Taiwan Semiconductor Manufacturing Company Limited (TSM) Q1 2013 Earnings Report, Transcript and Summary

[Chinese] Welcome to TSMC's First Quarter 2013 Earnings Conference and Conference Call. This is Elizabeth Sun, TSMC's Director of Corporate Communications and your host for today. The event is webcast live via TSMC's website at www.tsmc.com. [Operator Instructions] As this conference is being viewed by investors around the world, we will conduct this event in English only. The format for today's event will be as follows: First, TSMC's SVP and CFO, Ms. Lora Ho, will summarize our operations in the first quarter, followed by our guidance for the current quarter. Afterwards, TSMC's Chairman and CEO, Dr. Morris Chang, will provide his general remarks and a couple of key messages. Then, we will open the floor to questions. For those participants on the call, if you do not have a copy of the press release, you may download it from TSMC's website at www.tsmc.com. Please also download the summary slides in relation to today's earnings conference presentation. Before we begin, I would like to remind everybody that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the Safe Harbor notice that appears on our press release. And now, I would like to turn the podium to TSMC's CFO, Ms. Lora Ho.

Thank you, Elizabeth. Good afternoon, everyone. Thank you for your participation. Today, my presentation will start with financial highlights for the first quarter 2013 and followed by the guidance for the second quarter. Our first quarter revenue turned out better than guidance, due to a more favorable exchange rate and stronger mobile demand and TSMC's strong position in 28-nanometer. Compared to the previous quarter, first quarter revenue increased 1% to TWD 133 billion. On the margin side, gross margin was 45.8%, down 1.4 percentage points from the fourth quarter last year. Higher capacity utilization and the NT dollar depreciation contributed to a favorable impact to our gross margin by 3 percentage points. However, the inventory valuations adjustment on a quarter-over-quarter basis has impact our gross margin by 4.4 percentage points. This is consistent with the accounting treatment under R.O.C. GAAP number 10 as before. First quarter operating margin was 33.5%, down 1.8 percentage points from the prior quarter due to higher opening expense for Fab-15 capacity expansion for 28-nanometer. Nonoperating items was again of TWD 1.3 billion for this quarter, much better than the loss in the fourth quarter of 2012. The difference is mainly due to the lack of impairment charges of TWD 1.5 billion we took in the fourth quarter last year for certain invested companies. Also in the first quarter, we disposed a portion of these mixed shares at a profit of TWD 0.7 billion versus its value on our book. Overall, our EPS was TWD 1.53. ROE was 21.3%. As you may have noticed, we are preparing financial statements in accordance with Taiwan IFRS starting this year. To illustrate the difference, we prepared a comparison table under both R.O.C. GAAP and the Taiwan IFRS, using 2012 number as example. I would like to highlight a few…

I would like to make some comments on industry outlook and on 1Q and 2Q. And then on our technology department, 28, 20 and 16, as well as specialty technologies. And finally, I will talk about this year's capital expenditures. For the industry outlook, for the full year 2013, our estimate of global GDP remains unchanged at about 2.6% growth. At C [ph] market, we now estimate to be about 4% growth, that is up from the 3% earlier estimate, due mainly to memory price recovery. Fabless industry, our estimate of fabless industry growth remains unchanged at 9%. Foundry industry, we now estimate to grow at 10%. That is up from 7%, mainly due to the raise of our own estimate of our growth. TSMC growth will be much higher than the foundry industry growth of 10% that I've mentioned earlier. On inventory. Supply-chain inventory at the end of 4Q was close to seasonal, was reasonably balanced, close to the seasonal norm. And we expect it to remain slightly above seasonal in all 4 quarters this year. And basically, we feel that it is reasonably -- it will be reasonably balanced every quarter this year. On 1Q and 2Q of 2013, we are pleased with our first quarter performance and we are encouraged by the business prospects of the second quarter, as Lora has already guided. Both quarters are strong, good and seasonal. We attribute our strength to, first, mobile-related applications whose demand remains strong and TSMC's strong position in 28-nanometer technology, which has become widely adopted by many mobile-related applications. Now a few comments on 28-nanometer. I said before and I say again, our capacity and output continue to ramp up aggressively. Both our production and our revenue of 28-nanometer wafers in 2013 will triple the production and revenue…

Okay. this concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to limit your questions to 2 at a time to allow all participants an opportunity to ask questions. Questions will be taken both from the floor and from the call. Should you wish to raise your question in Chinese, I will translate it to English before our CEO or CFO answers your questions. [Operator Instructions] Now let's begin the Q&A session.

Okay. Our first question comes from the floor and it would be from Goldman Sachs' Donald Lu.

Chairman, Lora and Elizabeth [indiscernible], or maybe I should use English here. My -- first, congratulations on the very good second quarter guidance. That was -- first question is on the 20-nanometer. Some investors and also your customers has been talking about the cost per transistor might start to go up after double patterning, et cetera, et cetera. And what's your view on this impact to the supply chain, your profitability and your customer's profitability and the adoption of technology, et cetera? And my second question is on TSMC's -- I just want to hear from you, again, the return and the profitability for future nodes, 20-nanometer and 16-nanometer.

Thank you. Transistor cost -- 20-nanometer transistor cost. Yes, the slope of decrease of transistor cost has been less in -- between 28 and 20, the decrease in transistor cost is less than the decrease in transistor cost between 40 and 28. But now, the important point is that the value we offer in any new node is, I think, gradually shifting, or has already gradually shifted from just pure lower transistor cost to the performance aspects, performance speed and power. And, of course, in addition to speed and the power we also offer value in quick access to market and services and those things. So I think that there is -- there are opportunities for regaining the slope, the decrease of transistor cost. There are opportunities for regaining it. For instance, I think that the recent developments in EUV, I think, are encouraging and they do have -- they will have an important role to play if those developments continue to progress. Now on the other hand, the value that the new node offers in performance and power is exactly what our customers are looking for, and I think that yes, the transistor cost reduction has not been so great, but I think the acceptance of the 20-nanometer SoC will prove -- and I'm certain of the exceptions. I think exceptions will prove that the customers are still getting value. As to the return profitability of 20 and 16, while I'm not prepared to offer you a quantitative measure of the return profitability but I will repeat our goals in this 5-year period, I think starting 2012 to 2017, I guess. That is 10% per year compounded annual growth in profit before tax and return on equity of 20% or higher. And those targets, I announced 2 years ago, I believe, and we remain fully committed to those.

Okay. Next question also comes from the floor from Deutsche Bank's Michael Chou.

Chairman, so Q1 and Q2 sales momentum is stronger than seasonal. So do you think the inventory risk will be rising by the end of Q2, or you think the demand can digest most of the output?

Inventory, whether or not inventory risk is going up, whether the supply chain can digest the inventory this year.

Well, our inventory estimate is that it will only be slightly above seasonal in the next few quarters. Is that the question?

But can you give some color for Q2 outlook by segments?

Pardon me?

Q2 outlook by segments.

The outlook of different applications in Q2, segment by segment?

Yes.

Maybe, Lora?

Yes, I can comment. Okay. I just announced the guidance, that we will grow 17%. For us, every segment will grow across the board, with the communication will grow the most, mainly because the mobile devices.

Okay. The second question is -- I'm sorry...

This is your third, so we have to guard you on that one, okay. Okay. Next question also come from the floor and that will be from Citibank's Roland Shu.

Chairman, first question is the on-base [ph] application processor has been widely adopted on the mobile applications. So can we get your view about how do you think about on-base [ph] application processor on PC, notebook or even server? Do you think that will be happen very soon or that will maybe take longer than expected time?

Whether or not on-base [ph] processor can enter into the PC space, that's your question.

I think that, certainly, it's a -- I think it's a possibility and I think, in fact -- well, it's a possibility but that's not the first thing that comes to mind. I think the first thing that comes to mind is servers, though, and so I think some of our customers have already -- are already targeting that. Some of our customers are already targeting that.

So before this adoption, the ramp-up, what kind of the barrier or what kind of bottleneck on technology need to be resolved? And how TSMC can help to boost this kind of conversion going forward?

The what?

So your question is, what kind of bottleneck that exists from our current customers' offering that if they want to go to PC, right? And what TSMC can do to help them de-bottleneck.

I don't think I can answer that question very well. I think that it is something that I can't answer that question very well at this point.

You have a follow-up?

Yes. I think the follow-up is, I think for -- recently, TSMC just had that news release talking about tape out with ARM Cortex-A57 for 64-bit application processors. So this is going to tape out TSMC's 16-nanometer FinFET technology. So my question is, except for this Cortex-A57 16-nanometer FinFET, do you have any 64-bit ARM application processor on 20-nanometer or 28-nanometer technology?

Well, just as I said, I think the question is whether we have any 64-bit applications in -- just as I said, I think some of our customers are targeting the server segment. So that's where the ARM, where the ARM architecture is, so...

Okay. I think we should go to the call for the next question. So operator, could you proceed to the first caller on the line?

Dan Heyler from Bank of America Merrill Lynch.

I had 2 quick questions. First, Dr. Chang, in terms of the cadence where you've noted a pick-up, especially on 16-nanometer and your very strong growth in mobile processing and SoCs, what does this mean for the N minus 1 fab? You're noting significant growth in your specialty processors. Are we seeing a cadence pickup in the mid-tier technologies, such that you'll be able to keep the N minus 1 fabs relatively full going forward?

So Dan, your question is, in our leading-edge technology, the cadence has become faster and therefore, your question is, whether or not the same cadence become faster for the N minus 1 technologies.

For the N minus 1?

N minus 1. So 1 or 2 generations cadence, the bigger nodes. So 20-nanometer and then 28-nanometer is N minus 1. 40 is N minus 2.

Well, cadence, I mean in the past, it was 2 years. I mean, that's history already. So how does that become faster? I don't understand.

Yes. So that's exactly my question. So I'm wondering if there will be an excess capacity in the middle end of your technologies, with the leading edge picking up and the mainstream technology remaining at 2 cadences.

Oh, the specialty technologies, will they migrate faster than they used to? If they are -- right now it looks like some of these specialty technologies are skipping nodes. They're migrating from 0.18 to 90, perhaps, rather than to 0.13 first and then 90. But they are going from 0.18 to 90 directly and they are going from 0.13 to 65 directly, skipping a node. That's happening. I don't know whether that's what you're asking or not. Is it?

No, I'm -- well, the crux of the question is, how do you keep your mainstream fab full? When everything is moving at the same pace, you can keep your mainstream fab full. So will there be a challenge to keeping your middle-end technology fab full? How do you address it?

Well, I think you have come very close to the heart of my management problem, all right? Keeping the mainstream fabs full is almost as important as advancing the leading edge and, well, I understand your question but if I tell you the answers, I will be telling these to my competitors, also. So I'm not going to do that, yes. But basically, we can already guess -- I mean, why do I talk about these specialty technologies? The specialty technologies will keep the mainstream fabs full, hopefully.

So Dan, you don't have the second question, right? All right, so then we come back to the floor. The next question comes from the floor of Barclays' Andrew Lu.

Dr. Morris Chang and Lora, I have 2 questions. Last time you mentioned 20-nanometer in next year revenue will be larger than 28-nanometer last year. How about the percentage in each quarter, which means the total percentage revenue will be higher compared to year 2012? For example, last year, first quarter percentage, 28-nanometer is about 5%, but 22% by Q4.

I'm not going to go into those details. But I repeat, I'll reiterate, I stand by what I said last time. That is, that the volume -- of production volume upward of 20-nanometers next year will be greater than 28-nanometer in 2012. That, I stand by. But as to the percentage and so on...

My second question, earlier you also mentioned, migrate to 16-nanometer, I think that will be faster than the normal upgrade cycle, about 2 years from 20 to 16. How about from 16 to 10?

No, it's going to be the same -- the same old slow cadence, 2 years.

Back to 2 years? Won't be longer?

Back to 2 years, yes.

Okay. Our next question comes from the floor of Crédit Suisse, Randy Abrams. Randy Abrams - Crédit Suisse AG, Research Division: I wanted to go back to the second quarter guidance. It's well above most of the industry. Last quarter, I think you suggested first quarter was stronger, so it set up a higher base. So maybe what's driving the increased optimism? Whether it's marketing...

Presently, I found myself to be mistaken. Yes, 3 months ago, I thought the higher base -- in fact, even 6 months ago, yes, 6 months, I thought the higher base of the first quarter would make a -- significant growth in the second quarter very difficult. But presently, I found myself to be too pessimistic. Randy Abrams - Crédit Suisse AG, Research Division: Maybe to follow on that, what drove that change where now you think you're mistaken from your prior view? Was it market share gains or you saw incremental market momentum? And do you think sustainability into second half, where the last couple of years, we saw a bit of dip after a strong first half?

Well, it is basically mobile products and market share gain. The mobile products, actually, there is 1 piece that I must mention, which we did not foresee very clearly 6 months ago. That's the China piece. The China piece is playing a pretty important factor in my present surprise. Randy Abrams - Crédit Suisse AG, Research Division: The second question, on 450-millimeter, it's pretty far out, but we're seeing Intel already spend about $2 billion to $3 billion on a 450-millimeter shell, when do you expect...

On what?

On 450-millimeter. Randy Abrams - Crédit Suisse AG, Research Division: 18 inch. Yes, when is your expectation you'll have to start spending CapEx? And do you have any view on fab location at this stage?

Have we started to spend CapEx yet?

Not yet, we have to start to spend R&D. And we have a small team working on 450 but not really to start on the meaningful way in CapEx.

I think his question is when do we plan to start CapEx? Well, I think our schedule is 2016, isn't it?

In the 2016 time frame. That will be the time we start to spend CapEx.

So it's too early yet. Well, yes.

Randy has a part that -- you also asked for the location, right? Randy Abrams - Crédit Suisse AG, Research Division: Yes. If you could suggest what you're thinking about for fab location?

Well, actually, we -- didn't we mention that when we acquired the -- the Zhunan land, yes.

Yes, the land we acquired in Zhunan is the R&D site for the 450 but not the production site.

Not the production site, yes. It's going to take some time. These things take a little longer now than they used to. I mean, each advanced node of technology and the increase in the wafer diameter it's -- well, you understand that I think that only 3 companies can afford to follow it through now, maybe 3, certainly including us. And because these things take more time and require more resources, both money and people, talents. So the 450, certainly, is going to take longer, the transition from 300 to 450, is certainly going to take longer than the transition from 200 to 300.

Okay. I think our next question, we should take our next question from the call. Operator, please proceed with the first caller on the line.

Mehdi Hosseini from SIG.

Dr. Chang, thanks for providing some color on the number of CapEx for 20-nanometer. Korea, could you elaborate on the types of customers or the number of customers or any color that you could provide on the 20 tape out that you have so far for 20-nanometer? And I have a follow-up.

The -- some color on the customers for 20-nanometer?

The types of customers that add up to 20 tape outs for 20-nanometer.

The type of customer, yes. I was going to give you those. Now we have the -- our traditional leading-edge users, the graphics, the FPGA customers and -- but recently, we have added another class of customers, that's smartphones and tablets. So the mobile product users. So those are the main classes of customers that will provide the 20 tape outs for our 20 SoC.

And the follow-up I have is actually on 20-nanometer. Can you help me understand how important is interposer to the economics that 20-nanometer would provide, economics versus cost versus -- cost versus economics. Does interposer really make a big difference to what your customers could get out of 20-nanometer?

That would be related to our CoWoS, right, the Silicon Interposer?

Yes.

What was the question anyway?

What kind of benefit or importance that the interposer, the CoWoS, is to our customers at 20-nanometer compared to their cost?

It's -- basically, it's just -- the answer is, the integration -- it's kind of a Moore's Law on the circuit board, so denser packaging.

So if interposer is not ready, is that going to make a big difference to your customer as they evaluate cost and benefits?

Is that going to be an important element when our customer evaluate the 20-nanometer?

I think so, yes. Well, we already have customers using it, yes. And I believe they evaluated them that way, yes.

All right. Now we can come back to our floor. The next question comes from HSBC's Steven Pelayo.

The foundry industry is roughly thought to be about $40 billion, $50 billion. If it grows about 10% this year, that's equal to about what you're going to grow. So what does this mean for the rest of the industry? Are we seeing the competition even get to become a second-source opportunity at 28-nanometer this year, or is there nothing left over for them?

I have enough on my plate, myself, to worry about other foundries, so...

Well, maybe, I should ask it this way. Are second sources starting to become viable enough that maybe they could cause some pricing pressures, at some point, in 28-nanometer this year?

Well, yes. There are second sources on 28-nanometers. But second sourcing is not as simple in the foundry business, though. You have to work with -- the customer has to work with the foundry for quite a long time before you can use the foundry. So this is not a commodity business. A commodity business, you can set up second sources very quickly but this business, you can't. But still, having said all that, yes, there will be second sources. But if you look at 28-nanometer and I know that the analyst reports, the press have been talking about second sources and competition for TSMC in the 28 node. But the fact of the matter is that this year, which is the second full year of production and ramp-up for us, even this year, we see relatively little competition, and we will still have a very high market share of 28-nanometer.

And just as a quick follow-up question. You talk about 28-nanometer growing each quarter. You talk about it tripling year-on-year. You give capacity numbers for 300-millimeter and 200-millimeter. I'm curious, what is your 28-nanometer capacity today and what do you think it'll be at the end of the year? I don't know, on a monthly wafer starts per month, something like that.

Do we reveal that, Lora?

We do not reveal that.

All right, then, if you could just remind me, what -- when is 20-nanometer revenues going to start being 1%, or 2%, or 3% of revenue? I think -- I forgot if you guys could comment on it?

20-nanometer?

Yes.

20 has not started.

20 has not started.

I know. When do you expect?

When? First half next year.

You're asking about 20?

Well, she answered already, she said first half of next year. You actually set a pretty low hurdle. You said 2% or 3%, right?

What quarter will be the first few percentage of revenues will come from 20-nanometer for TSMC?

What quarter will be the first 2% quarter?

Well, based on our current estimation, it will be roughly second quarter 2014.

All right. Next question also comes from the floor, and that will be from Morgan Stanley's Charlie Chan.

ASML yesterday commented that they are making good progress in EUV throughput. We are wondering if TSMC is seeing a similar trend and then, will that -- TSMC change the timing of adopting the EUV technology? And lastly, can management give us a sense, what is the cost comparison between EUV and then non-EUV, for example, at the 16-nanometer?

Can you repeat that question?

Okay. So you are asking us to update on the EUV progress, right? And then, the second part is the cost of using EUV versus not using EUV at 10-nanometer, or..?

Yes. Maybe current nodes or the future nodes.

Current nodes, we don't use EUV. Future nodes, it depends on when EUV will be available. But your question is comparing the cost with or without EUV?

Yes, because it ups [indiscernible] to your timing of whether you want to advance your EUV adoption timing. If you want to use it for 16-nanometer then a competitors percentage should be at 16-nanometer [indiscernible]?

Well, all right, I think I understand the gist of the question. So let me try to answer. Actually, that question is probably more appropriate at the ASML Analyst Call but let me try to answer what I can, anyway. EUV recently has had a breakthrough. The wattage, the power source wattage has now gone up to 40 watts. Now that will allow a throughput of some 30 wafers, 30-something wafers per hour, which is still far from enough. To make it economically desirable, we'll need over 100 wafers per hour throughput. And that will require a power source of more than 100 watts. So there is still some distance to go but ASML is optimistic. And certainly, I mean, we certainly cheer them on. We are a stakeholder of ASML in more than one way. We are a financial investor, we are also going to be a significant user and, hopefully, significant beneficiary of the EUV. So -- but all right, so now when are we going to use it? I think if we are optimistic, I would say that we will be using it, at least partially, in the 10-nanometer node. And anyway, I think I have given as much answer as I can. Yes.

Okay. Next question will actually be coming from the line. Operator, could you please proceed to the next caller on the line?

Brett Simpson from Arete Research.

For Dr. Chang, I just wanted to ask, on Intel, we saw this Altera deal and there's rumors Cisco is going to follow. And Intel, on the call yesterday, was talking about this foundry strategy where they're crawling today but they want to walk and then run over the next couple of years. I'm just interested, how do you view Intel as a long-term competitor, and how do you see their efforts to open up their fabs to selected customers?

I didn't hear the entire question, so...

So your main question is, how do we see Intel as a competitor?

I think he also asked about Altera, didn't he?

Yes.

So it's Altera, Intel as a competitor, and...

No. Altera going to Intel for 14-nanometer. So Intel also announced the plan to come into foundries, so how we see them as compared to...

Anyway, 2 parts of the question, one is Altera, the other is Intel, right? Altera. I very much regret Altera's decision to work on the 14-nanometer with Intel even though the financial impact is relatively small and Altera remains a major and valued partner of TSMC's. We have gained many customers in the last few years but I really hate to lose even a part of an old one. We want them all, really. I regret it and because of this, we have thoroughly critiqued ourselves. If there was a thing like an investigative commission on what happened, we had it. And there were, in fact, many reasons why it happened and we have taken them to heart. And it's a lesson to us and I don't think that we -- at least, we'll try our very best not to let similar kinds of things happen again. Now -- but I just want to emphasize that Altera remains a major and valued partner of ours and I say this with Altera's concurrence. Now Intel as a competitor. I noted Intel's yesterday's analyst call, in which they said they have now gone from crawl to walk and then will go on to run and that's all very beautiful in metaphor. And then they also raised several conditions or several criteria -- 3 criteria but there's really nothing new in those and they have said those before. And I still view Intel as a selective picker among customers. As a foundry competitor, they will pick their targets and so on. And I don't view them as a general competitor because they have already said, through the 3 criteria that they used, they have already said themselves that they will not be a general or front [indiscernible] competitor. But they are a very serious competitor to our customers. I mean that, really, I would say, applies even greater pressure on us than they as -- than Intel as a direct foundry competitor. They are a very serious competitor to our customers. Our customers rely on us. That is very serious pressure and we respond to that pressure, of course. We respond to that pressure. Yes.

Maybe just a follow-up question, Dr. Chang. On the mobile guidance you've given for second quarter, is this all being driven by existing customers or are you seeing something new in your customer mix going forward?

Your question is whether or not the strength of our second quarter business comes from new customer, or...?

In mobile, yes.

In mobile, or?

No. Well, I mean obviously, something comes from new customers every quarter, but not big. We get new customers all the time but big customers are not -- I mean, second quarter strength is not due to 1 or 2 big, new customers, no.

That's very helpful. Can I maybe just ask 1 final question for Lora on depreciation, just to get a sense for -- or through the rest of this year, how might depreciation trend? Because it was flat in the March quarter.

Okay. You are asking about the March quarter. In general speaking, with the CapEx guidance, Chairman was talking about TWD 9.5 billion to TWD 10 billion. We expect whole-year depreciation will go up around 23% on a year-over-year basis.

Okay. Now we are coming back to the floor. The next question comes from Daiwa's Eric Chen.

Dr. Morris Chang, my first question regarding to the gross margin. I saw that for the 28-nanometer process, we took negative 4 to 6 quarter to have the higher than average gross margin. So for the 20-nanometer process, do we have any schedule or internal trend?

So Eric, your question is how long will it take for TSMC 20-nanometer to reach corporate gross margin? How long will it take, right?

Yes.

Let's see. I think that we have kept some statistics on these sort of things. Very interesting. I think it takes -- it took 6 quarters. 40-nanometer took 6 quarters -- 8 quarters, 7 or 8 quarters and 28-nanometer is taking about 8 quarters. And so you asked how long is 20-nanometer going to take? Well, I only have history to guide me, all right?

So how do you think, I mean, take a guess. How you think -- how long that will take?

He wants you to take a guess.

Well, I think history is my best guess.

So that means the 7 to 8 quarters?

Yes.

And my second question, regarding to the CapEx. And we see the CapEx is pretty centralized on the top 3, the semiconductor maker, and I will say that for this year, probably over 70%, even 75% of CapEx among these 3. So how you think about this kind of CapEx, the centralization? And how does TSMC look at the other 2 competitors in terms of CapEx? And then, the final is regarding to the 28-nanometer process competitor. So Lora, if you don't mind, could you give us an idea in terms of the Q1, Q2 and the whole year 28-nanometer process, the competitor?

So Eric, you have question here on CapEx. You said that you observed CapEx tend to be centralized on the top 3 players, and...

Top 3 what?

Players, the top 3 companies.

Intel, Samsung and TSMC.

Intel, Samsung and TSMC. So your question is, whether or not this is -- what the implication of these trends to...

; Yes. I would like to know this kind of CapEx, the centralization, will [indiscernible] own node? And how does TSMC, and to be asked anyone going forward in the amount of 3 [ph]?

Well, it looks like only 3 companies can afford to keep investing. That's what's you're referring to, really.

Right. So what's the other 2? Any strategy, I mean, for the CapEx we see the...

You mean TSMC versus the other 2?

Yes, in terms of the CapEx.

Whether our CapEx will react to the other 2 companies' CapEx?

Whether what?

Our CapEx will be responsive -- responding to our competitors.

No. Our CapEx is only responsive to our own needs.

Okay. So we are not going to see that CapEx over competition going forward?

CapEx war?

Over competition.

I never, never, never in the last 20 years engaged in any CapEx war with anybody. In the last -- ever since we started the company, our CapEx has always been responsive to just one thing, that is our own needs. Nevermind what the other companies spent.

So Lora, your 28-nanometer.

You're asking the capacity for each quarter for 28-nanometer?

Roughly an idea, yes.

I'm afraid I cannot tell you. But what I can tell you is, we are growing our capacity each quarter for 28-nanometer. But in terms of overall CapEx spending, I think we're towards the end of spending by the end of this year for 28-nanometer.

Okay. And very quick, the 28-nanometer processing, same quarter in turnover quarter-on-quarter competitive growth, how many percent will it be, roughly?

I think it will be in line with our revenue growth. Okay?

All right, if there's no other questions, we will end our investors conference and conference call right now. Thank you very much for coming, and we'll see you next quarter.

Thank you.