Good afternoon, everyone. Welcome to TSMC's Second 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.

Thank you, Jeff. Second quarter revenue increased 2.7% sequentially in NT dollar terms or 2.9% in U.S. dollar term. Our second quarter business was supported by continued strength in HPC and automotive-related demand. Gross margin decreased 2.4 percentage points sequentially to 50% mainly due to N5 dilution, the slower rate of cost improvement and the absence of positive inventory revaluation. Total operating expenses slightly increased TWD1.47 billion. Therefore, operating margins decreased 2.4 percentage points sequentially to 39.1%. Overall, our second quarter EPS was TWD5.18 and ROE was 27.3%. Let's move on to revenue by technology. 5-nanometer process technology contributed 18% of wafer revenue in the second quarter, while 7-nanometer accounted for 31%. Advanced technologies, which are defined as 7-nanometer and below, accounted for 49% of wafer revenue. Moving on to revenue contribution by platform. Smartphone decreased 3% quarter-over-quarter to account for 42% of our second quarter revenue. HPC increased 12% to account for 39%. IoT decreased 2% to account for 8%. Automotive increased 12% to account for 4%. And DCE decreased 12% to account for 4%. Moving on to the balance sheet. We ended the second quarter with cash and marketable securities of TWD871 billion. On the liability side, current liabilities decreased TWD14 billion mainly due to the decrease of TWD23 billion in accrued liabilities and others, partially offset by the increase of TWD6 billion in dividend payable. Long-term interest-bearing debt increased by TWD134 billion mainly as we raised TWD137 billion of corporate bonds during the quarter. On financial ratios. Accounts receivables turnover days increased 2 days to 42 days, while days of inventory also rose 2 days to 85 days primarily due to N5 wafer prebuild. C. C. Wei: Thank you, Wendell. We hope everybody is staying safe and healthy during this time. First, let me start with TSMC's long-term growth outlook and investment plan. We are witnessing a structural increase in underlying semiconductor demand as the multiyear megatrend of 5G- and HPC-related applications are expected to fuel massive increase in computation power and greater need for energy-efficient computing, which will require leading-edge technologies. COVID-19 has also fundamentally accelerated the digital transformation, making semiconductors more pervasive and essential in people's lives. With our technology leadership, manufacturing excellence and customer trust, we are well positioned to capture the structural growth from the favorable industry megatrend with our differentiated technologies. We now expect our long-term revenue CAGR from 2020 to 2025 to be near the high end of our 10% to 15% CAGR range in U.S. dollar terms. In the near term, we continue to observe both short-term imbalances in the supply chain driven by the need to ensure supply security, as well as a structural increase in long-term demand. While the short-term imbalance may or may not persist, we expect our capacity to remain tight throughout the year and into 2022 fueled by strong demand for our industry-leading advanced and special technologies. For the full year of 2021, we now forecast the overall semiconductor market excluding memory to grow about 17% while foundry industry growth is forecast to be about 20%. We now expect -- for TSMC, we are confident we can outperform the foundry revenue growth and grow above 20% in 2021 in U.S. dollar terms.

Thank you, C.C. Let me start by making some comments on our near-term demand and inventory. We concluded our second quarter with revenue of TWD372.1 billion or USD 13.3 billion, slightly above our guidance mainly due to better demand from HPC, IoT and automotive-related applications than our forecast 3 months ago. Moving into third quarter 2021. We expect our business to be supported by strong demand for our industry-leading 5-nanometer and 7-nanometer technologies driven by all four growth platforms, which are smartphone, HPC, IoT and automotive-related applications. On the inventory side, we expect our fabless customers' overall inventory to exit second quarter of '21 at a healthy level. We expect our customers and the supply chain to gradually prepare higher levels of inventory in the second half of the year as compared to the historical seasonal level, given the industry's continued need to ensure supply security following supply chain disruptions due to COVID-19 and uncertainties brought about by geopolitical tensions. Next, let me talk about our profitability. Our second quarter gross margin of 50% was slightly below the midpoint of our guidance, mainly due to an unfavorable foreign exchange rate. Our gross margin guidance provided 3 months ago was based on exchange rate assumption of USD 1 to TWD28.4, whereas the actual second quarter exchange rate was USD 1 to TWD28.01. This created about 0.5 percentage point difference in our actual second quarter gross margin versus our original guidance. In other words, if the exchange rate had maintained at $1 to TWD28.4, our second quarter gross margin would have been 50.5%. Based on the exchange rate assumption of USD 1 to TWD27.9, we have just guided third quarter 2021 gross margin to increase by 0.5 percentage point sequentially to 50.5% at the midpoint, mainly due to better back-end profitability.

Thank you. Thank you, Wendell, and good afternoon, everyone. Today, I will talk about TSMC's global manufacturing footprint. TSMC's mission is to be the trusted technology and capacity provider for the global logic IC industry for years to come. TSMC always treat our customers as partners. We do not compete with our customers. We grow our business by unleashing our customers' innovations and enabling their success. We earn our business by providing solid values, by providing industry-leading technologies, the world's largest logic capacity and efficient and cost-effective manufacturing to our customers, while maintaining trusting relationship with them. In our capital investment, our responsibility as TSMC management is to make the best decision in the interest of the company and our customers. And our fiduciary duty is to our shareholders. As the need for semiconductor infrastructure security has increased in recent years, we are expanding our global manufacturing footprint to sustain and enhance our competitive advantages and to better serve our customers in the new geopolitical environment. In Taiwan, we are building capacity for N5 and N3 in Tainan Science Park. Due to the strong customer demand, we have further expand -- plan to expand in Northern, Central and Southern Science Parks in Taiwan. And Taiwan will continue to be the home base and center of R&D for TSMC. As the initial phase of volume production of a leading-edge technology has to be in close proximity and closely coupled with R&D fab due to massive collaborative engineering activities, our leading node will continue to be ramped in Taiwan as well.

Thank you, Chairman. This concludes our prepared statements. . Now let us begin the Q&A session. Operator, can we please proceed with the first caller on the line?

Yes. The first caller, Gokul Hariharan, JPMorgan.

Maybe my first question I would focus on the semiconductor supply and demand. C.C., you mentioned that you're expecting demand to still remain extremely strong, supply to remain tight through end of this year potentially through next year as well. Could you talk about when do you expect supply and demand to come back into balance? Do you see a situation where your customers go into a bit of an inventory correction mode sometime soon? Or you think that the structurally higher inventory is something that is likely to last for a much longer period of time than what originally the market thought?

Okay. Gokul, let me summarize your first question. Gokul's first question is on the semiconductor supply-and-demand outlook. He notes that C.C. has said our capacity will be tight throughout this year and into 2022. All right. So let me summarize Gokul's question again on semiconductor supply and demand. He's wondering or asking, when do we see supply and demand in the semiconductor coming back into the balance? Is there a risk of an inventory correction anytime soon? And how long can a higher level of inventory continue? C. C. Wei: Well, Gokul, let me answer your question. Let's -- let me share with you our perspective on the shortage right now. The current semiconductor capacity shortage is being driven by both a structural increase in long-term market demand and also a short-term imbalance in the supply chain due to uncertainties from COVID-19 and geopolitical tensions. And that may or may not persist. We do not rule out the possibility of an inventory correction in the future, but we expect our capacity will remain tight throughout this year and extend at least into 2022. But let me share with you also that even inventory correction to occur, we believe it will be less volatile than previous downturn as our underlying structural megatrend of 5G-related and HPC application will continue. Do I answer your question?

Gokul, does that address your first question? And do you have a second question, if so? All right. Operator, I think we have lost Gokul. Maybe we'll move on to the next caller first.

Yes. The next one is Randy Abrams, Credit Suisse.

I wanted to ask the question on the trend per cost purchase. I'm curious for 5-nanometer and 3-nanometer, how you're seeing continued improvement? And for the customer motivation to migrate, how is that shifting between how our power performance density and cost versus also increasingly seek out the back-end system level integration?

Okay. Randy, let me summarize your first question. Your first question is on the cost per transistor. Randy wants to know what is the cost per transistor trend at 5-nanometer and 3-nanometer. And also what do -- I guess, Randy, your question is sort of what are the customers value -- or evaluate when they look at the technologies as well. Is that correct?

Yes, right, between power performance density versus cost and also back-end integration.

Okay.

Randy. No, actually, today, the technology has getting more complex than the simple scaling. Actually, we work with our customers very closely and we have been working on the performance and energy efficiency of a technology. And we took several approach, of course, material innovation, transistor structure innovation. We also, in recent generations, we work heavily on the design and technology cooperation. And also more recently, we work on the 3D ICs. All these innovation combined is to deliver the value for our customer to improve the system performance and the system's power efficiency. And TSMC is at the forefront of delivering this value and as evidenced by the strong demand and continued technology migration at 5 to 3. So in doing that, we believe we can continue earn a proper return for our investment. Thank you, Randy.

Okay. My second question, if you could give an update on the 5-nanometer where there's still some dilution. Maybe how much dilution you see in the second half? And update on how you see it trending towards the corporate gross margin. And as we look to next year, if 3-nanometer ramping late in the year, could we see a favorable like a bit more of a sweet spot for profitability as you have a more mature 5-nanometer, but not a new node yet ramping?

Okay. Randy, this is Wendell. Let me answer your question. The 5-nanometer contribution to our revenue will be much higher this year compared to last year. Therefore, we expect that the margin dilution from N5 this year will be between 2 to 3 percentage points. Now we also expect that N5, like previous nodes, will be able to reach -- the margin will reach the corporate average in about 7 to 8 quarters. With respect to 2022, it's a bit too early to talk about it. But we believe our long-term gross margin target of 50% continues to be achievable.

Okay. Thank you, Randy.

Yes. And the next one is Bruce Lu from Goldman Sachs.

I think a lot of your customers have like different supply chain management policies. A lot of your customers find a long-term contract with other flagship players with favorable pricing and payment terms. So does TSMC expect this will become the new norm for the foundry industry? Does the industry-wide profit pool become bigger and the earnings fluctuation will be less in the future?

Okay. Thank you, Bruce. Let me summarize your first question. Bruce's first question is about the semiconductor supply chain, and he's observed that a lot of customers seem to be signing long-term contracts with foundries. So do we expect this to be the new norm in the future? And could this drive a bigger or larger industry-wide profit pool? C. C. Wei: Okay. Bruce, let me answer the question. We are not able to comment on specific business terms with customers. However, we are working closely with our customers on different ways to secure their commitment. And customers understand our effort to support their growth. And if 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 maintained.

Okay. Thank you. Bruce, do you have a second question?

Sure. So I want to focus on the automotive like automotive revenue is roughly less than $2 billion in 2020 out of a $40 billion market. So with most of the automotive IDM companies are passive in capacity expansion, this creates a big addressable market for the foundry. So what is the addressable market for foundry in the automotive semi industry in the coming years?

Okay. Bruce, let me summarize your second question. So Bruce is asking about the longer-term outlook for the automotive industry. Automotive was less than $2 billion of our revenue last year, but he points out the automotive TAM is about $40 billion. So he's wondering about sort of the long-term TAM addressable market for TSMC. C. C. Wei: Well, let me answer again. We are quite positive on the long-term trend of the semiconductors in -- semiconductor content in automotive as the trend towards safer, greener and smarter vehicles will continue to drive silicon content increase as well as the demand for advanced and specialty technology, both.

Okay. Thank you. Does that answer your question, Bruce?

Can we have somehow like some specific numbers? Can you quantify that a little bit? C. C. Wei: We cannot forecast in the future very accurately, right? I mean that is so dynamic. But let me assure you that silicon content will be very important and will be increased.

Thank you, Bruce.

The next one is a question from Gokul Hariharan, JPMorgan.

Yes, let me ask my second question. My second question is that on your 3-nanometer business, clearly the market is expecting you to make a lot more inroads into the HPC segment. One of your existing IDM customers have redoubled their efforts to get back into the foundry business. Could you talk a little bit about how you manage this kind of what they call as coopetition, competition as well as cooperation? How does TSMC think about this when it thinks about capacity allocation, given that HPC is now becoming a very important driver for growth and this is probably the biggest HPC customer out there in terms of revenue size?

Okay. Gokul, thank you for your second question. Let me summarize. He points out that HPC is -- seems to become a larger and larger contributor or driver, particularly at our 3-nanometer. He also points out that existing IDM customers are redoubling their efforts. So his question is really how -- I guess, how do we manage this duality? How do we manage the capacity allocation and the relationship? C. C. Wei: Okay. Let me answer the question by, first, you mentioned that IDM is our important customer. Let me say that. And we are collaborating in some area and might competing in other areas. But let me explain again, TSMC is everybody's foundry and we support all our customers openly and fairly. We will allocate the necessary engineering resources to ensure all of our customers' product success, both existing customer and for the future customer. How to plan our capacity to support? Actually our capacity planning is based on the long-term market demand and thus is underpinned by the industry megatrend, okay? Does that answer your question, Gokul?

So maybe just one follow-up. Do you require a lot more assurance in terms of demand, solidity and demand longevity from IDM customers given they are also competing on process technology with you compared to pure fabless?

So Gokul is asking whether that we would require more longer-term assurance from our IDM customers versus our fabless customers. C. C. Wei: We will not specifically comment on that and a certain customer, okay?

Okay. Thank you, Gokul, and sorry for the disruption.

Next one to ask question, Charlie Chan from Morgan Stanley.

First of all, I would like to thank TSMC for the donation of the vaccine to Taiwan. I think that ensures business continuity of TSMC and also global semi supply chain. And my question is about still your gross margin trend. So first of all, is your 3D IC or advanced packaging. It seems like Wendell mentioned that the back end product possibilities are getting better in 3Q, I'm not sure if that's true. So the question is that currently the percentage of advanced packaging of TSMC's revenue. And would the company still believe 3D packaging can still outgrow the wafer business and whether they would create a margin dilution. So the first question is about the 3D advanced packaging.

Okay. Let me summarize your question, Charlie. You're asking about our 3D IC business. Charlie is asking the back-end profitability that Wendell cited improvement in the third quarter. Is this the truth? And then also sort of the outlook for our 3D IC business in comparison to our overall business. And whether 3D IC business will dilute our profitability.

Okay, Charlie. First of all, the gross margin that I was talking about, improvement of back-end service in the third quarter actually is a seasonal factor. As you know, our back-end services has seasonality. So second half normally has a higher gross margin. But longer term, we expect its margin to continue to improve although it's still not as high as a gross margin of our wafer revenue. But it has a lower asset capital intensity, therefore, the returns from back-end services is satisfactory. In terms of a revenue percentage, we expect back-end services to account for about 8% of our total revenue this year. And in the next 5 years, we expect it will grow slightly higher than the corporate average.

Okay. Thank you, Wendell. Do you have a second question, Charlie?

Yes, I do. Yes, my next question is about the long-term gross margin trend, right? Because we did create something better about the thrust -- or the higher CapEx intensity, your bargaining power against different vendors and those customers. So really, I want to ask this openly to company management. Does TSMC believe you has acquired the monopoly of the leading edge in the industry, why and why not? And if yes, do you think you have the monopoly, why cannot TSMC charge higher wafer price to cover the increase of CapEx intensity? And lastly, if the company were to -- need to choose between the margin sustainability and also the market share, what would be your choice?

Okay. Let me summarize your second question, Charlie. So Charlie is asking about the long-term gross margin trend and bargaining power. And he is wondering -- he wants to know whether we see or believe we have a monopoly at the leading edge or not. And if we do, I guess, part of your question, Charlie, is then how -- pricing. And then also, if we have to choose between market share and profitability, how should we choose? C. C. Wei: Charlie, let me answer this question. First, we do have a very high market share on the leading-edge technology node. But our pricing strategy is strategic, and we don't take an optimistic approach. And it's far away from you say that we try to bargaining power. In fact, we work with our customers closely and we want to help them to be successful while we get a proper return. That's all I can answer for you for our pricing. And looking ahead, we continue our practice, try our best to hear our customers, to grow and we want to get the proper return. So that's why we are firming up our wafer pricing. And we are confident that we can get our gross margin about 50% or above in the long term.

Okay. Thank you. All right.

The next one to ask a question, Nicolas Gaudois from UBS.

Earlier, you referred to your expansion of capacity at 28-nanometer in Nanjing for about 16,000 -- 15,000 since the launch. If the demand in trailing edge is effectively structurally higher and tying up to the leading edge, are we going to see additions in capacity in the trailing edge becoming more of a recurring feature for TSMC as for the rest of the industry going forward?

Okay, Nick, let me try to summarize your first question. So Nick is saying that, as Chairman said, we are expanding our capacity in Nanjing for 28-nanometer. So his question is that do we see the demand at the trailing edge or the mature nodes becoming structurally higher? And then will we consider or add capacity in those trailing edge nodes?

Okay. This is Mark. Let me answer your question. Our strategy more recently in mature node is to work closely with our customer to develop specialty technology solution. This is not described by the numbers. Actually, we are leading in many 28-nanometer specialty technologies and we can meet their requirement and create differentiated long-term value for them. And we expect this structural demand will continue. And of course, we'll focus on our investment on specialty technology to support that. So for the manufacturing greenfield, no expansion. And we do not rule it out, we will build case by case as long as the economics can justify and customer commitment can be secured.

Okay. Thank you. Nick, do you have a second question?

Yes. Very quickly. Thank you, Jeff. A clarification effectively for the investment in the U.S., you say the equipment moving in, in H2 '22 and then production in Q1 '24. A reasonably long runway, I guess. So is that because new fab you need to obviously run preproduction, that qualification time will be reasonably low?

Yes. We make -- prepared a little bit longer preparation time just because that's a new semiconductor environment for our operations. But of course, we will continue to compress the schedule as much as we can, yes.

Okay. Thank you, Nick.

The next one to ask question, Roland Shu from Citigroup.

My first question is for your Japan R&D center. So it is stated more than 20 Japanese companies will work with you at your 3D IC R&D center in Japan. So I want to know what are the roles and responsibility for every party including yourself in this Japan R&D center? And also, do you plan to start 3D IC packaging mass production in Japan one day? And do you plan to build a wafer fab in Japan for foundry business going forward?

Okay. Roland's first question. He wants to know about our 3D IC research center in Japan. There's more than 20 companies involved according to him. So what are the roles and responsibilities of that? And also, will we build a packaging -- 3D IC packaging integration facility in Japan. And will we consider a wafer fab in Japan. So three parts to this question. C. C. Wei: Okay. Roland, let me answer your question. First, yes, there is more than 20 membership to join this Japan's 3D ICs research center. In fact, what's the role and responsibility, TSMC is in charge of this one. And we also -- in technology, we're also in charge of the integration for all the major partner together, too, so that we can be successful in the most advanced packaging technology, which includes TSMC's 3D IC and some of our partner's advanced material and our partner's most advanced substrate technology. Everything put together, which is necessary for the future HPC's application that we needed. Do we have a plan to mass production in the 3D IC in Japan? It's not in our current planning yet, okay? And what's your next question?

And will we consider...

How about the wafer fab, yes. C. C. Wei: The wafer fab, we are -- actually, let me say that we are -- we do not rule out any possibility. And in Japan, we are in due diligence process now to expect to do that wafer fab, let me say that, clearly.

Okay. My second question is Mark also said that the key concern to build a fab overseas is considering the cost gap, and you are working with the governments to close the cost gap. However, recently there were some noises in the U.S. to request the U.S. government to invest wisely in domestic companies to support U.S. authority. So it changed U.S. government's plan and lead to fail to close the cost gap to TSMC in U.S. operations. And how are you going to close the cost gap if there is no adequate support from the U.S. government, okay?

Okay. Let me summarize Roland's second question, it's about our U.S. fab and the cost gap. He points out that recently there is some discussion for U.S. incentives to invest in domestic companies. And so therefore, if this were to be the case, how would that affect TSMC and how would we manage the cost gap?

Okay. Roland, right? This is -- I think this current event is still developing. You know that in U.S., the originally proposed CHIPS for America Act has gained bipartisan support. And we are very happy that in the Senate they passed a bill of U.S. Innovation and Competitive Act already passed in Senate. Right now it's in the hands of the House of Representatives, and we are very optimistic that they will gain bipartisan support. The reason we -- the bipartisan support for this is to create a level playing field for the semiconductor fab investment in U.S. so that there will be a renewal fab industry in the U.S. Of course, how well it can be done in operation up to each company to do the operating well. And we are still learning the cost structure in the U.S. But in the meantime, in addition to take on this level playing field opportunities and further on, the operating cost will have to be shared with our customers. So that's a part of our firm up price -- firming up pricing, including the increased global manufacturing footprint. So we believe then...

So pricing from the other companies?

Yes. Yes. We believe in that way, we can continue to sustain our profitability as before.

Okay. Thank you, Chairman, and thank you, Roland.

Next one, we have Laura Chen from KGI.

Can you hear me?

Yes, we can hear you, Laura.

First of all, I just want to ask about our global expansion plan. Can you give us more details about our expansion plan other than like the advanced node in the U.S.? Mark just mentioned that the 28-nanometer will be the sweet spot. So will we expand more other than China or other regions, would we consider that? And how would that impact our already announced USD 100 billion CapEx for the next 3 years? That's my first question.

Okay, Laura, let me summarize your question. Your question is about our global manufacturing footprint. And I think your question is on our mature node. Do we have plans for further expansion of mature nodes in different locations? And if so, how will this affect our CapEx in the next few years?

Yes.

Yes. Laura, I think several project is still under planning. We do not rule out the possibility in Japan. Actually, C.C. just mentioned that we are in the due diligence process now to have a specialty technology fab in Japan. But of course, the decision is still too early to disclose because the final decision will be based on our customer needs, operating efficiency, evaluation and cost economics. So for those projects, we have not included into the $100 billion CapEx budget. Okay?

Okay, very clear. Yes. And my second question is about the 16- and 12-nanometer. We know that the current supply is also quite tight and the client demand is very strong, particularly for the RF transceiver, et cetera. So I'm just wondering, do we also have the plan to expand on 16 and 12?

Okay. And so Laura's second question continues to more specifically on 16-nanometer/12-nanometer. The supply continues to be tight, demand is very strong. Do we have any plans to expand at this node? C. C. Wei: Well, Laura, let me answer this question. We -- again, this is the kind of mature node for TSMC. And we will expand our capacity with the customer's commitment. And also, we have to consider the economics. And so if everything is positive, in fact, we will consider to expand the capacity to support our customer actually.

Okay. Thank you, Laura.

Next one, we have Brett Simpson from Arete Research.

I had a question on gross margins. I guess you've talked about 50% gross margin as a long-term target for quite some time now. And I understand there's been FX headwinds and the 5-nanometer ramp as a headwind. If I look at your big fabless customers, they are delivering structurally much higher gross margins as a result of accessing your leading-edge capacity, particularly in the last 12 months when your gross margins are going down. And I'd just like to ask, given your position in the industry, do you really think 50% is an appropriate level of return. And doesn't your position warrant some structural margin expansion at the gross margin level?

Okay. Brett's first question is about our gross margin. He notes that, of course, we have been facing headwinds from the foreign exchange rate and also the 5-nanometer ramp, which carries some level of dilution. But he points out that our customers' gross margin is, particularly in the last 12 months, has been structurally higher than ours. So he wants to know, given our position, do we think 50% is achievable? Why would it not be something structurally higher? Is that correct, Brett?

Yes. Really whether 50% is an appropriate level of return. Given everyone else is right -- is seeing -- delivering higher gross margins, why wouldn't TSMC look for structurally higher gross margins as well like everyone else?

Right. This is Wendell. Let me answer your question. First, if we look at shorter term, you talked about in the last 12 months. Foreign exchange does play a very big role in the gross margin between last year and this year, year-to-date. Last year, the dollar against NT rate was $29.43 in average. This year, year-to-date, is 28% -- somewhere around 28%. That creates a 2 percentage point difference in gross margin, i.e., if the gross margin -- I mean, if foreign exchange rate stays where it was last year, we would be having a 52% gross margin in the second quarter already. And also, I talked about the dilution from the N5 this year, another 2 to 3 points. So with all these negatives, we still can make 50% in second quarter and the third quarter. That's the short term. Now longer term, the investment that we are making is for future business growth. At some point of time, at the beginning, we made -- the short-term advanced technology is getting more and more challenging in cost. Then we are working closely with our customers to firm up the wafer pricing and also working with our suppliers to ensure that the cost improvement can be delivered. And with all these efforts, we still think that 50% is a good target and is achievable.

Okay. Thank you, Wendell. And maybe just a follow-up. I noticed your China business grew from 6% of sales in Q2 -- sorry, 6% of sales in Q1 to 11% of sales in Q2. Can you talk about some of the drivers that delivered that upside? And long term, how do we think about China scaling within your business? Do you think we'll get back to sustainably double-digit percent of sales? And what would drive that?

Okay. So Brett's second question is related to our China business. He notes in the near term that China contribution has gone from 6% in the first quarter to 11%. So he wants to know what is driving this. And then he asked a longer-term question, which is how should we think about our China business over the next few years? And can it return or sustain at an improving or double-digit level? C. C. Wei: Well, Brett, let me summarize it. I think China remains a very strong and growing market. And we have developed a large customer base in China and we'll work with them to grow our business and expect our business from China will continue to increase in all the market sector that we're talking about, smartphone, HPC, IoT and automotive also.

And also, Brett's question on the improvement from China from 6% in first quarter to 11% in the second quarter. What is driving that?

That's because mainly the HPC platform.

Okay. Thank you, Brett.

Next one, we have Charles Shi from Needham & Company.

Can you guys hear me?

Yes, we can hear you fine, Charles.

So I want to ask the first question really is about your -- the adoption of your most leading-edge node process. Historically, if I understand correctly, your smartphone platform seems to lead the adoption in the past, at least especially in the first year of the production ramp. And I think you did say that the high-performance computing will be coming increasingly important. So the question really is about 3-nanometer, which we are about a year away from the mass production. So could high-performance computing, from what you see today, really play a bigger role or even like the leading role in the first ramp of the 3-nanometer especially in the first year? Do you even see like high-performance computing could eventually be like the actual lead adopter of the leading-edge nodes going forward?

Okay. So Charles' first question is about the drivers of leading node adoption. He notes in the past traditionally it's mainly come from smartphones, but HPC also seems to becoming more important. So his question is on N3, do we expect HPC to play a bigger role in the ramp of N3, particularly in the first year? And could N3 become the first adopter or primary adopter? C. C. Wei: Charles, This is C.C. Wei. Let me answer this one. The N3's first year is ramping up, still smartphone plays the biggest role. Of course, your observation is correct. HPC application is also important and getting more and more important. And in fact, HPC will be our largest revenue driver in the next 5 years. So in the N3 node, in addition to the smartphones, we do expect the HPC's application will become important also. Did that answer your question?

Yes, yes. Excellent. So maybe the second question, I still want to touch upon your global expansion. I know you probably are tired of answering that already, but forgive me, I'm going to ask another one. You announced your fab in Arizona, which the technology node will be 5-nanometer. And as I understand, 5-nanometer wafers will very likely require your in-house advanced packaging solutions like either InFo, CoWoS or maybe even SoIC going forward. So -- but your current packaging facilities, as I know, are 100% in Taiwan. And I can imagine that if you are really out putting wafers in Arizona for 5-nanometer, based on your current footprint, you got to ship those back to Taiwan for packaging then send back to your customers, which I can imagine it could be a little bit challenging in terms of cost, logistical efficiency. So I'm not going to ask for a specific plan, but do you kind of foresee maybe you want to set up an advanced packaging facility in Arizona in the near future?

Right. So Charles, I'll shorten your second question. I think Charles is asking in Arizona we are building wafer capacity with 5-nanometer. Will we also consider setting up 3D IC integration capacity or -- in Arizona as well?

This is Mark, Charles. I understand your concern. But if you look at the current industry landscape, what you just said is nothing new. Everyone have their wafer produced in one location and packaged in another even including major chip producer in U.S. So for that matter, we do not see and compose any logistical difficulties. We just continue to evaluate, and currently, we do not have that 3D IC fab in Arizona at this point.

Okay. Thank you, Chairman. Thank you.

Next one, we have Andrew Lu from Sinolink Securities.

Can you hear me?

Yes.

Okay. My first question is regarding 3-nanometer ramp-up for second half, starting from second half next year. I recall the 7-nanometer ramp-up in year 2018 second quarter with some revenue contribution and the 5-nanometer in second quarter last year, year 2020. But it seems like 3-nanometers clearly some delay for second half next year. So I want to ask, is that because the technology difficulty, we cannot ramp up in second quarter or we don't have a big customer to use 3-nanometer at the beginning stage that's why we push back the ramp-up in second half next year? That's my first question.

Okay. So Andrew's first question, let me summarize, is asking about our 3-nanometer ramp. He notes that 5-nanometer and 7-nanometer in the past few years basically ramped in the middle of the year. And 3, we said the ramp will be in second half of next year. So what is the reason behind this? C. C. Wei: Andrew, you have a very good observation and you calculate that, yes, about 3 to 4 months is a delay as compared with 5-nanometer. Yes, 3-nanometer technology actually is very complicated and in both processing technology and also the customers' product design. So we work with a customer, and finally, we decided to ramp up in the second half of next year. And this is -- we decided with our customer with the best fit their need.

Okay. Thank you.

Okay. My second question is recently, we believe NXP, Infineon, Renesas earlier either have some power down outage and also the fire resulting their second quarter or first quarter utilization down to almost 0. And recently, we are hearing these guys are back in the utilization rate to 100%. The wafer output may start appear in Q4. Most of these companies are leading company in automotive semiconductor, including MCU. Do we have some concern once these customers are ramping up their own fab and that will result the next year or starting from Q4 the order automotive semiconductor to us will be reduced -- largely reduced.

Okay. So Andrew's second question is on automotive. He asked as IDMs ramp up their production in the second half, are we concerned or do we have concerns that heading into the end of this year or into 2022, that TSMC's automotive customers will greatly reduce their orders to TSMC? C. C. Wei: Andrew, this is C. C. Wei again. Let me answer your question. A very short answer is that, no, we don't have any concern. The reason is very simple: because of we offer the technology and our customer working closely with us. And for some of the technologies, mostly in the leading edge -- not the leading edge, I'm sorry, it's 55-, 40-nanometer and 28-nanometer that our customers need TSMC's support and the demand will continue to grow. And so we don't worry about once they bring up their fab and then TSMC's demand will be decreased. The answer is no, and remember tight and actually very tight in 2022 also.

Okay. Thank you, Andrew.

Next one, we have Martin Lau from FSSA.

The first question relates to politics. I counted, you said geopolitical risk 5x during today's call. And someone mentioned about the fact it seems in Taiwan with vaccines are getting even more political. I just wonder, for management, how concerned are you with politics? It seems the U.S. sometimes is fighting against China; China, Taiwan. And what things have you thought about to mitigate, if anything, such political risk? And also, are your customers concerned when COVID happened in Taiwan, when China from time to time threatened a war against Taiwan? And are your customers concerned that they are so reliant on you?

Okay. So Martin's question is about politics and geopolitical risks. He notes that geopolitical is talked about more and more. His observation is that vaccines in Taiwan has become a political issue as well. So he wants to know how does TSMC manage or mitigate the political risks looking at U.S., China, Taiwan relations? And do our customers have concerns on things such as the recent increase in COVID or the threat of invasion from China? And how does TSMC manage these risks?

Martin, this is Mark. Thank you for asking. First of all, the -- on the recent COVID situation in Taiwan, I think the confirmed case has dropped. And of course, the current -- first priority is to get the people of Taiwan get vaccinated upon variants keep coming. But I'm really grateful that us and YongLin and Foxconn come together a humanitarian donation of 10 million doses of vaccine to the Taiwan government, CDC -- Taiwan's CDC and be able to vaccinate our people in Taiwan because TSMC's employees is inevitably embedded in the community of Taiwan, and that is important. The reason -- you know this could be political in the beginning. But at the end, we completed the contract and we did get support from all sites. So I don't think at the end it's as political anymore. Otherwise, this donation wouldn't be successful. In a global sense, the geopolitical development is continuing. I think this is a challenge for every company. Every company's management has to deal with it. And -- but I think in the new administration from U.S., I think the development is more predictable, more rule-based. So as long as rule-based, I think it's better for every company to adopt it, too. So that also prompted the talk I have earlier -- given earlier that the global manufacturing footprint may need to do an adjustment for our customers. Our customers in different countries, their infrastructure, supply security -- semiconductor-related infrastructure, supply security maybe come up a higher priority, and we do that adjusted to it. But of course, there is the customers' needs that we are adjusting it to upon the greater geopolitical development. As to the invasion of China, let me tell you, nobody -- I mean, everybody wants to have a peaceful Taiwan Strait. And because not only -- because it is to every country's benefit, but also because of the semiconductor supply chain in Taiwan, no one wants to disrupt it. You have a COVID -- only a COVID situation already made a major disruption for the global economy. And I don't think that any disability in Taiwan Strait is any country wish to make it happen. So I'm optimistic on that. Thank you, Martin.

Thank you. Do you have a second question, Martin?

Yes, yes. Can I follow up with a second one? It's kind of related. You mentioned about global manufacturing. You also mentioned about the need to maintain your highest -- most of your technology in Taiwan because of the proximity to R&D. My understanding is majority of our engineers are from Taiwan. As you go for this global manufacturing, do you -- can you talk about how you are changing, say, for example, your talent acquisition? Like how you try to get more people outside Taiwan so that maybe over time you can become more manufacturing outside Taiwan? And also maybe on the Board, I mean, because when I look at the Board, it remains largely Taiwanese. The recent two additions, one is Dr. Kung who's a minister; Yancei who's from Delta. Do you see also the Board maybe changing, become more international? Or maybe if I try to propose some debate, have a Mainland Chinese on the Board?

Okay. So Martin's second question is around talent and Board composition. Basically, with our expanding manufacturing footprint, he wants to know what is our strategy to attract more global talent for TSMC. And his secondary question is also on the Board. His observation is that we -- the new Board members are primarily only from Taiwan. So will we consider board members from other countries?

The question about talent. Indeed, I think the -- we have been advocating the talent development in semiconductor field in Taiwan as well as in U.S. I think semiconductor industry talent has been underdeveloped over the past decades. And today, everyone look at the semiconductor as a key economic drivers. So the talent needs to be connected with that. And in Taiwan, I think we have been advocating the government to set up the high-end advanced research colleges across the Taiwan major universities. In the U.S. and also President Biden talked about semiconductor human infrastructure, that is to develop the human talent in semiconductor through the vast investment of the R&D. So this is catching up in every place. And particularly in Taiwan, we get very strong support from the local government to be able to continue to supply talents in Taiwan. As far as the Board member, yes, Delta, Yancei has joined a year ago. And more recently, I think this coming shareholder meeting, we will nominate Rafael Reif, who is the President of MIT, to come to our Board. And we hope that it will go through the blessing of our shareholders. And that is a major increment of our corporate governance and particularly in the area of the talent development. We invited chair -- Board members upon they have a very strong knowledge and experience on the corporate governance. And that is we will continue to look for without any differentiation about nationalities.

Okay. Thank you, Chairman. Thank you, Martin.

Yes. The last one to ask question, Krish Sankar from Cowen and Company.

I had two of them. The first one, you spoke about investing $100 billion in CapEx over the next three years and how that does not include specialty process nodes like the ones in Japan. I'm kind of curious, can you just be more specific on how much you plan to invest in your U.S. Arizona fab or fab clusters over the next 3 years in terms of CapEx? And then I have a follow-up.

Okay. So Krish's first question is about our investment in CapEx looking at the next three years. He wants to know how much specifically we are investing in Arizona?

Okay, Chris. We have announced that the Arizona project will be a $12 billion project. This was announced last year.

Got it. And can you just elaborate how much you plan to do it over the next 3 years?

So he wants -- Chris is asking how much will we invest in Arizona over the next 3 years.

Yes, $12 billion.

Next three.

Next three.

Close to.

Well, basically, is next 3 years is about $8 billion.

Okay. All right. Perfect. And then a quick follow-up. Your auto revenues was 4% of total revenue. What technology node is that predominantly?

Okay. So Krish, your second question is on automotive. It was 4% of our revenue. What particular specific nodes is automotive using? C. C. Wei: Krish, I think I have mentioned that automotive MCU is the biggest one that we have and it's in 55-, 40- and 28-nanometer with the majority still in 55 and 40. And in the next 2 to 3 years, it will be moved to 28-nanometer. That's in our current plan, and we are working with our customers on that.

Thank you, C.C. Thank you, Krish. 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 4 hours from now. And the transcript will become available 24 hours from now. Both of these, which will be available through TSMC's website at www.tsmc.com. So thank you for joining us today. We hope everyone continues to stay safe and healthy, and we hope you will join us again next quarter. Thank you, and have a good day or good evening.