accounts payable and the increase of TWD 61 billion in accrued liabilities and others.
Long-term interest-bearing debt increased by TWD 150 billion, mainly as we raised TWD 157 billion of corporate bonds during the quarter. On financial ratios, accounts receivable turnover days remained at 40 days, while days of inventory increased three days to 88 days. Now, let me make a few comments on cash flow and CapEx. During the fourth quarter, we generated about TWD 378 billion in cash from operations, including TWD 80 billion prepayment from customers, spent TWD 236 billion in CapEx, and distributed TWD 71 billion for first quarter 2021 cash dividends. Bonds payable increased by TWD 157 billion due to the bond issuances. Overall, our cash balance increased TWD 211 billion to TWD 1.1 trillion at the end of the quarter.
In U.S. dollar terms, our fourth quarter capital expenditures total $8.46 billion. Now, let's look at the recap of our performance in 2021. We saw a strong growth in 2021 as our technology leadership position enabled us to capture the industry's megatrend of 5G and HPC. Our revenue increased 24.9% in U.S. dollar terms to reach $57 billion. In NT dollar terms, revenue increased 18.5% as the NT appreciated by 5% during the year. Such unfavorable foreign exchange rate also impacted our gross margin by about 2 percentage points. In addition, N5 dilution also created a headwind to our margin. However, as we continue to drive cost improvements, we were able to achieve gross margin of 51.6% and operating margin of 40.9% in 2021.
Overall, full year EPS increased 15.2% to 23.10 TWD, and ROE was 29.7%. On cash flow, we spent TWD 839 billion in CapEx, while we generated TWD 1.1 trillion in operating cash flow and TWD 273 billion in free cash flow. We also paid TWD 266 billion in cash dividends in 2021. I have finished my financial summary. Now, let's turn to our current quarter guidance. We expect our business in the first quarter to be supported by HPC-related demand, continued recovery in the automotive segment, and a milder smartphone seasonality than in recent years.
Based on the current business outlook, we expect our first quarter revenue to be between $16.6 billion and $17.2 billion, which represents a 7.4% sequential increase at the midpoint. Based on the exchange rate assumption of one US dollar to 27.6 NT, gross margin is expected to be between 53% and 55%, operating margin between 42% and 44%. Lastly, our 2022 effective tax rate is between 10% and 11%. This concludes my financial presentation. Now I will move on to key messages. I will start by making some comments on our 2022 capital budget and depreciation. Every year, our CapEx is spent in anticipation of the growth that will follow in the future years. We are witnessing a structural increase in underlying semiconductor demand underpinned by the industry megatrend of 5G-related and HPC applications.
In 2021, we spent $30 billion to capture the strong demand and support our customers' growth. In 2022, our capital budget is expected to be between $40 billion-$44 billion. Out of the $40 billion-$44 billion CapEx for 2022, between 70% and 80% of the capital budget will be allocated for advanced process technologies, including 2-nanometer, 3-nanometer, 5-nanometer, and 7-nanometer. About 10% will be spent for advanced packaging and mask making, and 10%-20% will be spent for specialty technologies. Our depreciation expense is expected to increase by low-to-mid-teens% year-over-year in 2022, as newly incurred depreciation will be partially offset by other nodes rolling off depreciation. With this level of CapEx spending in 2022, we reiterate that TSMC remains committed to a sustainable cash dividends on both an annual and quarterly basis.
Now, let me turn the microphone over to C.C.
Thank you, Wendell. We hope everybody is staying safe and healthy during this time. First, let me start with our 2022 outlook. We expect 2022 to be another strong growth year for TSMC. For the full year of 2022, we forecast the overall semiconductor market, excluding memory, to grow approximately 9%, while foundry industry growth is forecast to be close to 20%. For TSMC, we are confident we can outperform the foundry revenue growth and grow between mid- to high-20s% in 2022 in U.S. dollar terms. Our 2022 business will be fueled 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, IoT, and automotive.
Entering 2022, we expect the supply chain to maintain a higher level of inventory as compared to the historical seasonal level, given the industry's continued need to ensure supply security. While the short-term imbalance may or may not persist, we continue to observe the structural increase in long-term semiconductor demand underpinned by the industry megatrend of 5G and HPC related applications. We also observe the higher silicon content in many end devices, including automotive, PCs, servers, networking, and smartphones. As a result, we expect our capacity to remain tight throughout 2022, as we believe our technology leadership will enable TSMC to capture the strong demand for our advanced and specialty technologies. Next, let me talk about TSMC's long-term growth outlook and profitability. We are entering a period of higher structural growth.
As the technology becomes more pervasive and essential in people's life, and the digital transformation accelerates, the semiconductor industry value in supply chain is increasing. As we embark upon the 5G era, an intelligent and more connected world will fuel massive requirement for computation power and prepare greater need for energy-efficient computing, which demand greater use of leading-edge technologies. The multi-year megatrend of 5G and HPC related applications will drive multi-unit volume growth, and more importantly, substantial semiconductor content enrichment in HPC, smartphone, automotive, and IoT applications. To address the structural increase in the long-term market demand profile, TSMC is working closely with our customer to plan our capacity and investing in leading-edge and specialty technology to support their demand.
At the same time, we are committed to achieve a sustainable and proper return that enable us to invest to support our customers' growth and deliver long-term profitable growth for our shareholders. Over the last three years, we have raised our CapEx spending from $14.9 billion in 2019 to $30 billion in 2021, as we invest in anticipation of the growth that will follow. During the same period, our revenue in US dollar term has increased up from $34.6 billion in 2019 to $56.8 billion in 2021, or 1.6 times, and our EPS by 1.7 times. Looking ahead, as the world's largest reliable and effective capacity provider, with our technology leadership, manufacturing excellence, and customer trust, we are well-positioned to capture the growth from the favorable industry megatrend with our differentiated technologies.
We expect our long-term revenue to be between 15% and 20% CAGR over the next several years. In U.S. dollar terms, of course, fueled by all four growth platform, which are smartphone, HPC, IoT, and automotive. With the increasing need for computation, HPC will be the strongest driver of TSMC's long-term growth and expected to be the largest contributor in terms of our incremental revenue growth, with the CPU, GPU, and AI accelerators are the main growth area for our HPC platform. As we invest in leading-edge and specialty technology to support our customers' demand, we continue to face manufacturing cost challenges due to increasing process complexity at leading-edge node, new investment in mature node, expansion of our global manufacturing footprint, and rising materials and basic commodity costs.
We are continuing to work closely with our customer to support their growth, and our pricing strategy will remain strategic, not opportunistic, to reflect our value creation. We will also work diligently in our own fab operation and with our suppliers to deliver on cost improvement. By taking such actions, we believe a long-term gross margin of 53% and higher is achievable, and we can earn a sustainable and proper return of greater than 25% ROE through the cycle. Thus, even as we shoulder a greater burden of CapEx investment for the industry, we can continue to invest to support our customers' growth and deliver long-term profitable growth for our shareholders. Now I will talk about N5, N4P, and N4F status. As our N5 enter its third year of ramp, demand continues to be very strong, driven by smartphone and HPC applications.
Our N5 has proven to be the industry's most competitive leading-edge technology. To further enhance our N5 family's performance, power, and density improvement for next-wave five-nanometer products, we also introduce N4P and N4F technologies. N4P offers 11% performance boost as compared to N5, with 22% improvement in power efficiency and 6% density gain. N4P is designed for easy migration from N5, with its products taping our schedule for second half 2022. We also introduce N4X as an offering specially optimized for workload-intensive HPC applications. N4X will offer much more circuit performance boost over N5 and be expected to enter risk production in first half 2023. With our continuous enhancement of our N5 process technologies, we expect demand for our N5 family to continue to grow in the next several years, and for N5 family to be a large and long-lasting node for TSMC.
Next, let me 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 production will start in second half of 2022. We continue to see a high level of customer engagement at N3 and expect more new tapeouts for N3 for the first year as compared with N5. N3E will further extend our N3 family with enhanced performance, power, and yield. We also observe a high level of customer engagement at N3E, and volume production is scheduled for one year after N3. Our three-nanometer 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 our N3 family will be another large and long-lasting node for TSMC. Finally, let me talk about our mature node capacity strategy. TSMC's strategy at a mature node is to work closely with our customer to develop specialty technology solutions to meet customers' requirement and create differentiated and long-lasting value to customers. We expect the multi-year industry megatrend of 5G and HPC and the higher silicon content in many end devices to drive increasing demand at mature node for certain specialty technologies. We forecast 28-nanometer will be the sweet spot for our embedded memory applications and our long-term structural demand at 28-nanometer to be supported by multiple specialty technologies. In support of our specialty technology strategies, we are expanding our 28-nanometer manufacturing capacity at sites in China, Japan, and Taiwan.
Our capacity expansion is based on customers' need, business opportunities, operating efficiency, and cost economics considerations. We believe the expansion of our mature node capacity will enable us to better serve our customers' need and reach global talents in our differentiated specialty technology, will enable us to capture the demand generated from the industry megatrend, and deliver long-term profitable growth for our shareholders. This concludes our key message. Thank you for your attention.
Thank you, C.C. This concludes our prepared remarks. 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 their 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 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 queue, please press zero two. Now we will proceed to the Q&A session. Our Chairman, Dr. Mark Liu, will be the host.
Hello everyone. This is Mark Liu. I want to send my regards to every one of you during this pandemic, and wish we have a happy and successful 2022. Thank you.
Thank you, Chairman. Let's begin the Q&A session now. Operator, can we please proceed with the first caller on the line?
The first to ask questions, Randy Abrams from Credit Suisse.
Okay, yes, thank you. Congratulations on the results and the outlook and margins. First question on the growth outlook. When we compare the growth expectation mid- to high-20%, versus if I roll up the fabless and IDM customers, they're about mid-teens growth. So your outgrowth looks much wider than most years. Could you break it down a bit more, the factors between share gain, pricing moves, and also, if there's any component of inventory build in there?
Okay, Randy, let me summarize your first question. I believe your question is referring to the 2022 growth outlook.
Right.
Right. Randy is saying TSMC's guidance of mid-20s to high-20s%. His calculations show that the fabless industry is growing maybe around mid-teens%. We will outgrow the fabless. He is wondering what is driving this outgrowth. Is it share gain? Pricing? Are there other factors such as inventory built into this, and if we can each share?
Okay, let me answer the question. This is C.C. Wei. Actually, the growth in 2022 is all the above you just mentioned. It's a share gain, it's a pricing, and also it's, you know, unit growth. Did I answer your question?
Yeah, mostly. Maybe just two quick follow-ups to that. If you could break the growth by platform, and if you could indicate just how much inventory, like how much you think your customers wanna put in place more inventory, like how big a component you think that factor is.
Okay. Randy Abrams's follow-up is, can we give our 2022 growth by platform outlook, and then how much of a role is inventory build playing in this growth?
Hi, Randy. This is Wendell. Let me answer the platform question. In 2022, we expect the HPC and automotive to grow faster than the corporate average. IoT, similar. Smartphones, close to the corporate average. That's the platform growth.
Randy's second part is how much of a role is inventory build playing in this?
Well, Randy, as CC mentioned that the key message is we expect the inventory level to remain high, higher than before for a longer period of time. We're not able to quantify that factor.
Okay. No, I understand. Hey, if I could ask a second question. You disclosed about the. You're getting more prepayments, and a lot of customers have been disclosing those. If you could talk about the strategy behind this as far as the main objectives of the program and the protection you're looking for if we go into a downturn, what the scenario would be. If you could disclose kind of that, the strategy behind that.
Randy, the prepayment. We work closely and diligently with the customer to plan the capacity, including receiving their prepayments for capacity support. We will continue to work with them to determine the best way to support them. Such commitments or prepayment will strengthen our cash position and help mitigate our capital risk in capacity. Now, talking about securing commitment, we always work closely with our customers, and we believe that technology leadership, manufacturing excellence, and earning customer trust are the best or the most effective way to secure customer commitments.
As long as we plan our capacity well based on the structural increase in the long-term market demand profile, we believe our utilization and profitability can be well protected.
Okay, thank you, Wendell. Randy, does that address your second question?
Yeah. It was more about in a downturn, I guess, is it also an assurance in terms of volume or some cash flow, like to protect the amount you have to outlay? Also kind of an insurance policy of some sort.
Well, Randy, I think, as I just mentioned, the best way is to work with the customers and gain their commitment through our technology leadership, manufacturing excellence, and therefore earning their trust. If we plan our capacity well, based on the structural demand in the industry, then I think our utilization and profitability can be maintained.
Okay.
Okay. Great. Thank you, Wendell.
Thank you.
Yeah. Thank you, Randy. Operator, can we proceed to the next participant on the line, please?
Next one to ask questions, Bruce Lu from Goldman Sachs. Go ahead, please.
Hi, thank you for taking my question. Congratulations for the great result. I think management mentioned about a $1 trillion semiconductor market size in 2030. Can we have more color about the foundry market size by then? Or for the key growth drivers such as HPC, can TSMC provide a more quantitative forecast, for example, like what is the addressable market for Arm-based CPU in 2025 or 2030?
Okay. Bruce Lu, let me summarize your first question. Bruce Lu is saying, we have talked about a 1 trillion foundry market by 2030. He wants to know what can be the key drivers here, and also the role of HPC, including Arm-based, in this outlook.
Okay. Bruce Lu, this is Mark Liu. We don't have a very specific forecast for 2030 to share with you. $1 trillion is our model, and so has been quoted by general industry comments. However, we do believe the high growth approach the number is happening. We believe the semiconductor industry growth will continue fueled by the C.C. Wei just mentioned structural megatrend, 5G and high-performance computing. Also our leading-edge technology provide the most energy efficient technology for computation and accelerate the digital transformation for the next several years. But if we want me to comment on the foundry industry, it's pretty clear the foundry industry growth will be higher, because in addition to fabless company, IDM outsourcing will continue increasing in a fast growth rate.
Most importantly, system companies will grow particularly faster during this period of time. In that sense, we believe the foundry growth, foundry industry growth will have a good year, better than other sectors.
Bruce Lu is also asking about the role of HPC in this growth outlook, including the Arm.
Arm is a new phenomenon. I think the CPU architecture no longer being dominated by one architecture. Multiple architecture provided their better integration with software, provide a much wider application of CPUs. That, no matter what the CPU, which architecture they are, currently we are engaging to all CPU architecture customers.
Okay. Bruce, does that answer your first question?
Yes, I try to ask a follow-up for the addressing for the HPC. I mean, within that HPC, we understand that this is the fastest growing segment. Can you provide some more like, you know, different, like, you know, growth magnitude within the HPC? Which, you know, part of the HPC is growing faster and how big would that be in 3-5 years?
Well, Bruce, this is C.C. Wei again. No, we cannot disclose all the detail on each segment. For example, CPU, GPU and AI accelerator, which one is more, you know, what kind of percentage. In fact, they are all growing, but you know, for the specific percentage, we don't comment right now.
Okay, Bruce?
Understood. Thank you.
Okay.
Yes.
Do you have a second question?
Yes. The second question is regarding to the 28 nanometers. As we know that 28-nanometer is the biggest node among all the process node within TSMC, and which has suffered low utilization rate in 2018 and 2019. With the recent announcement in Nanjing or Japan or even in Kaohsiung, we suggest another 50% of capacity growth in 28-nanometer alone in TSMC. Not to mention like your industry peer are all aggressively in expanding the 28 nanometers. Can you provide like you know more like growth driver and how can you feel comfortable to expand the capacity in this magnitude in the coming years when you know a lot of like you know financial industry were talking about like down you know down cycle or a cycle peak over supply in 2023?
Okay. Bruce, the second question is on 28-nanometer. He notes that we are, you know, as C.C. said, expanding our capacity in 20-nanometer in various locations. But he also notes that, you know, other foundry peers in the industry are also expanding as well. He wants to know the risks, I guess, of oversupply.
Well, good question. As you pointed out in 2018, 2019, we have a low utilization rate. It's just a little bit above 80%. Right now, we do observe that our long-term structural demand at 28-nanometer was to be well supported by multiple specialty technologies such as a CMOS image sensor for multi-camera trend and better non-volatile memory application and other specialty technologies. To be in one word, actually, the enrichment in the silicon content in many end devices that developed in recent years helped to support this demand.
Okay, Bruce.
Great. Thank you.
Thank you. Operator, can we move on to the next participant, please?
Next one to ask question, Charlie Chan from Morgan Stanley.
Hello, Charlie. Are you on the line?
Yes. Hello, can you hear me okay?
We can hear you now.
Hello?
Please go ahead.
Thank you. Happy New Year and congratulations for good results. To management, my first question is about the macroeconomic risk, right? For example, inflationary pressure on the consumer tech demand, and also stay-at-home demand has been strong for two years, how management would consider that could fade away and impact your PC, TV release timings. Lastly, the crypto mining demand has been very volatile, right? Now it seems like crypto price also falls down. I'd like to know, management, whether you consider those kind of macro risks into your full year revenue growth forecast. Thank you.
Okay. Charlie's first question is related to the macro outlook, looking at concerns of inflation, work from home fading, consumer demand, and also crypto volatility. What impact could this have on end demand for PC, TV, and semiconductors, and have we considered this into our outlook? Maybe CC can address this.
Yes. Well, Charlie, we expect the supply chain to maintain a higher level of inventory for a longer period of time, given the industry's continued need to ensure the supply security. We also observe the end market momentum in certain segment may slow down or adjust in terms of units. The increasing silicon content in many end devices is a more important factor in supporting the strong semiconductor demand and will continue. Even if just a correction were to occur, we believe it could be less volatile for TSMC due to our technology leadership position. The structural mega trend demand of 5G related in HPC application, you know, the substantial increase in silicon content, that will make sure we expect our capacity to remain very tight throughout 2022. You also mentioned about cryptocurrency.
Yes, we have factored that one in.
Okay, great. Charlie, do you have a second question?
Yes, if I may. Second question is maybe to Wendell, that first quarter revenue sees a great sequential growth, right? May I know that those price hike benefits mostly reflected in 1Q or there will be still some, you know, price hike benefits in second quarter? Also, can you please repeat the depreciation growth guidance year-on-year for 2022? Thank you.
Okay. Charlie, let me summarize your second question. His question is directed to Wendell. In terms of the first quarter revenue sequential growth, how much of a price hike is reflected in this? And will a price hike be reflected in subsequent quarters?
Charlie, we're not able to break down those numbers for you. We continue to work closely with our customers, meet their demand, and also sell our value. That's what we're focusing on.
Yeah. The depreciation guidance, Charlie, this year is low to mid-teens year-on-year increase. Thank you.
Okay. Great. Thanks.
Okay. Thank you. Operator, can we move on to the next participant, please?
Next one to ask questions, Gokul Hariharan from J.P. Morgan.
Happy New Year, and thanks for taking my questions. First of all, could you give us an update on how we think about CapEx going into the next couple of years as well? Given the growth expectation is also now higher in the 15%-20% range, do we expect CapEx to peak out this year or should it continue to grow into next year? Maybe, Wendell, could you also talk a little bit about how we think about capital intensity? When do we expect that to peak, and what is more like the steady state as we get to the end of this particular growth period, 2020 through 2025? That's my first question.
Okay, Gokul. Let me summarize. Allow me to summarize your first question. Gokul was asking about our CapEx outlook. He notes that our growth is now higher at 15%-20% CAGR for the next several years. His question is, how should he think about the CapEx? Is 2022 the peak? How does CapEx look for the next few years, and also in terms of the long-term capital intensity?
Okay. Hi, Gokul. This is Wendell. Regarding CapEx number, this year, we guided TWD 40 billion-TWD 44 billion. Now, remember that we spend the CapEx in a given year for the growth prospect in the next several years. Whenever we think the growth outlook is good, we will continue our disciplined investment going forward. We're not able to give a guidance on CapEx beyond 2022 today. Also the capital intensity. Whenever we enter into a higher period of growth like today, the capital intensity will be high, which is appropriate. Now, if the growth were to slow down, the capital intensity will decline accordingly. From what we can see at this moment, longer term, maybe mid-30s remains appropriate level.
Okay. Thank you. Does that answer your first question, Gokul? Hello, Gokul, are you on mute?
Yeah, sorry. Thanks for answering the first question. My second question is about one of your IDM customer, which has been a top ten customer for TSMC, but now they are also reentering the foundry market aggressively and wanting to compete head-to-head with TSMC. How does TSMC navigate this situation, given the market is also expecting this customer to be a bigger revenue contributor for TSMC in the HPC area over the next couple of years? Just wanted to understand how management thinks about this business relationship, given that for some of the other IDM customers who are foundry competitors, TSMC doesn't do much direct business with them.
Okay. Gokul, let me summarize your second question. Gokul's second question is in terms of a specific IDM customer of TSMC, but this customer is also entering foundry. You know, his question is how do we navigate the relationship given that in the past, we do not really work with IDMs who compete with us?
Okay, Gokul, this is C.C. Wei. Let me emphasize that we always operate in good faith and support all our customer openly and fairly. The IDM customer has been the same. That's also TSMC's good customer. We also understand that IDM customer has their own plans for future insourcing, and we already have taken this into our capacity planning consideration. Did I answer your question?
Is there any ways for you to protect your longer-term growth when you deal with this kind of a customer? Just wanted to understand how it differs from your traditional fabless customers where I think that insourcing question is not really on the table.
Well, as I said, we have already taken this into our capacity planning consideration. Our capacity planning is based on the long-term market demand profile, underpinned by the industry megatrend of 5G and HPC, and the semiconductor content re-enrichment in various end devices. We do not depend on any one single customer or product. Okay.
Okay, Gokul.
Okay, that's clear.
Yeah.
Thank you very much.
Thank you, Gokul. Operator, can we move on to the next participant, please?
Next one to ask questions, Roland Shu from Citigroup.
Hi, good afternoon, and congrats for your very good result and outlook. First question from me is on your C.C. You also mentioned you always build your capacity according to customers' long-term demand profile. In the past, from time to time, you overbuild capacity for some nodes due to customers' focus always change. How confident you are this time to collectively spend a huge CapEx from 2021? And why do you think customers' focus at this time are more real than any time before? And also, what's your take of this semiconductor cycle? Will this cycle change or will it risk your CapEx spending plan going forward? Thanks.
Okay, Roland. So Roland's first question is really, you know, C.C. said that we build our capacity according to the long-term demand profile. His question is though, you know, customers' forecasts can change, so how confident can we be, and how do we see this current cycle playing out? Is that correct, Roland?
Yes.
Okay. Let me answer this question. What is the difference? Actually, this time, we see a structural increase in long-term market demand due to the multi-year industry mega trend of 5G, HPC, and digitalization. As well as some of the short-term imbalance that interruption of the supply chain brought about by COVID-19, and geopolitical tension. Let me say that, how confident we are. Very confident because as I said, long-term structural increase in the content and in the unit. The discount we had, the content increase as one of the important factor which we never reported before, and it was driven that the semiconductor industry to at a higher utilization rate. Because we have a very good technology leadership.
Yeah. I have also a question, part of the question is for the cycle. What's your take on the semiconductor cycle now?
Well, we cannot predict the cycle, right? Even if there is a cycle coming, we do believe that TSMC with its technology leadership and excellent manufacturing and the customers' trust will be a better position in the upturn or downturn cycle.
Okay. Understood. Thanks. Yeah, for my second question, you know, with your faster revenue growth and better margin going forward, for TSMC apparently needs more capable and experienced management and employees for continuous growth going forward. However, some of your senior management are approaching legal retirement age, so how are you going to retain the experience and the valuable management going forward? Also, by the way, what's the progress of your talent recruiting for your R&D and the manufacturing staff worldwide? Thanks.
Okay. Roland's second question is that his question really is around our talent. Part of it is our senior management. How do we keep you know our experienced and senior executives. Also, as we expand our global manufacturing footprint, what is the progress of our talent recruitment of engineers and R&D?
Roland, this is Mark. You're right. Our capable executives are our treasures, and they will bring the company forward. As long as they're energetic and they will want to contribute for this company, there's no forced retirement. For those top-notch executives, we will work with them if they want to stay. Yeah.
Also how about our talent recruitment, globally as we expand?
Yes. Talent recruitment, it is currently our focus, as the company deals with this fast-paced expansion. We particularly emphasize on the overseas recruiting. As you can see, our expansion into U.S. manufacturing and also in Japan manufacturing and are the vehicle that we will be able to reach for more global talents through those operations, and they may extend to local R&D. That is a part of our strategy.
Okay.
Okay. Thanks.
Thank you, Roland. Operator, can we move on to the next participant, please?
Next one, we have Brett Simpson from Arete Research.
Yeah. Thanks very much. My first question is for Wendell. Wendell, can you share with us how much prepayments and government subsidies TSMC received in 2021, and how we should think about prepayments and subsidies in 2022? Also, how do you account for this, particularly the subsidies, how do you account for this in the P&L going forward? Thank you.
Okay.
Okay. Oh, sorry.
Yeah, go ahead.
Sorry. Just to summarize Brett's question. His first question, he wants to know how much prepayments and government subsidies that we received in 2021, how much do we expect in 2022, and how do we account for these in our financial statements?
Okay. Brett, let me talk about the prepayment first. At the end of last year, we have received a total of $6.7 billion in prepayment, and you can, those are included in the financial statement as temporary receipts from the customers.
Going forward, you can look at our quarterly financial statement and find those numbers. As to subsidies, you know, different countries have different incentives, and they come in different forms. Some of them are related to asset reduction, some of them offset expenses, and some of them are tax reductions. We follow that. We use different accounting treatment to record that in the financial statements.
Okay. Thank you, Wendell. I guess the prepayments this year will be higher than TWD 21. Is that a fair statement?
Well, I cannot share with you the details, but we expect that there are more.
Okay. Thank you. My second question, looking at the smartphone segment of your business, it grew 8% in 2021, which is, you know, well below some of your big customers. You talk about silicon content drivers. 5G has been a really big silicon content driver in 2021, and it will continue to in 2022. So I'm just wondering, can you maybe just share with us your perspective on what happened in 2021 in smartphones? Why only 8% growth and how should we think about smartphones and the positive silicon content drivers from 5G this year?
Okay. Brett's question is focused on the smartphone. He notes that our smartphone platform grew 8% in NT dollar terms year-over-year in 2021. His question, I guess, Brett, if I'm hearing you correctly, your question is sort of, you know, what is driving the slower growth in the smartphone despite the higher silicon content?
Well, let me answer the question.
Right.
I mean, that's in terms of NT dollars, of course it's only 8%, but in terms of the US dollar, it's much higher. All right? Then the silicon content continues to increase every year. We expect that will be one of the major contributors to TSMC's growth continuously.
Okay, thank you very much.
Yeah. I think, let me add, the global smartphone unit growth last year is about 6%. You see some of the companies' smartphone revenue may grow. It could be due to the pricing. Our pricing strategy, as you understand, is strategic, not opportunistic. We grow with the smartphone units in our business.
Thank you very much.
Okay. Thank you, Brett. Operator, can we move on to the next caller, please?
Next one is Charles Shi from Needham & Company.
Hey, thank you for taking my question. Just a very quick clarification. C.C. Wei, I think you mentioned about the N3, the number of tape-outs in the first year. You expect that to exceed the number of N5 tape-outs in the first year. Since now you have both N3 and N3E, is a comment on original N3 alone or N3 family, which includes both N3 and N3E?
Okay. Charles' first question is that, he wants to clarify in terms of tape outs. When we say N3 tape outs are greater than N5 in its first year, is this N3 alone or N3 plus N3E?
Why you say N3 for right now? Because of N3E, we are, technology will be ready soon, and, the mass production will be one year later. Now most of the tape out, if all of them are N3.
Thank you. Now I want to ask my question on potentially the gross margin profile for the N3. I know it's gonna start to meaningfully contribute to revenue next year. You've said in the past that the newest process node will take about 7-8 quarters to reach the corporate average. Now this time, I think one difference about N3 is that you have both smartphone and HPC being supported at the beginning of the launch of the N3 node. My guess is that the revenue for smartphone and HPC ramp could really go parallel instead of one after another, and which could really help your volume, your revenue.
Could you comment on whether that 7-8 quarters of the margin dilution of the newest process node may be starting from N3, it will become shorter? That's my question. Thank you.
Okay. Thank you, Charles. Please allow me to summarize your second question. Charles' second question is looking at the N3 profitability and gross margin. He observes in the past, a new node typically takes seven to eight quarters to reach the corporate average. At N3, we have developed both smartphone and HPC platform. If these ramp in parallel, could N3 actually reach the corporate average in a shorter duration of time?
Okay. Charles Shih, this is Wendell Huang. It is too early to say when N3 can reach the corporate average gross margin at this stage, as the volume production has not started yet. However, the initial outlook for every new node always looks challenging, and the increasing process complexity of leading nodes, such as N3, brings even greater challenges. We will continue to work on selling our value and cost improvement to ensure that we earn the right profitability and returns.
Okay, Charles, does that answer your second question?
Yes. Yes, indeed. Thank you very much.
Okay. Thank you. Operator, can we move on to the next participant, please?
Next one to ask question, Mehdi Hosseini from SIG.
Yes, sir. Thanks for taking my question. The first one actually is a follow-up on the gross margin. For the next quarter, you're guiding to 53%-55%, but your longer term gross margin is around 63%. Does that mean that as we look beyond the Q1, there is a downward trend in gross margin? And also, how should we think about the long-term operating margin? And I have a follow-up.
Okay. Mehdi's first question is around gross margin. He notes our first quarter gross margin guidance is 53%-55%. But Mehdi, let me clarify. Our long-term gross margin guidance is 53% and higher gross margin. Nonetheless, Mehdi is asking, therefore, are we implying saying, you know, the margin outlook for the rest of the year will come down or decline.
Okay, Mehdi, this is Wendell Huang. We're not prepared to talk about gross margin outlook for the subsequent quarters of the year. Please be reminded that there are six factors affecting our profitabilities. Those factors include the ramp and development of our advanced technology, price, cost, mix, utilization, and foreign exchange rate. The foreign exchange rate is something that we cannot control and hard to predict. Summarizing all those factors together, we're saying that the long term we expect our gross margin to be 53% and higher. As to operating margin, we're not giving out the operating margin guidance as of now.
Okay. Thank you. Quick follow-up. Can you also help us understand how the mix of revenue by process technology like 7, 5, and 3 would change in 2022 versus 2021?
Okay. Mehdi's second question is asking revenue contribution by node. He is asking that N7, N5, and N3, what contribution will that be in 2022, and how is that different versus 2021?
Okay, Mehdi, we're not prepared to give the breakdown for the revenue contribution by node today. The N5 will continue to ramp, so we expect the revenue contribution to continue to rise throughout the year.
I think we have said before, Mehdi, that N3 will begin the volume production in second half 2022, and you will start to see the revenue contribution in 2023.
Okay, great. Congratulations on great execution.
Okay. Thank you, Mehdi. Operator, can we move to the next participant?
The next one to ask question is Sunny Lin from UBS.
Hi. Good afternoon. Thank you for taking my question. Happy New Year. My first question is on N3. You mentioned that the number of tape outs for N3 is higher than N5 in the first year. Now that we are getting pretty close to the mass production, I want to get your sense on the overall capacity outlook for N3 compared with N5. Also, if you could give us any sense of N3 revenue contribution in the first year of mass production.
Okay, Sunny. Please allow me to summarize your question. Her first question is on N3. Will N3 have a higher capacity or scale than N5?
This is C.C. Wei. I would like to say I'm not able to comment on the specific capacity by node. With a strong level of customer interest and engagement, N3 will be another large and long-lasting node for TSMC, just like our N5 and N7.
Sunny is also asking about the revenue contribution of N3.
It's still too early.
Yeah.
To talk about it.
Sure.
Yeah.
No problem. A follow-up would be that if we think about the N5 ramp-up, I think the second wave customer adoption seems to only occur in 2022, which is about two years after mass production. Looking at N3 with higher contribution and faster adoption of HPC, could you help us think about what the ramp up could be like? Would the scale pick up a bit faster because of HPC? Thank you.
Okay. Sunny's second question is also on N3, and she notes her question is that, you know, in N5, the second wave of adoption occurred two years after the initial production. With N3 and the higher and greater interest from HPC customers, would N3 see a faster ramp up?
Well, all I can say right now is, our customers' engagement on N3, N3E are quite, you know, more than what we observed in the N5. However, how to quantify that ramp up is too early to say. The engagement is very strong, all I can say that.
Sure. Maybe a very quick follow-up, is the equipment supply. Are you seeing any potential bottleneck for you to ramp up a larger capacity for N3 next few years, especially considering potential disruptions of ASML's EUV tools?
Okay. Last question from Sunny is that in terms of our capacity expansion, are we seeing any bottlenecks or potential bottlenecks in our capacity plans, particularly in terms of ASML's EUV tools?
Sunny, this is a very good question. All I can say right now is that 2022 we are okay, and now we are working on 2023 so that we can ramp up the capacity to meet customers' demand.
Got it. Thank you very much.
Thank you, Sunny. Operator, let's move on to the next participant, please.
Next one, we have Sebastian Hou from Neuberger Berman.
Hello, can you hear me?
Yes, we can.
Hello. Yes, thanks for taking my question. I only have one. This is more of a mathematical question. If I look past TSMC there, your pack, how long did you translate your CapEx to your revenue? It looks around like 1.7 times, which means that you've used double your CapEx in three years, you'll probably double your revenue in five years or triple or more similar to that kind of equation. I'm curious about this time around whether that kind of relationship or translation time will hold. And then if I look at your CapEx numbers in 2019 is where you start to move our CapEx aggressively, and in three years, you triple CapEx.
In four years, you quadruple your CapEx. If that equation continue to hold, then does that mean that your by 2023 or 2024 your revenue will exceed $100 billion, and by 2025, your revenue will be around like $130 billion-$140 billion, which means that your long-term or next three, four years revenue CAGR not just 15%-20%, but actually 25%-30%.
Okay. Sebastian, his question really is about the relationship between CapEx and revenue. He's asking, on his calculations of course, how long does it take CapEx to translate into revenue? He observed in the past that, you know, CapEx doubling in 3 years led to revenue doubling in about 5 years. Now he wants to apply it to this time to say, well, if CapEx has tripled from 2018, then can he expect TSMC revenue to be $100 billion? Sebastian, you said in 2024-
3-4.
Oh, 3-4.
Yeah, around that level. Yeah.
2020, so a TWD 100 billion level in kind of 2024, TWD 130 billion-TWD 140 billion in 2025.
Okay. Sebastian, this is Wendell. Well, frankly speaking, we didn't do that kind of math ourselves. What our guidance is, as C.C. mentioned earlier today, is 15%-20% CAGR in the next several years. With CAGR doesn't mean that every year we'll grow that kind of a numbers. Also, it's next several years. It's a longer term numbers. I would not try to use the mathematics that you just used. It's things are not that simple, I think.
Yeah. Thank you. Definitely not that simple, but maybe, let me simplify the questions a bit or put another way, is that the CapEx to revenue translation time, do you see any change on that now versus 5 or 10 years ago? Is it getting slower or faster or the same kind of the pace?
Yeah. Go ahead.
Yeah. I think Sebastian's asking, so, to try to simplify it in terms of CapEx, to revenue, are we seeing it, at the same pace as in the past? Is it getting longer? Is it getting shorter? I think Chairman can address this.
Answer this. I think, because of the equipment lead time is much longer, and also the technology complexity is longer. That might increase a little bit in terms of the lead time leading to revenue.
Okay. That's all from me. Thank you.
Okay. Thank you. Thank you, Sebastian. Operator, can we move on to the next participant, please?
Next one, we have Laura Chen from KGI. Go ahead, please.
Hi. Thank you. Happy New Year, and thanks Manny for taking my question. I have a question on your overseas expansion. Just wondering, will TSMC consider a joint venture in Europe, like what you announced in Japan? What's your current preference? Still prefer 100% own or you would be more open for potential joint venture like what TSMC did a long time ago in WaferTech and in Singapore with NXP before? That's my first question.
Okay. Laura, let me summarize. Your first question is about our overseas fabs, and she's wondering will we consider joint venture as a future model going forward, whether it's in Europe or elsewhere? As this is something that we have done with WaferTech in the past and also recently with Japan. Is JV the new model going forward?
Yes. Thank you.
Yes. Laura, this is Mark. To answer your first question about Europe, this is still very early stage we are assessing. To start a overseas fab, there are many considerations. Among them, the top few is, first is our customers' needs. In this current planning in the Japan fab, indeed, it was a joint venture. We haven't done joint venture for many years, and we think this joint venture is also a special case. Typically, every TSMC fab, no matter where it's located, will serve all the customer from around the world. This Japan joint venture will also the same. However, in Japan, we have a very large customer who has a single technology.
We can also leverage their operating manufacturing experience in Japan, which helps us ramp in the learning curve. That made us make the decision of a joint venture fab with Sony where we have a majority share. This is a special case, and we typically consider proceeding with a solely owned fab with 100% ownership. Yeah.
Yes. Thank you. That's very clear. Also my second question is, specifically on the N6-based RF transceiver. I recall that in your comments back in June last year, you mentioned that the N6-based RF transceiver for 5G. Can you give us more update, as we know that there is not much expansion on 16 or 12 nanometer, which are the major technology now for a 5G RF transceiver. With the limited supply, just curious about that N6-based 5G RF transceiver. Would that become the mainstream or, what's TSMC's capacity plan in this area, and also the client engagement for the N6-based RF? Thanks.
Okay, Laura, let me try to summarize. She had questions about RF transceivers for 5G. She wants to know, there doesn't seem to be any major capacity expansion. What is TSMC's strategy for RF transceivers for 5G? And also N6. You're talking N16 or N6? Sorry, Laura.
Uh, N6-
N6.
Because right now most of the
Right
RF transceiver are in N16, up to my understanding.
Right.
Yeah.
She wants to update.
There is not much expansion, right?
... on our N6, RF transceiver, strategy.
Yeah. Thank you very much.
Okay, Laura.
Yes.
We always are working with our customer closely, right? The customer make their decision to choose which technology node and to match their product design best. You are right. Right now, transceiver is starting moving from 28-nanometer to N16 and now moving to N6. We are extending our capacity to meet the demand. That's all I can say. Does that answer your question?
Okay. Yes, thank you. I think, can I follow up that will we expect that, N6-based RF will be the majority sometime, say, in 2023 or 2024?
Laura's follow-up is, can we expect N6 RF to be the majority by 2023 or 2024?
Well, Laura, I should not comment on that. This is between TSMC and TSMC's customer.
I got it. Thank you very much.
Okay, thank you. Operator, in the interest of the time, I think we'll take the final two questions.
Krish Sankar from Cowen and Company.
Yeah. Thanks for taking my question, and congrats on the really strong results. My first question is on gross margins. Wendell, you said long-term gross margin about 53%. The last couple of quarters, you said it'll be over 50%. Is it fair to assume that the price increases are the big reason for this increase in gross margin, and are these structural or these cyclical? And is there some other variable in play given the gross margin improvements, since it's interesting that CapEx is going up, but the depreciation is not having an impact longer term on the gross margins? That's my first question.
Okay. Krish, let me summarize your first question. Krish notes that Wendell and C.C. said our long-term gross margin target last time we said 50% and higher. Now, today we said 53% and higher. Is this because of price? And is this a cyclical element only, or is this something structural in terms of a higher 53% and higher long-term gross margin target?
Yeah. Well, let me share with you that we're talking about long-term, so several years down the road. I think that shouldn't be a cyclical issue. Previously it's long-term 50% and higher. Now, it's long-term 53% and higher. Okay. That's the difference. We are working closely with our customers and suppliers to both sell our value and drive our cost improvements. This is the result of all these efforts together.
Got it. All right. My second question is kind of a two-part question. It's a CapEx and an OpEx question. You spoke about $40 billion-$44 billion in CapEx. Can you just tell us how much of the CapEx is gonna be for growth? How to think about OpEx growth relative to inflation?
Okay. Krish's second question, first, is in terms of the CapEx guidance that we gave this year, $40 billion-$44 billion. He wants to know what portion is for overseas capacity as compared to Taiwan.
Okay. For this year, out of the TWD 40 billion and TWD 44 billion, we really cannot comment on the split between overseas and Taiwan. We normally split or give the breakdown guidance between advanced and mature specialty nodes.
The second part is on OpEx. What is the impact? Although we're not giving an operating margin guidance, but rising global inflation, we're trying to recruit talent overseas, what is the outlook for our OpEx?
Well, we try very hard, work very hard to control our operating expenses and to achieve operating leverage. You can probably see from the track records in the past few years and past few quarters that we're always able to achieve operating leverage. That's something that we will continue to work very hard on.
Got it. Thank you very much.
Okay. Thank you, Krish. Operator, in the interest of time, we will take the last participant's questions.
Last one to ask questions, Frank Lee from HSBC. Go ahead, please.
All right. Thank you for taking the question. Just had a question, maybe if I can follow up on the CapEx. I think a year ago you guys gave a longer-term three-year CapEx target. I understand that this time your CapEx is only given for 2022. Just wanted to understand, is there more uncertainties about giving a longer term CapEx that is why we're not seeing maybe a longer term target being given? I think last quarter you guys had also talked about your CapEx intensity, and it's gonna be more than 50% in 2021, but long-term target you set in the 30s%.
I believe last quarter, I think it was implied that it was not gonna come down to the thirties in the next two or three years. Is that still the case? That's the first one.
Okay. Frank's first question is on CapEx. He said we had talked about $100 billion CapEx in the next three years. His question is, "Why are we not guiding for next three years CapEx now? Is there some concern we have?
Okay. Frank, we will not comment on CapEx outlook beyond 2022 to date. Please be reminded that every year our CapEx is spent in anticipation of the growth that will follow. As long as our growth outlook looks good, then we will continue our investment approach, the disciplined investment approach to support our customers and capture the growth opportunities. As regards to the-
Okay.
Yeah, the capital intensity.
Sorry.
Yeah. As I said earlier that in the period of high growth like today, it is quite normal or it is appropriate for the capital intensity to be high. If our growth were to slow down-
The capital intensity is expected to decline accordingly. From what we can see at this moment, you're asking several years from now. Probably, if a normal situation, we're probably in the mid-thirties.
Is that likely in the next two or three years, or it's gonna be further out the way that
Long term, several years.
Okay.
I think we'll give you year by year updates.
Okay. Thank you.
Yes. Okay?
Okay. Great. I guess my second question is, I think a consistent message on this call was really about the semi content, structural semi-content growth. I think you guys have mentioned this, a couple of times on this call. Over the past year, though, I guess just wanting to get your sense of if the semi content growth is something that seems like, you know, has been well understood, but has there been a incremental surprise in terms of the semi content, for you guys over the past year that's a more positive, structural long-term growth? If so, can you share maybe by what application or are we seeing content growth in mature nodes as well?
Okay. Frank's second question is on the semiconductor enrichment. His question is what, you know, what are we seeing? Have we seen upside in the last year? In terms of what specific applications are driving the semiconductor content enrichment. Maybe CC can address this last question.
Yeah, sure. I will give you a live example. If you look at the new car being introduced in the market, you'll find out that the semiconductor content has been dramatically increased. The same happened to, you know, data center, server and et cetera, etc.
Okay.
Sorry to step up, Robin. Is it fair to say then that the semi-content growth is something that is structured across all nodes, including advanced as well as some of the specialty areas?
Frank's question then with the semiconductor enrichment, is this only for the leading nodes or is this also hold true at the mature nodes as well?
Well, it's across all the technologies that I can say. I give you one example again. If you look at the ADAS, that's a leading-edge technology. If you look at the power management, that's a mature technology.
Okay. All right, Frank. Thank you.
Okay. Thank you.
Thank you, CC. Thank you, Frank. 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 1 hour from now. The transcript will become available 24 hours from now. Both are gonna 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. Happy New Year, and we hope you join us again next quarter. Goodbye, and have a good day.