Taiwan Semiconductor Manufacturing Co. Aktie

Good afternoon, everyone, and welcome to TSMC's Third Quarter 2026 Earnings Conference Call. This is Jeff Su, TSMC's Director of Investor Relations and your host for today. 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. [Operator Instructions] The format for today's event will be as follows: First, TSMC's Senior Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the third quarter 2026, followed by our guidance for the fourth quarter 2026.

Then Mr. Huang and TSMC's Chairman and CEO, Dr. C.C. Wei, will jointly provide the company's key messages. That will be followed -- then we will open the line for the Q&A session. 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 the 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. And 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. Thank you for joining us today. My presentation will start with financial highlights for the third quarter of 2026. After that, I will provide the guidance for the fourth quarter of 2026. Third quarter revenue increased 17.6% sequentially in TWD, supported by strong demand for our leading-edge process technologies. In U.S. dollar terms, revenue increased 16.2% sequentially to $46.7 billion, above the high end of our third quarter guidance. Gross margin decreased by 1.3 percentage points sequentially to 66.4%, primarily due to margin dilution from 2-nanometer ramp.

Total operating expenses accounted for 7.8% of net revenue. Operating margin was 58.7%. Overall, our third quarter EPS reached TWD 29.43 and ROE was 45.5%. Now let's move on to revenue by technology. 2-nanometer process technology contributed 8% of wafer revenue in the third quarter, while 3-nanometer, 5-nanometer and 7-nanometer accounted for 33%, 28% and 10%, respectively. Advanced technologies, defined as 7-nanometer and below, accounted for 79% of wafer revenue.

Moving on to revenue contribution by platform. HPC increased 22% quarter-over-quarter to account for 68% of our third quarter revenue. Smartphone increased 10% to account for 21%. IoT increased 5% to account for 4%. Automotive increased 23% to account for 4%, while DCE declined 12% to account for 1%. Moving on to the balance sheet. We ended the third quarter with cash and marketable securities of TWD 4 trillion or USD 126 billion. On the liability side, current liabilities increased by TWD 110 billion quarter-over-quarter, mainly due to the increase of TWD 46 billion in current portions of bond payable and bank loans and TWD 26 billion in dividend payable.

In terms of financial ratios, accounts receivable turnover days decreased by 1 day to 28 days. Inventory days decreased by 5 days to 82 days, primarily due to the increase in wafer shipment. Regarding cash flow and CapEx. During the third quarter, we generated about TWD 920 billion in cash from operations, spent TWD 443 billion in CapEx and distributed TWD 156 billion for fourth quarter 2025 cash dividends.

Overall, our cash balance increased TWD 488 billion to TWD 3.6 trillion at the end of the quarter. In U.S. dollar terms, our third quarter capital expenditures totaled $13.8 billion. I have finished my financial summary. Now let's turn to our current quarter guidance. Based on the current business outlook, we expect our fourth quarter revenue to be between USD 52.3 billion and USD 54.3 billion, which represents a 14% sequential increase or a 58% year-over-year increase at the midpoint.

Based on the exchange rate assumption of USD 1 to TWD 31.8, gross margin is expected to be between 65% and 67%, operating margin between 56.5% and 58.5%. This concludes my financial presentation. Now let me turn to our key messages. I will start by talking about our third quarter '26 and fourth quarter of '26 profitability. Compared to second quarter, our third quarter gross margin decreased by 130 basis points sequentially to 66.4%, primarily due to the margin dilution from the initial ramp-up of our 2-nanometer technology.

We have just guided our fourth quarter gross margin to decrease by 40 basis points to 66% at the midpoint, primarily due to the continued margin dilution from the steep ramp-up of our 2-nanometer technology as well as dilution from overseas fabs. These impacts are expected to be partially offset by cost improvement efforts, including productivity gains and across node capacity optimization. While the cost of overseas fabs remain higher, thanks to the company's overall larger scale and cost improvement efforts, we now expect the gross margin dilution from the ramp-up of our overseas fabs to be closer to 2% for the full year 2026.

As we put more effort to support the strong structural demand from our customers, more fabs will be ramping up in the U.S. Thus, we expect the dilution from our overseas fabs to widen to 3% to 4% in the next few years. We will leverage our increasing size in Arizona and work on our operations to improve the cost structure. We will also continue to work closely with our customers and suppliers to manage the impact. Overall, with our fundamental competitive advantages of manufacturing technology leadership and large-scale production base, we are confident that TSMC continues to be the most cost-efficient manufacturer in every region that we operate.

Next, let me make some comments on our 2026 capital budget. At TSMC, a higher level of capital expenditures is always correlated with higher growth opportunities in the following years. Given the continued robust multiyear demand from our customers, including the emerging Agentic AI market, we reiterate our full year 2026 capital budget to be between USD 60 billion and USD 64 billion. We continue to expect our CapEx dollars investment from 2026 to 2028 to be even more significantly higher than the past 3 years as we continue to invest heavily to support our customers' growth.

Additionally, while the tool supply is tight, we always collaborate closely with our tool suppliers well in advance to prepare the capacity, just like our customers collaborate with us well in advance to plan our capacity. Even as we invest for the future growth with this level of CapEx spending in the next few years, we remain committed to delivering profitable growth to our shareholders. We also remain committed to a sustainable and steadily increasing cash dividend per share on both an annual and quarterly basis. Now let me turn the microphone over to C.C.

Thank you, Wendell. Good afternoon, everyone. First, let me start with our near-term demand outlook. We concluded our third quarter with a revenue of USD 46.7 billion higher than our guidance in U.S. dollar terms, driven by strong demand for our leading-edge process technologies. Moving into fourth quarter 2026, we expect our business to be supported by continued strong demand for our leading-edge process technologies, including the steep ramp-up of our 2-nanometer technology.

Looking ahead, we continue to observe consumer and price-sensitive end market segment are being challenged due to the impact of rising component prices and macroeconomic uncertainties. As such, we are being prudent in our business planning going into 2027, while focusing on the fundamentals of our business to further strengthen our competitive position. Having said that, demand for our leading-edge technologies continue to be extremely robust, supported by our strong technology differentiation and broad customer base. Therefore, we now expect our full year 2026 revenue growth to be around 44% year-over-year in U.S. dollar terms.

Next, let me talk about the AI-related demand outlook. The AI market continues to be very strong. Recent developments are very positive to AI's long-term demand outlook. We observe increasing AI demand fueled by the rising adoption of Agentic AI as we transition from chatbot to AI assistants. In addition, we are observing great adoption of open source AI models. This will drive greater efficiency and help lower the barrier to future AI usage and result in more and more demand for leading-edge silicon.

AI-related demand is being supported by strong underlying demand for AI accelerators. In addition, the emergence of Agentic AI, CPUs and the continued demand for networking chips are driving further silicon demand. Therefore, semiconductor will remain essential in the AI megatrend. The value of semiconductors in the global supply chain is set to continue to increase in our view. This will create more value opportunities for our customers to capture and more value opportunities for TSMC as well.

No matter which approach is taken, it will require use of the most advanced semiconductor process technology. With our technology leadership, manufacturing excellence, global capacity expansion and customer trust, we are well positioned to capture the full multiyear structural AI-related demand. Therefore, all the AI players are working with TSMC.

Next, let me talk about TSMC's capacity expansion strategy. As a foundry, our first responsibility is to provide our customers with the best technologies and effective capacity support to unleash their innovations. To address the structural increase in the long-term semiconductor market demand profile, TSMC collaborated closely with our customers and our customers' customers to plan for our capacity. In addition, we are now going one step further and sending our own teams of people to check on the data center build-out progress and power supply situation in the U.S.

The feedback is so far so good, and the CSPs has been well prepared. Internally, TSMC employed a disciplined capacity planning system to assess the market demand from both a top-down and bottom-up approach. This is a continuous and ongoing process. Given the fundamental complexity of leading-edge technologies and the design-in and lead time involved, we already have a very good idea of customers' multiyear product road map and production plans. Thus, we continue to hold strong conviction in the AI megatrend.

We are moving full speed ahead to increase our capacity to meet our assesses demand and stepping up our CapEx investment to support our customers' future growth. TSMC is expanding at an unprecedented pace with a near 20 fabs worldwide in progress. We are building 13 leading edge and advanced packaging fab in Taiwan over the next several years and another 5 to 6 fabs overseas. Our new node, N2 has already entered high-volume manufacturing in the fourth quarter of 2025 with good yield. N2 is ramping successfully in multiple phases at both Hsinchu and Kaohsiung site.

We expect our N2 family technologies annual capacity to more than double in 2027, supported by strong demand from both smartphone and HPC AI applications. We are also leveraging our manufacturing excellence to generate more N2 wafer output, further increasing our capacity support to customers. In addition, we have already begun preparing the capacity for our A14 technology. Risk production is well on track for 2027 and volume production will start in 2028. Thus, TSMC's semiconductor technologies and manufacturing will continue to play a pivotal role in supporting the global semiconductor industry while unleashing our customers' innovation.

Now let me talk about TSMC's global N3 capacity expansion update. As we said before, historically, we do not add additional capacity to a node once it reaches its targeted capacity. However, based on our assessment, we are executing well on our global capacity plan to add 3 new fabs to support a robust multiyear pipeline of demand for 3-nanometer technologies. In Taiwan, volume production for our new 3-nanometer fab in Tainan Science Park is well on track to begin the first half of 2027.

In addition, driven by the strong demand from our customer, we have decided to add one additional 3-nanometer fab in Tainan. Volume production is scheduled for the first half of 2028. In Arizona, we have completed the construction of our second fab and are in the process of tool moving and installation. Volume production will begin in third quarter '27, slightly ahead of our previous schedule. In Japan, we are progressing well to our plan to utilize 3-nanometer in our second fab and volume production is scheduled in 2028. In addition to all the new fabs, we continue to convert 5-nanometer tool to support 3-nanometer capacity in Taiwan. We are also leveraging our manufacturing excellence to greater productivity across our fab in all locations to generate more wafer output.

We are also focusing on capacity optimization across nodes, which including flexible capacity support among N7, N5 and N3 node. Thus, we are using multiple levers to do everything we can whenever we can, as fast as we can to maximize the support to our customers. We have no intention to leave food on the table for others, and we want to capture every piece of business possible. This is concluding our key message, and thank you for your attention.

Thank you, C.C. This concludes our prepared statements. [Operator Instructions] Now let's begin the Q&A session. Operator, can we proceed with the first caller on the line, please?

Yes. First one, Laura Chen, Citi.

I appreciate C.C. for sharing with us the TSMC capacity expansion strategy. So may I follow up more on that given the increasing aggressive AI semiconductor demand, how does TSMC now see the AI revenue CAGR growth? And how does TSMC reconcile customers' capacity requirements with some of the potential concern on the data center power availability and also maybe funding recently and other critical components such as HBM, ABF advanced packaging and also in the networking side, we see the tightness of laser as well. So if these kind trends create a mismatch between semiconductor supply and also actual AI development, how would TSMC adjust your capacity allocation and also the strategy to minimize the potential overbuilding in the supply chain? That's my first question.

Okay, Laura. Laura has a very involved first question. I think she's asking how do we see the AI demand outlook? The AI demand, as C.C. said, seems to be very strong, but she's worried or concerned. How do we reconcile the strong demand of chips, AI chips versus risks related to whether it be power supply, I think also like different bottlenecks within the supply chain, et cetera. How do we make our assessment of what is the appropriate CapEx and capacity we need to prepare? Is that correct, Laura?

Yes. Thank you.

That's a long, long question. Laura, let me share with you. We have a very sophisticated capacity planning process, and we also have a very disciplined on our consideration. And let me share with you, we put every possibility we can talking about the CSP's CapEx, data centers build-out, financial, power supply and also other component supply, et cetera, into our consideration. And so we make our own judgment. And then based on our assessment the demand, we still think the demand is higher than the supply, and we are working very hard to narrow the gap. Did I answer your question, Laura?

Sure. Thank you very much. So far from our -- like a comprehensive study in the supply chain, we are still quite positive on the demand in the longer term, right? So in that case, can you also give us like more color about that what -- any update on your AI CAGR growth? I recall previously, we're talking about like high 50%.

Yes. So part of Laura's question is then based on our assessment, we had previously guided our AI CAGR -- AI accelerator CAGR, which is a narrow definition to grow mid- to high 50s percentage from 2024 to 2029. So Laura wants to know, do we have any update to that number today?

Laura, all I can say right now is the demand is very extremely robust, okay? And we did not see any sign of slowdown. Whether it's a high 50 CAGR, it doesn't matter. It probably will be higher.

Okay. Thank you very much. Yes. And my second question is also about like the -- given the very strong demand, the CapEx cycles, we see that definitely this round will be much higher than before. And TSMC also announced that earlier, you will adopt a High-NA EUV around 2030. So should we think about that, that will be further improved the productivity that will be more meaningful for further like accelerating revenue growth in the longer term? And also given that the CapEx intensity are also getting much higher. So how should we think about that the pricing power and also the long-term gross margin and the capital efficiency returns? That's my...

Okay. That's actually 2 questions, but okay, just very quickly. So part of it is on with the strong demand and our CapEx cycle, how do we view the adoption of High-NA, which we said we will in 2030? And then her question is on profitability, long-term outlook.

Okay. Let me answer that question. Yes, for the long-term productivity, we continue to improve it and High NA definitely is a very good tool. And so far, I think our plan is to use it in 2030. Again, our CapEx will be much higher than today. And so you're talking about the gross margin or the profitability can be increased or not. Let me say that. There's no ceiling for TSMC's profitability, but a dramatic increase will not happen overnight either. So our philosophy or our strategy always working with our customer, and we want to make them be successful and then we can get our value.

Operator, can we move on to the next participant, please?

Next one, Gokul Hariharan, JPMorgan.

My first question, again, on capacity addition. It looks like the shortage situation next year could be even more than 2026. Just wanted to hear C.C. your view on this. And I think it's been quite some time since we have seen TSMC add capacity at more than 10% every year. I think if you look at your annual report, you see like about 10 years that you have added capacity at more than 10%. Given this accelerated build-out, can we start getting wafer capacity growth closer to 10% or even higher than 10% in the next couple of years? Any comment on like what is your planned wafer capacity growth in the next couple of years? That's my first question.

Okay. So Gokul's first question is regarding the capacity addition and the demand. So -- his question first is the tightness in the wafer supply, do we think that will be even more or greater in 2027 than 2026? And then along with that, how do we see our overall capacity addition? Although you say 10%, I think every year we disclosed in the annual report, it's been about mid- to high single digit and sometimes even lower. But anyways, regardless, those are the 2 questions. So first, on just the demand outlook and then the supply, will next year be even more tight than this year?

I guess I mentioned the demand is extremely robust and getting higher and higher. Of course, we have our own judgment, and we look at every angle. As I said, we talk to customer, customers and customer, and then we do our own one more step to check the data centers build-out. All the results we put into our judgment and our assessment, the result is the demand is very, very high. And so the gap between demand and supply are still quite big, and that's why we put all our strengths to increase the capacity, as I just mentioned. Whether it's bigger or the gap become bigger or the gap become narrower, we are working very hard. I cannot answer this question right now because my customers continue to increase their demand.

Okay. Gokul, I think it's a bit too early to talk about capital capacity increase percentage going forward. But let me just remind you that the increase every year, although the number is -- even though the number may be small, but the increase in advanced packaging -- sorry, advanced nodes are really the main driver for the capacity increase, okay?

Fair enough. Okay. Secondly, could you -- I think you alluded to more aggressive overseas expansion plans in the U.S., especially. And I think recently, there's been a lot of discussion with potentially a new fab cluster in Texas as well as I think one of your customers also talked about discussions with TSMC as well. So C.C., could you talk a little bit about your overseas plans, like the more accelerated kind of capacity build-out? And also any engagement with SpaceX or Tesla, given your customer did call you out last week in terms of some engagements you have? Like what are the broad contours of this kind of engagement likely to be? Is it similar to what you've done with some of the customers in Japan and ESMC as well?

Okay. Thank you, Gokul. So Gokul's second question, he notes that C.C. said, we're expanding at an unprecedented place, doing everything we can as fast as we can to increase the capacity. So this question is particularly to the U.S. What is the status or update? There's been reports that we are looking to expand in Texas. There's been reports that one of our customers, we are engaging in discussions. He wants to know what kind of framework or parameters of any cooperation may be or any comment on either of those 2 things?

Okay. Gokul, -- we work very hard to expand our capacity in U.S. This is necessary because of all our customers, most of our large customer in U.S. and they requested TSMC to increase the capacity. And also U.S. government are very helpful. So that's a place we want to expand. So recently, we did expand our capacity consideration to see outside Arizona, whether we can get another chance to expand our capacity. And Texas is one. So we send the teams to Texas to do all the evaluation. And we also talked to one of our customers who try to build a fab, Terafab, let me be more specific. And we discuss with them to see whether that we can be utilizing that actual capacity also. So we joint discussion, if we have any solid conclusion, we'll share with you. But today, and that's it, yes, from Arizona to Texas, yes.

Okay. But all these are going to be TSMC run fabs, right? I think you're not becoming part of the fab project or anything like that. Could you just help us understand what are the principles that you're using here? Is it the similar principles that you use for JASM or ESMC?

Okay. So Gokul's question is in particular to the potential cooperation with this customer. Would it be [ TSMC own? ] What's the principles that we have, if anything?

We are still in discussion. Let me say that. It's not either way, it will be possible, but the fab got to be run by TSMC.

Operator, can we move on to the next participant, please?

Yes. Next one, Krish Sankar, TD Cowen.

I have 2 of them. First one, C.C. some customers talk about a secondary wafer market where one customer passes their wafer allocation to another. Is this like a very small portion of the business? Is it customer specific? Or do you see that as an emerging trend where customers whose demand is we pass some of the allocation to stronger customers? And then I had a quick follow-up.

All right, Krish. Thank you. So Krish's first question is asking some of our customers are talking about a resell a secondary wafer market where they can pass their allocation on. So is this a very small portion of our business? Or is this more of a trend for us to do? And is this something that we allow our customers to do?

Well, let me answer that. To say whether this is a new strategy, new methodology, the answer is no. The principle of TSMC is we do not allow our customer to sublet or auction of the capacity. But in certain cases, customers can collaborate through a mask licensing agreement, but this is not our common practice. And the amount is very, very small. Krish, did I answer your question?

Yes. And then a quick follow-up. You did touch upon it the High-NA agreement with ASML from 2030 onwards. I'm just curious, historically, TSMC was more ahead of the curve with the EUV adoption. It looks like by the time you adopt High-NA EUV will be 2030, 2031, but some of your competitors have already been using it by then. So do you see that as a disadvantage? Or you're not worried about competition getting ahead with High-NA EUV?

Okay. So Krish's second question is looking at High-NA. He notes some of our competitors are both using -- already adopted or will adopt High-NA earlier than TSMC. So his question, does this put us at a competitive disadvantage or risk compared to our competitors in terms of technology leadership?

Krish? Let me answer that. We always evaluate a new tool. And if we want to put into the mass production, it got to be the best technology with a very reasonable cost and most important is the best maturity, okay? So we say in 2030, we'll start to use it because we also think at that time, High-NA will be mature. And we're also working very closely with ASML to establish a 12-inch photomask by 2031. And at that time, we think that the high cost will be reasonable and for TSMC to make it into the mass production.

And this is a very good tool. Let me emphasize that. And in the smaller geometry, it definitely has some advantage. Whether that our technology because we are using it later than our competitor, so disadvantage to TSMC, no, because we are confident that we are still the best company to offer the most advanced technology to our customers.

Krish, does that answer your question? Operator, can we move on to the next participant, please?

Charlie Chen, Morgan Stanley.

Also to hear very strong AI semi demand amid some recent stock market correction. So my first question is really about your leading-edge capacity expansion strategy. As far as I know, one of your IDM customers is expanding CapEx very aggressively. So there are some concern about this customer may in-source more and outsource to you, right? So first of all, I'm not sure how you're going to deal with this sort of business opportunity? And secondly, I remember last time that C.C., you mentioned about A14 could be a bigger node than N2. May I know whether that assume the big IDM company will outsource more CPU like server CPU at A14?

Okay. So Charlie's first question is around IDM competitor. He wants to know very specifically if we see this competitor moving to in-source more. So his question is then our prior comment of A14 being a larger, bigger node than N2 has even been -- does this include the server CPU from this IDM competitor, which...

Well, Charlie, we treat every customer fairly. So whether it's IDM or it is purely fabless, we treat them the same. But of course, we take into consideration of the internal capacity. So we calculate our support, we calculate our capacity buildup according to our own judgment. Actually, in our A14, we have seen more and more customers actually higher than the 2-nanometer activities so far. And we expect that the A14 will be a bigger node than our current N2 technologies.

I see. So it doesn't need to consider this IDM will also outsource more. We can still have A14 to be a bigger node than N2. Is that the right understanding?

Well, again, Charlie wants to know very specifically this customer outsourcing plan, which you probably may need to ask them, but...

Well, Charlie, as I said, to be frank with you, I don't comment on a specific customer, but let me guarantee you that we take everything into consideration.

Do you have a second question, Charlie?

Yes, I do. Thanks Jeff. So the second question is kind of follow-up about your advanced packaging. From where we've seen that the retail size is getting much bigger C.C., you said last time or even 2 quarters ago about your response, Sow, 3.5D. So lots of initiatives is going on. So my question is that do you think those are kind of a proposed SoW, 3.5D can really catch customers' very rapid development of their larger reticle size chip, like 14 reticle? And secondly, if some customer, they initially choose to use Intel's EMIB-T, and for some reason, cannot get to sufficient demand (sic) [ supply ]. Are you okay to bail out or support customers for the CoWoS version or capacity?

All right. So Charlie's second question is focused on the advanced packaging. I guess, Charlie, if I understand your question, summarize it, is, of course, customers are moving to larger and larger reticle size 14x. So he wants to know TSMC's road map and solutions and how competitive we can be against -- I think he mentioned specifically our competitors' EMIB-T. So how do we view our solution? Is it as competitive as our competitor? That's the first part of this question.

All right, Charlie. Actually, I still think that TSMC's advanced assembly or they say, packaging technology are very, very competitive. But the most important thing is that we're using that one to help us help our customers product can be introduced in the market smoothly. That's our #1 priority. And so we have to offer them the best technology with enough capacity. So far, let me share with you today's situation. We are in the real-world production of 5x reticle size. I don't believe that anybody can do that today, except at TSMC. And the yield is excellent. So our customers are very satisfied with the result. And then we are working on the 9x reticle size and moving into the 14x reticle size.

Of course, we are also working on the CoWoS and to make sure that in the future, we can offer a reasonable cost reduction. But so far, the CoWoS is still the #1 choice for our customer. As time goes by, yes, our competitors offer a very competitive solution. Still, I want to say that today's situation, we are the only one that make sure that the advanced packaging is a 5x reticle size and still, we are very competitive. I want to emphasize that.

And then Charlie's question, if the competitor cannot deliver on their solution, Charlie, I think you're saying we will need to be something like a backup plan.

I try to avoid to answer it directly, but let me say that. Yes, we will be #1 priority. I already indicated that TSMC's customers product can be introduced into the market smoothly. That's the #1 priority, okay? And that gives you enough hint.

Yes, indeed. I hope you don't give any food to others, including the advanced packaging.

All right, operator, let's move on to the next participant, please.

Next one is Sunny Lin, UBS.

So my first question is to have a bit more color on how we should think about CapEx beyond 2026. Recall a few quarters ago, you mentioned if you are doing your job well, then you may be able to drive CapEx to grow a bit less than top line. But now obviously, the overall cloud demand has further strengthened and you're also willing to support your clients with more substantial expansion. And therefore, I wonder, at this point, when you look at maybe the coming 2 years, would you expect maybe CapEx to grow a bit faster, maybe in line with sales or even a bit faster just like for 2026? And then meanwhile, obviously, the tool price is also increasing given a very tight supply. And therefore, going to 2027, would the higher tool price also be a driver for high CapEx as well?

Higher -- sorry. What was the last part of your question, Sunny? Higher tool price, higher what?

So yes, I wonder when we think about 2027 CapEx, should we also consider tool price inflation to be a meaningful variable for a lot higher CapEx compared with this year.

Understood. Okay. Sorry. Okay. So Sunny's first question is about our CapEx beyond '26. really, she's asking '27, '28, the next 2 years. With the strong extremely robust demand that C.C. indicated, question is around our capital intensity in the next 2 years. How do we see the capital intensity trending revenue growth versus the CapEx growth, part one. And then also the impact of tool inflation on the CapEx outlook as well.

Okay. Sunny, this is Wendell. It's a bit too early to share the CapEx number beyond this year. But as we said in the prepared remarks, we have very strong conviction in the AI megatrend. Therefore, our CapEx from 2026 to 2028 is expected to be significantly higher than the past 3 years. And this is a step-up of our investment to support the customers' growth. We have no intention to leave any food on the table for others. The tool price inflation is real. The tool price, material, labor costs, et cetera. However, the majority of the CapEx is always related to the long-term demand. So that will be the composition going forward. And capital intensity-wise, as we always said, if we do our job right, the increase in revenue will outpace the increase in CapEx, and that is still the case from what we are looking at.

Got it. If I may have a very quick follow-up on the CapEx side. Earlier, Wendell, you raised the margin dilution from overseas to, I think, 3% to 4% starting from 2027, given faster overseas expansion. But for the U.S., you're still so far focused on Arizona. So does that imply for Arizona, the expansion for the second phase will happen very quickly or you are actually accelerating the expansion for the third phase or even the other phases as well? Is that therefore the reason for higher margin dilution for next year?

Okay. So Sunny's question is kind of about the puts and takes to next year's margin. She does note, as Wendell said in our prepared remarks, that the overseas fabs will widen a bit to 3% to 4% in the next few years. So she wonders what's driving that wider dilution. C.C. did say we are pulling in the schedule for the second fab for 3-nanometer. Her question is, are we also pulling in third fab? What's leading to this wider dilution? And maybe -- yes.

Okay. Sunny, as we mentioned, we are doing everything we can, wherever we can to increase our capacity to meet the strong demand from the customers. And we are also accelerating everywhere, including in the U.S. We did mention before that the later part of the 5-year period, the dilution -- margin dilution from overseas fab will be 3 to 4 percentage points, and that still remains the case. So it's 3 to 4 percentage points in the second part of the 5-year period.

I see. My second question, I want to follow up.

That was 2 questions, Sunny. So there's already 2 questions. Can we -- sorry, but we probably just to accommodate as many as we can, we're going to move on. Thank you for your understanding. Operator, can we move to the next participant, please?

Next one, Haas Liu, Bank of America.

Congrats on the great results. My first question would be on your view regarding the potential shift in AI demand mix. You mentioned in the prepared remarks that the AI accelerator chipset demand is proliferating into generative AI demand, including CPUs and networking switches. We understand the overall AI investment is still pretty strong and even trending stronger across the board. But can we get more clarity that are you seeing your customers' demand shifting towards more efficient compute solutions, especially monetization gets more important? And how should we think about the incremental opportunity for your manufacturing and packaging business?

So Haas's first question is on the AI-related demand. His question is, are we seeing a shift in the AI-related demand more and more to CPU or networking switches and away from AI accelerators? Or any type of comment?

Well, all I can say is AI demand is extremely robust, okay? Whether it's come from accelerator or more from the CPU side, it's not for me to make a comment. Although we did see the demand from Agentic AI, CPUs continue to increase. Yes, that is a trend today. But other than that, everything is very robust. And I want to emphasize again, in our judgment, we consider all the parameters and the demand actually is still far above our capability today to supply, and we are working very hard to narrow the gap. That's all I can say.

Okay. That's very clear. And my second question is just regarding the yield learning curve going forward for your dual technology tracks strategy going forward. So historically, I think you have been ramping up the yield of a new node through producing the chipsets with smaller die size and simpler design to improve the learning curve of a new node. But for the future node families supporting both conventional power delivery for the applications like mobile and consumer applications and Super Power Rail technology for HPC and AI chipsets, would operating two process architectures at the same time make the learning curve more challenging or increase the manufacturing complexity? And how should we think about the implication to your gross margins from the yield and also cost control perspective?

Okay. So Haas's second question is -- his question is really about ramping a new node, the learning curve of yield and cost and therefore, the implication to gross margin, really, I think, is what he's driving at. And his question is that in the past, a new node is mostly ramped with smaller die size and simpler design. But -- so today, we have both the smartphone and also with the AI and HPC applications coming in. Does that increase the complexity of the node process architecture, make the learning curve more challenging or difficult and therefore, the implication to the margin?

Well, today's situation actually is a little bit different from previous year. Yes, we now have a very strong demand from smartphone and also AI HPC applications. And so the smaller die size per se or the big die size in the accelerator, they happen at the same time. And this is to TSMC's advantage, let me say that, because we are good. We have been prepared for this kind of thing. And our results are very excellent and customers are very happy to work with us. And so the learning curve, yes, but we are doing a reasonable job. And so all the customers are very happy today.

Okay. Operator, let's move on to the next participant, please.

On the line Macquarie.

I have a question on the power as gating factor. So I want to understand have power grid constraints led you reduce your customer forecast in Q3? And also, could this tighter and tighter power grid surprise your customer actually accelerating adoption of your Super Power Rail products for the better performance per watt?

Okay. So Arthur's first question is about the power supply or electricity supply. Is it a gating factor? And have we seen that lead to a reduction in customers' forecast? And with the tighter -- on the other side or on the other hand, with the tighter power electricity supply, have we seen more customers' adoption of Super Power Rail technologies?

Well, let me update that. Definitely, the power supply situation is one of the consideration factor that we are looking into. And that's why also we do one more step to double check whether the power supply situation in the U.S. And we have done all the study and so far, so good, let me say that. And whether this kind of situation change our customers' behavior and move more into the backside power rail? The answer is no. Whether you use the backside power, those kind of metallization scheme is based on your need for your product, okay? So some of the accelerator might be using it, some of the products like mobile phones, they don't use it. So it depends on customers' product. Did I answer the question?

Yes.

Do you have a second question?

So just to recall the previous discussion on the Agentic AI demand on the CPU. So I wonder according to our previous framework, demand forecast framework, the mid- to high teens growth, included the CPU or not? And if we hear our clients increase the CPU forecast significantly, when is the next time we refresh our AI demand growth numbers?

Okay. So Arthur second question is on -- his question is really on our -- we already stated by anyway, our AI accelerator definition. He wants to know with the emergence of Agentic AI and CPU, does that AI accelerator definition include CPUs or not? And if when we will provide inclusion for him?

Let me answer the question. It's not included in our AI demand calculation. The reason is very simple because we cannot identify which CPU user in the conventional server CPU or AI server CPU. And also that it depends on our customers' application and their own allocation also. So no. So far today, we stick to our initial very narrow definition on the AI demand.

Understand. And so if we have seen a really strong CPU-driven demand, when we will revise up the...

Arthur, I think we already said very clearly, it's not part of the AI accelerator definition. It's a narrow definition, but it is certainly part of our business, okay?

Okay. Operator, in the interest of time, can we take question from the last participant on the line, please?

Sahej Singh, Stifel Financial.

Thank you for the question. Into 2027, you have N3 turning accretive and now that's crossed over, N2 is still in its dilutive ramp and the overseas dilution is widening. So on the offset side, there have been reports of significant price increases across your mature and advanced nodes. Could you provide some color into perhaps some of that conversation and where pricing does move, does it typically apply across the existing contracted book? Or is it mainly being applied to incremental capacity above what customers have already committed to?

Yes. Okay. So the first question is talking about -- I guess you're asking about the potential puts and takes for next year? Or you're asking specifically on pricing, sorry? For pricing.

Specifically on pricing.

Okay. So it's a very specific question that there's reports that we are increasing the price. He wants to know, does it apply across the existing contracted commitments from customers? Or is it mainly for the incremental capacity?

Well, we have discussed this kind of thing before that our pricing strategy is very strategic. It also depends on the market segment, depending on the location and depends on the technology and something. Once we offer the flexibility of the physical location like in the U.S., it will be a little bit different from the one we produce in Taiwan. So no matter what, well, we try to get our value to be recognized, and we will give our shareholders the maximum value of our technology.

Understood. For the second question, something you said in your prepared remarks, stuck out to me was we will not be leaving food on the table. And earlier, I think it was maybe 1 or 2 weeks ago, there was the announcement via GlobalFoundries that they would be providing silicon interposers to TSMC. And so would love some color into how are you thinking through these strategic decisions in terms of procuring from some of the specialty foundries versus providing in-house?

Okay. So his second question is, again, he notes that recently, we had an announcement with GlobalFoundries and interposers. So how do we view these partnerships? And what is our approach strategically?

Let me answer the question. We designed a deal with GlobalFoundries -- and this is part of our overall strategy in the back-end supply. Actually, our #1 priority is to help our customers to be successful in their market. We want them to be fully supported by all the supply, including the front end and back end. Definitely, we have a little bit shortage on the back end, and we'd like to work with our so-called partner with GlobalFoundries so that we can both TSMC and GlobalFoundries can support TSMC's customer for their product to be introduced with enough support on the capacity.

Okay. Thank you for your questions. Thank you for all the participants. This does conclude our question-and-answer session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within 30 minutes from now. The transcript will become available 24 hours from now, and you can find both will be available through TSMC's website at www.tsmc.com. So again, thank you for joining us today. We hope everyone continues to stay well. Have a good autumn, and we hope you will join us again next quarter. Thank you, and have a good day.

Good afternoon, everyone. And welcome to TSMC's Second Quarter 2026 Earnings Conference and Conference Call. This is Jeff Su, TSMC's Director of Investor Relations and your host for today.

Today's event is being webcast live through TSMC's website at www.tsmc.com, where you can also download the earnings release materials. [Operator Instructions]

The format for today's event will be as follows: First, TSMC's Senior Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the second quarter 2026, followed by our guidance for the third quarter 2026. Afterwards, Mr. Huang and TSMC's Chairman and CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then we will open both the floor and the line for the question-and-answer session.

As usual, I'd 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.

And now I would like to turn the microphone over to TSMC's CFO, Mr. Wendell Huang, for the summary of operations and the current quarter guidance.

Thank you, Jeff. Good afternoon, everyone. Thank you for joining us today. My presentation will start with financial highlights for the second quarter of 2026. After that, I will provide the guidance for the third quarter of 2026.

Now let's move on to revenue by technology. 2-nanometer process technology contributed 3% of wafer revenue in the second quarter; 3-nanometer, 5-nanometer and 7-nanometer accounted for 30%, 33% and 11%, respectively; advanced technology, defined as 7-nanometer and below, accounted for 77% of wafer revenue.

Moving on to revenue contribution by platform. HPC increased 20% quarter-over-quarter to account for 66% of our second quarter revenue. Smartphone decreased 4% to account for 22%. IoT increased 4% to account for 5%. Automotive increased 15% to account for 4%. DCE increased 5% to account for 1%.

Moving on to the balance sheet. We ended the second quarter with cash and marketable securities of TWD 3.5 trillion or USD 110 billion. On the liability side, current liabilities increased by TWD 144 billion quarter-over-quarter, mainly due to the increase of TWD 58 billion in accounts payable and the increase of TWD 48 billion in accrued liabilities and others.

In terms of financial ratios, accounts receivable days increased by 3 days to 29 days. Inventory days increased 7 days to 87 days, primarily due to the ramp of N2 technology.

Regarding cash flow and CapEx. During the second quarter, we generated about TWD 783 billion in cash from operations, spent TWD 496 billion in CapEx and distributed TWD 156 billion for third quarter 2025 cash dividend. Overall, our cash balance increased TWD 99 billion to TWD 3.1 trillion at the end of the quarter.

In U.S. dollar terms, our second quarter capital expenditures totaled USD 15.7 billion.

I finished my financial summary. Now let me -- let's turn to the current quarter guidance. Based on the current business outlook, we expect our third quarter revenue to be between USD 44.6 billion and USD 45.8 billion, which represents a 12% sequential increase or a 37% year-over-year increase at the midpoint.

Based on the exchange rate assumption of USD 1 to TWD 32, gross margin is expected to be between 65% and 67%, operating margin between 56% and 58%. This concludes my financial presentation.

Now let me turn to our key messages. I will start by talking about our second quarter '26 and third quarter '26 profitability. Compared to the first quarter, our second quarter gross margin increased by 150 basis points sequentially to 67.7%, slightly ahead of our guidance, primarily due to cost improvement efforts and a slightly higher overall capacity utilization rate, partially offset by dilution from our overseas fabs.

We have just guided our third quarter gross margin to decrease by 1.7 percentage points to 66% at the midpoint, primarily as we expect the steep ramp-up of our 2-nanometer technology to dilute our gross margin by about 3 to 4 percentage points. This dilution is expected to be partially offset by very strong demand for our leading-edge technologies and continued cost improvement efforts, including productivity gains and across node capacity optimization.

Looking at the second half of the year, given the 6 factors that determine our profitability, there are a few puts and takes that I would like to share.

First, we expect the steep ramp-up of our 2-nanometer to dilute our gross margin by about 3 to 4 percentage points in the second half of the year. Furthermore, as the scale of our overseas expansion grows, we continue to forecast the gross margin dilution from the ramp-up of overseas fabs in the next several years to be 2% to 3% in the early stages and widen to 3% to 4% in the latter stages.

On the other hand, demand for our leading-edge technologies is very strong. In addition, we continue to leverage our manufacturing excellence to generate more wafer output and drive greater across node capacity optimization in our fab operations to support our profitability.

Finally, we have no control over the foreign exchange rate, but that may be another factor.

Next, let me talk about our 2026 capital budget. At TSMC, a higher level of capital expenditures is always correlated to higher growth opportunities in the following years. With our strong technology leadership and differentiation, we are well positioned to capture the multiyear structural demand from the industry megatrends of 5G, AI, and HPC.

Given the continued strong structural demand from our customers, including the newly emerging Agentic AI market, we have decided to raise our full year 2026 capital budget to be between USD 60 billion and USD 64 billion as we continue to invest heavily to support our customers' growth. We always collaborate closely with the tool suppliers well in advance to prepare the capacity, whether it is a strong up cycle or down cycle, just like our customers collaborate with us well in advance to plan our capacity. Thus, we do not foresee any bottlenecks to our capacity expansion plans.

About 70% to 80% of the 2026 capital budget will be allocated for advanced process technologies. About 10% will be spent for specialty technologies and about 10% to 20% will be spent for advanced packaging, testing, mask-making and others.

Even as we invest for the future growth with this level of CapEx spending in 2026, we remain committed to delivering profitable growth to our shareholders. We also remain committed to a sustainable and steadily increased cash dividend per share on both an annual and quarterly basis.

In 2025, we paid TWD 467 billion in cash dividends, up 28.6% year-over-year as TSMC shareholders received a total of TWD 18 cash dividend per share. In 2026, they will receive TWD 24 per share, up another 33% year-over-year, and we expect a continued and increasing cash dividends per share in 2027 as well.

Now let me turn the microphone over to C.C.

Thank you, Wendell. Good afternoon, everyone. First, let me start with our near-term demand outlook. We concluded our second quarter with revenue of USD 40.2 billion at the high end of our guidance in U.S. dollar terms driven by strong demand for our leading-edge process technologies.

Moving into third quarter, we expect our business to be supported by continued strong demand for our leading-edge process technologies, including the steep ramp of our 2-nanometer technology.

Looking ahead, we observe consumer and price-sensitive end market segments are being challenged due to the impact of rising component prices and macroeconomic uncertainties. As such, we are being prudent in our business planning while focusing on our fundamentals of our business to further strengthen our competitive position.

Having said that, AI-related demand continues to be extremely robust. The AI megatrend continues to drive the need for more and more computation, which supports the robust demand for leading-edge silicon. Our customers and customers' customers, who are mainly the cloud service providers, continue to provide us with a very strong signal and positive outlook. Thus, our conviction in the multi-year AI megatrend remains very high. Supported by our robust technology differentiation and the broad customer base, we now expect our full year 2026 revenue growth to be slightly above 40% year-over-year in U.S. dollar terms.

Now let me talk about the acceleration of Agentic AI. The AI market continues to be very dynamic. The emergence of Agentic AI is leading to a resurgence in the role of CPUs in AI data centers, which drives more silicon demand in addition to AI accelerators. We believe this is positive for TSMC as no matter what CPU approach is taken, whether it's x86, ARM-based, or RISC-V architecture, they are almost all TSMC's customers. We are already collaborating closely with our CPU customers and working to support them with the most advanced technologies and necessary capacity, so they can capture the Agentic AI market opportunities.

Next, let me talk about TSMC's capacity expansion strategies. To address the structural increase in overall long-term semiconductor market demand profile, TSMC collaborates closely with our customers and our customers' customers to plan our capacity.

Given the fundamental complexity of leading-edge technologies and the design-in and lead time involved, we also have a very good idea of their multiyear product road map and production plans. This is important because it takes more than 5 years to develop the technology and product, prepare the capacity, and ramp it up to high-volume production.

Internally, TSMC employs a disciplined capacity planning system to assess the market demand from both a top-down and bottom-up approach. This is a continuous and ongoing process. Based on our assessment, we are stepping up our CapEx investment to increase our capacity, to support our customers' future growth.

Now with a strong collaboration and support from our leading U.S. customers and the U.S. federal state and city governments, we would like to announce an additional USD 100 billion investment in Arizona. This is to build several more semiconductor logical wafer fab for 2-nanometer and below technologies as well as advanced packaging fabs to support the strong multiyear demand from our leading U.S. customers.

We believe this investment will help to further foster the development of the U.S. semiconductor ecosystem, strengthen the supply chain, and support an increasing number of high-tech, high-paying jobs in the United States. At the same time, we are building 13 leading-edge and advanced packaging fab in Taiwan over the next several years, and we will continue to further invest in Taiwan.

Therefore, TSMC's semiconductor technology and manufacturing work continue to play a pivotal role in supporting the global semiconductor industry while unleashing our customers' innovation.

Now let me talk about the current N3 capacity expansion. We are executing well on our global plan to add 3 additional 3-nanometer fabs, one in Taiwan, one in Arizona, and one in Japan to support the robust multiyear pipeline of demand for 3-nanometer technologies.

In addition to all the new fabs, we continue to convert 5-nanometer tools to support 3-nanometer capacity in Taiwan. We are also leveraging our manufacturing excellence to drive greater productivity across our fab in all locations to generate more wafer output. We are also focusing on capacity optimization across node, which including flexible capacity support among N7, N5 and N3 nodes.

In summary, we are using multiple levers to do everything we can, wherever we can, however we can to maximize the support to all our customers.

Now let me talk about our mature node strategies. TSMC's strategy at mature node has not changed. Our first priority is to fully support our customers. And now we continue to increase, not decrease our mature node capacity in the higher value-added segment. For example, we are increasing our mature node capacity through JASM Fab 1 in Japan for CMOS image sensor application and ESMC in Germany for automotive and industrial applications.

In today's market, outside of specific areas such as power management IC and CMOS image sensor, the mature node demand in other commodity area is not that strong. Thus TSMC will continue to focus on the higher value-added and strategic segment by ensuring we have a necessary capacity to support our customers' growth.

Finally, let me talk about our A14 status. As I mentioned a few minutes ago, the complexity of leading-edge technology continue to increase. The lead time to develop a new technology such as A14, building the capacity, and then ramping it up now takes 5 to 7 years. There are no shortcuts.

Our A14 technology representing the second generation of nanosheet transistors and deliver another full node stride from N2 with performance and power benefit to address the incessant need for high-performance and energy-efficient computing.

Compared with N2, A14 will provide 10 to 15 speed improvement at the same power or 25 to 30 power improvement at the same speed and close to 20% chip density gain. A14 technology development is on track and progressing well.

Internal product-like vehicle demonstrated close to 90% device performance and close to 90% 256 megabits SRAM yield. We are observing a strong level of customer interest and engagement from both smartphone and HPC AI applications and customer now tape-out activity is ongoing and ahead of schedule. Pre-production will start in 2027 and volume production is scheduled for 2028.

With our strategy of continuous enhancement, we also introduced A13 and A12 as extension of the A14 family. A13 represents a further advancement of A14, achieving an over 6% die area saving through an innovative 97% optical shrink. Through continuous design technology co-optimization, A13 also drive further performance and power efficiency improvement. A13 design rule are backward compatible with A14 to ensure smooth IP migration.

We also introduced A12, which will bring our innovative superpower rail technology to the A14 platform for superior performance, power, and area benefit. Both A13 and A12 are scheduled for volume production in 2029. We believe A14 and its derivative technologies will propel our A14 family to be an even larger and long-lasting node for TSMC than N2. Just like 2-nanometer technology is a larger and longer-lasting node than 3-nanometer, and here further extend our technology leadership position well into the future.

This concludes our key messages, and thank you for your attention.

Thank you, C.C. So this does conclude our prepared statements. [Operator Instructions] Please note that we will conclude today's meeting at around 3:10 or so. So we will try to get in as many participants your questions as possible. But if we're not able to, we do apologize in advance, and thank you, everyone, for your patience.

So operator, well, let's begin the Q&A session. We'll take the first 2 questions from the floor, and then we'll go online. Maybe again, left, center, right. Maybe we'll take the first question, Sunny Lin from UBS, please.

Congrats on the very strong performance and outlook. So number one, I want to double-click on the CapEx. So, very encouraging CapEx outlook. And I do think it's essential that TSMC showcase a stronger determination in capacity expansion, given the stronger demand and the very tight supply. And so beyond 2026, I think every large client also wonder how aggressive TSMC is planning for CapEx. And so back in the COVID super cycle, TSMC did provide a 3-year CapEx outlook back then. And so I wonder, at this point, will it be possible for you to share any color, maybe for the coming 3 years CapEx? Thank you.

Okay. So Sunny's first question is regarding CapEx. She does believe it's important, essential to show our determination to support our customers with these large CapEx investments. So she wants to know, do we have a 3-year CapEx guidance '26, '27, '28, similar to what we did back in 2021?

Okay. Sunny, we do not have a number to share with you. But as you know, we invest CapEx this year for the future business opportunity. And as long as there are business opportunities, we will not hesitate to invest. As you can hear from our prepared remarks that we -- our conviction in the megatrend, AI megatrend multiyear is very strong, and we are stepping up the CapEx, including increasing this year's CapEx. Last time, we said our CapEx in the next 3 years will be significantly higher than the CapEx in the past 3 years. Now is the -- the CapEx in the next 3 years will be even more significantly higher than the past 3 years.

Yes. Sorry, maybe let me follow up on CapEx from another...

Second question.

Yes, sure. So you just announced additional $100 billion CapEx in the U.S., and I think that's pretty important for you to secure the business in the U.S. as well. And so now with total $265 billion CapEx in Arizona, what's your current plan to bring on the capacities in Arizona in the coming few years?

Okay. So Sunny's second question is on to, C.C. said, investing additional $100 billion in Arizona based on the strong demand from our customers. So the total investment now is $265 billion. What is the schedule time frame or the plan for this investment? Is that correct, Sunny? Yes. Okay.

Sunny, the schedule will depend on the market situation. You know that. So today's situation, the megatrend is so strong so that we announced additional $100 billion investment in Arizona. How many fabs? Many. So actually, let me say that, say probably, additional 4 more fabs will be built.

And that's combining front- and back-end?

Yes.

Okay. Thank you. Let's go to the middle. We have Charlie Chan from Morgan Stanley. We'll go left, middle, right from where I sit.

So first of all, congrats for a very strong outlook. My first question is really about the foundry competition. I understand that there's no shortcut for a newcomer like TerraFab, but how about like Samsung Foundry, right? They got a huge profit from memory business. Intel got a U.S. policy support. So I'm not sure how TSMC is going to address those competition, because apparently, several U.S. companies are engaging with those industry peers. And recently, actually yesterday, ASML just announced to expand the EUV capacity for 2028. Would TSMC worry that your competitors to take more slots and build a large capacity in the future to compete with you in the leading-edge business?

Okay. So Charlie's question is competition from 2 angles. One, he does note, C.C. said foundry competition, no shortcuts, but he says according to the news, many of our customers are engaging with our foundry competitors. One of them in Korea is making huge amounts of money these days. Another one may have the U.S. government policy support. So the first part of his question, how do we see the competition and threat of customers moving to our competitors, number one.

Well, let me say that. Yes, one of my competitors in South Korea, they make a huge amount of money, and I'm jealousy about it. And then the other one in the U.S., they got a very strong U.S. government support. We also got the government support, by the way, although we don't announce it.

However, let me share with you, as we said, there is no shortcut. What does that mean? Meaning that in the semiconductor industry, you have to go back to fundamental. Government's help is welcome. Really, we also appreciate that. A lot of money, of course, that's nice to have.

But the most important thing as we continue to say is the technology, manufacturing, and customer trust. These 3 fundamental never change. For my 30-some years, 40 years career, it's always the most important thing. And that's always the TSMC's secret recipe to win the business.

So from my -- from the competition point of view, choosing a technology, ramping it up is not buying a milk from 7-Eleven. Well, I'm using that -- I'm quoting the sentence for my customer, anyway. It says that you're choosing a kind of a technology partner, it is no shortcut. You need to understand the technology. You need to really utilize it using the test chip, and then something, and work together, and then prepare the capacity and ramp it up. That's why I would say, it takes about 5 years. It's not that today, you think this milk is better, you go to the next store, it's a 7-Eleven. You don't like it, you go to another store. No. So that's my answer, Charlie, is it's -- agree?

Yes. So I hope you can buy more milk, so other people can get it. Yes. So let me switch gears to a more exciting side. C.C. just said you see a very strong signals from customers, customers' customers. You also revised up the full year revenue guide. Are you ready to revise up the 5-year revenue CAGR, especially this AI semi CAGR? I remember it was like high 50%. But here comes the question, right, that Agentic AI demand is so strong, CPU is a great opportunity for TSMC. But how about those kind of memory cost increase, right? It's kind of a big chunk of this kind of AI CapEx. So, what's the update of the AI semi CAGR? And how should we look at the contents of this AI semi related to TSMC's growth?

Okay. So Charlie's second question is regarding the AI-related demand. We do continue to see very strong and positive signals from our customers, customers' customers. We revised up our full year. So his question is around our AI CAGR guidance that we gave in January on a 5-year period, mid- to high-50s CAGR growth. He's wondering if there's any update to that. Agentic AI, a new opportunity, what is our definition of AI accelerator? Do we include that? And what is the CAGR?

Charlie, if you read our message that we continue to invest more. We increased the CapEx with a good reason. So if you're asking about the AI's CAGR, let me give you not a number, but it's stronger and stronger and stronger. So we don't give you the number today because it continue to increase. So we don't know how to answer this question, but stronger than what we said before.

Thank you, Charlie. All right. Let's move to this side. Maybe we'll take the question from Arthur from Macquarie.

First, congrats on the strong execution and performance. My question is regarding the new advanced packaging technology. We noticed that especially the EMIB-T is gaining traction. So how will TSMC react this request?

Okay. Sorry. So Arthur's first question is on advanced packaging and competition. I guess, very simply put, EMIB-T, in his view, is gaining traction. So how do we see the competitive threat from this?

Well, let me say that, our packaging capacity is so tight that now it's limited by customers' growth. So we welcome that additional flexibility in the market. And so that will help TSMC's front-end wafer business growth, which is a majority part of TSMC's business. The technology looks good, according to the newspaper. And we hope they will be successful and so that share some of the loading from TSMC. Today, we're working very hard to shorten the gap between the demand and the capacity. And so as I said, we welcome to have this additional alternatives, and so the flexibility for my customer.

That makes a lot of sense. So a follow-up. So as this is a new technology, right? So if your clients, they ask your support, and our value is support our customers' success, right, how TSMC will handle this special request?

Sorry, your question is...

So if these technologies have some small problem, and then ask our company to support. So how our company accommodate it?

So Arthur's question is, if there is some issues with this technology, do we have a -- are we -- I don't know how to say it. Is there an alternative plan?

Let me answer the question. Our #1 is to support our customers' success. So whatever that we can hear about our customers' business, we want to win. Does that answer your question?

Okay.

Let's come back. We'll take one more here, and then we'll go online and then back to the room. Gokul?

First question on, maybe since you are not wanting to give a longer-term numerical guidance, could you talk a little bit about the philosophy of how you are expanding capacity? Obviously, customer feedback, customers' customers feedback is important. Do you also consider competitive pressure? Because as an outright market leader, having like undersupply for a very long period of time is not really desirable for TSMC, right? You probably want a market which is more balanced. So when you think about your capacity expansion, like how long do you think it takes to fulfill the demand as you see right now? That's one.

And second, chips, obviously, is the current shortage. But there is also a lot of discussion about data center delay, power capacity being available. So could you also share some thoughts on how you are layering in that kind of concerns because you don't want your chips to be available, but having to wait for the data center deployment to happen. So just to understand how that goes into your planning framework as well.

Yes. Thank you. So Gokul's first question is, again, how do we plan our capacity and determine the capacity expansion plan. Certainly, we take into consideration the demand, multiyear demand from our customers and customers' customers. But do we also consider the competitive pressures from competitors building capacity? Is that part of our calculus to expand the capacity, one? And then also, what about things outside of chips like data center delays or power, these type of deals?

Gokul, that's a good question. Definitely, every time when we think about the business, we consider the competition. That's number one. And then, we look at where we are. And then we decide a bottom-up and then top-down's assessment of those demand. Those are the typical thing, I mean, in our daily life. So we make a lot of judgment and then we'd be more careful. We talk to customers and customers' customer. Those are the CSPs. And then, we get all their input for the demand. And then we make a judgment.

Now remember that I believe every customer tell me the truth, everyone. You put all the truths together, it's not the truth. So we have to make some of the judgment. You know what I mean, since you are laughing. Because all the customers are very aggressive, right? That's the CEO's job. CEO got to be aggressive.

So they give me the number of their demand, and I believe they try their best to tell me the truth. So I put all together, all the truths together is not a truth. Mark down that word.

So yes, we do a very careful judgment. May not be correct, may not be correct, but we did carefully and because this is a big money, right? This year, we say we increased the CapEx from $52 billion to $56 billion, now $60 billion to $64 billion. And you bet, that will continue to increase. It's a big money. So we do it carefully.

So we did all the assessment and that lead to your second question. Are we sure that we deliver the chips to our customer, and they were not put into inventory? So we -- actually, we are checking the AI data centers progress, the building, the location, the demand, the racks, we're checking all that to make sure that TSMC chips will not be put in inventory. That answer your question?

Yes, that's clear. So C.C., so do you -- you still believe even end of next year, we are still going to be running short of supply even with this elevated capacity build-out plans?

You want me to give you a guarantee, right? Let me say that, I believe from this day on all the way to probably 2029, 2030, the demand is very strong. Whether in between there's a dip or not, I'm not very sure. But the trend is so robust that I believe we are witnessing a kind of a new industry. I would like to say the new industry called AI industry, which is so common in our daily life because it's going to affect our automotive, affect the humanoids, robot, and also impact to all the industry. So by the amount of money we put in, I mean, including all the CSPs, this alone is a very important new industry to the world. And so the demand will be there. And the fundamental thing is semiconductor chips, and most of them in TSMC.

My second question is on your profitability. So C.C., you joke that you are definitely jealous of your memory competitor on their margins. But it definitely feels like profitability-wise, longer-term foundry, especially leading-edge foundry should be higher than memory, looking at number of competitors out there. So as you are investing for a lot of your customers, how is that discussion going? Because you are no longer the most profitable semiconductor manufacturing company at this point in time. So you probably have less pressure in terms of passing on your value and capturing your value right now compared to maybe 1 year back.

Okay. So Gokul's second question is on profitability and pricing to a certain extent. Of course, some of the memory makers are making very good profitability and margins today, but he notes the role of foundry could be even more value in TSMC's role as well. So what should be the right way to think about the long-term profitability for your foundry? Should it be better? And then I guess, really pricing into this, what type of pricing approach do we want to take?

Yes, Gokul, your question actually is simple. What is the wafer pricing strategy for TSMC, and what kind of gross margin we should have?

The higher, the better, of course. But we are a partner, a partner meaning that I said many times, our customers got to be successful. I don't want to squeeze them out from the market. And besides, we are a very trustable company with our customers. So we don't suddenly increase our price by, which I like to have, 4x or 5x. You cannot survive for that kind of -- for your customer to survive for that kind of a price increase.

So we earn our value, and we make sure that our profit, our gross margin is enough for our long-term sustaining expansion. That's to the benefit of my customer and TSMC also. That's our philosophy. So yes, I'm really jealous about the memory companies, 86% gross margin. 86%? About 68%, I would be happy about that. But all right. Anyway, so I answered the question. We are very trustable.

Thank you. Operator, can we take the next 2 questions from participants on the line, please?

Yes. Now it's James Fontanelli from Arete.

Yes. Could I ask about the risk that you see around customer concentration, as AI demand continues to significantly outgrow other end markets? I think your exposure to your top 5 customers is becoming meaningfully larger than at any point in your history. So I'd just like to understand how you see that risk.

Okay. So Jim's first question is risk around customer concentration. We have large customers that are getting larger. Are we worried that we have too many big customers or the customer concentration?

No, that's not our concern. Besides what you say the customers are growing bigger and bigger, we are very happy about it. And some of the customers also growing very fast. So it's not -- Jim, it's not what you said that the bigger customer is growing bigger and bigger. No. I mean that's -- there's a lot of new player in the AI industry.

Do you have a second question?

Yes. So we're seeing your direct customers put capital into both financing, investing and investing in AI demand. Is that something that TSMC is considering?

So Jim's second question, he knows some of our customers are helping to invest in their customers. Jim, if I understand you right correctly, you're asking if TSMC, if this is an approach we would take to invest in our customers. Is that correct? Or financing and investing?

Yes, in the end-to-end customers, not your direct customers.

Right. So in customers' customers as well.

To answer, Jim, to answer your question directly, every company has a different consideration and every company has a different strategy. So far, no, TSMC don't do this kind of financial arrangement because of, we think we are working with current customer with the current model smoothly and also successfully.

Okay. Thank you, operator. Can we move on to the next participant on the line, then we'll come back to the floor.

Next one to ask question, Mehdi Hosseini from SFG.

Two, from my end. I want to go back to the $100 billion investment in U.S. Is there any way you can give us some time line? Is it over the next 3 years, 5 years? How should we think about the progression of this $100 billion investment in the U.S.? And I have a follow-up.

So Mehdi's first question is around the announcement today, additional $100 billion investment in the U.S. In terms of the CapEx time frame, is it in the next 3 years, in the 5 years? Do you have any schedule or time frame to share about this additional $100 billion?

We do have a plan, but let me share with you actually the progress or the schedule. Most of the time, it depends on the market situation and our customers' demand. So if you ask me to give you a firm schedule, no, we don't have it today. But we do have a plan. And we speed it up. We try to speed it up as fast as possible.

So the message is, you're flexible, but also you're expediting the investment in U.S. Is that correct?

So -- well, I think C.C. said we're trying to move as fast as we can, but everything is based on our customer needs.

Yes. We're also moving the new fabs and the facilities in Taiwan as fast as possible. And the same thing, we try to bring up a new fab in the Japan as fast as possible. Because of the situation today is the demand and the supply, the gap is so big. So we are working very hard to narrow the gap.

Do you have?

I want to as a follow-up. Actually, I want to dive into the compute part of HPC. And I want to ask you about networking switches. And in that context, when should we expect the COUPE platform to have a material contribution to your top line?

So Mehdi's second question, very specific. He wants to know for our COUPE platform, when will it have a very meaningful contribution to the business?

We start the production right now, and it will be ramped up. As time goes by, I think the AI data center need to lower down the power consumption and increase the bandwidth of the communication channel. So I believe the COUPE will continue to increase the demand, and then will become a fairly important technology in the next few years.

Okay. Thank you. Let's come back to the floor. We'll take the next question from Citibank, Laura Chen.

And my first question is also about a very promising outlook as TSMC raised the CapEx and also the growth outlook for this year. And particularly, I think C.C., you mentioned about the Agentic AI and the CPU growth potential. But can you give us more update among that AI, different kind of chips between GPU, accelerators or CPU? What you see the growth potential and your visibility?

So Laura's first question is around sort of the outlook. We obviously raised the CapEx and growth outlook for this year. She wants to know within the AI, the outlook for Agentic AI and CPUs versus AI accelerators, GPUs, et cetera. How do we see these segments?

Laura, I don't think I can give you a very specific number. But let me share with you. All of them are in TSMC. And they're also using the same kind of leading-edge technologies. We're working with our customers to allocate the wafer, the supply to balance the CPU, GPU, XPUs ratio, okay?

Okay. That's -- yes, that makes sense. And then my second question is also about the advanced packaging. We know that during the symposium, TSMC previously already announced a 14x reticle CoWoS road map to enable larger AI packaging. But at the same time, we also noted that TSMC, maybe last month in Japan, you showed the substrate development for CoWoS to enable some of the glass technology. So I'm just wondering if you can give us more like technologies progress update on the different kind of technology for glass core or glass substrates or glass carrier. What's the progress at TSMC right now?

So Laura's second question is on advanced packaging. She notes, as we said, we road map to even larger than 14x reticle size with CoWoS. But she also wants to know the technology process in newer areas like glass substrates, glass cores, what is the progress and status?

We are -- let me say that. Today, the majority is still CoWoS, right? And we are developing that alternative to try to lower down the cost. And we also work with substrate vendor so that our customer can have their product be in the market. The progress, we're building a pilot line that I announced a few quarters ago. And it takes about another 1 year to be mature, so we can put it into the production with our customer. Okay.

Thank you. Let's move to this side of the room. Bank of America, Haas Liu.

And congrats on the great results. My first question is regarding your CapEx and sales. You gave pretty solid CapEx outlook for this year and also said the CapEx outlook in the next couple of years will continue to be pretty significant. And you also raised this year at 40% plus. So would you be able to provide your next couple of years' sales growth outlook, try to quantify it?

And relatedly, I think also on that topic is whether you can just try to break down which part of the demand you are seeing as the key driver for you to raise your CapEx and also for this year's demand? Is it still mostly driven by cloud computing or it is proliferating to edge computing? Or to some extent, is it also related to your equipment supply chain is raising their price as well?

Okay. That's several questions in one. So I'm going to take that as 1.5 questions at least. But basically, Haas is asking, with the CapEx increase and the revenue increase this year, I think he's trying to look at intensity, but he wants to know what about the revenue guidance for the next several years. Yes, I'll stop there for now.

Okay. Let me answer that question. Because of the revenue corresponding to our investment, right, because we know we forecast our demand, and then we make an assessment, and then we do the CapEx. Next few years is going to be a very good business for TSMC. That's all I can say.

And then the other part, so what's the key driver? Is it cloud AI? Is it edge AI? Is it cloud vendors are increasing the price?

It's all AI related, everything.

Okay. Yes.

Do you have a quick follow-up?

Yes. I think it is more on your long-term strategy because a lot of people have actually been asking about your CapEx and also competition on the front-end. But I would actually say that if on the back-end competition is rising, especially coming from Intel EMIB-T, are you worried that your value add for your overall foundry business across front-end manufacturing to the back-end packaging business, the value add could actually be cannibalized with growing competition?

Okay. So Haas's second question is around the competition in advanced packaging. If our competitors are able to gain traction or business with things like EMIB-T, would that be the gateway or an entry way into more competitive threat on the front-end logic wafer side? So does advanced packaging lead to front-end wafer?

Haas, let me answer that. The front-end's wafer business and the back-end's business are 2 different things, right? If they are the same, then you can expect ASE become the front-end competitor also. It's 2 different things. And I also say that since our capacity in the back-end is so in shortage mode, the gap is bigger. And so, I welcome that the competitor offers some of the flexibility to my customer so that their front-end wafer can be put into the package, and that help TSMC's front-end wafer business. That's our attitude.

Operator, let's take one more from the online, and then we'll wrap up with back in person.

Next one to ask question, Robert Sanders from Deutsche Bank.

You recently stated that High-NA tools are too expensive. But could you please discuss how your customers are considering the impact of die stitching challenges from a smaller field size with High-NA. Could that actually slow the adoption of High-NA even if the tech improves or the tech gets more productive? And I have a follow-up.

Well, Rob's first question is a very specific technology around High-NA adoption. He wants to know the customers' feedback on the challenges with die stitching. Is this an impediment or barrier to High-NA adoption in our view?

You got a very detailed understanding of the High-NA. Today, the fuel size is only one half. And we put that one into our consideration of the manufacturing cost and something. Again, let me answer this question quickly. We are -- whether we use a High-NA or not, actually, one High-NA is a very good tool. Let's assume that, right? We understand it's a very high performance. But then, TSMC make it clear that we work with ASML and try to make it more suitable for manufacturing in terms of the cost and in terms of the maturity. So we always consider that technology maturity and the cost, and whether we use it or not, okay?

Okay. Thank you, C.C. Do you have a second question, Robert?

Yes. Just a quick follow-up. I think all of us on this call are assuming that the unconstrained demand for 3-nanometer and below is sort of 30% to 50% above your ability to supply. Is it, in fact, much larger than 30% to 50% above? It feels like it might be based on what you're saying because I think all of us are assuming it sort of solvable over the next 3, 4 years, but it sounds like the number could be much larger.

Well, those are your numbers, but Robert is asking the demand in excess of supply. Is it 30% to 50%? Is it something even larger? Do we have a number to share?

No, we don't have a number to share because of -- let me say that, the gap is very big. Sorry, I don't want to make a comment on the memory, but a very big gap.

Okay. We have about 9 minutes left. We'll come back to the floor with any questions. Let's take one from here. Evelyn Yu from Goldman.

Because we mentioned a lot on that we're going to step up our capacity growth. I'm just trying to quantify here because I noticed that during your symposium that you actually mentioned about 2-nanometer family capacity growth will be growing at around 70% CAGR from '26 to '28 and N3 plus and N5 to grow by 25% CAGR from '22 to '27. So I was just wondering, are those numbers still right assumptions today? Are we seeing actually any changes over the past quarter? And how should we compare with the non-supporting demand out there?

Okay. So Evelyn's first question is around capacity growth. She knows during the symposium, we did share some 5-year CAGR growth numbers for 2-nanometer around 70% CAGR and then 3-nanometer around, I guess, I can't remember the exact, but 25%. So are those numbers still the same? Or has it changed now that our CapEx and stuff?

Did we say that in Technology Symposium? We showed the chart. Okay. Now it's bigger. That's what I say.

Do you have a second question?

Okay. Very good direction. All right. My another question touched based on the advanced packaging side. Because you always bundle the advanced packaging CapEx together with testing, mask making, and others, that's around 10% to 20% of total CapEx.

And so one thing I'm trying to figure out here is that how much of that actually goes to advanced packaging alone? And because given that advanced packaging is capital intensive, less capital intensive versus front-end, so how should we think about the gap between its pricing revenue share and its CapEx share over the next few years?

And well, I think finally is that, as it becomes more important, how should we think of -- maybe you should consider breaking it out as a separate CapEx item going forward?

Okay. So Evelyn's question is around advanced packaging. She wants to know when we guide for the CapEx, of course, we guided in a bucket of packaging, testing, mask making and others together. Why do we not separate out just into packaging specifically? Her suggestion is, we should. But I think more -- so that's part of it, number one, the CapEx breakdown.

Evelyn, let me say that. We try very hard to make sure that our CapEx number is correct, but with the flexibility between the front-end and the back-end. Sometimes we have a bottleneck. So we put more money to buy the bottleneck tools. And sometimes it is in the front-end, sometimes it's in the back-end. But in the ballpark, the percentage is just like a Wendell share with everybody.

For long term, I mean, that's the back end, is about...

10% to 20%.

10% to 20%, that's a big range anyway. So that we -- what I can say is that's still 10% to 20%. Because of -- as I said, actually, I'm very honest to tell you that as time goes by, some of the customers' product need more tester. You cannot believe that. I mean, so the tester in shortage, so we have to put more CapEx in the tester or in the packaging or in other areas. So that's why we cannot very specifically say which area we put how much of the CapEx.

That's too specific. Yes. Okay. With the last participant, KGI, Felix Pan, thank you for being patient.

So my first question is regarding to the CapEx revision. So from year-to-date, so TSMC raised the CapEx guidance by almost USD 10 billion. So can you give me some color where is the upside from? How you guys see the difference from 6 months ago? Is that from like CPU accelerator or memory components or back-end CoWoS expansion? Just the upside, how we see things differently from 6 months ago?

Okay. So Felix is noting in January, we guided for $52 billion to $56 billion. In April, we said closer to $56 billion, now $60 billion to $64 billion. So we have increased the CapEx guidance. What is driving this? Is it Agentic AI only? Is it packaging? Is it AI accelerator?

Well, simply put, the most important reason is because of the demand continued to increase, and we feel the pressure from the customer to drive TSMC, not drive actually, to cooperate with TSMC for the capacity increase. That's one of the major reasons. The second reason is inflation. Now we buy the tools with inflation price, okay? You understand what I say.

Okay. So my second question is about the mature node. So people always focus on AI leading nodes, but it seems like mature nodes also seeing a very strong demand recovery and also some supply issue as well. So how do you guys see the demand-supply dynamic and pricing for the mature node? Because apparently, there's some impact from the AI crowding out effect. But mature nodes still largely depends on the consumer demand. So customer demand is still weak. So how do you guys see the demand supply dynamic for mature nodes?

Thank you. So Felix's second question is on mature node. He notes there's lots of talk that mature nodes are seeing a strong demand recovery and the supply is very tight. So mature node pricing is very favorable or strong. So he wants to know how do we see the mature node supply-demand situation?

Actually, the mature node cover a lot of different segments. Only the one which related to AI is in shortage, which is the most important one, is the #1 is power management IC, because all the AI data centers need a lot of power management. And those are the mature node technology like 0.18 micron, 90-nanometer or something like that. Those are in shortage definitely.

And also the sensor portion because of -- you need a lot of sensor to detect the environmental information and put into the AI data center to analyze it. Other than that, other area, just like you pointed out, the consumer product is not in a high demand. And so other segment is not so strong demand. And as I pointed out in my statement, other area, no, it's not so much of, say, in a lot of shortage, not at all.

Okay. Thank you. Thank you, C.C. Thank you, Wendell. Thank you, everyone. This does conclude our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within 30 minutes from now. The transcript will become available 24 hours from now, and both are going to be available through our website, again at www.tsmc.com.

If some of you were not able to ask your question, please feel free to reach out to TSMC IR, and we will follow up with you. So thank you, everyone, for joining us today. We hope everyone continues to stay well.

Have a good summer, and we hope you'll join us again next quarter. Thank you, and have a good day.

Good afternoon, everyone, and welcome to TSMC's First Quarter 2026 Earnings Conference Call. This is Jeff Su, TSMC's Director of Investor Relations and your host for today. 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. [Operator Instructions].

The format for today's event will be as follows: First, TSMC's Senior Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the first quarter 2026, followed by our guidance for the second quarter 2026. Afterwards, Mr. Huang and TSMC's Chairman and CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then we will open the line for the Q&A session.

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. So please refer to the safe harbor notice that appears in our press release.

And 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. Thank you for joining us today. My presentation will start with financial highlights for the first quarter 2026. After that, I will provide the guidance for the second quarter 2026. First quarter revenue increased 8.4% sequentially in NT supported by strong demand for our leading-edge process technologies. In U.S. dollar terms, revenue increased 6.4% sequentially to USD 35.9 billion, slightly ahead of our first quarter guidance. Gross margin increased 3.9 percentage points sequentially to 66.2%, primarily due to cost improvement efforts, a high capacity utilization rate and a more favorable foreign exchange rate. Operating margin improved 4.1 percentage points sequentially to 58.1% due to operating leverage. Overall, our first quarter EPS was TWD 22.08 and ROE was 40.5%.

Now let's move on to revenue by technology. 3-nanometer process technology contributed 25% of wafer revenue in the first quarter, while 5-nanometer and 7-nanometer accounted for 36% and 13%, respectively. Advanced technologies, defined as 7-nanometer and below, accounted for 74% of wafer revenue.

Moving on to revenue contribution by platform. HPC increased 20% quarter-over-quarter to account for 61% of our first quarter revenue. Smartphone decreased 11% to account for 26%. IoT increased 12% to account for 6%. Automotive decreased 7% and accounted for 4%, and DCE increased 28% to account for 1%.

Moving on to the balance sheet. We ended the first quarter with cash and marketable securities of TWD 3.4 trillion or USD 106 billion. On the liability side, current liabilities increased by TWD 256 billion quarter-over-quarter, mainly due to the increase of TWD 129 billion in accrued liabilities and others and the increase of TWD 82 billion in accounts payable. On financial ratios, accounts receivable turnover days was flat at 26 days. Days of inventory increased 6 days to 80 days, reflecting the ramp-up of our 2-nanometer technology and strong demand for our 3-nanometer technology.

Regarding cash flow and CapEx. During the first quarter, we generated about TWD 699 billion in cash from operations, spent TWD 351 billion in CapEx and distributed TWD 130 billion for second quarter 2025 cash dividend. Overall, our cash balance increased TWD 268 billion to TWD 3 trillion at the end of the quarter. In U.S. dollar terms, our first quarter capital expenditures totaled USD 11.1 billion.

I have finished my financial summary. Now let's turn to our current quarter guidance. Based on the current business outlook, we expect our second quarter revenue to be between USD 39.0 billion and USD 40.2 billion, which represents a 10% sequential increase or a 32% year-over-year increase at the midpoint. Based on the exchange rate assumption of USD 1 to TWD 31.7, gross margin is expected to be between 65.5% and 67.5%, operating margin between 56.5% and 58.5% Also, in the second quarter, we will need to accrue the tax on the undistributed retained earnings. As a result, our second quarter tax rate will be around 20%. We continue to expect the full year tax rate to be between 17% and 18%.

This concludes my financial presentation. Now let me turn to our key messages. I will start by talking about our first quarter 2026 and second quarter 2026 profitability. Compared to fourth quarter, our first quarter gross margin increased by 390 basis points sequentially to 66.2%, primarily due to cost improvement efforts, a higher overall capacity utilization rate and a more favorable foreign exchange rate. Compared to our first quarter guidance, our actual gross margin exceeded the high end of the range provided 3 months ago by 120 basis points, mainly due to a higher-than-expected overall capacity utilization rate and better cost improvement efforts. We have just guided our second quarter gross margin to increase by 30 basis points to 66.5% at the midpoint, primarily driven by a higher overall utilization rate and continued cost improvement efforts, including productivity gains, partially offset by dilution from our overseas fab.

Looking ahead to the second half of the year, given the 6 factors that determine our profitability, there are a few puts and takes I would like to share. As we have said before, the initial ramp-up of our 2-nanometer technology will start to dilute our gross margin in the second half of this year, and we expect between 2% and 3% dilution for the full year of 2026. Furthermore, as the scale of our overseas expansion grows, we continue to forecast the gross margin dilution from the ramp-up of overseas fabs in the next several years to be 2% to 3% in the early stages and widen to 3% to 4% in the latter stages.

In addition, given the recent situation in the Middle East, prices for certain chemicals and gases are likely to increase. Based on our current assessment, there may be impact to our profitability, but it is too early to quantify the impact. On the other hand, we will continue to leverage our manufacturing excellence to generate more wafer output and drive greater cross node capacity optimization in our fab operations to support our profitability. Also, N3 gross margin is expected to cross over to the corporate average in second half 2026. Finally, we have no control over the foreign exchange rate, but that may be another factor.

Next, let me talk about the materials and energy supply update given the recent situation in the Middle East. TSMC operates a well-established enterprise risk management system to identify and assess all relevant risks and proactively implement risk mitigation strategies. In terms of material supply, TSMC's strategy is to continuously develop multi-source supply solutions to build a well-diversified global supplier base and to improve the local supply chain. For specialty chemicals and gases, including helium and hydrogen, we source from multiple suppliers in different regions, and we have prepared safety stock inventory on hand. We are also working closely with our suppliers to further strengthen the resiliency and sustainability of our supply chain. Thus, we do not expect any near-term impact on our operations for material supply.

In terms of energy, TSMC worked closely with Taipower and the Taiwan government to ensure a stable and sufficient energy supply. With the recent situation in the Middle East, the Taiwan government has announced it has secured sufficient LNG supply through at least May. The government has also said it is actively working on securing further LNG supply, diversifying sourcing to other regions and other power backup plants. Therefore, we do not expect any near-term disruption or impact to our operations.

Finally, let me talk about our 2026 capital budget. At TSMC, higher level of capital expenditures is always correlated with higher growth opportunities in the following years. With our strong technology leadership and differentiation, we are well positioned to capture the multiyear structure demand from the industry megatrends of 5G, AI and HPC. We now expect our 2026 capital budget to be towards the high end of our range of between USD 52 billion and USD 56 billion as we continue to invest heavily to support our customers' growth. Even as we invest for the future growth with this level of CapEx spending in 2026, we remain committed to delivering profitable growth to our shareholders. We also remain committed to a sustainable and steadily increasing cash dividend per share on both annual and quarterly basis.

Now let me turn the microphone over to C.C.

Thank you, Wendell. Good afternoon, everyone. First, let me start with our near-term demand outlook. We concluded our first quarter with revenue of USD 35.9 billion, slightly above our guidance in U.S. dollar terms, driven by strong demand for our leading -edge process technologies. Moving into second quarter 2026, we expect our business to be supported by continued strong demand for our leading-edge process technologies. Looking ahead, we are very mindful of the impact of rising component prices, especially in consumer and price-sensitive end market segment. In addition, the recent situation in the Middle East also brings further macroeconomic uncertainties. As such, we are being prudent in our business planning while focusing on the fundamentals of our business to further strengthen our competitive position.

Having said that, AI-related demand continues to be extremely robust. The shift from generative AI and the query mode to agentic AI and command and action mode is leading to another step-up in the amount of tokens being consumed. This is driving the need for more and more computation, which supports the robust demand for leading-edge silicon. Our customers and customers' customers, who are mainly the cloud service providers, continue to provide us with their very strong signal and positive outlook. Thus, our conviction in the multiyear AI megatrend remains high, and we believe the demand for semiconductors will continue to be very fundamental. Supported by our robust technology differentiation and broad customer base, we maintain strong confidence for our full year 2026 revenue to now grow by above 30% in U.S. dollar terms.

Next, let me talk about our N2 capacity expansion plan. Our practice is to prioritize the land in Taiwan to support the fast ramp of our newest node due to the need for tight integration with R&D operations. Today, our new node, N2, has already entered high-volume manufacturing in the fourth quarter of 2025 with good yield. N2 is ramping successfully in multi phases at both Hsinchu and Kaohsiung site, supported by strong demand from both smartphone and HPC AI applications. With our strategy of continuous enhancement such as N2P and A16, we expect our N2 family to be another large and long-lasting node for TSMC.

Now let me talk about TSMC's global N3 capacity expansion plan. Historically, we do not add additional capacity to a node once it has reached its target capacity. However, as a foundry, our first responsibility is to provide our customers with the most advanced technologies and necessary capacity to unleash their innovations. Based on our assessment, to meet the strong demand in AI application, we are stepping up our CapEx investment to increase our N3 capacity. Thus, we are now executing a global capacity plan to support the robust multiyear pipeline of demand for 3-nanometer technologies, which are used by smartphone, HPC AI, including HBM-based dies, automotive and IoT customers.

In Taiwan, we are adding a new 3-nanometer fab to our GIGAFAB cluster in Tainan Science Park. Volume production is scheduled for the first half of 2027. In Arizona, our second fab will also utilize 3-nanometer technologies. Construction is already complete and volume production will begin in the second half of 2027. In Japan, we now plan to utilize 3-nanometer technology in our second fab and volume production is scheduled in 2028.

In addition to all the new fabs, we continue to convert 5-nanometer tool to support 3-nanometer capacity in Taiwan. We are also leveraging our manufacturing excellence to drive greater productivity across our fab in all locations to generate more wafer output. We are also focusing on capacity optimization across nodes, including flexible capacity support among N7, N5 and N3 nodes. Thus, we are using multiple levers to do everything we can, wherever we can, however we can to maximize the support to all our customers across all platforms. Also, let me emphasize that while the capacity is tight, we do not pick and choose or play favorites among our customers.

Next, let me talk about our mature node strategy. TSMC's strategy in mature node has not changed. Our focus is to build high-yield capacity for specialized technologies rather than just normal capacity. For example, we are increasing our mature node capacity such as in JASM Fab 1 in Japan for CMOS image sensor application and ESMC in Germany for automotive and industrial applications. Meanwhile, we have a plan to wind down our Fab 2, which is 6-inch fab; and Fab 5, which is 8-inch fab; focus on gallium nitride and use available space to optimize the support for leading-edge applications. Even with our Fab 2 and Fab 5, we still have enough capacity to fully support our existing customers. In summary, our strategy will be to continue to optimize our capacity mix within mature nodes and focus on the higher value-added and strategic segment, while ensuring we have a necessary capacity to support our customers' growth.

Finally, let me talk about our A14 status. Featuring our second-generation nanosheet transistor structure, A14 will deliver another full-node stride for N2 with performance and power benefit to address the sensible need for high-performance and energy-efficient computing. Compared with N2, A14 will provide 10 to 15 speed improvement at the same power or 25 to 30 power improvement at the same speed and close to 20% chip density gain. Our A14 technology development is on track and progressing well. We are observing a high level of customer interest and engagement from both smartphone and HPC applications. Volume production is scheduled for 2028. Our A14 technology and its derivative will further extend our technology leadership position and enable TSMC to capture the growth opportunities well into the future.

This concludes our key message, and thank you for your attention.

Thank you, C.C. This concludes our prepared statements. [Operator Instructions]. Now let's begin the Q&A session. Operator, can we proceed with the first participant on the line, please? Thank you.

First one to ask questions, Haas Liu from Bank of America.

Congrats on the solid results and guidance. I would like to start with your 3-nanometer gross margin outlook. You just mentioned the node is going to cross the corporate average gross margin in second half this year, which is now at mid-60 percentage levels. And we understand the technology is in severe undersupply backed by strong AI demand, and you already forecasted the capacity expansion through conversion and greenfield through 2028. Would you be able to discuss more in detail on what kind of applications are driving such strong business for you and convince you to expand more?

And the other thing on 3-nanometer as well is just, the node started to ramp from fourth quarter 2022, which means some of your equipment will be fully depreciated by 2027. Should we expect the node margins to be trending even higher with very solid utilization and also pricing trend?

Okay. So the first question from Haas Liu of Bank of America, it's 2 parts on 3-nanometer. First, as C.C. described, we are executing a plan for expanding 3-nanometer capacity. So he wants to understand what are the applications to drive such a strong multiyear looking ahead pipeline of demand for 3-nanometer since it's already been around in volume production since late '22. That's the first part of his question.

Let me answer that. I think the application is simple. It's still the HPC AI applications. Does that answer your question?

Okay. Yes. That is the first part. And the second part is...

And the second part of this question is on the gross margin for 3-nanometer. His question is really, what is the gross margin outlook for 3-nanometer? Will it crossover in the second half of this year? To what level? And then once it becomes fully depreciated, what happens to the margin?

Okay. This is Wendell. We expect the N3 gross margin to reach and cross the corporate gross margin level in the second half of this year. And we don't have a number to share with you, but after the full depreciation as our previous notes, the gross margins are generally very high.

Okay. Haas, I'll take that as a 1.5 questions. So if you have a quick follow-up for your second question?

Yes. And the other, I think, just a 0.5 follow-up is probably just on the CapEx. You revised up to the high end of your guidance for USD 52 billion to USD 56 billion for this year. Compared to 3 months ago, what gives you the incremental confidence when you discuss with your customers and also customers' customers regarding the demand outlook to support your stronger or the upper half of your guidance for the CapEx this year?

Okay. Thank you, Haas. So his second question is he notes that indeed, we have this time guided to the high end of our CapEx range versus January. So what incrementally is driving this revision to the CapEx? What gives us the confidence to go to the high end of the USD 52 billion to USD 56 billion range?

Well, again, this is C.C. Wei. Let me answer this question. A very simple answer is, the demand are very robust, especially from the HPC and AI applications. And also, we try very hard to speed it up and pull in all the equipment as we can. Still, our supply is very tight. Demand is continuing to increase. And so we continue to work with our suppliers to speed it up. And that's why we are towards our high end of CapEx forecast.

Okay, Haas, does that answer your question?

Yes.

Next one to ask question, Gokul Hariharan, JPMorgan.

My first question on your comments on demand. Clearly, demand is even better than what you predicted back in January, C.C., and you also raised the CapEx. Now all your customers seem to be telling everybody they can tell that wafers still remain the biggest constraint. So given your expanded 3-nanometer capacity plan and faster CapEx, C.C., what is your expectation that how long this supply constraint is likely to last? Do you have any visibility of when you can kind of bring some kind of balance here based on what you hear from customers? And as a strategy, do you also plan to build out a more clean room space, because that seems to be a little bit of a constraint right now to bring on the capacity quickly. That's my first question.

Okay. Gokul, please allow me to summarize your first question. So his question is directed for C.C. He notes that the demand seems to be even stronger than our forecast in January. We have also raised the CapEx, and customers continue to say they need more chip supply. So with our capacity plan, do we have a forecast or expectation of how long the constraint can last? And will we have a strategy to build up clean room space first? Is that correct, Gokul?

That's right. Yes.

Okay. Gokul, let me answer the question. Again, it's very simple, because demand continues to be robust and the number continues to be increased, and we double check with our customers, customers' customers, or those CSPs. They gave us a very positive outlook, right? And so we have to speed it up with our buildup of clean room and buying the tools. And so we are working with construction and we are working with our equipment suppliers. And so we want the pulling forward of our forecast schedule. That's a simple answer. Because of AI, it's so strong.

Any read, C.C., on when we can kind of meet this demand? Or do you think the next couple of years is still going to be very challenging to meet -- that supply is still going to be running below demand, let's say, into '27 also?

So Gokul would like to know when the supply can meet the demand? Do we have a forecast or a time frame?

Gokul, you know we are -- it takes 2 to 3 years to build a new fab. And with the current schedule, we believe that '27, we will announce it anyway when we enter '27, but let me say that, it takes time to build a new fab, it takes time to ramp it up. And so we expect this to continue to be very tight. So that's why we just announced that we try to build 3 new N3 fab to meet the demand.

Okay. That's very clear. So '27 is also very tight. My second question on competition. So obviously, you have the traditional competitors, Samsung, Intel. But one of your customers, Elon Musk, also announced Terafab Initiative recently. What is TSMC's perspective on this initiative? They have also been a customer of yours, and they recently signed a deal with Samsung a few months back. So what is TSMC's response here now that they are also trying to kind of build chips on their own? How are you trying to win back this customer? Like, C.C., what is your perspective here?

Okay. So Gokul's second question is on competition. He notes that we have competition and then recently, a competitor, or he notes that this Terafab. So he wants to know what is our perspective on this initiative. This customer has also been a customer of TSMC, but has also signed a deal with one of our other competitors, Samsung. So Gokul would also like to know what is our perspective on the Terafab? And what is our view on winning back this customer's business?

Well, Gokul, actually, both Intel and Tesla, they are TSMC's customers. But again, they are our competitors, and we view Intel as our formidable competitor and do not underestimate them. But having said that, there are no shortcuts. The fundamental rules of the foundry game never change. They need the technology leadership, manufacturing excellence and customer trust, and most of all, the service, which has been mentioned by Jensen; thank you for his wording.

Again, let me say that it takes 2 to 3 years to build a new fab, no shortcuts. And it takes another 1 to 2 years to ramp it up. Again, that's the fundamental of foundry industry. And whether we try to win them back, actually, they are still our customer. And we are very confident in our technology position, and we work very hard to capture every piece of business possible. Gokul, did I answer your question?

Okay. That is very clear. So do you think your faster ramp-up of capacity can kind of win some of these customers back, because the reason seems to be mostly about capacity tightness rather than any other kind of big reasons, right? So is that your evaluation that this is probably the most important thing to win some of these customers back?

Okay. So Gokul's final question is then in winning customers back, his concern is because our capacity is tight. Is that the reason we are losing customers? And so can we win customers back?

Well, again, let me emphasize, it takes 2 to 3 years to build a new fab. So in this time, we are also building a new fab to meet our customers' strong demand. No shortcuts. So anyway, the capacity is very tight, as I said, but we are working hard to make sure that we can meet customers' demand.

Next one, we have Charlie Chan from Morgan Stanley.

Congratulations for very, very strong results again. So I think I would also address the competition topic from a little bit different angle. So as you can see that those AI customers, they are developing a much larger reticle size chips, right? And some customers are considering to use eMIPs because it's a kind of substrate base, more suitable for circular larger size of chip design. So I'm not sure what's the TSMC's strategy to address this competition. And more strategically, is TSMC comfortable to open up your compute die to your competitors, for example, Intel to do the package? What's the kind of thought process behind?

All right, Charlie, thank you. So Charlie's first question is also related to competition. He notes that AI customers are seeking for larger and larger reticle sizes. So he wants to know what is our assessment of the competitive threat from solutions such as like eMIP? And what's our strategy to address this competition? Will we be willing to open up our front-end wafer and let someone else do the packaging basically?

Well, Charlie, today, TSMC is supplying the largest reticle size packaging. And yes, we understand that our competitors also offer very attractive technology, but we welcome that so our customers can have more choices and then we can do more business with our customers. That's our attitude. But saying that, we don't leave any business on the table. We are working very hard to meet all our customers' demand. We also are developing very large reticle size packaging technologies. We are working with all the customers. And so far, so good.

C.C., I have a follow-up on this. When you mentioned about larger size packaging technology, are you referring to CoPoS or CoWoS-L 3.5D? Or do you think 3D stacking can resolve this kind of panel expansion problem?

So Charlie is asking a follow-up. So he wants us to comment on, for larger reticle size, is it CoWoS-L, is it panel level? What exact detailed solutions are we doing?

Charlie, so far today, we have very large reticle sized CoWoS. Of course, we are also working on CoPoS. And together, we try to make sure that we give enough capacity to support our customer with a reasonable cost. So that's why we build a CoPoS pilot line right now and expect production a couple of years later. But today, the main approach or the main supplier is still a large-sized CoWoS. And together with System on Wafer technology, we think TSMC gives our customers the best options for their product in the market.

Got it. So yes, I would take, we don't need to worry too much about this [indiscernible] competition.

So my second question is actually about your long-term CapEx plan. C.C., as you said that it takes 2 to 3 years to build a new fab. So you definitely have that visibility, right? So I remember back in 2021, management also provided 3-year CapEx guidance as USD 100 billion given very strong demand. I'm not sure if TSMC can provide a little bit longer-term CapEx guidance? Because as you said, right, the equipment supply is also pretty tight. Yesterday, ASML reported very, very strong results. So you said the EUV supply is an issue. And secondly, would the management provide kind of a long-term CapEx guidance to investors?

All right, Charlie, that's a lot of questions. But the second one then on CapEx and building capacity. Again, Charlie notes C.C.'s comment, capacity is not born overnight, it takes time. So he would like to know, besides this year's CapEx, which we have already said at the high end, can we provide a guidance for the next 3 years CapEx like we did back in 2021 in terms of the dollar amount?

Okay. Charlie, we don't have a number to share with you. But look at it this way. In the past 3 years, our total CapEx was USD 101 billion. This year, we're already saying CapEx is towards the high end, which is USD 56 billion, which is already over 50% of the past 3 years in total. So we have a strong conviction in the AI megatrend. So we expect the CapEx in the next few years, in the next 3 years, will be significantly higher than the past 3 years.

And then the final part of Charlie's question, with such a long lead time, are we concerned about securing tools or bottlenecks and such?

Well, Charlie, in TSMC's culture, we're always working with our suppliers, because we view them as our partners. So we continue to work with them, especially for those ASML, Applied Materials, Lam Research, et cetera. So, so far, we are very happy with their supportive. That's all I can tell you.

Next one, we have Sunny Lin from UBS.

Congrats on the steady results. So my first question is, again, to follow up on CapEx. So if you look at from 2024 to 2026, so in this, call it, AI cycle, TSMC has been able to keep capital intensity at a healthy level of 30% plus, given very strong technology leadership and operating leverage. I understand the company doesn't really have a specific target on capital intensity. But for the coming few years, given the very strong revenue ramp of leading edge, how should we think about the revenue growth compared with CapEx growth? Should we think top line will remain steady and therefore, CapEx could grow in line or even below? What's the best way for us to think about it?

Okay. Sunny, thank you for your question. So please allow me to summarize. Sunny's first question is on, well, I think CapEx and really capital intensity. She notes, in the past few years, we've been able to keep capital intensity around the 30-something percent level. She notes that we don't have a specific capital intensity target per se, but her specific question, looking ahead the next several years, how do we see revenue growth versus CapEx growth? Is it likely to be higher, flat, lower? And therefore, what type of intensity does that imply? Is that correct, Sunny?

Yes. Thank you very much, Jeff.

Okay, Sunny. So in the past few years, as you correctly pointed out, the revenue growth outpaced the CapEx growth. That's because if we do our job right, then we will continue to see that happen in the next several years. The revenue growth outpaced the CapEx growth, okay? Now therefore, we do not expect, in the next several years, a sudden surge in capital intensity.

I see. Maybe a very quick follow-up. A lot of questions on competitions already. But also from a competition point of view, given a very tight supply at TSMC's side in recent years, would TSMC actually consider maybe spending CapEx a bit more, so that clients won't need to diversify given the tight supply?

All right. So Sunny's 1.5 question is, in terms of the CapEx, will we consider accelerating or spending more given the competitive threat from the competitors? If there's not enough capacity, then our customers will go to competitors. That's your question, correct?

Yes. Thank you, Jeff.

Well, Sunny, we're repeatedly saying that we prepare the capacity to meet customers' demand, not because of our competitor or not because of other considerations. The most important one is our customers' demand and they work with TSMC and so we plan our capacity and so our capital expense. Sunny, did I answer your question?

Yes. Yes, very clear. So maybe my 0.5 question. And so if we look at this year, earlier, you just guided a bit higher than 30% growth for top line. But indeed, there's ongoing supply tightness. And so for 2026, how much upside could you realize for top line? And at this point, have you started to see some impact of consumer end demand and therefore, on your demand coming from smartphone and PC?

Okay. So Sunny's second question is regarding 2026 full year outlook. She notes now that we have increased the guidance to above 30%, how much more upside can there be? Well, maybe the first part also, how do we see the impact from the memory price hike to the end market? And how do we see, with above 30%, is there more upside?

Well, Sunny, memory price hike definitely has some impact to price sensitive end market, especially in PC and smartphone market. We did see a little bit softer market. But to share with you, all the high-end smartphones continue to do better, and this is to TSMC's advantage. And as you're asking about how much higher than above 30% year-over-year growth, we will share with you in July, how about that, that we will have a more accurate or a more precise number to share with everybody.

No problem.

Next one, Jim Fontanelli at Arete.

So my first question is to do with demand. So you commented earlier in the call that demand continues to outstrip supply for leading edge capacity. And obviously, you just delivered a very strong print and guide for gross margins. So against this backdrop, has management's thinking changed about the sustainable margin structure and what appropriate longer-term returns might be for the business?

Okay. So Jim's first question is asking on the margin structure. He notes, as we said that demand continues to be extremely robust and very strong. So how does this change? I think your question is our view on the long-term margin profile and the return profile. Is that correct?

That's correct.

Okay, Jim. As we said in the last earnings calls, we've revised up our long-term margin target and ROE target. From 2024 to 2029, we're now saying the gross margins will be 56% and higher through the cycle, and we're looking at ROE of high 20% through the cycle. That's what we're currently looking at, and that's already higher than before.

And that thinking is not changing against the backdrop where other parts of the AI supply chain are clearly starting to print super normal returns? That doesn't impact how you think about margin structure for the next 2 or 3 years?

Yes, Jim, the long-term planning is an ongoing and continuous process. So we do that all the time, and we will update you when there is a change.

Okay. And my second question is, it looks like the Arizona site is becoming more strategic in terms of leading edge commitment for TSMC, particularly with the recently added second parcel of land. Could you talk about how you see mid- to long-term capacity opportunity and also how confident you are that the U.S. fab economics will match Taiwanese produced wafers?

Okay. So Jim's second question is on our Arizona fab expansion plans. He notes that it is becoming more and more strategic. We have recently, as we said, acquired a second large piece of land. So what is the plan or the purpose behind this? And then what is the profitability or margin outlook as well?

Well, Jim, let me answer the question. We acquired the second land because we need it. We want to build more fabs in Arizona. And this is actually to meet the multiyear demand from our leading edge U.S. customers. And again, let me emphasize again that we are working very hard to speed it up. We already gained a lot of experience in Arizona. And so now we are much more confident than last year that we can make it a good progress and moving aggressively forward. And we expect we can improve the cost structure, of course.

Next one, Bruce Lu from Goldman Sachs.

I think I want to follow up on Jim's question for the profitability. I think earlier last year, when I asked why TSMC did not raise the profitable target when TSMC continued to sell the value. I think C.C. told me that to focus on the above version of 53% and above. I think last quarter, we raised it to 56% and above. So the question is that do you believe the current profitability fully reflects TSMC's value? So I'm guessing C.C. might ask me to focus on the higher portion of the profitability target again. So the real question is that given the uniqueness of the dominant position for TSMC, it's not easy to find a perfect benchmark for TSMC's profitability. So can you tell us how we should think the profitability benchmark for TSMC? Or what is the best way to see TSMC value to be fully reflected into the gross margin and operating margins?

Okay. Bruce's first question is, he wants to know what profitability benchmark he should be looking at, and whether we believe our current profitability level fully reflects TSMC's true value.

Well, Bruce, actually, you asked about our pricing strategy. Let me say that we always view our customers as our partners. Of course, we know our value; of course, we know our position, but we also view our partners as very important business partners, so that we don't change our pricing dramatically or something like that. We just try to make sure that our customers can be successful in their market. And at the same time, we grow together, and we also earn our value, so that we can continue to expand our capacity to support them. That fundamentally is, number one, our customer got to be successful. That's our consideration, number one, and we grow together. And again, there's a keyword please pay attention to. Customer is our partner.

Okay. So if your customers continue to be successful, maybe in a couple of quarters, we can see the higher profitability target again.

Bruce, what's your second question?

Okay. My second question is that management has been guiding that AI accelerator revenue to grow about like mid- to high 50s CAGR in (sic) between 2024 and '29. So how does TSMC plan and forecast AI-related demand? I mean, does TSMC incorporate metrics such as total consumption growth in your assumption? Because the recent consumption in the first quarter is definitely accelerated and faster than earlier expectation. Do we see the changes for the AI accelerated revenue growth in the coming years?

Okay. So Bruce's second question is on our AI accelerator long-term CAGR guidance, which, yes, we have guided mid- to high 50s. He notes with the strong token growth and demand for tokens, do we have any changes to this long-term guidance?

Bruce, actually, I think I say now that it's a very strong demand, and we continue to receive a very positive signal from our customers and customers' customers. And so what you say is whether we change our CAGR on AI accelerator? Actually, we continue to see strong demand, but again, let me say that it is toward higher 50s of CAGR that we observe.

Next one to ask question, Laura Chen from Citi.

May I take more details on TSMC's strategy in advanced packaging? And what will be the business model working with your OSAT partners, as we see that there are various different solutions provided by your peers and also the OSAT makers, yet TSMC is also expanding more in the advanced packaging. So how would TSMC work with your customers' planning on their advanced node wafer demand, but also align with their advanced packaging demand at TSMC?

Okay. So thank you. Laura's first question is on advanced packaging. She would like to know, we work with customers, collaborate with customers to plan our front-end wafer capacity. How do we work with the customers to plan the advanced packaging capacity is what she would like to understand, and also in the context of working with our OSAT partners on the advanced packaging businesses.

Well, Laura, our priority actually, again, is to support our customers, right? And whenever we can or wherever we can, we want to make sure that their product can be -- the demand of their product can be met by TSMC's front-end and high-end packaging. So we certainly, let me say that our advanced packaging capacity is very tight also. So we have to work with our OSAT partners. We hope that we can increase the capacity to support our customers. Let me emphasize again, we support our customers. So we try very hard to increase our own capacity also. But certainly, it just has been very tight. And so that's what's our situation today.

Sure, sure. Understood. My second question is also about advanced packaging. As C.C. highlighted before many times that AI chips are going into super chips with very large die size and TSMC now working at the biggest reticle in the world. But at the same time, there's potential technical challenges such as warpage. So do you think that the following road map like SoIC or like CoPoS can solve this kind of technical issue? And based on TSMC's technology road map, do we see any technology like SoIC or CoPoS will be a bigger ramp in a couple of years, can solve this problem?

Okay. So Laura's second question is also related to advanced packaging, AI and larger reticle sizes post potential technical challenges such as warpage. So she would like to know how do we see SoIC or panel-level packaging? What's the key to solving these issues? And what is the outlook in the next several years?

Well, Laura, you are good. Actually, that's all the challenges that we have in advanced packaging technology. Mechanical stress, which is very top challenge to the electrical engineering, like I am. However, we accumulated a lot of experience already today because we have supplied most of the leading edge and in packaging area. And we continue to increase the die size and continue to meet all the challenges from the mechanical stress, like you said, actually the warpage or the thermal limitation. A good challenge. And we like it. The harder the better, because of TSMC's strength in technical engineering, and we have confidence that we can work with our customers to solve all the issues and continue to move on.

So should we expect that SoICs, TSMC may introduce that earlier to solve this kind of a challenge, because we already have the learning curve and already have the products in production. So that should go faster than other technologies, I suggest.

So Laura's question is very specific. Yes, on SoIC, how do we see that developing, I guess?

Well, we work with our customers, and we meet their demand, and that's all I can tell you. Speed it up or slow down? No, no, no, no. We work with our customers to meet their demand.

Operator, in the interest of time, can we take the questions from the last participant, please?

Next one to ask question, Charles Shi from Needham.

TSMC's definition of AI revenue includes data center GPU, AI accelerator, HBM-based stack. Maybe I left out a few others, but it specifically excludes data center CPU. I think you made that definition very clear for a couple of years now. But with the CPU, there's more and more conversation about CPU now becoming part of the AI infrastructure, especially for agentic workloads. Any chance for TSMC to maybe provide us revised numbers for AI revenue and maybe AI revenue growth, CAGR projection going into 2029, 2030. And maybe hopefully give us some sense of how the historical AI revenue numbers would have been if some of the data center CPU numbers, especially for agentic AI workloads are included there. That's my first question.

Okay. Thank you, Charles. So Charles' first question, please let me summarize, is regarding our definition of AI accelerator, which is, of course, we have said GPU, ASIC and HBM controllers for training inference in the data center. He notes now with the agentic AI, he wants to know, will we start to include CPUs in this definition? If so, can we provide the historical data with CPU included? And what would the AI accelerator guidance be if it includes CPU?

Charles, certainly, CPUs become more and more important in today's AI data center. But actually, let me share with you -- this is a good question, by the way. Let me share with you that we are not able to identify which CPU goes where, right? It's PC or desktop or it's AI data center. So today, we still not include the CPUs in our AI HPCs calculation. Someday later, we might consider.

Charles, do you have a second question?

Thanks, C.C. Yes. Maybe it's kind of also tied to the recent development in overall AI infrastructure, how things have been evolving. So NVIDIA, of course, they recently added more CPU content to the overall Vera Rubin SuperPOD, but I think that most people are focusing on that brand-new LPU. They recently added -- we understand and appreciate that the TSMC is very strong in CPU and we will definitely participate in that upside in CPU, but the LPU business, the acquired business, well, for historical reasons, it's still at your competitors, Samsung Foundry. And I think Investors are looking at that and seeing that maybe looks like Samsung Foundry finally made the first 2 inroads into AI. So any thoughts from TSMC side, how should we think about whether and how TSMC will win back that LPU business or any future difference chip business coming from your customers? Yes, give us some thoughts there, we would appreciate that.

Okay. Charles' second question is a very specific question about a very specific customer and very specific product, which is we typically do not comment on, but he wants to know for this customer's LPU product, which he notes is made at one of our competitors. How do we see this business going to the competitor? Do we have plans to win this LPU business back in the future?

Charles, I think Jeff already gave me enough warning, very specific and very specific customer, very specific area. Let me answer your question. We are working with our customer for their next-generation LPU anyway. And we are very confident in our technology position, and we will work hard to capture every piece of business possible. How about that?

Very good. Thank you, C.C. That's very good color.

Okay. Thank you, Charles. Thank you, C.C. Thank you, Wendell. This concludes our prepared statements -- sorry, I should say 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 30 minutes from now, and the transcript will become available 24 hours from now. Both are going to be available through TSMC's website at www.tsmc.com. So again, thank you, everyone, for taking the time to join us today. We hope you continue to stay well, and we hope you join us again next quarter. Goodbye, and have a good day.

Good afternoon, everyone, and welcome to TSMC's Fourth Quarter 2025 Earnings Conference and Conference Call. My name is Jeff Su, TSMC's Director of Investor Relations and your host for today.

Today's event is being webcast live through TSMC'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 Senior Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the fourth quarter 2025, followed by our guidance for the first quarter 2026. Afterwards, Mr. Huang and TSMC's Chairman and CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then we will open both the floor and the line for the question-and-answer session.

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.

And now I would like to turn the microphone over to TSMC's CFO, Mr. Wendell Huang, for the summary of operations and the current quarter guidance.

Thank you, Jeff. Good afternoon, everyone. Thank you for joining us today. My presentation will start with financial highlights for the fourth quarter of 2025 and a recap of full year 2025. After that, I will provide the guidance for the first quarter of 2026. Fourth quarter revenue increased 5.7% sequentially in NT, supported by strong demand for our leading-edge process technologies. In U.S. dollar terms, revenue increased 1.9% sequentially to TWD 33.7 billion, slightly ahead of our fourth quarter guidance.

Gross margin increased by 2.8 percentage points sequentially to 62.3%, primarily due to cost improvement efforts, favorable foreign exchange rate and high capacity utilization rate. The operating expenses accounted for 8.4% of net revenue compared to 8.9% in the third quarter of '25 due to operating leverage. Thus, operating margin increased sequentially by 3.4 percentage points to 54%. Overall, our fourth quarter EPS was TWD 19.5 and ROE was 38.8%.

Now let's move on to revenue by technology. 3-nanometer process technology contributed of 28% of wafer revenue in the fourth quarter, while 5-nanometer and 7-nanometer accounted for 35% and 14%, respectively.

Advanced technologies, defined as 7-nanometer and below, accounted for 77% of wafer revenue. On a full year basis, 3-nanometer revenue contribution came in at 24% of 2025 wafer revenue, 5-nanometer, 36% and 7-nanometer, 14%.

Advanced technologies accounted for 74% of total wafer revenue, up from 69% in 2024. Moving on to revenue contribution by platform. HPC increased 4% quarter-over-quarter to account for 55% of our fourth quarter revenue. Smartphone increased 11% to account for 32%. IoT increased 3% to account for 5%. Automotive decreased 1% to account for 5%, while DCE decreased 22% to account for 1%.

On a full year basis, HPC increased 48% year-over-year. Smartphone, IoT and automotive increased by 11%, 15% and 34%, respectively, in 2025, while DCE remains flat. Overall, HPC accounted for 58% of our 2025 revenue. Smartphone accounted for 29%. IoT accounted for 5%, automotive accounted for 5% and DCE accounted for 1%.

Moving on to the balance sheet. We ended the fourth quarter with cash and marketable securities of TWD 3.1 trillion or USD 98 billion. On the liability side, current liabilities increased by TWD 182 billion quarter-over-quarter, mainly due to the increase of TWD 95 billion in accrued liabilities and others and the increase of TWD 61 billion from the reclassification of bonds payable to current portion.

In terms of financial ratios, accounts receivable days increased by 1 day to 26 days. Inventory days remained steady at 74 days. Regarding cash flow and CapEx, during the fourth quarter, we generated about TWD 726 billion in cash from operations, spent TWD 357 billion in CapEx and distributed TWD 130 billion for first quarter '25 cash dividend.

Overall, our cash balance increased TWD 297 billion to TWD 2.8 trillion at the end of the quarter. In U.S. dollar terms, our fourth quarter capital expenditures totaled TWD 11.5 billion.

Now let's look at the recap of our performance in 2025. Thanks to the strong demand for our leading-edge process technologies, we continue to outperform the foundry industry in 2025. Our revenue increased 35.9% in U.S. dollar terms to TWD 122 billion or increased 31.6% in NT dollar terms to TWD 3.8 trillion. Gross margin increased 3.8 percentage points to 59.9%, mainly reflecting a higher capacity utilization rate and cost improvement efforts, partially offset by an unfavorable foreign exchange rate and margin dilution from our overseas fabs. With operating leverage, our operating margin increased 5.1 percentage points to 50.8%. Overall, full year EPS increased 46.4% to TWD 66.25 and ROE increased 5.1 percentage points to 35.4%.

In 2025, we generated TWD 2.3 trillion in operating cash flow, spent TWD 1.3 trillion or USD 40.9 billion on capital expenditures. As a result, free cash flow amounted to TWD 1 trillion, up 15.2% from 2024. Meanwhile, we paid TWD 467 billion in cash dividends in 2025, up 28.6% year-over-year as we continue to increase our cash dividend per share.

TSMC shareholders received a total of TWD 18 cash dividend per share in 2025, up from TWD 14 in 2024, and they will receive at least TWD 23 per share in 2026. I have finished my financial summary.

Now let's turn to our current quarter guidance. We expect our business to be supported by continued strong demand for our leading-edge process technologies. Based on the current business outlook, we expect our first quarter revenue to be between USD 34.6 billion and USD 35.8 billion, which represents a 4% sequential increase or a 38% year-over-year increase at the midpoint.

Based on the exchange rate assumption of USD 1 to TWD 31.6, gross margin is expected to be between 63% and 65%, operating margin between 54% and 56%. Lastly, our effective tax rate was 16% in 2025. For 2026, we expect our effective tax rate to be between 17% and 18%. This concludes my financial presentation.

Now let me turn to our key messages. I will start by talking about our fourth quarter '25 and first quarter '26 profitability. Compared to third quarter, our fourth quarter gross margin increased by 280 basis points sequentially to 62.3%, primarily due to cost improvement efforts, a more favorable foreign exchange rate and a higher overall capacity utilization rate.

Compared to our fourth quarter guidance, our actual gross margin exceeded the high end of the range provided 3 months ago by 130 basis points, mainly as we delivered better-than-expected cost improvement efforts.

In addition, the actual fourth quarter exchange rate was USD 1 to TWD 31.01 as compared to our guidance of USD 1 to TWD 30.6. We have just guided our first quarter gross margin to increase by 170 basis points to 64% at the midpoint, primarily driven by continued cost improvement efforts, including productivity gains and a higher overall capacity utilization rate, partially offset by continued dilution from our overseas fab.

Looking at full year 2026, given the 6 factors, there are a few puts and takes I would like to share. On the one hand, we expect our overall utilization rate to moderately increase in 2026. N3 gross margin is expected to cross over to the corporate average sometime in 2026, and we continue to work hard to earn our value. In addition, we are leveraging our manufacturing excellence to drive greater productivity in our fabs to generate more wafer output.

We are also increasing a cross-node capacity optimization, which includes flexible capacity support among N7, N5 and N3 nodes to support our profitability. On the other hand, as the scale of our overseas expansion grows, we continue to forecast the gross margin dilution from the ramp-up of overseas fabs in the next several years to be between 2% to 3% in the early stages and widen to 3% to 4% in the latter stages.

Furthermore, the initial ramp-up of our 2-nanometer technology will start to dilute our gross margin in the second half of the year, and we expect between 2 to 3 percentage -- percent dilution for the full year of 2026. Finally, we have no control over the foreign exchange rate, but that may be another factor in 2026.

Next, let me talk about our 2026 capital budget and depreciation. At TSMC, a higher level of capital expenditures is always correlated to the high-growth opportunities in the following years. With our strong technology leadership and differentiation, we are well positioned to capture the multiyear structural demand from the industry megatrends of 5G, AI and HPC.

In 2025, we spent USD 40.9 billion as compared to USD 29.8 billion in 2024 as we began to raise our level of capital spending in anticipation of the growth that will follow in the future years. In 2026, we expect our capital budget to be between USD 52 billion and USD 56 billion as we continue to invest to support our customers' growth. About 70% to 80% of the 2026 capital budget will be allocated to advanced process technologies.

About 10% will be spent for specialty technologies and about 10% to 20% will be spent for advanced packaging, testing, mask making and others. Our depreciation expense is expected to increase by high teens percentage year-over-year in 2026, mainly as we ramp our 2-nanometer technologies.

Even as we invest in the future growth with this level of CapEx spending in 2026, we remain committed to delivering profitable growth to our shareholders.

Finally, let me talk about TSMC's long-term profitability outlook. As a foundry, our biggest responsibility is to support our customers' growth, and we always view them as partners. Having said that, we are in a very capital-intensive business. In the last 5 years alone, our CapEx totaled USD 167 billion. Our R&D investments totaled USD 30 billion. Therefore, it is important for TSMC to earn a sustainable and healthy return as we continue to invest in leading -edge specialty and advanced packaging technologies to support our customers' growth. Today, we face increasing manufacturing cost challenges due to the rising cost of leading nodes.

For example, the cost of tools are becoming more expensive and process complexity is increasing. As a result, the CapEx dollar required to build 1,000 wafer per month capacity of N2 is substantially higher than 1,000 wafer per month capacity for N3. The CapEx per k cost for A14 will be even higher.

We also faced additional cost challenges from expansion of our global manufacturing footprint, new investments in specialty technologies and inflationary costs. These all lead to a higher level of CapEx spending. As a result, in the last 3 years, our CapEx dollars amount totaled USD 101 billion, but is expected to be significantly higher in the next 3 years. Having said that, we continue to work closely with our customers to plan our capacity while sticking to our disciplines to ensure a healthy overall capacity utilization rate through the cycle.

Our pricing will remain strategic, not opportunistic to earn our value. We will work diligently with our suppliers to drive greater cost improvements. We will also leverage our manufacturing excellence to generate more wafer output and drive greater a cross node capacity optimization in our fab operations to support our profitability. By taking such actions, we believe a long-term gross margins of 56% and higher through the cycle is achievable, and we can earn an ROE of high 20s percent through the cycle.

By earning a sustainable and healthy return, even as we shoulder a greater burden of CapEx investment for our customers, we can continue to invest in technology and capacity to support their growth while delivering long-term profitable growth to our shareholders. We also remain committed to a sustainable and steadily increasing cash dividends per share on both an annual and quarterly basis.

Now let me turn the microphone over to C.C.

Thank you, Wendell. Good afternoon, everybody. First, let me start with our 2026 outlook. In 2025, we observed robust AI-related demand throughout the whole year, while non-AI end market segment bottomed out and saw a mild recovery. Concluding 2025, the Foundry 2.0 industry, which we define as all logic wafer manufacturing, packaging, testing, mask making and others increased 16% year-over-year.

Supported by our strong technology differentiation and broad customer base, TSMC's revenue increased 35.9% year-over-year in U.S. dollar terms, outperforming the Foundry 2.0 industry growth. Entering 2026, we understand there are uncertainties and risk from the potential impact of tariff policies and rising component prices, especially in consumer-related and price-sensitive end market segment.

As such, we will be prudent in our business planning while focusing on the fundamentals of our business to further strengthen our competition position. We forecast the Foundry 2.0 industry to grow 14% year-over-year in 2026, supported by robust AI-related demand. Underpinned by strong demand for our leading-edge specialty and advanced packaging technologies, we are confident we can continue to outperform the industry growth.

We expect 2026 to be another strong growth year for TSMC and forecast our full year revenue to increase by close to 30% in U.S. dollar terms. Next, let me talk about the AI demand and TSMC's long-term growth outlook. Recent development in the AI market continue to be very positive. Revenue from AI accelerator accounted for high teens percent of our total revenue in 2025.

Looking ahead, we observe increasing AI model adoption across consumer, enterprise and sovereign AI segment. This is driving need for more and more computation, which supports the robust demand for leading-edge silicon. Our customers continue to provide us with a positive outlook.

In addition, our customers' customers who are mainly the cloud service providers are also providing strong signals and reaching out directly to request the capacity to support their business. Thus, our conviction in the multiyear AI megatrend remains strong, and we believe the demand for semiconductor will continue to be very fundamental. As a foundry, our first responsibility is to fully support our customers with the most advanced technology and necessary capacity to unleash their innovations. To address the structural increase in the long-term market demand profile, TSMC works closely with our customer and our customers' customer to plan our capacity.

This process is continuous and ongoing. In addition as process technology complexity increases, the engagement lead time with customers is now at least 2 to 3 years in advance. Internally, as we have said before, TSMC employs a disciplined capacity planning system to assess the market demand from both top-down and bottom-up approaches.

We focus on the overall addressable megatrend to determine the appropriate capacity to build. Based on our assessment, we are preparing to increase our capacity and stepping out our CapEx investment to support our customers' future growth.

We are also putting forward the existing fab schedule to the extent possible, both in Taiwan and in Arizona. We will also leverage our manufacturing excellence to drive greater productivity in our fabs to generate more output, convert N5 capacity to support N3 wherever necessary and focus on capacity optimization across node to maximize the support to our customers.

Based on our planning framework, we raised our forecast for the revenue growth from AI accelerator to approach a mid- to high 50s percent CAGR for the 5 years period from 2024 to 2029. Underpinned by our technology differentiation and broad customer base, we now expect our overall long-term revenue growth to approach 25% CAGR in U.S. dollar terms for the 5-years period starting from 2024.

While we expect AI accelerators to be the largest contributor in terms of our incremental revenue growth, our overall revenue growth will be fueled by all 4 of our growth platform, which are smartphone, HPC, IoT and automotive in the next several years. As the world's most reliable and effective capacity provider, we will continue to work closely with our customers to invest in leading-edge specialty and advanced packaging technologies to support their growth. We will also remain disciplined in our capacity planning approach to ensure we deliver profitable growth for our shareholders.

Now let me talk about TSMC's global manufacturing footprint update. All our overseas decisions are based on our customers' need as they value some geographic flexibility and a necessary level of government support. This is also to maximize the value for our shareholders. With a strong collaboration and support from our leading U.S. customers and the U.S. federal, state and city government, we are speeding up our capacity expansion in Arizona and executing well to our plan. Our first fab has already successfully entered high-volume production in 4Q '24.

Construction of our second fab is already complete and tool moving and installation is planned in 2026. Due to the strong demand from our customers, we are also putting forward the production schedule and now expect to enter high-volume manufacturing in the second half of 2027.

Construction of our third fab has already started, and we are in the process of applying for permits to begin the construction of our fourth fab and first advanced packaging fab. Furthermore, we have just completed the purchase of a second large piece of land nearby to support our current expansion plan and provide more flexibility in response to the very strong multiyear AI-related demand.

Our plan will enable TSMC to scale up an independent giga-fab cluster in Arizona to support the need of our leading-edge customers in smartphone, AI and HPC applications. Next, in Japan, thanks to the strong support from the Japan Central prefecture and the local government, our first specialty fab in Kumamoto has already started volume production in late 2024 with very good yield.

The construction of our second fab has started and the technologies and ramp schedule will be based on our customers' need and market conditions. In Europe, we have received strong commitment from the European Commission and the German federal state and city government, construction of our specialty fab in Dresden, Germany is progressing in our plan.

The ramp schedule will be based on our customers' need and market conditions. In Taiwan, with support from Taiwan government, we are preparing multiple ways of 2-nanometers fabs in both Hsinchu and Kaohsiung Science Park. We will continue to invest in leading edge and advanced packaging facilities in Taiwan over the next few years.

By expanding our global footprint while continually invested in Taiwan, TSMC can continue to be better to be the trusted technology and capacity provider of the global logic industry for years to come.

Last, let me talk about N2 and A16 status. Our 2-nanometer and A16 technologies lead the industry in addressing the insatiable demand for energy-efficient computing and almost all the innovators are working with TSMC. N2 successfully entered high-volume manufacturing in 4Q 2025, at both our Hsinchu and Kaohsiung site with good yield.

We are seeing strong demand from smartphone and HPC AI applications and expect a fast ramp in 2026. With our strategy of continuous enhancement, we also introduced a N2P as an extension of N2 family. N2P features further performance and power benefits on top of N2 and volume production is scheduled for the second half of this year.

We also introduced A16 featuring our best-in-class superpower rail or SPR. A16 is best suitable for specific HPC products with complex signal route and dense power delivery network. Volume production is on track for the second half 2026. We believe N2, N2P, A16 and its derivatives will propel our N2 family to be another large and long-lasting node for TSMC, while further extending our technology leadership position well into the future.

This concludes our key message, and thank you for your attention.

Thank you, Wendell. Thank you, C.C. This does conclude our prepared statements.

[Operator Instructions] So now let's begin the question-and-answer session. I think we'll take the first few questions from the floor here. So why don't we start over here with Gokul Hariharan from JPMorgan.

So C.C., it definitely feels like you have heard what your customers have said to you over the last 3, 4 months. Could you give us a little bit more color on what you're hearing from your customers' customers on demand because this is a very big step-up in the capacity commitment. There is definitely a lot of concern in the financial market, especially about whether we are in a bit of a bubble. And obviously, you are the one who is putting up all the capital in this industry. So you've definitely considered this very careful as well. So give us a little bit more detail in terms of what you're hearing from the customers and your views on the cycle, given if you think about typical semiconductor cycle, we've already probably lasted a little longer than usual cycles, but this is definitely doesn't feel like a typical semiconductor cycle.

Okay. Gokul, let me summarize your question for the benefit of those online and those in-person. So again, Gokul's question is really, he would like to hear C.C.'s views about the overall AI-related demand and the semiconductor cycle. So again, Gokul notes that as Wendell and you said, we are substantially stepping up our CapEx to support the customers. But he does say there is concerns about an AI bubble and risk. So part of Gokul's question is how -- what is the feedback? Any color we can share about what type of discussions and feedback we're getting from both customers and the customers' customers that C.C. mentioned. And how long do we think this cycle can last?

Okay. Gokul, you essentially try to ask us, say, whether the AI demand is real or not. I'm also very nervous about it. You bet because we have to invest about USD 52 billion to USD 56 billion for the CapEx, right? If we didn't do it carefully, and that would be big disaster to TSMC for sure. So of course, I spend a lot of time in the last 3, 4 months talking to my customer and end customers' customer. I want to make sure that my customers demand are real. So I talked to those cloud service providers, all of them.

The answer is that I'm quite satisfied with the answer. Actually, they show me the evidence that the AI really help their business. So they grow their business successfully and healthy in their financial return. So I also double check their financial status. They are very rich. That sounds much better than TSMC.

So no doubt, I also asked specifically that what's application, right? I mean that's -- for one of the hyperscalers, they told me that, that helped their social media software. And so the customer continue to increase. So I believe that. And with our own experience in the AI application, we also help to our own fab to improve the productivity. As I mentioned, 1 time say that 1% or 2% productivity improvement, that is free to the TSMC.

And that's where also our gross margin is a little bit satisfied even if this very high post period of time. And so all in all, I believe in my point of view, the AI is real, not only real, it's starting to grow into our daily life. And we believe that is kind of -- we call it AI megatrend, we certainly would believe that. So you -- another question is can the semiconductor industry to be good for 3, 4, 5 years in a row, I'd tell you the truth, I don't know.

But I look at the AI, it looks like it's going to be like an endless, I mean, that for many years to come. No matter what, TSMC stick on the fundamental technology leadership, manufacturing excellence, and we work with customers to get their trust. And I think that fundamental thing position TSMC to be very good future growth, let me say that, 25% CAGR as we projected, and we used to be conservative. You know that.

My second question is on the U.S. expansion. You're pulling in some of the capacity in response to customers. You're already starting planned for the Phase 4. There's a lot of media reports about TSMC, you might have to build more fabs in the U.S. How should we think about U.S. expansion in principle over the next few years? I think previously, you had talked about reaching 20% or even 30% of 2-nanometer capacity in the U.S. eventually, the total capacity would be in the U.S. Could you give us a little bit more detail about how that is progressing? And when could we get there in terms of the 30% or even 20% capacity?

Okay. So Gokul's second question is about our overseas expansion, particularly in the U.S. He knows that C.C. said, we are pulling in the schedule for fab 2 earlier. We're starting the application for the fourth fab. And so his question is partly around recent reports that we intend to build more fabs in Arizona. So his question is how should we or how is TSMC thinking about the future expansion in Arizona.

And we have said in the past that around 30% of our 2-nanometer and more advanced capacity would be based in Arizona once we complete scaling out to this independent giga-fab cluster, so what is the time frame, more timetable for that? How quickly can we get there?

That's a long question. We built a fab in Arizona, and we work hard. So today, everything, even the yield or defect density is almost equal to Taiwan. And due to the strong demand, as I just answered from the AI stronger, that's a megatrend. All my customer and AI customers in the U.S., so they ask a lot of support from the U.S. fab.

So because of that, we have to speed up our fab expansion in Arizona. In Taiwan also actually, we increased most of the capacity in Taiwan. No doubt about it because this is the most adjacent one we can progress very well. In the U.S., we try to speed it up and the progress is very good. We got the help from the government. But still, we have to meet all the requirement for the permits, for those kind of things. And so both in Taiwan and in Arizona, we speeded up our capacity expansion to meet the AI demand.

I can always say one word. The capacity is very tight. We work very hard to narrow the gap so far. Probably this year, next year, we have to work extremely hard to narrow the gap, okay? We just bought a second land in Arizona. That gives you a hint. That's what we plan to do because we need it. We are going to expand many fabs over there and this giga-fab cluster can help us to improve the productivity, to lower down the cost and to serve our customers in the U.S. better.

Okay. Thank you, Gokul. Let's move over here next to Laura Chen from Citibank, please.

Thank you, C.C. and Wendell for very comprehensive outlook briefing and also congratulate for the great results. Of course, we see that the AI semiconductor growth has seen very strong growth. And I believe all of your customers and customers' customers very desperate to add more capacity support from TSMC. But I'm just wondering how does TSMC evaluate the potential power electricity supply for data center. So other than that, the chips we can discuss with our customers, I think for the overall infrastructure buildup for data center, a lot of factors also very important. Just want to understand more how does TSMC evaluate those key factors for the AI infrastructure buildup? That's my first question.

Okay. So Laura's first question is around the AI demand. She notes, again, as we said, AI megatrend and the growth is very strong and customers, customers' customers and ourselves are strong believers. But when we do our planning, how do we balance this against the other considerations? Do we look at things, for example, I think Laura's question is powering electricity grid availability to basically assess is this part of our -- included as part of our planning process, do we factor such things in?

Well, Laura, let me tell you first. I worry about the electricity in Taiwan first. I need to have a lot of enough electricity, so I can start to expand the capacity without any limitation. But talking about build a lot of AI data center all over the world, I use one of my customers' customers I answer because I ask the same question.

They told me that they plan this one 5, 6 years ago already. So as I said, those cloud service providers are smart, very smart. If I knew that, I will -- anyway. So they say that they work on the power supply 5, 6 years ago. So today, their message to me is silicon from TSMC is a bottleneck and ask me not to pay attention to all others because they have to solve the silicon bottleneck first.

But indeed, we look at the power supply all over the world, especially in the U.S. Not only that, we also look at the who support those kind of power supply like a turbine, like the nuclear power plant, the plant or those kind of things. We also look at the supplier of the rack. We also look at the supplier of the cooling system, everything. So far, so good. So we have to work hard to narrow the gap between the demand and supply from TSMC. Did that answer your question?

That's great to know that it will not be the constraint for the further AI developments. Yes. And my second question is on the leading-edge advanced packaging. And Wendell, can you remind us that what would be the revenue contribution last year for the advanced packaging overall? First of all, we see that -- I recall that in the past that the CapEx for leading-edge advanced packaging is roughly about 10%. Yes. But now it could be up to like 20%. So I'm just wondering that for the expansion, can you give us more detail about what kind of the plans you are looking for.

Will you focus more on like 3DIC, SoIC? Or you also start to work on more advanced like panel based in the longer term? I also think before we talk about that, we'll work more closely with OSATs partner on the leading-edge advanced packaging. So just wondering what kind of the process will be the key expansion plan in the space.

Okay. So Laura's second question is more related to advanced packaging. What was the revenue contribution of what we call the back end, which is advanced packaging testing as a whole in 2025. And then she notes the CapEx, actually, this year, I believe, Wendell, we guided 10% to 20% of CapEx, which is the same as last year. But anyways, she wants to know what is the focus of this CapEx? Is it on 3DIC? Is it on SoIC packaging solutions, is on panel level? Sort of what is the key areas we're focusing on relative to the CapEx?

Okay. Laura, the revenue contribution last year from advanced packaging is close to 10%. It's about 8%. For this year, we expect it to be slightly over 10%. Okay. We expect it to grow in the next 5 years, higher or faster than the corporate. And the CapEx, yes, you're right, in the past, it's about 10%, lower than 10%. Now we're saying advanced packaging together with mask making and others accounted for between 10% to 20%. So you can see that the investment amount is higher. And we're investing in areas in advanced packaging where our customers need. So the areas that you mentioned, basically, we continue to invest.

Thank you, Wendell. Okay, let's move on to Charlie Chan from Morgan Stanley here.

So first of all, amazing results and guidance. Congratulations to the management team. So my first question is about outside of AI, what do you see for those end markets, right? You talked about the memory costs, et cetera. So can you give us some your underlying assumption for PC shipments, smartphone shipments, et cetera? And also in your HPC, there are some other business like networking and general servers. Can you comment about the growth potential for those segments?

Okay. Charlie's first question is very specific. Well, generally, he wants to know about how do we see the non-AI demand, especially in the context where the certain component costs such as memory costs are rising. So he wants to know what do we see the impact on the PC and smartphone markets in terms of shipments. He's also asking very specifically, what about networking, what about general server, each of these different segments.

Charlie, those -- although we say it's call non-AI, but actually that's related to AI, you know that, right? Because the networking processor, you still need to have AI data to scale up or scale out. Those are the networking switches or those kind of things still grow very strong. As for PC or the smartphone, to tell the truth, we expect higher memory price. So we expect the unit growth will be very minimal.

But for TSMC, we did not feel our customer change their behavior. And we look at it and then we found out that we supply most of the high-end smartphones. The high-end smartphone is less sensitive to the memory price. So the demand is still strong. Using one sentence, I'd like to say we still try very hard to narrow the gap. We have to supply a lot of wafers to them also.

I think that's very consistent with your 5-year CAGR outlook for all the 4 segments. And my second question is about the Intel's foundry competition. I think U.S. President seems to be very happy with Intel's recent progress. And even mentioned 2 of your key customers, right, NVIDIA, Apple may have a sound partnership with Intel Foundry. Should we concern about this so-called competition? And what TSMC can really do to mitigate or avoid potential market share loss at those key U.S. customers, not limited to the 2 customers I just mentioned.

Okay. So Charlie's second question is on the foundry competition and competition from a U.S. IDM. He knows U.S. President is very happy with the progress. A couple -- 2 of our key customers. He also was mentioned. So his question is fundamentally, is there a concern or risk going forward of market share loss for TSMC to our foundry competition?

Well, kind of a simple question, I should say, no. Let me explain a little bit because in these days, it's not a money to help you to compete, right? I also like whoever you just mentioned, to invest on Intel, I like them to invest on TSMC also. But the most fundamental thing is let me share with you. Today's technology is so complicated. So once you want to design a very complete or advanced technology, it takes 2 to 3 years to fully utilize that technology.

That's today's situation. And so after 2 to 3 years of preparation, you can design your product. Once you get your product being approved, it takes another 1 to 2 years to ramp it up. So we have a competitor, no doubt about it. That's a formidable competitor.

But first, it takes time; two, we don't underestimate their progress. But are we afraid of it? For 30-some years, we're always in a competition with our competitors. So no, we have confidence to keep our business grow as we estimate.

Thank you, C.C. All right. Let's take the next 2 questions online in the interest of time. Operator, can we take the first call from the line, please?

First question on the line, Macquarie.

First, congrats, very strong performance. My first question is about the global capacity plan. Recently Taiwan local news report that TSMC could exit the 8-inch business and mature node, 12-inch to convert into the advanced packaging. And the investors is keen to know if this is true. And the decision is based on what kind of key factor, i.e. C.C. just mentioned about the power tightness or it's ROI concern?

Okay. So Arthur's first question is about basically mature node. Our strategy on mature node. He knows a local news has been reporting that TSMC is exiting 8-inch and 12-inch businesses and converting the capacity to advanced packaging. So he wants to know if this is true. And if so, what are the reasons behind the power constraints, ROI, et cetera, et cetera?

Good question. Indeed, we reduced our 8-inch wafers capacity and 6-inch. But let me assure you that we support all our customers. We discuss with our customers and to do this kind of resources more flexible and more -- what is the word we say optimize, which I should. Optimize the resources to support our customer.

But let me assure you also to my customer, well, we continue to support them. We will not let them down. If they have a good business, we continue to support that even in the 8-inch wafer business.

Okay. Arthur, do you have a second question?

Yes. My second question is regarding the consumer and demand outlook. So C.C. also mentioned that the memory price actually inflation and he also pushing up the cost of the consumer electronics. So investors actually are concerned about the further demand softness in this year and next year or particularly next year. So can management comment about what your client or your clients' client, how to resolve this memory tightness or we call memory urgency issue?

Okay. So Arthur's second question is on the impact from the memory price increase and the demand softness. I believe this question really because C.C. already shared the impact this year. He wants to know what is the impact for 2027?

For TSMC, no impact. As I just mentioned, most of my customers now focus on high-end smartphone or PCs. So those kind of demand has less sensitive to the components price. So they continue to give us a very healthy forecast this year and next year.

Okay. Thank you, C.C. All right. Let's -- operator, let's move on to the next participant from the line, please.

Next one, Brett Simpson, Arete.

My question is really on AI. I mean, TSMC has been supply constrained for your AI customers, I think, since 2024, and it sounds like 2026 is another year where we're going to see challenges. Do you think the CapEx you've laid out for this year. TWD 52 billion to TWD 56 billion, could that mean that we start to see supply and demand more in balance in 2027? Any thoughts there just in terms of how you're thinking about that capacity plan? And does it alleviate the supply bottlenecks that we see today?

And as part of this, from a supply perspective, we hear TSMC is finding it quite challenging to develop enough engineering talent quick enough, both in the U.S. and in Taiwan. Can you talk more about this trend? And what's the scale of the labor shortage of foundry engineers at the moment?

Okay. So Brett's first question is related around AI and our capacity. So he notes, the supply looks to continue to be tight in 2026. But with the significant step-up in our CapEx to support the customers, TWD 52 billion to TWD 56 billion, do we expect the supply/demand or the gap, so to speak, to be more balanced in 2027?

And then is engineering resources, fab engineers a constraint or a bottleneck for us in making these expansions, whether in Taiwan or the U.S.?

Okay. Let me answer this question first. If you build a new fab, it takes 2 and 3 years -- 2 to 3 years to build a new fab. So even we start to spend the TWD 52 billion to TWD 56 billion, the contribution to this year almost none and to 2027, a little bit. So we actually are looking for 2028, 2029 supply.

And we hope at that time that the gap will be narrow. For 2026 and 2027, we are focused on the short-term more output. Actually, our productivity continue to increase. Our people has an incentive because of one of the TSMC's incentive is to satisfy customer. It's not because of our financial results are good, but we want to let customer feel that TSMC is trusted that whenever, they have a good opportunity to grow, we will support it.

So in 2026, 2027, for the short term, we focus on the productivity improvement, which we've done quite a good result because of, Wendell just mentioned that we can have a good financial result because of that. But that's not our incentive -- that's our incentive, but that's not our purpose. Our purpose is to support our customers. So 2026, 2027 for the short term, we are looking to improve our productivity. 2028, 2029, yes, we start to increase our CapEx significantly, and it will continue this way if the AI demand megatrend as we expected.

Brett -- thank you, C.C. Brett, do you have a second question?

Yes, I do. That was very clear. I guess my second question is about pricing. And if I look at 2025, this was the second consecutive year where TSMC's wafer ASPs were up around 20%. And as leading edge becomes a bigger portion of the mix and also you feed through price increases. When we factor in the ramp of more expensive overseas fabs, is 20% ASP -- wafer ASP increases the new normal for TSMC? Typically, you have an annual price negotiation about this time of the year. And so I'm trying to understand how you project ASPs in '26. And is your March quarter guidance factoring in price increases at leading edge?

Okay. So Brett's question is on pricing. He notes that our -- which he looking at the blended wafer price is increasing close to 20% according to his estimates. Of course, that's blended both on price and mix, but it's a leading edge and also we have mentioned earning our value. So he wants to know is the new normal going forward?

This is a tough question. I'll get the CFO to answer.

Okay. Every new node that we have a price. The price will increase. The blended ASP will increase I think they continue this way in the past and will continue with the way going forward. But Brett, I think you're asking about the contribution from pricing to the profitability.

Now as we mentioned before, the profitability, there are 6 factors affecting the profitability. And price is just one of them. And of course, we continue trying to earn our value. But in fact, in the last few years, the pricing benefits to the profitability was just enough to cover the inflation cost from tools, equipment, materials, labor, et cetera. There are other factors contributing to the higher profitability.

The first one will be a high utilization rate. As the demand is so high and as our disciplined approach to capacity planning, the utilization rate supports our high profitability. The other 1 will be our manufacturing excellence. As C.C. said, we continue to drive increasing productivity to generate more wafer output. Also, we continue to drive optimization capacity among nodes, which includes converting part of the N5 to N3.

It also involves cross support from different nodes from the mature nodes to the more advanced nodes. That is a very important advantage of TSMC. So with all these efforts, we're able to maintain a good, healthy, sustainable return profitability so that we can continue to invest to support our customers' growth.

Okay. In the interest of time, we'll take 2 more questions from the floor and 1 more from the line. So we'll go here, Sunny Lin, UBS and then...

Very strong results. Congratulations. So number one, if we look at the company, very different versus in the past from many angles. But if we look at the ramp from new node, now you can generate actually higher revenue from new node in year 4 or even year 5 of mass production versus in the past, new node like peak revenue in the second or even third year of mass production. And so could you help us understand what this new trend, what's the financial implications? And then what does that imply for you to operate or even compete differently versus in the past?

So Sunny's first question, I think maybe is related, well, to our technology differentiation, but she knows that when we ramp a new -- in the past, when we have a new node after a few years, sort of the revenue comes down a bit, but she notes that nowadays, we can still enjoy very high revenue from a node even after in its fourth or fifth year. So her question is what are the financial implications from this and also from a, I believe, competitive dynamics?

If I can answer, say we are lucky. Actually, if you -- if you look at the semiconductors product, right now, the trend is you need to have a lower power consumption always and then high-speed performance. And for TSMC, our technology depreciation becomes more and more clear, we have both benefit. We have a high speed, and we have a low power consumption. And so our leading edge customer, the first wave, the second wave, the third wave continue to come and so that sustain the demand for a long, long time.

That's a difference. Of course, this one, you need to have technology leadership, and which the technology leadership much easier to say. But every year, you have to improve. As we said, we have N2, N2P and then you won't be surprised, and the third one will be N2 something and continuously. And so that one give us the benefit and to support our customers continuous innovation.

And so they continue to stay with TSMC. And so their product can be very competitive in the market. So that answers the question say that once we got the peak revenue and did not decrease, it's continuous because second wave, third wave customers continue to join.

Thank you very much, C.C. And then maybe a question on 2 nanometer, which you should see meaningful revenue coming through in 2026. And so in the past, you guide like how much a new node will contribute to sales for the year. And so any expectations on the revenue contribution from 2-nanometer in 2026? And then I recall in terms of process migration, a few years ago, there were a lot of concerns on increasing cost per transistor. And that obviously is not declining from 5-nanometer? But then now looking at 2-nanometer, I think process migration seems to be reaccelerating even for smartphone and PC and then with larger demand coming from high-bandwidth compute. And so maybe based on your feedback from clients, maybe for smartphone and PC clients, why are they reaccelerating process migration into 2-nanometer?

Okay. So Sunny's second question very quickly in 2 parts, 2 nanometers, as we said, is a fast ramp in 2026, very strong customer interest and demand. So what -- do we have any revenue percentage to guide for in 2026?

Yes. Sunny, the 2-nanometer will be a bigger node than 3-nanometer from the start, okay? But it's less meaningful nowadays to talk about the percentage of revenue contribution when the new node starts because the corporate, as a whole, the revenue has become much bigger than before. So yes, revenue dollar, it's a bigger node. But percentage-wise, less meaningful.

Okay. And then the second part of Sunny's question from a technology perspective, as she noted increasing cost per transistor, as we said, CapEx per k going higher. So the question very simply, what's the value? What's driving smartphone, HPC customers actually to see -- we're seeing a widening out of the adoption of N2. So what is the value that is providing that the customers are willing to adopt N2?

I already answered the question, right? Because now the whole product is looking for lower power consumption and high-speed performance. And our technology can provide that value. I also say that every year, we improve. So every year, they adopt the same -- even the same name of the same node, their products continue to improve. So that provides the value. It's -- if you say that the cost per transistor is increasing, I saw the cost per transistor, the performance compared to call the CP value is increased, is much better. So that customer stick with the TSMC. Our headache right now, if I can call it a headache, is a demand and supply gap. We need to work hard to narrow the gap.

Operator, can we take the last call from the line, and we'll take one last one from the floor.

Next one, Krish Sankar, TD Cowen.

Okay. Krish, are you there? I guess not. Then let's just take the last -- not call -- sorry, the last question from Bruce Lu from Goldman Sachs.

Thank you for letting me to ask the last question. Hopefully, it's not that difficult. So I think one of the key -- I understand that TSMC is trying very hard to increase the capacity. AI revenue is growing like 15% a year, 15% plus a year. But token consumption for the last few quarters is 15% a quarter. So the gap is still there, right? I mean that's why [indiscernible] was talking about the chip war. So can you share with us that in your assumption when you provide 50% plus AI revenue growth, what kind of token consumption you can support? And how many gigawatts power in terms of the chips you can support in your assumption when you provide this kind of 5 years revenue guidance for AI?

Okay. So Bruce's first question is around our AI CAGR. Actually, to be correct, we have guided for the AI CAGR to grow mid- to high 50s CAGR in the 5-year period from 2024 to 2029. So that is the official guidance we have provided just today. Bruce's question is, in this guidance, what is our assumption basically assuming about the token growth behind this type of CAGR? What is our assumption in terms of translating to how much gigawatts of data center can we support and other specific assumptions behind our guidance?

Bruce, you got me. I mean that's -- I also try to understand what is the tokens of growth. But my customers, their product improvement continue to increase. So from -- it's a well-known from Hopper to Blackwell to Rubin, that almost double, triple their performance. So the one they can support the tokens of growth or the one they can continue to support the compute power is enormous.

And so I lose the track to be frank with you. And for gigawatt, I want to see that how much of TSMC can make the money from the gigawatt rather than say that how much we can support. Today, from my point of view, still the bottleneck is TSMC's wafer supply. Not the power consumption, not yet. So we also look at carefully. To answer your question, say that TSMC's wafer can support how much of the gigawatt, still not enough. They still have abundant of power supply in the U.S.

Okay. My next question is for the CapEx, right? I want to double check with what I just heard that C.C. was talking about like 2027, the CapEx will be more for the productivity improvement and '28, '29 may be meaningfully higher. So I do recall that in 2021, TSMC provided at 3 years for $100 billion CapEx to support that structural growth. Now the demand is even stronger. Based of that, can we do 3 years $200 billion of CapEx for the next 3 years. The math sounds doable.

Okay. So well, first, a clarification because C.C. was talking about this year, we have substantially stepping up our CapEx investment, but C.C. also mentioned it takes 2 to 3 years to build capacity. So in terms of -- Bruce's question, do we say 2027 significant step up in CapEx, I think we're saying it takes time to -- for that capacity to come out. So that's the first part.

Yes. I think Bruce, what C.C. said was the productivity was our main focus in '26 and '27 because when we start to invest the fab, the volume production will not come out until '28 and '29. So the dollar amount invested today is for 2 years or even in the future. And CapEx dollar amount, as I said, in the last 3 years, $101 billion in the next 3 years, significantly higher. I'm not going to share with you the number, but significantly higher.

So I think Wendell has addressed at least both parts of Bruce's question. Okay? So again, thank you. So again, thank you, everyone. This does conclude our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within 30 minutes from now. The transcript will become available 24 hours from now, and both are available or will be available through our TSMC's website at www.tsmc.com. So again, thank you, everyone, for taking the time to join us today. We certainly would like to wish everyone a Happy New Year. We hope everyone continues to stay well, and you will join us again next quarter. Thank you. Goodbye, and have a good day.

Good afternoon, everyone, and welcome to TSMC's Third Quarter 2025 Earnings Conference Call. This is Jeff Su, TSMC's Director of Investor Relations and your host for today.

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 Senior Vice President and CFO, Mr. Wendell Huang, will summarize our operations in the third quarter 2025, followed by our guidance for the fourth quarter 2025. Afterwards, Mr. Huang and TSMC's Chairman and CEO, Dr. C.C. Wei, will jointly provide the company's key messages. Then 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.

And 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. Thank you for joining us today. My presentation will start with financial highlights for the third quarter 2025. After that, I will provide the guidance for the fourth quarter 2025.

Third quarter revenue increased 6% sequentially in NT as our business was supported by strong demand for our leading-edge process technologies. In U.S. dollar terms, revenue increased 10.1% sequentially to $33.1 billion, slightly ahead of our third quarter guidance.

Gross margin increased 0.9 percentage points sequentially to 59.5%, primarily due to cost improvement efforts and a higher capacity utilization rate, partially offset by an unfavorable foreign exchange rate and dilution from our overseas fabs. Accordingly, operating margin increased 1.0 percentage points sequentially to 50.6%. Overall, our third quarter EPS was TWD 17.44, up 39% year-over-year, and ROE was 37.8%.

Now let's move on to revenue by technology. 3-nanometer process technology contributed 23% of wafer revenue in the third quarter, while 5-nanometer and 7-nanometer accounted for 37% and 14%, respectively. Advanced technologies, defined as 7-nanometer and below, accounted for 74% of wafer revenue.

Moving on to revenue contribution by platform. HPC remained flat quarter-over-quarter to account for 57% of our third quarter revenue. Smartphone increased 19% to account for 30%. IoT increased 20% to account for 5%. Automotive increased 18% to account for 5%. And DCE decreased 20% to account for 1%.

Moving on to the balance sheet. We ended the third quarter with cash and marketable securities of TWD 2.8 trillion or USD 90 billion. On the liability side, current liability decreased by TWD 101 billion quarter-over-quarter, mainly due to the decrease of TWD 112 billion in accrued liabilities and others as we paid out 2025 provisional tax of TWD 136 billion.

In terms of financial ratios, accounts receivable turnover days increased 2 days to 25 days. Days of inventory decreased 2 days to 74 days due to strong shipment in N3 and N5.

Regarding cash flow and CapEx. During the third quarter, we generated about TWD 427 billion in cash from operations, spent TWD 287 billion in CapEx and distributed TWD 117 billion for fourth quarter '24 cash dividend. Overall, our cash balance increased TWD 106 billion to TWD 2.5 trillion at the end of the quarter. In U.S. dollar terms, our third quarter capital expenditures totaled $9.7 billion.

I have finished my financial summary. Now let's turn to our current quarter guidance. Based on the current business outlook, we expect our fourth quarter revenue to be between USD 32.2 billion and USD 33.4 billion, which represents a 1% sequential decrease or a 22% year-over-year increase at the midpoint.

Based on the exchange rate assumption of USD 1 to TWD 30.6, gross margin is expected to be between 59% and 61%, operating margin between 49% and 51%.

This concludes my financial presentation.

Now let me turn to our key messages. I will start by talking about our third quarter '25 and fourth quarter '25 profitability. Compared to second quarter, our third quarter gross margin increased by 90 basis points sequentially to 59.5%, primarily due to cost improvement efforts and a higher overall capacity utilization rate, partially offset by margin dilution from our overseas fabs and an unfavorable foreign exchange rate.

Compared to our third quarter guidance, our actual gross margin exceeded the high end of the range provided 3 months ago by 200 basis points, mainly as the actual third quarter exchange rate was $1 to TWD 29.91 compared to our guidance of $1 to TWD 29.

In addition, we also delivered better-than-expected cost improvement efforts. We have just guided our fourth quarter gross margin to increase by 50 basis points to 60% at the midpoint, primarily driven by a more favorable foreign exchange rate, partially offset by continued dilution from our overseas fabs.

While the cost of overseas fabs remain higher, thanks to the company's overall larger scale, we now expect the gross margin dilution from the ramp-up of our overseas fabs to be closer to 2% in the second half of 2025. For the full year 2025, we now expect it to be between 1% to 2% as compared to 2% to 3% previously.

Looking ahead, we continue to forecast the gross margin dilution from the ramp-up of our overseas fabs in the next several years to be 2% to 3% in the early stages and widen to 3% to 4% in the latter stages. We will leverage our increasing size in Arizona and work on our operations to improve the cost structure. We will also continue to work closely with our customers and suppliers to manage the impact.

Overall, with our fundamental competitive advantages of manufacturing technology leadership and large-scale production base, we expect TSMC to be the most efficient and cost-effective manufacturer in every region that we operate.

Now let me make some comments on our 2025 CapEx. As the structural AI-related demand continues to be very strong, we continue to invest to support our customers' growth. We are narrowing the range of our 2025 CapEx to be between USD 40 billion and USD 42 billion as compared to USD 38 billion to USD 42 billion previously.

About 70% of the capital budget will be allocated for advanced process technologies, about 10% to 20% will be spent for specialty technologies, and about 10% to 20% will be spent for advanced packaging, testing, mass making and others.

At TSMC, a higher level of capital expenditures is always correlated with higher growth opportunities in the following years. Even as we invest for the future growth with this higher level of CapEx spending in 2025, we remain committed to delivering profitable growth to our shareholders. We also remain committed to a sustainable and steadily increasing cash dividend per share on both an annual and quarterly basis.

Now let me turn the microphone over to C.C.

Thank you, Wendell. Good afternoon, everyone. First, let me start with our near-term demand outlook. We concluded our third quarter with revenue of USD 33.1 billion, slightly above our guidance in U.S. dollar terms, mainly due to the strong demand for our leading edge process technologies.

Moving into fourth quarter 2025, we expect our business to be supported by continued strong demand for our leading edge process technologies. We continue to observe robust AI-related demand throughout 2025, while non-AI end market segment have bottomed out and are seeing a mild recovery.

Supported by our strong technology differentiation and broad customer base, we now expect our full year 2025 revenue to increase by close to mid-30s percent year-over-year in U.S. dollar terms.

While we have not observed any change in our customers' behavior so far, we understand there are uncertainties and risk from the potential impact of tariff policies, especially in consumer-related and price-sensitive end market segment. As such, we will remain mindful of the potential impact and be prudent in our business planning going into 2026 by continuing to invest for the future megatrend.

Amidst the uncertainty, we will also continue to focus on the fundamentals of our business, that is technology leadership, manufacturing excellence and customer trust, to further strengthen our competitive positioning.

Next, let me talk about the AI demand outlook and TSMC's capacity planning process disciplines. Recent developments in AI market continue to be very positive. The explosive growth in token volume demonstrate the increasing consumer AI model adoption which means more and more computation is needed, leading to more leading-edge silicon demand.

Companies such as TSMC, we are leveraging AI internally to drive greater productivity and efficiency to create more value. As such, enterprise AI is another source of demand. In addition, we continue to observe the rising emergence of sovereign AI.

We are also happy to see continued strong outlook from our customers. In addition, we directly received very strong signals from our customers' customers, requesting the capacity to support their business. Thus, our conviction in the AI megatrend is strengthening and we believe the demand for semiconductor will continue to be very fundamental.

As a key enabler of AI applications, TSMC's biggest responsibility is to prepare the most advanced technologies and necessary capacity to support our customers' growth. To address the structural increase in the long-term market demand profile, TSMC employs a disciplined and thorough capacity planning system. Externally, we work closely with our customers and our customers' customer to plan our capacity.

We have more than 500 different customers across all the end market segments. In addition, as process technology complexity increases, the engagement lead time with customer is now at least 2 to 3 years in advance. Therefore, we probably get the deepest and widest look possible in the industry.

Internally, our planning system involve multiple teams across several functions to assess and evaluate the market demand from both top-down and bottom-up approach to determine the appropriate capacity to build. This is especially important when we have such high forecasted demand from AI-related business.

As the world's most reliable and effective capacity provider, we will continue to work closely with our customers to invest in leading edge specialty and advanced packaging technologies to support their growth. We will also remain disciplined and thorough in our capacity planning approach to ensure we deliver profitable growth for our shareholders.

Now let me talk about TSMC's global manufacturing footprint update. All our overseas decisions are based on our customers' need as they value some geographic flexibility and necessary level of government support. This is also to maximize the value for our shareholders.

With a strong collaboration and support from our leading U.S. customers and the U.S. federal, state and city governments, we continue to speed up our capacity expansion in Arizona. We are making tangible progress and executing well to our plan.

In addition, we are preparing to upgrade our technologies faster to -- and to a more advanced process technologies in Arizona, given the strong AI-related demand from our customers. Furthermore, we are close to securing a second large piece of land nearby to support our current expansion plans and provide more flexibility in response to the very strong multiyear AI-related demand.

Our plan will enable TSMC to scale up through an independent giga fab cluster in Arizona to support the needs of our leading-edge customers in smartphone, AI and HPC applications.

Next, in Japan, thanks to the strong support from the Japan central, prefectural and local government, our first specialty fab in Kumamoto has already started volume production in late 2024 with very good yield. The construction of our second fab has begun, and the ramp schedule will be based on our customers' needs and market conditions.

In Europe, we have received strong commitment from European Commission and the German federal state and city government. Construction of our specialty fab in Dresden, Germany, has also started, and we are progressing smoothly with our plans. The ramp schedule will be based on our customers' need and market conditions.

In Taiwan, with support from the Taiwan government, we are preparing for multiple phases of 2-nanometer fab in both Hsinchu and Kaohsiung Science Parks. We will continue to invest in leading edge and advanced packaging facilities in Taiwan over the next several years.

By expanding our global footprint while continuing to invest in Taiwan, TSMC can continue to be the trusted technology and capacity provider of global logic IC industry for years to come.

Finally, let me talk about our end-to-end A16 status. Our 2-nanometer and A16 technologies lead the industry in addressing sizable demand for energy-efficient computing, and almost all innovators are working with TSMC. N2 is well on track for volume production later this quarter. With full year, we expect a faster ramp in 2026, fueled by both smartphone and HPC AI applications.

With our strategy of continuous enhancement, we also introduced N2P as an extension of our N2 family. N2P feature further performance and power benefits on top of N2 and volume production is scheduled for second half '26.

We also introduced A16 featuring our best-in-class Super Power Rail or SPR. A16 is best suited for specific HPC product with compressed signal routes and dense power delivery networks. Volume production is on track for second half '26. We believe N2, N2P, A16 and its derivatives will propel our N2 family to be another large and long-lasting node for TSMC.

This concludes our key message, and thank you for your attention.

Thank you, C.C. This concludes our prepared statements. [Operator Instructions] Now let's begin the Q&A session. Operator, can we please proceed with the first caller on the line. Thank you.

First one, Gokul Hariharan, JPMorgan.

Great results again. So on the AI front, C.C., I think you have met with pretty much everybody who is driving the Gen AI revolution over the last couple of months. And as you said, everybody seems to be a lot more positive. I think we gave a guidance of mid-40s data center AI growth CAGR earlier this year until 2029. Anything that you see which should kind of change that number? Definitely feels like the growth today seems to be much stronger.

And related to that, you did talk about the very detailed capacity expansion planning that TSMC does. In past technology cycles, TSMC CapEx has gone up significantly to prepare for the next upgrade or next leading-edge node. But in this cycle, TSMC revenues have grown 50% from the previous peak in '22, CapEx has only grown about 10%.

So how should we think about the CapEx over the next couple of years? I know that you're not giving numerical guidance yet, but I just wanted to understand like are we looking at much higher CapEx in the next couple of years, given all these conversations you've had. And I have a follow up after that.

Okay. So Gokul's first question, sorry, Gokul, let me summarize for everyone's benefit. So again, he wants to know, firstly, related to the AI-related demand that TSMC works with many, if not everyone, who is doing AI. And many of the customers seem to be even more positive today. So I guess he would like to ask C.C. sort of what are we seeing or hearing from our customers.

And then we had previously said that the next 5 years from 2024 to '29, we expect AI accelerator to grow at a mid-40s CAGR. Is there any update to this? I think this is the first part, then I'll get to the second part on CapEx.

That's a long question, isn't it? Gokul, the AI demand actually continue to be very strong, it's more -- more stronger than we thought 3 months ago, okay? So in today's situation, we have talked to customers and then we talk to customers' customer.

So the CAGR we -- previously we announced is about mid-40s. It's still -- it's a little bit better than that. We will update you probably in beginning of next year. So we have a more clear picture. Today, the number are insane.

And then the second part of Gokul's question are related to CapEx. He notes that in the past, when TSMC sees opportunities for higher growth, past cycles or past instances, we would step up the CapEx significantly to prepare to drive the future growth.

But he notes, this cycle, actually, though, while CapEx is increasing, the revenue is increasing even faster. So his question really, I think, how do we see this playing out over the next few years, both in terms of the CapEx spend and the growth relative to the revenue growth.

Okay. Gokul, every year, we spend the CapEx based on the -- our business opportunity in the following few years. As long as we believe there are business opportunities, we will not hesitate to invest. And if we do our job right, the growth of our business, of our revenue should outpace the growth of the CapEx. And that's what we have been delivering in the past few years.

Now going forward, assuming we're doing -- still doing a very good job, then we will continue to see that happening again. So a company of our size, the CapEx number, it's unlikely to certainly go up significantly in any given year. When we continue to invest and our growth is outpacing our CapEx growth, then you see the growth like what we have done in the past few years.

Understood. I know that it is unlikely to drop, but it is also likely to grow quite a bit given what C.C. mentioned in terms of every customer asking you and every customers' customer requesting you for capacity addition, right?

Yes. As I said, a higher level of CapEx is always going to be correlated with a higher growth opportunity. So as C.C. said, next year looks to be a healthy year, and we are confident on the mega trend that we'll continue to invest.

Yes. Maybe one more follow-up question from me. C.C., I think last year also, you gave us an indication of how much CoWoS capacity you would be building. I think you talked about 2x, of doubling the CoWoS capacity. It clearly feels like even that is not enough.

Could you give us some idea about how much capacity would you be building next year just to get some idea about what you are seeing in terms of AI demand? And also just to get some understanding of TSMC's data center AI exposure, I think last year, we talked about mid-teens revenues. Where do we end up this year? Do we end up close to like 30% of revenues coming from AI?

Okay. So Gokul, your second question, really, he wants to understand can we provide any detail or color on the CoWoS capacity plan for 2026 in terms of year-on-year increase. And also in terms of our definition of AI accelerated revenue, the narrow definition, how much will it contribute for 2025 revenue? Is it 30%?

Well, Gokul, this is C.C. Wei again. Talking about the CoWoS capacity, all I can say is continue the 3 months ago, we are working very hard to narrow the gap between the demand and supply. We are still working to increase the capacity in 2026. The real number, we probably update you next year. Today, all I want to say about the AI everything related like front-end and back-end capacity is very tight. We are working very hard to make sure that the gap will be narrow, but all I can say is we are working very hard.

Okay. Thank you, Gokul. I think we need to move on in the interest of time. So operator, can we move to the next participant, please?

Yes. Next one, Charlie Chan, Morgan Stanley.

And again, congratulations for very strong results, C.C., Wendell and Jeff. So my first question is really about your business demand. As C.C. just mentioned, your front-end demand is also very strong into next year. But one of your major customer said that more so is that -- I think his point is that doing maybe a system-level innovation in thermal, et cetera, can boost up more kind of performance.

So just a kind of a dumb question. How do we reconcile your very, very strong leading edge demand and the customers continue to migrate to your most advanced nodes? And also you continue to reflect value, whereas the customer continue to think that Moore's law is dead. Can we get some clarification from TSMC?

So Charlie's question is very specific although -- he wants us to comment on a customer saying Moore's law is dead but how do we reconcile this with a very strong leading edge demand into 2026 and also with system-level innovations?

Okay. Charlie, this is C.C. Wei. Yes, one of my customers, very important customer say Moore's law is dead, but what he meant is it's not only we rely on the chip technology anymore, now we have to focus on that whole systems' performance. So he want to emphasize the whole systems' performance rather than just talking about the Moore's law, which is not enough to meet his requirement.

So again, we work very closely with his people and to design our technology both in front-end and back-end and also in all the packaging to meet his requirement. That's all I can say.

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

Yes, I do, Jeff. Yes. So anyway, I would interpret it as so-called Moore's Law 2.0 that your co-COO, Mr. Cliff Hou also comes here during the SEMICON Taiwan. But anyway, thanks, C.C., for your commentary.

And my second question is actually a follow-up from last quarter's same question. Back then, I consulted you about China AI GPU demand, right, whether you can seize the market opportunity because China, certainly, they also expanding their AI infrastructure very rapidly.

But given the recent kind of back and forth between U.S. and China, whether China can really import NVIDIA GPU, would that kind of discount your potential long-term growth of the AI CAGR? Is that something that TSMC would worry about?

Okay. So Charlie's second question is related around AI demand and specific to China. With the sort of the export control and restriction, his question is, does that impact our ability to address the market opportunity and will this impact our AI CAGR growth if we are not allowed to fully serve China.

Yes. I think it's a little bit of both sides, meaning restriction from the U.S. but also China government's kind of discouragement to procure U.S. chip. Sorry for the interruption.

Well, Charlie, to speak the truth, I have confidence on my customers, both in graphic or in ASIC, they are all performing well. And so if the China market is not available, but I still think the AI's growth will be very dramatically and as I said, very positive, and I have confidence that our customers' performance, and they will continue to grow, and we will support them.

So even with limited opportunity from China for the time being, you are still confident that a 40% CAGR or even higher can be achieved in coming years?

You are right.

Operator, can we move on to the next participant, please?

Yes. Next one, Sunny Lin, UBS.

Congrats on the very strong gross margin. So my first question is how should we think about 2026. I understand we should get better color maybe into January, but just want to get some directional major puts and takes for gross margin trending going to 2026. Especially, how should we think about the gross margin impact from 2-nanometer ramp for 2026?

Okay. So Sunny's first question is regarding gross margin. She would like to know directionally, how do we see the gross margin for next year 2026 in terms of certain puts and takes. And also if Wendell is able to comment specifically, Sunny, sorry if I heard you right, on the N2 dilution impact, correct?

Yes, that's right.

Okay. That's her first question.

Okay. Sunny. Yes, it's too early to talk about 2026. But you already mentioned about the N2 dilution. And as all the new node, when they just come out, the N2 will have dilution in our gross margin in 2026. But at the same time, the N3 dilution is gradually coming down, and we expect the N3 to catch up to the corporate average sometimes in 2026.

The other factors include like overseas fabs dilution, which will continue and which we said that it will be about 2% to 3% dilution in the early stage of the next several years. That will also be there.

And also we all saw the dramatic foreign exchange rate movement in the earlier part of this year. There's no control. We don't know where that will be. But every percentage move of dollar against NT will affect our gross margin by 50 -- 40 basis points. So that just gives you some rough idea.

Got it. Sorry, if I may, yes, a very quick follow-up. And so on 2-nanometer, would the typical 2% to 3% dilution by new node for the first 7 to 8 quarters of mass production be a good reference for 2-nanometer as well for 2026?

Okay. So Sunny, a quick follow-up. She wants to know for the 2-nanometer dilution, if we're able to provide any detail. And can she still think about it in terms of 7 to 8 quarters or 6 to 8 quarters dilution to reach the -- time, sorry, to reach the corporate average?

Yes. Sunny, let me share with you. N2's structural profitability is better than the N3, okay? Secondly, it's less meaningful nowadays to talk about how long it will take for a new node to reach to a corporate average in terms of profitability. And that's because the corporate profitability, the corporate gross margin moves and generally, it has been moving upwards. So less meaningful to talk about that, okay?

Got it. No problem. That's very helpful. My second question is maybe for C.C. Thanks a lot for sharing with us the details on how you think about the capacity expansions and planning.

And so my question is now cloud AI is ramping a lot faster than the prior opportunities like smartphones and PCs. Yes, I think the demand for cloud AI is also may be harder to forecast.

So just want to maybe get a bit more color from you that now versus the prior rounds of capacity expansions, what is TSMC doing differently versus before? And how do you ensure that while you are ramping up the capacities more quickly while still having good risk control?

Okay. Thank you, Sunny. So Sunny's second question is regarding capacity planning and expansion. In a capital-intensive business, she notes this is very important. But in the past smartphone and PC megatrends; today, it's AI and cloud AI. She is wondering, does that make this planning process more difficult to forecast? And what are we doing differently or how do we forecast this to make sure that we are investing appropriately?

Sunny, indeed, right now because of -- I believe we are just in the early stage of the AI application. So very hard to make the right forecast at this moment. What do we do differently? There's a big difference because right now, we pay a lot of attention to our customers' customer.

We talk to and then discuss with them and look at their applications, be it in the search engine or in social media application, we talk with them and see how they view the AI application to those functions. And then we make a judgment about what AI going to grow. And so this is quite different as compared with before, we only talk to our customers and have an internal study, this is different. Did I answer your question?

Got it. Yes, yes, yes, and looking forward to the CapEx guide in January.

You're welcome.

All right. Thank you, Sunny. Operator, can we move on to the next participant, please?

Next one, Bruce Lu, Goldman Sachs.

I think Jensen talked about like $3 trillion to $4 trillion AI infrastructure opportunity by 2030, right? This compared to like 600 billion CapEx recent -- for this year implies for about 40% CAGR growth. This is similar to Jensen's guidance for the AI growth, right? But for me, first of all, what I want to know is what's TSMC's view for the AI infrastructure growth for the next 5 years? And what's TSMC's forecast for the token growth rate in the next few years?

TSMC used to provide like semi industry growth, foundry growth and how much TSMC can outperform the industry, right? Given the context that can we assume like TSMC AI-related revenue can track -- will track with the CapEx growth of AI or the major cloud service provider? Or should we expect even higher growth rate for TSMC considering you're potentially getting more value out of it?

Okay. Let me try to summarize your question, Bruce. He notes that one of our customers has highlighted a $3 trillion to $4 trillion infrastructure opportunity over the next few years compared to 600 billion current CapEx, implying a 40-something percent CAGR or growth rate, which is similar to ours.

Bruce's question is, he wants to know what is TSMC's forecast or view for AI infrastructure growth. He would also like to know what is TSMC's forecast or view for the token growth. And then what is TSMC's AI-related revenue growth? Can it track that of the cloud service providers? And his question is, should it be even higher -- shouldn't it be even higher given the value that we capture. That's actually several questions, but is that correct, Bruce?

That's right.

Well, Bruce, essentially just want to know that how accurate that we can predict the AI demand. We give you a number, roughly 40 -- the mid-40s is the CAGR, not including all the infrastructure build up and also align with our major customers' forecast for their view.

More than that, I think if we are talking about the tokens, the number of tokens increase is exponential. And I believe that almost every 3 months, it will be exponentially increase. And that's why we are still very comfortable that the demand on leading edge semiconductor is real.

And as I continue to say that we look at all the demand and look at our capacity expansion, we need -- TSMC need to work very hard to narrow the gap. That's what we are doing right now. Exact number that we probably will share with you in next year, so that when we have a very better, clear picture.

I just had a quick follow-up. I'll use that as my second question anyway. I think the question is that the token growth seems to be substantially higher than the AI-related revenue guidance on TSMC, right? So the gap is actually enlarging if you compound in the outer years, right?

That's why -- that's the differences between the -- what we see for the current TSMC outlook and the potential token consumption, right? So the gap is continue to see enlarging. How do we solve this? And do we really see that as a major issue?

Okay. So Bruce's second question, which is a follow-on from his first, is that the token growth is growing at a much higher rate or exponentially than TSMC's AI revenue growth, and this gap will only enlargen or widen in the next few years. So he wants to know -- sorry, Bruce, basically, what's the implication to TSMC or how do we see this? Is that correct?

That's right.

Okay, Bruce, you are right, you are right. The tokens and the number of tokens increase exponentially is much, much higher than TSMC's CAGR as we forecasted. And let me tell you that, first, our technology continue to improve. And so our customer moving from one node to the next node so that they can handle much more tokens number in their basic fundamental calculation. So that's one thing.

We progressed very well for one node into the other node, and our customer working with TSMC to continuously to improve their performance. And that's why when we say that we have about 40%, 45%, CAGR, then the token number are exponentially increased because of our customer and TSMC's technology combined that can handle much more or much efficient than before. Did I answer your question?

I see. So you believe your node migration plus your customer design change can fulfill or can meet the exponential growth for the token consumption?

Exactly.

Is that the conclusion?

Yes.

Sorry, Bruce, that was your second question. Operator, we need to move on. Thank you, Bruce.

Next on, Laura Chen, Citi.

Appreciate C.C. sharing your view on TSMC strategy on the AI capacity planning. I think along with very strong advanced node demand, I believe that advanced packaging like CoWoS is also one of the focus for your AI clients they are now looking for. I recall that TSMC previously also planned to expand advanced packaging in Arizona.

So can you give us an update here? And also, I mean, for the time being, just very stretched demand at the moment. So TSMC will work more closely with your OSAT partner to fulfill the strong demand at the same time? That's my first question.

Okay. Thank you, Laura. So her first question is on capacity planning. We have talked earlier on the call about the planning for leading nodes. She wants to understand also on the CoWoS capacity and specifically, I guess, advanced packaging in Arizona and how do we work with our OSAT partners.

Okay. We have announced our plan to build two advanced packaging fab in Arizona and to support our customers. But at the same time, actually, right now, we are working with one OSAT, a big company and our good partner, and they are going to build their fab in Arizona, and we are working with them because they're already breaking ground, and the schedule is earlier than TSMC's two advanced packaging fabs. And we are working with them. And our main purpose is to support our customer, and so we can many in the U.S.

Laura, do you have a second question?

Yes. Certainly, I mean, obviously, we see that the advanced node, advanced packaging are quite strong. And also at the same time, we are also seeing that the migration is also happening for N2 and N3. So just wondering that from the revenue growth perspective, I know it's still early to predict next year based on your guidance.

But I'm just wondering, will it be more driven by the ASP increase because of the technology migration? TSMC will be able to sell in your value or more that will be driven by the capacity or volume growth on both N2 ramp-up?

And also, C.C., you mentioned some of the mild silicon recovery. So that may also drive some of the volume growth into next year. So just wondering, like if you look at the growth outlook, that will be more driven by the technology upgrade ASP increase or also more like a volume? That's my second question.

Okay. So Laura's, again, second question is looking at 2026. She would like to understand what will be the key drivers of the growth. Is it more from the technology mix migration, things like N2? Is it more from ASP upgrade? Or is it more from just pure wafer volume growth?

Laura, all the above. All right? You knew it, right?

May I also follow up that because we see that actually N3 is very tight. And at the same time, we are also kind of expanding on N2. And C.C., you previously mentioned that you will migrate some of that to even N7, N6, and also N5G like -- but specifically on N3, do we also need to add more capacity into next year, newly added capacity?

Sorry, Laura is saying that will -- next year, will we continue -- sorry, Laura, if I understand correctly, will we need to add new capacity? Will we continue to do conversion? What will we do to support the very strong demand we see at leading edge next year?

Right. Yes.

Well, let me answer that question. We continue to optimize the N5, N3 capacity to support our customer. For the new building for the N3 capacity to expand, we put the new building for the N2 technology. That's today's plan.

Thank you, Laura. Operator, in the interest of time, we'll take the questions from the last two participants, please. Thank you.

Yes. Next one, Krish Sankar, TD Cowen.

My first one is, C.C., about 10 years ago, back in the smartphone days, TSM would talk about the revenue opportunity for TSM per phone. I was wondering in today's world, can you talk about how much 1 gigawatt of AI data center capacity could translate in terms of wafer demand or revenue for TSMC? And then I have a follow-up.

Okay. So Krish's first question, he noted in the past, in the smartphone megatrend, we talked about the content per phone opportunity for TSMC. So now with AI, is there a way to frame or quantify 1 gigawatt of data center capacity, what is the revenue opportunity for TSMC?

We -- recently, as I said, the AI demand continue to increase, and then my customer say that 1 gigawatt, they need about -- invest about 50 billion, how much of TSMC's wafer inside, we are not ready to share with you yet because of different from different projects, okay?

And then a quick follow-up.

Yes, excuse me, I just want to say that right now, it's not only one chip. Actually, it's many chip together to form a system, all right?

Got it, got it. Very helpful for that. Then a quick follow-up. Obviously, you first forecast long-term trend and then build capacity toward that. I'm kind of curious, when you look at the AI demand over the next several years, from a TSMC angle, does it matter whether it's -- that demand is coming through a GPU or an ASIC? Does it have an impact on your revenue or gross margin mix?

Okay. Thank you, Krish. So his second question is, again, with our business outlook. Again, we forecast the long-term trends. We plan our capacity, as C.C. said, in a thorough and disciplined manner. His question is, what are the implications, for example, of -- I believe you said GPU versus ASIC in terms of the AI market. Do we have a preference or what? Is there a difference for TSMC? Is that correct, Krish?

That's right. The impact to revenue and gross margin, whether it's a GPU or an ASIC.

Right. Okay.

Krish, no matter if it's GPU or it's an ASIC, it's all using that our leading-edge technologies. And from our perspective, we are working with our customer, and we all know that they are going to grow strongly in the next several years. So no differentiation in front of TSMC. We support all kinds of types.

Operator, can we take -- thank you, Krish. So we'll take the question from the final participant, please.

Last one, Arthur Lai, Macquarie.

Congrats on a strong outlook. I'm Arthur Lai from Macquarie. So my question is about competition. So C.C., you define the Foundry 2.0 market. And I wonder what's the strategic initiative that TSMC's undertaking to further strengthening your competitive landscape and also in this broader ecosystem. So some context. I got the question from the U.S. investor as your clients have announced they invest in Intel.

Okay. So Arthur's question is around competition. In the Foundry 2.0 landscape, what strategic initiatives, what things are TSMC focusing on to further strengthen our competitive advantage? I think the last part, Arthur, you're asking in the environment where one of our competitors in the U.S., how do we focus on the competition? Is that correct?

Yes, yes.

Okay. Let me answer that one. When we introduced our Foundry 2.0, we set a purpose that -- as I said, one of my customers say that the system performance is very important in these days, they're not only a single chip. And also -- let me share with you that our advanced packaging revenue is approaching close to 10% and it's significant in our revenue, and it's important for our customer.

So that's why we introduced Foundry 2.0 to categorize this foundry business. Not as usual, previously, we only look at the front-end portion. Now it's the whole thing, the front-end, the back-end and also important for our customer. That's why we introduced 2.0.

Talking about our competition in the U.S. Well, that competitor happened to be our customer, very good customer. So in fact, we are working with them to -- for their most advanced product. Other than that, I don't want to make any more comment.

Can I ask one more question?

Yes, you have two. So your second question, sure.

Yes. My second question is very quick on the end demand. So I recall, C.C., you, last time, mentioned that we should also monitor and worry about the prebuild, especially in the consumer electronics. And then this quarter, our number suggest that there's a Q-o-Q 19% growth in the smartphone. So my question is, do you still worry about the prebuild?

All right. So Arthur's second question is on smartphone. Do we -- are we concerned about prebuild or sort of, I guess, pulling prebuild from customers in that regard?

No. We don't worry about the prebuild because of -- when you have a prebuild, you have inventory. And in this stage, the inventory already go to the very seasonal level and very healthy. So no prebuild.

Okay. Thank you, C.C. Thank you, Arthur. Thank you, everyone. So 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 30 minutes from now. The transcript will become available 24 hours from now, and both are going to be available through TSMC's website at www.tsmc.com.

So thank you, everyone, for joining us today. We hope you all continue to stay well, and we hope you will join us again next quarter in early 2026. Thank you, and have a good day.