After that, I will provide the guidance for the fourth quarter of 2024. Third quarter revenue increased 12.8% sequentially in NT as our business was supported by strong smartphone and AI-related demand for our industry-leading 3-nanometer and 5-nanometer technologies. Gross margin increased by 4.6 percentage points sequentially to 57.8%, mainly reflecting a higher capacity utilization rate and cost improvement efforts. Due to operating leverage, total operating expenses accounted for 10.4% of net revenue.

Thus, operating margin increased by five percentage points sequentially to 47.5%. Overall, our third quarter EPS was TWD 12.54 and ROE was 33.4%. Now let's move on to revenue by technology. 3-nanometer process technology contributed 20% of wafer revenue in the third quarter while 5-nanometer and 7-nanometer accounted for 32% and 17%, respectively.

Advanced technologies, defined as 7-nanometer and below accounted for 69% of wafer revenue. Moving on to revenue contribution by platform. HPC increased 11% quarter over quarter to account for 51% of our third quarter revenue. Smartphone increased 16% to account for 34%.

IoT increased 35% to account for 7%. Automotive increased 6% to account for 5%. DCE decreased 19% to account for 1%. Moving on to the balance sheet.

We ended the third quarter with cash and marketable securities of TWD 2.2 trillion or USD 69 billion. On the liability side, current liabilities increased by TWD 31 billion, while long-term interest-bearing debt decreased by TWD 38 billion. This change was primarily driven by the reclassification of TWD 42 billion in bonds payable from noncurrent to current liabilities. In terms of financial ratios, accounts receivable turnover days remained steady at 28 days.

Inventory days increased by 4 days to 87 days primarily due to the prebuild of N3 and N5 wafers. Regarding cash flow and capex. During the third quarter, we generated about TWD 392 billion in cash from operations, spent TWD 207 billion in capex, and distributed TWD 91 billion for fourth quarter '23 cash dividend. Overall, our cash balance increased TWD 88 billion to TWD 1.9 trillion at the end of the quarter.

In U.S. dollar terms, our third quarter capital expenditures totaled $6.4 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 26.1 billion and USD 26.9 billion, which represents a 13% sequential increase or a 35% 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 57% and 59%, operating margin between 46.5% and 48.5%. This concludes my financial presentation. Now let me turn to our key messages.

I will start by talking about our third quarter '24 and fourth quarter '24 profitability. Compared to the second quarter, our third quarter gross margin increased by 460 basis points sequentially to 57.8%, primarily due to a higher capacity utilization rate and better cost improvement efforts, including productivity gains. Compared to our third quarter guidance, our actual gross margin exceeded the high end of the range provided three months ago by 230 basis points, mainly due to a higher-than-expected overall capacity utilization rate. We have just guided our fourth quarter gross margin to increase by 20 basis points to 58% at the midpoint.

This is primarily due to a higher overall capacity utilization rate in the fourth quarter, partially offset by continued dilution from N3 ramp-up, higher electricity prices in Taiwan and N5 to N3, to conversion costs. Next, let me talk about our 2024 capex. Every year, our capex is spent in anticipation of the growth that will follow the future years. And our capex and capacity planning is always based on the long-term market demand profile.

As the strong structural AI-related demand continues, we continue to invest to support our customers' growth. We now expect our 2024 capex to be slightly higher than USD 30 billion. Between 70% and 80% of the capital budget will be allocated for advanced process technologies. About 10% to 20% will be spent for specialty technologies.

And about 10% 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. And as long as our growth outlook remains strong, we will continue to invest. Now let me turn the microphone over to C.C.

C.C. Wei -- Chief Executive Officer

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 23.5 billion, above our guidance in U.S.

dollar terms. Our business in the third quarter was supported by strong smartphone and AI-related demand for our industry-leading 3-nanometer and 5-nanometer technologies. Moving into fourth quarter. We expect our business to continue to be supported by strong demand for our leading-edge process technologies.

We continue to observe extremely robust AI-related demand from our customers throughout the second half of 2024, leading to an increasing overall capacity utilization rate for our leading 3-nanometer and 5-nanometer process technologies. At TSMC, we defined several AI processors as GPUs, AI accelerators, and CPUs performing training and inference functions and do not include networking, edge or on-device AI. We now forecast the revenue contribution from several AI processors to more than triple this year and account for mid-teens percentage of our total revenue in 2024, supported by our technology leadership and broader customer base, we are well-positioned to capture the industry's growth opportunities. We now forecast our full year revenue to increase by close to 30% in U.S.

dollar terms. Next, let me talk about our global manufacturing footprint update. TSMC submission is to be the trusted technology and capacity provider of the global logic IC industry for years to come. All of 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 of our shareholders. In Arizona, we have received a strong commitment and support from our U.S. customers and the U.S. federal state and city governments and have made significant progress in the past several months.

Our plan to build three fabs will help create greater economies of scale as each of our fab in Arizona will have a clean room area that is approximately double the size of a typical logical fab. Our first fab entered engineering wafer production in April with 4-nanometer process technology. And the result is a highly satisfactory with a very good yield. This is an important operational milestone for TSMC and our customers, demonstrating TSMC's strong manufacturing capability and execution.

We now expect volume production of our first fab to start in the beginning of 2025 and are confident to deliver same level of manufacturing quality and reliability from our fab in Arizona as from our fabs in Taiwan. Our second and third fabs will utilize more advanced technologies based on our customers' needs. The second fab is scheduled to begin volume production in 2028, and our third fab will begin production by the end of the decade. TSMC will continue to play a critical and integral role in enabling our customers' success while remaining a key partner and enabler of the U.S.

semiconductor industry. Next, in Japan, thanks to the strong support from the Japan central, prefectural, and local governments, our progress is also very successful. Our first specialty technology fab has completed all process qualifications. Volume production was started this quarter, and we are confident to deliver the same level of manufacturing quality and reliability from our fab in Kumamoto as from our fabs in Taiwan.

The then preparation for our second specialty technology fab in Kumamoto has already begun and construction will begin in fourth quarter next year. This second fab will support our strategic customers for consumer, automotive, industrial, and HPC-related applications, and volume production is targeted by the end of 2027. In Europe, we have received strong commitment from European Commission and the German federal state and city government. Together with our JV partners, we had a groundbreaking ceremony in August for our specialty technology fab in Dresden, Germany.

This fab will focus on automotive and industrial applications utilizing 12, 16 FinFET, and 2028 process technologies. Volume production is scheduled to begin by the end of 2027. Under today's fragmented globalization environment, overseas fab costs are higher for everyone, including TSMC and all other semiconductor manufacturers. Having said that, we are leveraging our fundamental competitive advantage of manufacturing technology leadership and large-scale manufacturing base, thus TSMC will be the most efficient and cost-effective manufacturer in the regions that we operate, while continue to provide our customers with the most advanced technology and scale to support their growth.

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

Jeff Su -- Director of Investor Relations

OK. Thank you, C.C. This does conclude our prepared statements. Before we begin the Q&A session, I would like to remind everybody to please limit your questions to two at a time to allow all the participants an opportunity to ask their questions.

Should you wish to raise your question in Chinese, I will translate it into English before our management answers your question. [Operator instructions] Now, let's begin the Q&A session. Operator, can we proceed with the first participant on the line, please?

Questions & Answers:

Operator

The first one to ask questions is Gokul Hariharan from J.P. Morgan. Go ahead, please.

Gokul Hariharan -- Analyst

Yeah. Thank you. Good afternoon, C.C. Wendell, and Jeff.

My first question is on the AI investments and the growth that you see. Recently, obviously, there's been a lot of questions about ROI of GenAI investment, whether this could end up being a bubble. How does TSMC view this trend as you're making your capacity plans given you are enabling pretty much the processing capacity for pretty much everybody? And what gives you the confidence that this is going to be a more longer-run growth cycle? And relative to this, could C.C. also talk a little bit about what you think about the duration of this current semiconductor upcycle? Do you think it will continue into the next couple of years? Or are we getting closer to the peak of the cycle? That's my first question.

Jeff Su -- Director of Investor Relations

OK. Thank you, Gokul. Please allow me to summarize your first question. So, Gokul's first question has two parts.

The first part, I believe, is more focused on the AI-related demand, the ROI, and sustainability of this. He notes recently there's been a lot of questions about the return or ROI from generative AI investments. And therefore, how do we view this trend? Is there a worry that this demand sustainability may be a bubble? And very importantly, certainly, what gives us the confidence that this could be a more long-run sustainable demand cycle for AI?

C.C. Wei -- Chief Executive Officer

OK, Gokul, let me answer your question. Simply, whether this AI demand is real or not, OK, and my judgment is real. We have talked to our customers all the time, including our hyperscaler customers who are building their own chips. And almost every AI innovator is working with TSMC.

And so, we probably get the deepest and widest growth of anyone in this industry. And why it is real, because we have our real experience. We have been using the AI and machine learning in our fab and R&D operations. By using AI, we are able to create more value by driving greater productivity, efficiency, speed, qualities.

And think about it, let me use, 1% productivity gain that were almost equal to about TWD 1 billion to TSMC. And this is a tangible ROI benefit. And I believe we cannot be the only one company that have benefited from this AI application. So, I believe a lot of companies right now are using AI for their own improving productivity, efficiency and everything.

So, I think it's real. Did I answer your question?

Gokul Hariharan -- Analyst

That's clear. Yeah. Could you also talk a little bit about how this feeds into your view of this current semiconductor cycle also, you see?

Jeff Su -- Director of Investor Relations

So, Gokul, the second part is, that we believe the AI demand is real, but how do we view the overall semiconductor demand in cycle? I think Gokul, you're saying, do we think it's reached peak out already?

C.C. Wei -- Chief Executive Officer

OK. I forgot to answer this one. The demand is real, and I believe it's just the beginning of this demand, all right? As one of my key customers said, the demand right now is insane that it's just a beginning. It's a form of scientific engineering, OK? And it will continue for many years.

Jeff Su -- Director of Investor Relations

And then the overall semiconductor demand, I think.

C.C. Wei -- Chief Executive Officer

Overall semiconductor demand, except the AI, I think is everything stabilized and start to improve.

Jeff Su -- Director of Investor Relations

OK, Gokul?

Gokul Hariharan -- Analyst

Understood, yeah. That's clear. Thanks. My second question is more to do with capex.

I think, usually, if you look at the past cycles of strong demand uptick, TSMC's capex also starts to move up quite a bit. This time around, 2023 and 2024, your capex has been reasonably stable. It's still a large number, but not really growing a lot in the TWD 30 billion, TWD 31 billion range. How should we think about when we look forward into the next couple of years, what are you doing -- what are you planning for capex growth? Is there any reservations that TSMC has about the pace of growth that the capex is still a little bit lower? Or should we expect that the capex also should start accelerating given you're growing at 30% this year, and it looks like you're preparing for pretty strong growth the next couple of years as well?

Jeff Su -- Director of Investor Relations

OK. So, Gokul's second question is related to capex and looking out the next few years. So, he notes that in past cycles when demand is very strong, our capex starts to move up. But however, this time, even with what C.C.

described as very real and the beginning of a strong AI demand, these last two years, we kept our -- sorry, capex, it's not really been growing much. So, his question is, is that because we have any concerns or reservations around the demand sustainability? Or what will the capex start to look like the next several years? Will we have to -- we'll begin to step it back up? Is that roughly what you're asking, Gokul?

Gokul Hariharan -- Analyst

Yeah, that's right. Thanks, Jeff.

Wendell Huang -- Vice President, Chief Financial Officer

OK, Gokul, we do not have a number for 2025 capex today to share with you. However, we always use a disciplined and thorough system to determine the appropriate capacity to build. And we always review our capex plans on an ongoing basis. For TSMC, a higher level of capital expenditures is always going to be related with higher growth opportunities in the coming years.

And as long as our growth outlook remains strong, we will continue to invest. Now as C.C. said, next year looks to be a healthy year. So, it is very likely that our capex next year will be higher than this year.

And we will provide you more updates in January conference.

Gokul Hariharan -- Analyst

OK. That's very clear. Thanks. Thanks, Wendell.

I'll go back to the queue.

Jeff Su -- Director of Investor Relations

OK. Thank you, Gokul. Operator, can we move on to the next participant, please?

Operator

Yes. Sunny Lin from UBS. The line is open to you now.

Sunny Lin -- UBS -- Analyst

Thank you very much for taking my question. Good afternoon. Very strong execution on margin. And so, my first question is to follow up on the gross margin outlook.

How should we think about into 2025? Last quarter, management quantified overseas expansion to about 2% to 3% margin. But besides that, could you help us understand some of the other puts and takes? How should we think about the depreciation growth into 2025?

Jeff Su -- Director of Investor Relations

OK. Sunny, thank you for your question. So, Sunny's question first is on the gross margin outlook into 2025. She notes last time Wendell had already shared that overseas fab would dilute our gross margin probably by 2% to 3%.

She wants to know besides that, what other factors should we consider and also in terms of depreciation growth?

Wendell Huang -- Vice President, Chief Financial Officer

OK. Sunny, again, it's too early to talk about 2025 in detail. But there are a few things I can share with you, the puts and takes for 2025. Indeed, we expect the dilution from N3 ramp to gradually reduce next year.

We continue to sell our value. So, these things will help. Plus, next year will be a healthy growth year. So, utilization is also a positive.

On the other hand, as we said, there'll be 2 to 3 percentage point dilution from the overseas fabs when we begin to ramp them. At the same time, we are also converting some of our N5 capacity to N3 to meet the strong demand for N3. And also, don't forget there's -- we're ramping the N2 in 2026. There will also be some preparation costs for ramping N2.

And as we migrate every leading nodes, more and more advanced, this preparation cost will become bigger and bigger. Now you also know that the electricity cost has also risen very recently and second time in the year, 14% higher for TSMC in October. This is after 15% increase in 2022, 17% increase in 2023, and 25% increase in 2024. Basically, the price has doubled in the last few years.

So, next year, we think that the electricity price for us in Taiwan will be the highest in all the regions that we operate. And the higher electricity price, plus other inflationary cost is expected to impact our gross margin by at least 1%. And finally, of course, we do not have any control or cannot forecast foreign exchange rate movement. 1% of foreign exchange rate movement, dollar NT, will have an impact of 40 basis points to our gross margin.

Jeff Su -- Director of Investor Relations

Thank you, Wendell. And also Sunny, does that sort of give you a better understanding of the puts and takes for next year?

Sunny Lin -- UBS -- Analyst

Yeah, that's very helpful. Thank you very much, Wendell. So, a quick follow-up on if any view on depreciation increase into 2025. Should we expect similar magnitude as this year, which grew about 25%? And also a quick follow-up on the 2% to 3% gross margin dilution for overseas.

If I calculate correctly, that would imply your overseas fabs in the U.S. and Japan could be running at close to 0% or very low gross margin to start with. Would that be a conservative assumption that you are assuming at this point? Yeah.

Jeff Su -- Director of Investor Relations

OK. Well, that's two follow-ups. So, I might take that as the second question. But the first one is, again, she -- Sunny had asked about depreciation next year.

Wendell Huang -- Vice President, Chief Financial Officer

Yeah. Sunny, let us give you more depreciation guidance in the January conference, OK?

Jeff Su -- Director of Investor Relations

And then the second question really is about our overseas fabs as we ramp up Arizona and Arizona and Japan. Sunny, I believe your question is with that ramp and dilution of 2% to 3%, does that imply the overseas fabs profitability, is more closer to breakeven or something like that? Correct, Sunny?

Sunny Lin -- UBS -- Analyst

Right. Right. Thank you, Jeff.

Wendell Huang -- Vice President, Chief Financial Officer

Yeah. Sunny, the overseas fabs have basically a lower profitability than the fabs in Taiwan, mainly because of a smaller scale and of -- next year will be initial ramp and higher costs. So, it will have a lower profitability, but it will gradually improve over the years. Now, don't forget, in Arizona and in Kumamoto, we are ramping more than one phase, one fab.

So, when fab 1 begin to improve its profitability, the second phase comes in. And Arizona, when fab 2 improves, the third phase comes in, as similar situation in Kumamoto. And that's the reason why we're saying that in the next three or five years, we expect 2 to 3 percentage point dilution every year.

Jeff Su -- Director of Investor Relations

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

Operator

Next one, we'll have Charlie Chan from Morgan Stanley. Go ahead now.

Charlie Chan -- Analyst

Thanks, Jeff, C.C., and Wendell. Good afternoon. First of all, also congrats for your very, very strong results and guidance. So, my first question is really about your future buying power to your customers and vendors and what would that mean to your long-term gross margin target? Because right now, even without the price adjustment next year, your gross margin is already blow up, right, is already 57%, 58% higher than most of your customers.

So, with that, do you think you'll be a little bit too aggressive to further hike price to your customers? And I'm also wondering for those kind of mature nodes, what's your pricing strategy here? So, all I want to ask are the implications to, number one, your long-term gross margin level. And secondly, are you going to handle the potential antitrust risk, given your near monopoly and also prepared to do some price adjustments?

Jeff Su -- Director of Investor Relations

OK. So, I think Charlie's question is sort of we always say that we value selling and it's continuous and ongoing thing. So, he wants to know what does the status look like with both customers and vendors. And then what is the implication to long-term gross margin target? And are we concerned about sort of monopoly or things like that.

Let me start -- that, I think, is the first part, and then the mature part, I'll summarize later.

C.C. Wei -- Chief Executive Officer

OK, Charlie. This is C.C. Wei, let me answer your question. Sharing our value actually is a continuous and ongoing process for TSMC.

We view all our customers as partner. And the progress of this effort and sharing our value is so far, so good. Also, you are talking about we have more backing power with our equipment supplier. Again, we view them as our partners.

So, I don't use the backing power. We always work with them and move to the next step. And so, both TSMC and suppliers and customers are all working together. And with the purpose of that, we can have a TSMC's customer to be successful in their market, OK? That's my goal.

If our customers are doing well, we can be good also, OK? And you mentioned one thing that I have very big concern. You say that TSMC's gross margin now is higher than my customer. It's not true. You look at one of the biggest AI supplier, they have gross margin that I probably never be able to reach in my life.

But anyway, we are very happy to see them be successful. And we are in a different kind of business. We are a capital-intensive business. So, we need a very high gross margin to survive and to have sustainable and healthy growth.

That's why we set up our pricing strategy.

Jeff Su -- Director of Investor Relations

Yeah. And then a quick one, I think part of Charlie's question, the second part. Charlie, if I heard you right, is sort of the outlook for mature nodes in terms of pricing.

C.C. Wei -- Chief Executive Officer

No. I think Charlie, you're talking about the way that the antitrust concern or something like that. Did I hear you right?

Charlie Chan -- Analyst

Yes, I'm a little bit concerned, same as other investors as well.

Jeff Su -- Director of Investor Relations

So, yeah.

C.C. Wei -- Chief Executive Officer

OK. The antitrust, meaning that TSMC has a very high market share and with some of unnecessary competition of methodology. Let me assure you that the last time we proposed a new version of the foundry 2.0, which including the wafer manufacturing and packaging and testing, mass-making and all others. All these kind of things become more growing importance like packaging testing, mass-making.

Right now, probably is a little bit higher than 10% of TSMC's total revenue. That's one. And all my competitors, they are IDMs, in particular. They also have their packaging, testing, and mass making.

And so, I think Foundry 2.0 is better reflects TSMC's addressable market and our share is not -- it's probably around 30%. So, not dominant yet. We are big, yes, because we perform, we will. But no, it's not a kind of an antitrust concern.

It's not in our picture, actually. Charlie, did I answer your question?

Charlie Chan -- Analyst

Yeah. Yeah, thanks. Yes. So, just hitting a little bit, I guess, you won't stop your margin improvement until 75%.

But anyway, I will leave that kind of long-term margin target or guidance to the next caller or your next January guidance. But I do have a second question. Is that OK, Jeff, if I can ask a second one?

Jeff Su -- Director of Investor Relations

Quickly. Yes.

Charlie Chan -- Analyst

OK. The second one is about your IDM outsourcing opportunity, right, also part of your Foundry 2.0 because one of the major IDM opportunities that Intel and recently announced they want to spin off their foundry segments, right? So, number one, TSMC, do you really expect more outsourcing from Intel given this change? And even with TSMC consider to acquire part of the Intel's fab in the long term? So, I think that's that. And also quickly comment on news reporting about Samsung's IDM outsourcing opportunity. So, that's my second part of the question.

Jeff Su -- Director of Investor Relations

OK. So, Charlie, second part is on IDM outsourcing. He wants to know with the, I guess, U.S. IDM and Foundry 2.0, do we expect more outsourcing from this U.S.

IDM? And how do we plan for the capacity? Would we consider, I think, acquiring part of this IDMs fabs or manufacturing? And then how do we see outsourcing from Samsung?

C.C. Wei -- Chief Executive Officer

Well, that's a lot of questions. Let me answer one of the easiest one, are we interesting to acquire one of IDMs so far? The answer is no. OK. No, not at all.

Now let's talk about the business part. It always a customer's decision for their outsourcing strategy. But I look at the business of the IDMs -- one of the IDMs in California, which has been a very good customer to TSMC. And we continue to receive a sizable business from them to be frank with you.

Yeah. So, your question is whether that we continue to increase, that is too specific. So, let's wait for the next few quarters to answer your question.

Jeff Su -- Director of Investor Relations

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

Operator

Yes. Next one, Bruce Lu, Goldman Sachs. Go ahead, please.

Bruce Lu -- Goldman Sachs -- Analyst

OK. Thank you for taking my question. My question is going back to the longer-term growth outlook. I think TSMC has been guiding for 15% to 20% revenue CAGR from '21 to '26.

On the back of the insane AI demand, do we have an updated outlook for the revenue guidance beyond 2026? In addition, when TSMC guided in 2021, TSMC achieved 25% growth almost every year, except for 2023. So, what kind of like revenue growth outlook or the growth pattern in the next five years? Do we expect to be a more stable growth every year? Or do we expect to be a stronger growth for most of the year and the one year of weakness?

Jeff Su -- Director of Investor Relations

OK. So, thank you, Bruce. So, Bruce's first question is about our long-term growth outlook. He notes, yes, we did provide a long-term revenue growth CAGR of between 15% to 20% CAGR in U.S.

dollar terms from '21 to '26. So, he wants to know what is the updated revenue guidance, the outlook beyond 2026? And he also notes that when we provided this guidance in 2022, we basically were able to grow greater than 25% every year except for 2023. So, he wants to know, the next five years, will the growth be similar level every year? Or will it be like strong years followed by a down year, sort of the pattern of the growth?

C.C. Wei -- Chief Executive Officer

OK. Bruce, that's a big question. You ask us whether the next five years will be as good as the past five years from '21 to '24. Right now, we -- except the 2023, we have a very good growth.

As you said, it's always 20% to 30%. Next five years, it will be very healthy also to TSMC, but I don't have a long-term CAGR number to update you, but they will be very healthy also. That's so far, I can assure you.

Bruce Lu -- Goldman Sachs -- Analyst

So, a quick follow-up, does that insane AI demand is going to help you to grow slightly faster than before?

Jeff Su -- Director of Investor Relations

So, Bruce wants to know with the very robust AI-related demand, can we grow faster than before?

C.C. Wei -- Chief Executive Officer

I hope so. But as I said, today, we don't have a long-term CAGR number to share with you, OK?

Bruce Lu -- Goldman Sachs -- Analyst

OK. Let me switch gear to 2 nanometers and A16s. So, I think that the HPC demand for the thermostat is very strong. We do see more customer engagement for 2 nanometers.

But at the same time, we also see more chiplet design in 2 nanometers, which might effectively lower the 2-nanometer wafer area requirement. So, how should we think about your 2-nanometer capacity build versus the 3, 5, 7 in the past? And how do we see the A16 migration beyond 2 nanometers?

Jeff Su -- Director of Investor Relations

OK. So, Bruce's second question, as he said, is on 2-nanometer and A16. He notes certainly with HPC and AI related more and more engagement at 2-nanometer, but he also notes or his observation with chiplets, that could reduce the demand for 2-nanometer. So, he really wants to know sort of what is the capacity build or capacity outlook for 2-nanometer that we are looking at? And also how do we see the migration from 2-nanometer to A16.

Is that correct, Bruce?

Bruce Lu -- Goldman Sachs -- Analyst

Yes.

C.C. Wei -- Chief Executive Officer

OK. Let me answer this question. Yes, the chiplets has become kind of our HPC, especially our HPC customers strategy. It's going to reduce capacity for 2-nanometer because they become chiplets, the answer is no.

Actually, we have many, many customers interested in the 2-nanometer. And today, with the activity with TSMC, we actually see more demand than we ever dream about it as compared with N3. So, we are to prepare more capacity in N2 than in N3. And following by A16, again, A16 is very, very attractive for the AI servers' chips.

And so, actually, the demand is also very high. So, we are working very hard to prepare both 2-nanometer, A16's capacity. OK. Bruce, did I answer your question?

Bruce Lu -- Goldman Sachs -- Analyst

Yes, perfect. Thank you.

Jeff Su -- Director of Investor Relations

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

Operator

Now Brett Simpson from Arete. The line is open to you now.

Brett Simpson -- Arete Research -- Analyst

Yes. Thanks very much. My question was on the long-term planning around AI. Keen to understand how TSMC gets comfortable with customer demand for AI beyond 2025.

And I ask this because it takes a couple of years before you can build a fab. So, you need to be taking an early view on what does AI look like in 2026, 2027. So, how are you specifically cooperating on long-term plans for capacity with these AI customers? What commitments are these customers giving you? And I guess, historically, we've seen hyperscaler capex go through digestion periods. So, how do you derisk the capacity plans here for AI as we go through this really heavy demand period? Thank you.

Jeff Su -- Director of Investor Relations

OK. So, Brett's first question again is on the long-term planning around AI demand. Question is really for beyond 2025, given the lead times, how do we plan our capacity for the long term? How do we get comfortable around the customer demand related to AI beyond 2025? I think that is your gist of your question, Brett.

C.C. Wei -- Chief Executive Officer

OK. Brett. Yes. Let me say again that we did talk to a lot of our customers.

Almost every AI innovator are working with us. And that's including the hyperscalers. So, if you look at the long-term structure and market demand profile, I think we have some picture in our mind and we make some judgment, of course, and we work with them on a rolling basis. And so, how we prepare our capacity, actually, just like a Wendell said, we have a disciplined and the rough system to plan the appropriate level of capacity.

And that -- to support our customers' need also to maximize our shareholders' value. That's what we're always keeping our mind. OK?

Brett Simpson -- Arete Research -- Analyst

Thank you. Yes, that's great. Thanks, C.C. I guess my follow-up question, we've read a lot about Taiwan having energy challenges.

And this comes at a time when TSMC is preparing for a big node with 2-nanometer. So, my question is, are there any power challenges to overcome when you're building out your N fabs, especially in Shinshu and Gosheng? And does it make sense for TSMC to plan for nuclear power? I mean we see a lot of hyperscalers are planning for nuclear power in the U.S. to build out gigawatt facilities. How do you think about nuclear power in future for TSMC fabs? Thank you.

Jeff Su -- Director of Investor Relations

OK. Thank you, Brett. So, Brett's second question, he notes that Taiwan electricity and energy, there are a lot of challenges with this. So, how do we plan for this, especially when we're bringing on new nodes? What are the power challenges to overcome given the state of Taiwan's energy? And then would we consider even nuclear power for ourselves to help support?

C.C. Wei -- Chief Executive Officer

OK, Brett. Yes, we are building many fabs in Taiwan, and that require electricity, water and the land. We continue to work with the government. Actually, we have a very close communication with government and to tell them that our requirement, our plan.

And we got assurance from the government saying that they will support TSMC's growth and we believe that. And so, whether how they prepare for the electricity, from the nuclear power plant or from just some kind of other sources like green energy, something, we are not ready to share with you yet, but we got the assurance that we're going to get enough electricity support, including the water and the land.

Brett Simpson -- Arete Research -- Analyst

Thank you very much.

Jeff Su -- Director of Investor Relations

OK. Thank you, Brett. Can we move on to the next participant, please?

Operator

Next one to ask question, Krish Sankar from TD Cowen.

Krish Sankar -- Analyst

Yeah. Hi. Thanks for taking my question. The first one I had was, I'm quite curious on the non-AI demand.

How do you look at your wafer demand for PC and mobile into calendar '25? And have you seen any meaningful revision upwards or downwards on that? And then I had a follow-up.

Jeff Su -- Director of Investor Relations

OK. So, Krish's first question is really focused on PC and mobile demand, how do we see this demand going into 2025? Have we seen it improve or revision up or downwards? And how would we look at it going into next year?

C.C. Wei -- Chief Executive Officer

OK, Krish, the unit growth of PC and smartphone is still in a low single digit. But more important is the content. The content now, we put more AI into that chip. And so, the silicon area increased faster than the unit growth.

So, again, I have to say that for this PC and the smartphone business, not -- is gradually increased, and we expect to be healthy in the next few years because of our AI-related applications.

Krish Sankar -- Analyst

Got it. Got it. Very helpful. And then my quick follow-up is I'm kind of curious on your packaging side -- advanced packaging today is part of the non-wafer revenues.

Obviously, we're investing in CoWos and other technologies. How to think about that advanced packaging revenue growth over the next few years? And do you think at some point in the next couple of years, advanced packaging can reach corporate-level gross margins? Or would it always be below that? Thank you very much.

Jeff Su -- Director of Investor Relations

OK. Thank you, Krish. So, Krish's second question is on advanced packaging. So, we have been putting a lot of effort.

So, his question is, what is the revenue growth outlook for advanced packaging in the next few years? And also when or do we think it can reach the corporate average gross margin as well? So, maybe Wendell can address.

Wendell Huang -- Vice President, Chief Financial Officer

Yes, Krish, advanced packaging in the next several years, let's say, five years, will be growing faster than the corporate average. This year, it accounts for about high single digit of our revenue. In terms of margins, yes, it is also improving. However, it's approaching corporate, but not there yet.

Krish Sankar -- Analyst

Thank you very much.

Jeff Su -- Director of Investor Relations

OK, Krish. Operator, in the interest of time, we will take questions from the last two participants, please.

Operator

Yes. Now the line is open to Laura Chen from Citi.

Laura Chen -- Citi -- Analyst

Hi. Thank you for taking my question, gentlemen. And also congratulations from the strong performance. I'm just wondering that with the decent free cash flow increase in the recent quarter, and I believe that next year will be a decent year for TSMC to grow.

So, is there any opportunity for TSMC to consider increase the cash dividend in the near future? And how do you view the balance sheet of capital allocation between shareholder return and also the continued investment in advanced technology like 2- and 3-nanometer? That's my first question.

Jeff Su -- Director of Investor Relations

OK. Thank you, Laura. So, Laura's first question is related around -- her question really is looking at our free cash flow generation, noting that it has continued to increase. And then with next year being a healthy good year, you should continue to grow.

So, her question is, is there room? Or how do we see the cash dividend as related to free cash flow? And also, I think, Laura, really, her question was also balance sheet ramification. How do we balance the shareholder return interest set against what C.C. said, our capital-intensive industry and capex? So, two parts to it. First, on the cash dividend and then sort of on the balance sheet management.

Wendell Huang -- Vice President, Chief Financial Officer

Right. Laura. Certainly, our dividend policy, as we said before, is sustainable and steadily increasing, and steadily increasing when we are harvesting the investment that -- made in the past. And as the free cash flow increase, that means we are harvesting the investment in the past.

So, it's going to be steadily increasing. That's for the dividends. Balance sheet. We are -- the primary objective of our using our balance sheet or our cash resources is organic growth.

So, that will bring our shareholders the biggest return. And then whatever the free cash flow is left, then we will return part of those to our shareholders. That's always our policy.

Laura Chen -- Citi -- Analyst

OK. Yes, that's very consistent. But I think given our, like, rich cash flow generating capabilities, and also -- yes, I mean, we can probably be able to harvest. So, I think investor kind of expecting the dividends increase gradually.

Yes. And also my second question is about our Foundry 2.0 model. That has been discussed since last time. But could you share more details on how the Foundry 2.0 is being implemented? And from different spec like the traditional logic foundry business and advanced packaging and also maybe the IDM customers, can you give us kind of an idea what will be the like a growth outlook for different kind of segmentation in that new definition?

Jeff Su -- Director of Investor Relations

OK. So, Laura, your second question is on Foundry 2.0. So, I think her question is looking at the different components of this. She is asking if we can provide a growth outlook for the different components, such as the logic, wafer manufacturing, advanced packaging segment, IDM segment, etc., etc., the growth outlooks for each specific one.

Laura Chen -- Citi -- Analyst

Yes, appreciate it. Thank you.

C.C. Wei -- Chief Executive Officer

Laura, I think again, in this Foundry 2.0, among the content inside, the leading process nodes, advanced packaging has much stronger growth. And on those mature nodes and conventional packaging, that is not so rosy as advanced packaging and leading process node. Did that answer your question?

Laura Chen -- Citi -- Analyst

OK. Got it. Thank you.

Jeff Su -- Director of Investor Relations

OK. Thank you, Laura. Operator, can we move on to the last participant, please?

Operator

Yes. The last one to ask question, Rick Hsu from Daiwa.

Rick Hsu -- Analyst

Yeah. Hi. Thank you for taking my questions Very quick one from me. The first one, can you update your forecast for this year's global semiconductor revenue as memory? I remember you were talking about around 10% growth.

So, can you give us a new update and also share with us a preliminary outlook for next year?

Jeff Su -- Director of Investor Relations

OK. So, Rick's question is really, can we provide an update to our forecast for, SEM, whether it's semiconductor excluding memory industry growth and Foundry 2.0 for 2024 first, yeah.

C.C. Wei -- Chief Executive Officer

Well, our forecast stays the same, very similar to what we said last time. Of course, TSMC is, of course, is a little bit better than the last time we estimate. But overall, the whole industry is almost the same as we said in last quarter.

Jeff Su -- Director of Investor Relations

Yes. And then the second part of Rick's question, what about next year?

C.C. Wei -- Chief Executive Officer

Well, we continue to say that it's too early to make a comment on 2025 growth outlook, but we are going to share with you in the next quarter's earnings release, OK?

Rick Hsu -- Analyst

All right. Thank you so much. It's a little question as a follow-up, the second one. Can you share with us your CoWoS capacity buildup for this year and next year? I know you guys seem to have revised that up several times.

So, can you share the latest one?

Jeff Su -- Director of Investor Relations

Yeah. So, Rick, second question then is to update on our CoWoS capacity plan for both 2024 and 2025 to the extent that we can share.

C.C. Wei -- Chief Executive Officer

OK, Rick. In fact, we are putting a lot of effort to increase the capacity of the CoWos. Roughly, let me share with you, today's situation is our customers demand far exceed our ability to supply. So, even we work very hard and increase the capacity by about more than twice, more than two times as of this year compared with last year and probably double again, but still not enough.

But anyway, we are working very hard to meet the customers' requirements.

Rick Hsu -- Analyst

All right. Thank you so much.

Jeff Su -- Director of Investor Relations

OK. Thank you, Rick. Thank you, everyone. 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 of which you can find and are available through TSMC's website at www.tsmc.com. So, thank you, everyone, for joining us today. We hope everyone continues to be well, and we hope to see you again and you will join us again next quarter. Goodbye, and have a good day.

Duration: 0 minutes

Call participants:

Jeff Su -- Director of Investor Relations

Wendell Huang -- Vice President, Chief Financial Officer

C.C. Wei -- Chief Executive Officer

Gokul Hariharan -- Analyst

Sunny Lin -- UBS -- Analyst

Charlie Chan -- Analyst

Bruce Lu -- Goldman Sachs -- Analyst

Brett Simpson -- Arete Research -- Analyst

Krish Sankar -- Analyst

Laura Chen -- Citi -- Analyst

Rick Hsu -- Analyst

More TSM analysis

All earnings call transcripts

This article is a transcript of this conference call produced for The Motley Fool. While we strive for our Foolish Best, there may be errors, omissions, or inaccuracies in this transcript. As with all our articles, The Motley Fool does not assume any responsibility for your use of this content, and we strongly encourage you to do your own research, including listening to the call yourself and reading the company's SEC filings. Please see our Terms and Conditions for additional details, including our Obligatory Capitalized Disclaimers of Liability.

The Motley Fool has positions in and recommends Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.

Taiwan Semiconductor Manufacturing (TSM) Q3 2024 Earnings Call Transcript was originally published by The Motley Fool