[Non-English content]
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 2020, followed by our guidance for the fourth quarter 2020. Afterwards, TSMC's CEO, Dr. C.C. Wei and Mr. Huang, 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 in our press release. Now I would like to turn the call over to TSMC's CFO, Mr. Wendell Huang, for the summary of operations and the current quarter guidance.
Thank you, Jeff. Good afternoon, everyone. Third quarter revenue increased 14.7% sequentially in TWD or 16.9% in USD. As we saw strong demand for our advanced technologies and special technology solutions driven by 5G smartphones, HPC and IoT related applications. Gross margin increased 0.4 percentage point sequentially to 53.4%, mainly thanks to a much higher level of utilization, partially offset by the margin dilution from N5 ramp and an unfavorable exchange rate. The operating expenses increased by TWD 7.4 billion , mainly attributable to a higher level of development activities for N4 and N3 technologies, and one-time expenses to facilitate our expansion in Hsinchu. Therefore, operating margin slightly declined by 0.1 percentage points sequentially to 42.1%. Overall, our third quarter EPS was TWD 5.3 and ROE was 31.3%. Now let's move on to the revenue by technology.
Five nanometer process technology contributed 8% of wafer revenue in the third quarter, while seven nanometer and 16 nanometer contributed 35% and 18% respectively. Advanced technologies, defined as 16 nanometer and below, accounted for 61% of wafer revenue. In terms of revenue contribution by platform, smartphone increased 12% quarter-over-quarter to account for 46% of our third quarter revenue. HPC increased 25% to account for 37%. IoT increased 24% to account for 9%. Automotive decreased 23% to account for 2%. Digital consumer electronics decreased 24% to account for 3%. Moving on to the balance sheet. We ended the third quarter with cash and marketable securities of TWD 742 billion. On the liability side, current liabilities decreased by TWD 27 billion, mainly due to the decrease of short-term loans and the decrease of current portion of bonds payable.
Long-term interest bearing debts increased by TWD 146 billion, mainly as we raised TWD 145 billion of corporate bonds during the quarter. On financial ratios, accounts receivable turnover days decreased four days to 40 days, while days of inventory increased three days to 58 days, primarily due to N5 ramp.
Regarding cash flow and CapEx, during the third quarter, we generated about TWD 190 billion in cash from operations, spent TWD 99 billion in CapEx and distributed TWD 65 billion for fourth quarter 2019 cash dividends. Short-term loans decreased by TWD 17 billion, while bonds payable increased by TWD 136 billion, mainly due to the bond issuances. Overall, our cash balance increased TWD 137 billion to TWD 604 billion at the end of the quarter. In US dollar terms, our third quarter capital expenditures totaled $3.4 billion. I have finished my financial summary. Now, let's turn to our fourth quarter guidance.
Based on the current business outlook, we expect our fourth quarter revenue to be between $12.4 billion and $12.7 billion , representing a 3.4% sequential increase at the midpoint. Based on the exchange rate assumption of one U.S. dollar to TWD 28.75 , gross margin is expected to be between 51.5% and 53.5%. Operating margin between 40.5% and 42.5%. I will hand over the call to C.C. for his key messages.
Thank you, Wendell. Good afternoon, everyone. We hope everybody is staying safe and healthy during this time. Let me start with our near-term demand and inventory. We concluded our third quarter with revenue of TWD 356.4 billion, or $12.1 billion, which was above our guidance, mainly due to better demand across all our platforms than our forecast three months ago. Moving into fourth quarter 2020, we expect our sequential growth to be supported by strong demand for our industry-leading 5 nanometer technology, driven by 5G smartphone launches and HPC related applications. On the inventory front, we forecast our fabless customers' overall inventory to exit the year above the seasonal level as the supply chain continues to make efforts to ensure supply chain security and actively prepare for the new 5G smartphone launches.
Looking ahead, we expect our customers' overall inventory to remain above the historical seasonal level for a longer period of time, given the industry's continued need to ensure supply chain security amidst the lingering uncertainties. For the full year of 2020, although COVID-19 continued to bring some level of impact to the global economies, we also observed that COVID-19 is accelerating digital transformation while 5G and HPC related applications continue to drive semiconductor content enrichment. We now forecast the overall semiconductor market, excluding memory, to increase mid-single digit %, while foundry industry growth is expected to be close to 20% year-over-year. For TSMC, our technology leadership position enable us to capture the industry mega trend of 5G and HPC. We expect to outperform the foundry revenue growth and grow by about 30% in 2020 in U.S. dollar terms.
Let me talk about our N5 ramp-up and N4 progress. TSMC's N5 is the foundry industry's most advanced solution with the best PPA. N5 is already in volume production with good yield, while we continue to improve the productivity and performance of the EUV tools to further enhance our leadership in EUV technology. Due to the robust demand from 5G smartphones and HPC applications, we reaffirm N5 will contribute about 8% of our wafer revenue in 2020, and we expect even higher percentage in 2021. N4 will leverage the strong foundation of N5 to further extend our five nanometer family. N4 is a straightforward migration from N5 with compatible design rules while providing further performance, power, and density enhancement for the next wave five nanometer products. N4 risk production is targeted for 4Q 2021, and volume production in 2022.
With our continuous technology enhancement, we expect our 5-nanometer family to be a large and long-lasting node for TSMC. Now, I will talk about our N3 status. N3 will be another full node straight from our N5, with up to 70% logical density gain, up to 50% performance gain, and up to 30% power reduction as compared with N5. We have chosen FinFET transistor structure for our N3 technology to deliver the best technology maturity, performance, and cost for our customers. Our N3 technology development is on track with good progress. N3 will offer complete platform support for both mobile and HPC applications. Risk production is scheduled in 2021, and volume production is targeted in second half of 2022. Our 3-nanometer technology will be the most advanced foundry technology in both PPA and transistor technology when it is introduced.
Thus, we are confident our 3 nanometer will be another large and long-lasting node for TSMC. Finally, I'll talk about the TSMC 3DFabric. TSMC has developed an industry-leading and comprehensive wafer level 3D IC technology roadmap to enhance system-level performance. Our differentiated chiplet and heterogeneous integration technologies drive better power efficient and smaller form factor benefits for our customers while shortening their time to market. These technologies, including chip stacking solutions such as SoIC, as well as advanced packaging solutions such as InFO and CoWoS. We are consolidating this offering under one umbrella and naming it TSMC 3DFabric. As our industry continue to seek innovations to enhance system-level performance, 3DFabric will complement our advanced technology to unleash our customers' innovation.
We expect revenue from our back-end services, which include both advanced packaging and testing, to grow at a rate slightly above the corporate average in the next few years. Now let me turn the microphone over to Wendell.
Thank you, C.C. Let me start by making some comments on our profitability. Our 3rd quarter gross margin exceeded the high end of our guidance to reach 53.4%, mainly as we saw a much higher than expected overall capacity utilization rate in the 3rd quarter. That helped to offset the margin dilution from the initial ramp-up of our 5 nanometer technology. We have just guided 4th quarter gross margin to decline by 0.9 percentage points sequentially to 52.5% at the midpoint, primarily due to the margin dilution from the continued steep ramp-up of our 5 nanometer and a less favorable foreign exchange rate in the 4th quarter. Looking to 2021, we expect a strong ramp of N5 to contribute a higher percentage of revenue as compared to 2020. The yield rate of N5 continues to improve.
Similar to prior nodes, we forecast N5's gross margin to take seven or eight quarters to reach the corporate average level. Thus, N5 is expected to dilute our gross margin by about two to three percentage points for the full year of 2021. As a reminder, the following six factors determine TSMC's profitability, leadership technology development and ramp-up, pricing, cost reduction, capacity utilization, technology mix, as well as foreign exchange rate. Taking all these factors into consideration, we believe a long-term gross margin of about 50% is achievable. Now let me talk about our capital budget for this year. Our business outlook is supported by strong demand for our industry-leading advanced technologies and specialty technology solutions, driven by the industry megatrends of 5G and HPC-related applications.
In order to meet this demand and support our customers' capacity needs, we now expect our full year 2020 CapEx to be about $17 billion. Now I will make some comments on our corporate bond issuances and capital structure. The multi-year megatrends of 5G-related and HPC applications are expected to continue to drive strong demand for our advanced technologies in the next several years. Given the macroeconomic uncertainties this year, a current low interest rate environment, and ability to diversify our funding sources, TSMC's board of directors has so far approved the issuance of TWD 120 billion in corporate bonds and $4 billion in corporate bonds. Year to date, we have issued TWD 89.5 billion and $4 billion in corporate bonds with favorable pricing terms.
With our solid financial performance, strong balance sheet and cash position, and capacity to take on debt, we are able to aggressively invest in our future to enhance our technologies and capabilities. This enables us to continue to outgrow the semiconductor industry through the cycles. With our disciplined capital management, we remain committed to a sustainable cash dividends on both an annual and quarterly basis.
Thank you, Wendell. 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 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 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. Now let's begin the Q&A session. Operator, can we please proceed with the first caller on the line?
The first caller on the line is Gokul Hariharan, JPMorgan. Go ahead, please.
Congratulations on great quarter, and thanks for taking my question. My first question is on CapEx and capital intensity. Looks like this year you could come in around 36%, 37% capital intensity. Could we talk a little bit about how we should think about capital intensity and absolute CapEx as well, looking forward, at least on a directional basis? Feels like the investment cycle is still going to be pretty much intact going into next year.
Also, looking at some of the financial options in terms of bond rating, et cetera, that TSMC has undertaken. That is my first question. My second question is on N5. I think in previous calls, we had indicated that while N5 will be a long and large node, it may not have the same number of tape outs as N7 have had, which is probably the historical high. Is our view changing on N5?
Could we talk a little bit about will N5 exceed N7 in terms of wafer capacity as well as wafer revenue in the next two years or so? Thank you.
Okay, Gokul. Thank you very much. We'll take your questions one by one. Please allow me to summarize your question. Your first question relates to our CapEx and capital intensity. You point out that with the guidance that our capital intensity this year, in your estimation, is probably around 36%-37%. Your question is, how should we think about CapEx and capital intensity in the next few years? If we cannot give a quantitative number directionally, how do we see CapEx and capital intensity, and how does this tie in with our recent things like such as bond issuances and fundraising? How does that factor in? That's the first question. Maybe CEO Wendell can address.
Yes. Gokul, our capital intensity, as you are right, this year will be lower than 40%. In the next several years, longer term, we expect the capital intensity to be around mid-30 percentage point. Having said that, there may be years where capital intensity is higher if we see the strong demand for our technologies or capacity, and we decide to invest.
Okay. Your second question, Gokul, please allow me to summarize again, is really regarding to our 5-nanometer, that we have said that it's a long and large node, but that the number of tape outs of N5 versus N7 may be lower. Your question is, can N5 exceed N7? Do we believe 5-nanometer can be a bigger node than seven in terms of revenue and capacity?
Well, let me say that, we don't comment on how many tape outs so far, but we continue to see strong tape out activities at N5 from both HPC and the smartphone applications. The revenue for this year, we just mentioned, is at 8% of the wafer revenue. Next year, it will be even higher than close to or 20%, or something like that. The exact number, we are still not able to comment. I can assure you that our 5-nanometer family will be another big and long-lasting node for TSMC.
Okay. Thank you, Gokul. Why don't Operator, can we move on to the next caller, please?
Next one, we have Randy Abrams, Credit Suisse.
Okay. Yes, thank you. My first question I wanted to ask on, Wendell, you raised the gross margin. Originally, it was 50%. Could you discuss now where you're saying it could be above 50%, the factors driving that change? Could you clarify on the two to three point impact on 5-nanometer? I think you already have that impact. Does that apply for next year, pretty similar to the type of gross margin you're running now are potentially even better?
Okay. Randy, I summarize your question. Your first question is in regards to, I believe, our gross margin, and long-term gross margin. I think you're asking that, we raised our target. I think we, as Wendell said, 50% is achievable for us. You're also asking as part of that, the dilution from 5-nanometer, how will that impact our gross margin next year, and where should we, I guess, be thinking about gross margin for 2021?
Okay.
Randy, maybe let me answer this like this. We have a very high gross margins in the third quarter. We believe we will continue to have a pretty high margin in the fourth quarter. Main reason is that we are enjoying a very high utilization across almost all the nodes at this moment. The very high utilization may not continue forever. Our long-term growth target or long-term growth goal for our gross margin, it continues to be above 50%. In terms of dilution from N5, we see the dilution of N5 for next year to be around two to three percentage point, similar to previous nodes. Remember that the N5 will account for a much bigger percentage of our revenue next year. As we ramp up quickly, the dilution will continue to exist.
We are still expecting that it will reach the corporate margin in seven to eight quarters.
Okay. No, great. Thanks. I misunderstood. I thought I heard the word above for 50, but thanks for the clarification. Second question on the recent U.S. restriction on SMIC. I'm curious if you're seeing any additional diversification or inquiries for business, and given they're more on the mature nodes, how you're positioned, if you are seeing those, to take on business on the mature nodes.
Okay, Randy, let me just summarize your second question. Your second question is in regards to the recent restrictions on SMIC. Randy is wondering whether we are seeing any types of diversification or inquiries from customers in regards to business, and especially at the mature nodes.
Well, Randy, let me answer the question. Actually, we are still evaluating the impact to the semiconductor industry due to the ban on SMIC. Let me say that our capacity planning and all our CapEx are continued based on the long-term demand profile that is underpinned by the industry megatrend, such as 5G related and HPC application. All right? Does that answer your question?
Yeah. Maybe just one quick, but for the mature nodes which are running tight across the industry, just if it's so that there's an incremental surge, how well could you handle incremental business from, say, this type of piece if it were to come through?
Randy is asking if we were to see a surge in demand at the mature nodes, how ready or do we have capacity to take on or handle this type of surge demand?
Well, we continue to work with our customers dynamically, and we try our best to meet their demand. That's all I can say for today.
Yeah.
Okay, great. No, thank you.
Okay. Thank you, Randy. Operator, can we move on to the next caller, please?
The next one is Sebastian Hou from CLSA.
Thank you. Good afternoon, gentlemen. My first question is, I think besides the higher than usual inventory, which may be a new norm because of this supply chain fear of disruption, I'm curious about how does TSMC assess customers' overbooking or pull behavior and the magnitude? In particular, based on the recent smartphone OEMs aggressive procurement appetite, assuming Huawei is going to be dead next year, how do you assess that kind of potential overbuilt inventory risk that may potentially lead to a destocking correction sometime next year? This is my first question. Thank you.
Okay, Sebastian. Let me repeat or try to summarize your question. Your question is basically related to the inventory, and you want to ask how does TSMC assess the risk that there is overbooking in light of the restrictions on Huawei, and therefore, what type of levels or magnitude of inventory overbuild is there, and does this create the risk of inventory correction sometime next year?
Well, let me share with you our view on this inventory-related issues. First, I want to say that due to the pandemic, actually, the digital transformation has been accelerated, and that create a demand on 5G and HPC-related products. For the longer-term basis, we do expect our customers' overall inventory to remain above the seasonal level for a longer period of time. Majority partly because of they have some concern on industry's supply chain security and due to the uncertainties. That will be the high-level inventory will sustain, continue for longer period of time. That we can say that.
Okay, Sebastian, do you have a second?
Yeah. All right.
Sorry.
Yeah. Okay. Anyway, that wasn't actually what I'm looking for. Anyway, thank you for that, CC. My second question is, on the HPC business, apparently, I think, CC, you mentioned the primary market that you see a lot of growth this quarter and also continuous through next quarter, and driven by the accelerating digital transformation, you just said, led by the pandemic and work from home demand likely to stay for longer, and also the continuing market share gain from TSMC against IDM. When do you expect your HPC revenue exposure to cross over with smartphone revenue percentage? Possible to see that by end of next year or 2022? Thank you.
Okay. Sebastian, let me just summarize your question, which is regards to our HPC platform business. You point out that there is the trends of the accelerating digital transformation and the work from home, and also market share gains versus IDM. You want to know, when do we see our HPC platform revenue crossing over with smartphone or others to become the primary.
Okay. Let me answer the question. We do see HPC platforms of growth rate is higher among our four platform, which is smartphone, HPC, automotive, and IoT. In the next few years, we continue to expect, or we forecast that the HPC's growth will be higher than the corporate level. When you will cross over, I don't make any comment right now.
Okay.
Okay. Thank you.
Thank you, Sebastian. Operator, can we have the next caller, please?
Next one, we have Bruce Lu from Goldman Sachs. Go ahead, please.
Hi, good afternoon. I want to ask for the 5G penetration rate. What is the latest forecast for the total smartphone growth and the 5G penetration rate in 2020, and maybe a little bit color on 2021 as well. We also see that some of the telco is slowing down their 5G base station installation. What kind of impact we see at this moment?
Okay, Bruce. Your question is regards to 5G and smartphones. You want to know what is the smartphone growth and 5G penetration rate for 2020, as well as 2021. In light of the telecoms potentially slowing down the deployment. Correct?
Yes.
Okay. All right.
Yes. Thank you.
Let me answer the question. We continue to expect the faster penetration of 5G smartphone as compared to 4G. For this year, we still forecast a high teens penetration rate. Next year, even higher. Much higher, let me say that. That's all we have today.
Oh, any impact on the telco's business as well?
I think all countries and all regions are preparing to build up the infrastructure right now. I believe next year, even not 100% completed, but all the region, all the countries have a lot of 5G phone being introduced. That create a higher % penetration rate.
Okay. Understand. My next question is that, I'm a little bit surprised that China's revenue contribution only increased slightly from 22% to 23% in third quarter. Which region will we see the strongest growth in the fourth quarter?
Okay. Bruce, your question is regards to our revenue by geography. You want to know for the fourth quarter, which region will contribute the most growth in the fourth quarter.
Okay. Bruce, we're not prepared to comment on geographic allocation among revenues in the fourth quarter. I can share with you that we expect the platforms that will grow in the fourth quarter will be smartphone and automotive. The other two will likely to be down.
Understand that. Thank you.
All right. Thank you, Bruce. Operator, can we move on to the next caller, please?
Next one we have Sunny Lin from UBS.
Hi. Good afternoon. Thank you for taking my question. My first question is on 5 nanometer demand. Into next two to three years, what do you think revenue split could be by smartphone, HPC, et cetera? Do you think the mix could be a bit different from 7 nanometer?
Sorry. Can you repeat your question, Sunny? You broke up a little bit.
Sure, sure. No problem. Sorry about that. I wonder, for 5 nanometer demand into next two to three years, what does the management think the revenue mix could be by smartphone, HPC, et cetera? Would the product mix be a bit different from 7 nanometer?
Okay. All right. Let me summarize. Thank you, Sunny. Your question is regards to 5 nanometer. When we look out over the next three years, how do we see the demand of 5 nanometer, the mix changing in terms of smartphone, HPC, different platforms, and then how does this compare to 7 nanometer. Correct?
That's right. Thank you, Jeff.
Hi. We don't break it down or disclose the platform mix of certain nodes. We can share with you, as CC just mentioned, in the next several years, we expect HPC to be the largest contributor of our growth. That should give you some idea. These guys use advanced technologies.
Sure. Got it. My second question is that for this year, a key part of your growth in smartphones is driven by higher silicon content for 5G and your share gains. I wonder if you could walk us through how your average silicon content in smartphone may trend into 2021 and 2022. Thank you very much.
Okay. Sunny, your second question is regards to the silicon content in 5G phones. The silicon content increase in 5G phone, along with share gain, is contributing to our smartphone growth this year. She wants to know what is the silicon content outlook for 2021 and 2022.
This is pretty hard for me to answer because I cannot release all the information I got from my customer. Let me say that on the average, the 5G phone have about 30%-40% more silicon content as compared with the 4G. Did I give you some kind of idea?
Sure. I have a very quick follow-up. I wonder if you could give us some color regarding your expectation for your market share for smartphone into next two, three years.
Sunny is asking whether we can give some comment on our market share in 5G phones the next two to three years.
No. It's not very appropriate for me to give some kind of estimate right now. Let me say that as long as we have a technology leadership position, we are very confident that we are going to have a high market share.
Okay?
Sure. Got it. Thank you very much.
Thank you, Sunny. All right. Let's move on. Operator, can we move on to the next caller on the line, please?
Next, we have Roland Shu from Citigroup.
Hi. Good afternoon. My first question is, can you update the status of your license applications for shipment to Huawei? When do you expect to receive approval from U.S. government? Does your 4Q revenue forecast include any waiver shipment to Huawei? It's my first question. Thanks.
Okay. Roland, your question is regards to He wants an update of our license application status regarding Huawei. He also wants to know, does our fourth quarter guidance include any shipments to Huawei?
Roland, we are complying
Hi again.
fully with the regulations. We also noticed that there is a report saying that TSMC got the license. We are not going to comment on this unfounded speculation. We also don't want to comment on our status right now. For the 4Q shipment to Huawei, no. The ban, the regulation already say that after September 17th, zero.
15th.
September 15? Okay. Same.
Okay. Thank you. Okay. My second question is, how is the pricing pressure across the old technology node so far? Some of your foundry peers are considering to raise wafer ASP, given a very high utilization at 8-inch fab. Were you considering to follow to raise the pricing on 8-inch or on other mature technology node? Thanks.
Okay, Roland. Thank you. Your second question is regards to pricing pressure. Your note is that some of the foundry peers are considering to raise the 8-inch wafer price. You want to know, does TSMC plan to raise our 8-inch wafer pricing or also raise our pricing on the mature nodes?
Let me answer the question. The big answer is no. We continue to work with customers, and customer are our partners. For short-term supply shortage, definitely, we are not using this kind of opportunity to raise our price. Our wafer price, we are selling our values, our service to our customer, that including the technology, delivery, quality, everything. Certainly, TSMC is working with all the customer and view them as partners. We don't using this opportunity to raise our wafer price. Did I answer your question?
Yes. Thank you.
Okay. Thank you, Roland. Let's move on, operator, to the next caller.
Yes. Next, we're having Brett Simpson from Arete Research. Go ahead, please.
Thanks very much. I just had a question on your long-term capacity planning. You've laid out the view that we're going to see some structural tightness for the next couple of years in foundries, potentially. I'm just wondering if you see You have a very strong growth position in HPC, but you still have a very low market share in X86 or PC and servers broadly. I'm just wondering, if we do see Intel looking to outsource major CPU lines to foundry, it could be a large one-time boost to the foundry industry. Would TSMC be able to meaningfully support Intel's needs if there was a big one-time outsourcing, and would you be prepared to take capital intensity to much higher levels should the opportunity arise? Thank you.
Okay, Brett. Let me try to summarize your question. Your question basically is premised around our long-term capacity planning and pointing out that there's a structural tightness in Foundry, and we, TSMC, has a strong growth position. Your question specifically relates to X86 and Intel. If Intel were to outsource to Foundry, your premise is that this could be a one-time big outsourcing opportunity. How would we prepare or handle for this?
Well, let me say that we do not comment on the specific customers nor on the specific product. Let me say, our CapEx and capacity planning is based on the long-term demand profile that is underpinned by the industries that make a trend to meet our customers' demand. Intel is one of our important customers, and we continue to work with them.
Do you-
Okay, thank you. Maybe just a follow-up regarding your capacity plans over the near term. Are you planning to add any capacity at the mature nodes? Maybe not so much eight inch, but certainly sort of 28 nanometer or even 16 nanometer. Do you foresee putting any customers on allocation, given the backdrop with tightness at the moment? Thank you.
Okay. Brett, your second question is regards to our capacity plans in the near term, specifically at some of the mature nodes like 28 and 16 nanometer. Are we planning to add capacity and with the tightness, are customers on allocation?
Well, again, let me say that we plan our capacity to meet the customers' demand, whether it's a leading edge or mature node or specialties. We always work with customers dynamically and also work with them closely, so to plan our capacity. Definitely today, there are some shortages, but we are doing our best to serve our customers.
Okay. Thank you, Brett.
Thanks so much. Thank you.
Thanks a lot, Brett. All right. Operator, can we move on to the next caller on the line, please?
Next one to ask question, Charlie Chan from Morgan Stanley. Go ahead, please.
Thanks, and good afternoon, gentlemen. My first question is about your 2 nanometer progression, because I think couple weeks ago there was a news talking about you may see the 2 nanometer in mass production in 2024. Just want to get company's clarification about your progress here, maybe your technology roadmap, and that realistic timing for the mass production. Thank you.
Okay. Charlie's first question is in regards to our 2 nanometer. He says according to news reports that the production is going to begin in 2024. He wants to know whether we can share the technology roadmap requirements and the timing of our 2 nanometer.
Charlie, let me say frankly, we are not ready to make any comment on the two nanometer yet. All right?
Okay. Yeah, there seems to be some comments from your technology forum. Any reason why you can't disclose that to investor society yet?
No. I think, Charlie, all we have disclosed about our two nanometer is the location, which will be in Hsinchu. We have not commented on the technology specifications, the timing, or anything beyond that. That is you, as you said, according to your reading the news. That is not TSMC's comment. as CC said.
Okay
we are not prepared to comment on two nanometer.
Okay. No problem at all. My second question is maybe to Wendell about the gross margin trend follow-up. Based on your current depreciation table, when do you think the depreciation is going to peak, in the coming years or coming quarters, at what point? Also, I think you mentioned that the new node ramp is a key factor to the gross margin dilution. I think four nanometer is a part of the five nanometer family, right? Can we expect that in 2022, there's not going to be any kind of margin dilution from the four nanometer? Thank you.
All right. Charlie, your second question is regards to depreciation and gross margin. Charlie wants to know when do we expect depreciation to peak out on a quarterly or an annual basis. He also wants to know that would we expect dilution from four nanometer in 2022, given that four nanometer is an extension of our five nanometer. Should there therefore not be dilution from four nanometer?
Okay, Charlie. The first question, really difficult to answer, because if you continue to invest, you may not have a peak in depreciation. Just as if you continue to have a strong growth, you may not have a peak in your revenue. The second question, yes, we still expect that N5 family, the gross margin to reach corporate average in about seven or eight quarters, and which is sometime in 2022.
Great. All right.
Okay. That's very helpful. Thank you.
Thank you, Charlie. Operator, let's move on to the next caller on the line, please.
Right now, we're having Laura Chen from KGI. Go ahead, please.
Hi. Thank you for taking my question and congratulations for the good result. My first question is regarding the 3 nanometer. Can you give us an update on current engagement? We know that CC just mentioned we will have risk production next year and mass production probably on second half 2020. I'm just wondering, will it be smartphone or HPC go first? Okay. That's my first question. Thanks.
Okay. Laura, your first question is regards to our 3 nanometer. She wants to know what is the current engagement with customers, and then with the volume production targeted for second half 2022, is it going to be smartphone or HPC driven?
All right. Let me answer the question first. On the engaging with customer. We are engaging with more customer at a N3 as compared with a N5 and N7 at a similar stage. Okay. There's a lot of customers are working with us. Now, which one in the second half of 2022, which one will be the first product? Actually, in smartphone and HPC applications, both.
Okay. Thanks. My second question is about our supply chain and equipment procurement plan. I think given our positive outlook and continuous CapEx, do we plan to evaluate more local suppliers? I think given TSMC's leading position in the global foundry space, I think that give a good position to lead the localization equipment. Can you give us some color about what's your view on to buy more equipment from the Taiwanese supplier or current status of the total procurement % per year from Taiwanese vendor, something like that?
Okay, Laura. Your second question is regards to our vendor and supply chain procurement strategy. Your question is really, will we consider to use more local Taiwan suppliers? Do we have any type of percentage breakdown or anything like that? Correct?
Yes. Right. Thanks.
Okay. We develop the technology, well, we maintain the technology and the manufacturing based on the best performance and the best cost structure. We did not put where it came from, or we did not put the regions into consideration, to be frank with you. The best technology, the best manufacturing cost is what we count. We don't have any certain % limitation on which area or the equipment it came from. All right?
Okay. Does that answer your question, Laura? Okay. Thank you. Operator, let's move on to the next caller, please.
Next one would have Krish Sankar, Cowen and Company. Go ahead, please.
Hi. Thanks for taking my question. I have two of them. First one is on the mature nodes, i.e., 28 nanometer and above. Not currently, but over the next few years, how do you expect the revenue and wafer starts to trend on the mature nodes, especially as some of your customers start migrating to the leading edge? My second question is, in the past, you've spoken about converting some 28 nanometer plus capacity to 20 nanometer or so for IoT and other applications. Can you provide us an update on how this transition is going?
Okay. Thank you, Krish. Let me try to summarize your questions. Maybe I'll summarize the first one, and then we can summarize the second. Your first question is regards to our mature nodes, specifically 20 nanometer and above. You want to know in the next few years, what is the revenue outlook, and also the demand or wafer starts outlook, over the next few years, especially as customers may start to migrate to more leading nodes. What do we do at the 28 nanometer and above? What is the outlook?
Let me answer that specifically on the 28 nanometer. We continue to improve the technology, and now we offer 22 nanometers ultra-low power, and that's for IoT applications. We also work with the customer to migrate their product from 65, 55, to 45, to 28, and to 22. Today, the loading is not perfect yet, but we expect in one or two years, and then we expect the loading will greatly improve. To answer your question on all the mature node, we still are improving our technologies, and we still expect the growth.
Okay. Krish, just to clarify, the second question was in regards to 28, your question was conversion to 20. As CC said, we're converting 28 to 22, hopefully that also addressed your second question. All right?
Yes, it does. Thank you, Jeff. Thank you, CC.
Sure, Krish. Thank you very much. All right, let's move on, operator, to the next caller, please.
Next one, we are having Rick Shi from Cowen Securities. Go ahead, please.
Yeah. Hi, good afternoon, guys. Okay, my first question, I just want to make a little clarification about your CapEx for this year. I think Wendell said about it's going to be around $17 billion, or it's going to be over 17? Can you clarify on this? Also give us a little bit color about the CapEx for next year, please.
Your first question to clarify our 2020 CapEx. Is it about or above TWD 17?
Yeah, it's about 17.
Okay.
Yeah. I'm sorry.
Yes, please. Go ahead.
Yeah. Your second question is about 2021 CapEx. It's too early to discuss the 2021 CapEx at this moment. If we see strong demand, we will make the investment, because the CapEx investment in this year is always for the demand in the following years. If we see the following years have strong demand, we will invest.
All right. Thank you so much. I presume it's not a second question, just a follow-up, right? Can I ask one more?
Sure. Your second question, please.
Okay. Second question is about the inventory. I think CC did mention that right now because of the macro uncertainties, COVID-19, et cetera, so customers intend to keep their inventories above seasonal for a longer period of time. What if, because unless the uncertainty remains structural and it goes on forever, otherwise, one day when uncertainty are removed, you worry about your customers to unwind inventory and cause some business correction.
Okay, Rick, your second question is regarding inventory. Although there is macro uncertainty in COVID-19, someday this will be over. Does this worry us? Will we see a sudden sharp correction or inventory drop as a result?
Okay. Let me share with you again, our view on inventory. In fact, we don't worry too much about it because of the As I said, now because of a pandemic, the digital transformation has been accelerated. That create a lot of new demand, let me say that. Take a for example, now work from home, everybody buy a PC, every kid had to buy a PC. Just look at again on the 5G smartphone's benefit. The advantage on the bandwidth, on the speed, on the low latency, everything, and people are going to need it in this digital transformation. Even right now, we expect the inventory is higher than historical high level. The demand will pick up and in next year or 2022, we are confident that demand will pick up.
That minimize or mitigate the impact of the inventory correction that everybody has a doubt on their mind.
Okay.
Okay, great. Thank you so much.
Sure. Thank you, Rick. Operator, let's move on to the next caller, please.
Next one is Mehdi Hosseini from SIG.
Yes, sir. Thank you for taking my question. First one, if your customers are willing to have inventories above this average trend line, should we assume that your wafer shipment in the first half of 2021, specifically Q1, would also follow a better than seasonal trend? I have a follow-up.
Okay. Mehdi's first question is regarding to basically our first quarter. If customers are willing to hold a higher level of inventory, should we assume that wafer shipments in the first quarter will also be much better?
We are going to share with you in the first investor conference. All right? Right now, we are not ready to make any comment on 2021, especially the first quarter.
Okay. Your second question, Mehdi?
Okay. Sure. Can you please remind us how we should think about tape-out activity, specifically at N4 and N5, and how does it compare to N7? Any follow-up would be great.
Your question is the tape-out activity at N4 and N5 as compared to N7.
Well-
Yes, if there's any update.
Okay. The demand is very strong in N4, N5, and we are engaging many customers. The exact number of the tape-outs right now is all in our planning. I can share with you that customers' demand is very strong and will be continue to be strong for the next couple of years.
Okay. All right. Thank you, Mehdi. Operator, can we move on to the next caller, please?
Right now, we have Gokul Hariharan from JPMorgan.
Hi. Thanks for the follow-up question. There's been a lot of discussion on market share on leading edge. CC, could you comment a little bit on how do we think about TSMC market share in N7, which I think is probably like 85%, 80%, 80% or even higher, and compare that with what are we expecting for the N5 family, if we include N5 and N4? I had a second question as well. Thank you.
All right. Gokul's first follow-up question is in terms of market share. He wants to ask CC, what do we see in terms of our market share at seven nanometer, and what is our expectation or outlook at the five nanometer family?
Gokul, since I'll continue to say we have technology leadership, so I can share with you that we have very high % of market share. What exactly the number is not appropriate to announce it, because it's all our own estimate. Again, the most important thing is not the market share. The most important thing for us is continue to maintain the technology leadership, and we are focused on that.
Okay.
Okay. Just a first question on that. Could we say at least directionally if N5 market share in our own estimate higher than N7 or similar to N7?
Okay, the second question Gokul wants to ask is still on market share. Do we see directionally, will N5 market share be higher than that of N7?
They are very similar because we are always the technology leader. When we introduce the N7, we are the technology leader, and when we introduce the N5 this year in mass production, we continue to be the technology leader. They are very similar.
Okay. Thank you.
Understood.
Yeah. Thanks, Gokul.
Can I ask one more question?
I think, Gokul, sorry, that's two. Sorry, I would like to ask you to get back in the queue because we still have, I think, quite a few people.
Okay. All right.
Thank you. All right. Operator, let's move on to the next caller, please.
Next one, we have Randy Abrams from Credit Suisse.
Okay. Yeah. No, thanks for the follow-up. I wanted to ask on the R&D, it stepped up faster in the quarter. From this higher level, could you discuss the investment rate that you're expecting for R&D, say, as a % of sales? With the new advanced nodes and packaging investments start to increase the R&D intensity.
Okay. Randy's first question is that he noticed, or points out actually, that our R&D has increased or stepped up in the third quarter this year. He wants to know, given advanced packaging and to continue technology leadership, what is the R&D % of sales outlook that we should expect?
Randy, let me share with you that in the third quarter, the R&D expenses are higher because of our development activities in N4 and N3. Longer term, we're still expecting the R&D expense to be about 8% or slightly higher than 8% of our revenue.
Okay, great. Appreciate that. The second follow-up question I had, just on a couple of segments. Auto, I think you mentioned earlier about coming back. It was soft in the quarter. Could you discuss now as a growth driver from a low base, if you're finally seeing some of those content drivers for next one to two years, there could be a meaningful pickup even without auto, but from a content. The other side on consumer, which was quite weak, just despite a lot of work from home and consumer electronics coming through. If you could give color maybe on something happening in the consumer segment.
Randy's second question is really a little bit split into two, but he wants to know with the automotive business seeming to bottom out, how do we view our automotive platform as a growth driver or outlook over the next few years? Similarly, he also is asking about digital consumer. All right. Actually, let me comment on the automotive platform. Actually, the COVID-19 has a major impact on the automotive market, and supply chain this year have all been affected. We are seeing the sign of recovery in 4Q. In the longer term, the trend towards safer, greener, and smarter vehicle will continue to drive silicon content increase as well as the demand for advanced and specialty technology. Again, I want to emphasize, with our technology leadership, we are well-positioned to capture the opportunities.
The growth rate, it will continue to pick up but still behind the HPC's growth rate. For the digital consumer, it's kind of flat or it's a little bit growth that I can see to date. Did that answer your question, Randy?
Yeah. Just maybe the near term, I was surprised it was as much down, factoring in their stay-at-home consumer electronics demand. I don't know if anything, just specific or short-term in nature on that.
Okay. Actually, some of the product, because of stay at home or the work from home, some of the product we put into the HPC subcategory.
That's right. Okay. All right. Great. No, thank you.
Okay. Thank you, Randy. Operator, let's move on to the next caller, please.
The next one is Sebastian Hou, CLSA.
Hello, Sebastian. You may need to unmute.
Hey. Sure. Exactly. Thank you, Jeff. First question, let me try the overbooking inventory question in another way again, if I may. We understand the higher inventory is structural, led by COVID-19, but how about the higher inventory if that's led by customers' fear of foundry capacity tightness, which is now under supply almost everywhere from leading-edge to trailing-edge? Based on the past cycles' experience, the tighter the supply of any components, the higher the risk of supply chain overbooking enhances. I'm curious whether TSMC is seeing any gap between customers' ordering volume and your internal forecast on end demand, or it's not a concern at all, as all the strong orders are just a reflection of the real demand. Thank you.
Okay. Sebastian's question is around the inventory. While his view, some of the inventory may be related to COVID-19 and more structural or linger for a while, he wants to know, is there a concern, does TSMC have a concern that because the foundry is tight, that therefore their customers are doing a lot of overbooking or so-called double booking? Therefore, does this create a concern for TSMC when we look at our internal forecast for the end demand market versus customers' booking, that there is a large gap and risk of shortfall? Well, Sebastian, actually, in TSMC's view, all my customers are our partners. We work with them very closely. To, let's say that minimize the fear of overbooking. Because they don't have to be afraid of capacity shortage and then do the overbooking to TSMC.
No, we work with them as a partner, and both parties, all my customer work with TSMC and tell us their view on the market, and we share our view on the market with them also. This one minimize a lot of the possibility of overbooking. That's the way that TSMC working with our customers. They are all our partners. Did that answer your question, Sebastian?
Great. Thank you. Yes. That's very great answer. Thank you, CC. My second follow-up question is that we've seen the rising cross-strait relationship risk in recent months. I wonder if TSMC or your customers are concerned or discuss with you about the potential risk in production operation, as most of your fabs are located in Taiwan. If such heightened risk continue for longer than just months, whether TSMC will consider keep most of fab throughout in Taiwan or increase investment in the other regions. Thank you.
Okay. Sebastian, thank you. Let me summarize your second question. Your question is regarding that you observed the rising or growing risks in the cross-strait relationships. Therefore, for our customers, do they feel there's a heightened risk? Thus, is there a need for TSMC to, I guess, paraphrase, expand our manufacturing footprint into other locations, given the state of cross-strait relations in the next few years? Okay, Sebastian. In fact, TSMC will continue to focus on Taiwan. That's our center of R&D and majority of our production fabs will continue to be located in Taiwan, regardless of all the geopolitical tension or any kind of disruption. Did that answer your questions?
Yes. That's great. Thank you, C.C., and thank you, Jeff.
Sure. Thank you, Sebastian. All right, operator, let's move on to the next caller, please.
Next one we have Bruce Lu from Goldman Sachs.
Okay. The question is for the advanced packaging. What is the revenue growth for the advanced packaging in 2020? The growth rate seems to be very strong, the management also only guided for the future growth for the advanced packaging is only slightly higher than the corporate average. This is much slower than what we have in the past 2 to 3 years. Any reasons behind that? What's the profitability for the advanced packaging right now?
Okay. Bruce, your first question is regards to our advanced packaging business. You want to know what is the growth of the advanced packaging business in 2020, and also, what is the profitability of the advanced packaging. Yeah, Bruce, the growth of our advanced packaging in this year is close to the corporate, but not as high. In these next several years, we do expect that on a calendar basis, it will grow faster than the corporate average. In terms of margins, its margins is lower than the corporate. However, its investment intensity, capital intensity is lower. Therefore, on a return basis, ROIC basis, it is acceptable to us.
Okay. The next question is for the 28 nanometers. I want to clarify something. In the fourth quarter of 2019, I think the management showed a very high confidence that 28 nanometer utilization will go back to the corporate average, driven by the more applications such as CMOS, memory sensor, cachefly, et cetera. If my understanding is correct, management still expect it will be lower than the corporate average in the coming years in terms of utilization rate. Is that the right understanding right now?
Okay. Your second question, Bruce, is regarding our 20 nanometer.
28.
I'm sorry?
28.
28.
I'm sorry, 28 nanometer.
Yes, 28 nanometer. That you said that we had commented in the fourth quarter 2019 earnings result, January this year, that our 28 nanometer utilization would improve in one to two years' time, and to the corporate average. Now your question is, does that statement still hold true?
Bruce, let me say that the progress is a little bit slower than we expected. Still, in one to two years, the utilization rate of the 28 nanometer, particularly we advance it to 22 nanometer, will be reaching the corporate average.
I see. Thank you.
All right. Thank you.
Thank you.
Thank you, Bruce. All right. In the interest of time, we will take the question from the last caller, or last participant. Please.
The last one to ask question is Roland Shu from Citigroup. Go ahead, please.
Yes. On your N6 technology is with one more EUV layer insertion than N7+, but N4 is with reduced mask layers from N5, and it's with a simpler type process. Can you elaborate your technology development logic between N6 and N4, and also the target market for N6 and N4, and how will N6 and N4 contribute to your business, respectively, going forward?
Okay. Roland, your question is regards to N6 versus N4 positioning. You point out technology-wise, N6 has one more EUV layer than N7+, but N4 may have reduced mask layers versus N5 and with simplified process. You're asking, does N4 serve the same group or target the same group of customers as N6, or are they separate markets or targeting separate customers and applications?
Correct. Okay.
Roland, it's actually very hard to answer your question whether the N6 is the same kind of group of the N4. Let me give you some kind of idea. N6 is a kind of development, continued enhancement of the N7 or N7+. All the second wave of the customer will use N6 when they want to enter the seven nanometer family.
Because of that, they offer the better density, better performance, and better power consumption. Now, similar to N6, N4 is also, we continue to improve the N5, and we also observe that if we can reduce the mask count, we can improve the defect density, we can improve the cycle time, and at the same time, we also offer the better density, better performance, et cetera. Are they the same group? I cannot answer this question, but it's the same purpose.
We offer N6 to be the second wave of the N7 customer. We offer the N4 also to offer to the second wave of the customer of the N5.
Yeah.
Okay. Thanks. Yeah, a little bit complicated. Because for N4, is there any performance enhancement to N5? Because this is with the simplified process, and I can understand there is the improvement on this defect on this production cycle type. How about it from the performance point of view? Is there going to be enhancement than N5?
Okay. Your second question, Roland, continues to ask about the four nanometer. Will N4, does it carry any performance enhancement or PPA improvement as compared to N5?
Yes. The short answer is yes. We improve the density, we improve the performance, including the transistor performance.
Okay.
Okay.
Thank you, Roland.
Thank you.
Yep. Thank you very much. All right. This concludes our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within four hours from now. The transcript will become available 24 hours from now. Both of them are going to be available through TSMC's website at www.tsmc.com. Thank you, everyone, for joining us today. We hope everyone continues to stay safe and healthy, and we hope you will join us again next quarter. Goodbye, and have a good day.