Taiwan Semiconductor Manufacturing Company Limited (TSM) Q3 2013 Earnings Report, Transcript and Summary

[Chinese] Welcome to TSMC's Third 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 and today's event will be as follows: First, TSMC's SVP and CFO, Ms. Lora Ho, will summarize our operations in the third 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. With those participants on the call, if you do not yet 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 joining us today. I will start the presentation with financial highlights for the third quarter, and I will follow that by providing the guidance for the fourth quarter. In the third quarter, TSMC set another record in both revenue, and net income, thanks to our leadership in leading-edge technologies. In the third quarter, revenue increased 4.3% to reach TWD 163 billion. Gross margin was 48.5%, down 0.5 percentage points sequentially, due to the lower capacity utilization, offset by favorable inventory valuation adjustment. Total operating expenses was TWD 19.3 billion, slightly increased from the second quarter, mainly due to higher R&D expense for 20 SoC and 16 FinFET technology development. The operating margin was at 36.7%, down 0.3 percentage points from the second quarter. Non-operating items was a loss of TWD 0.3 billion in the third quarter, mainly due to the TWD 1.35 billion impairment charge associated with our investment of Stion. Stion is a U.S.-based solar company which we invested in 2010. The EPS impact for this event is about $0.05. Overall, EPS was $2 in the third quarter, and ROE was 26.8%. Let's take a look at our revenue by applications. After 5 consecutive quarters of growth, communication segment declined by 3% in the third quarter, as consumer -- customers start to manage down their inventories. Computer segment declined 18%, Consumer segment grew significantly in the third quarter, thanks to the strong demand from the game consoles. Meanwhile, Industrial and Standard-related revenues grew by 5% from the second quarter. Despite the volatilities across different applications, our 28-nanometer continued to grow and contributed 32% of total wafer revenue in the third quarter, up from 29% in the second quarter. Combined with 40-nanometer, the advanced technologies represented 52% of our total wafer…

Good afternoon, ladies and gentlemen. My message is outlined here on the screen. First, a few comments on third and fourth quarters of this year. Third quarter was another record quarter for TSMC, both in revenue and in EPS. In a period where sales of certain mobile products was slowing, we are pleased with our third quarter results, as they demonstrated once again TSMC's strong position in the leading-edge technologies, particularly in the 28-nanometer node. As I said 3 months ago, the fourth quarter may be a down quarter, lower than the third quarter, because we expected the supply chain to take serious actions to manage the inventory in the second half of this year. That has happened and is still happening. As the CFO has just indicated in the guidance, our fourth quarter revenue will decline by about 10.5% from the third quarter. This decline is mainly attributable to the softer demand for certain high-end smartphones and the inventory correction. We believe the decline is short term. Meanwhile, TSMC's structural profitability, our technological strengths and our close customers' bonds remain intact. And we are optimistic about 2014. Next, a few comments on industry outlook, supply-chain inventory and mobile products market. For the full year 2013, we estimate that semiconductor market, the world semiconductor market, will grow 4%. This is a bit higher than the 3% that we estimated last year -- last quarter, mainly due to the strength of the memory segment of the semiconductor industry. We again estimate that the fabless industry will grow 9%, that's unchanged from our last quarter estimate. We estimate that the foundry industry will grow 11%, unchanged from last quarter. And we now estimate that TSMC this year will grow between 17% and 18%, which of course, is much higher than the foundry…

Yes. This concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to limit the number of questions that you have to 2 at a time to allow all participants an opportunity to ask their 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 in English, before our CEO or CFO answers your question. For those of you on the call, if you would like to ask a question, [Operator Instructions].

Now, let's begin the Q&A session. Our first question comes from the floor, and it will be Bank of America Merrill Lynch, Dan Heyler.

Chairman, thanks for taking my question, and we hope you'll come back frequently and participate in the quarterlies. A couple of questions. First, on the 16-nanometer, you have made an accurate prediction of your position in 28 in 2 respects. You had said that TSMC would extend your lead at 28 several years ago, and that has occurred. And you'd also indicated that you would actually be with your partners more competitive than the 2 biggest foundries, who were at 22. So could you offer us to forecast of 20 and 16 on both metrics, number 1, will you extend your lead relative to your competition in your view? And number 2, where do you rest with your partners relative to the IDMs on 16 and 20?

The question as I understand it, is whether we would extend our lead in 20 and 16. On 20, well, definitely, I think on 20. On 16, I think the battle is still raging. On 20, I believe that we will start with a very high market share, just like we did on 28 and so on. And we'll keep that high market share for quite a while, several years. On 16, I believe that we do have the serious competitors, the gorillas that I mentioned, and -- but my goodness, we intend to prevail. But I'm not going to tell you now that we have already won, all right? I hope to tell you in, let's say, 1.5 years, I guess, that it's -- that we have already won. But on 16, the battle is still raging.

Could you offer a bit follow-up on that, a little bit color on 16, because obviously, it involves what TSMC does internally, it also involves your ecosystem partners in achieving your advantage. So maybe elaborate a bit on what more is going on at 16 to further, given how competitive it is, perhaps intense-ed alliances, or closer collaboration with the ARM ecosystem or perhaps more color there on how you can pull ahead?

Well, I think that what's going on with our competitors, well, I think...

Ecosystem partners. With our ecosystem partners on the 16-nanometer development plan.

With our ecosystem partners. Well, I don't think I'm an expert on that. I'm however, you know who our ecosystems are -- people are, and we do have Cliff, I think. Well, not I -- nor just I think, I know that we have Cliff that's in charge, responsible for that part of working with our ecosystem partners. And as you know, ARM is a very important one and -- but others -- and others -- other EVA and IP providers are also very important. But more importantly I think, that our large customers don't need the ecosystem as much as the middle-sized customers. And that's why the bond, our bond with the large customers becomes very important. And we have really, I believe, worked very hard. I -- even 15 years ago, I started to say that we had the 3 major strengths: Technology, manufacturing and customer partnership. And that's true today. Technology, manufacturing and customer partnership. And we later kind of rotated between the word partnership and the word customer's trust. To us, it means the same thing. So the large customers are most important, and they don't rely on the ecosystem as much. They do use it, they don't rely, because they have their own special stuff.

That's great. And second question for Lora would be on the EPS growth comments that you made earlier getting probably back, about back to trend, if not exceeding trend on the EPS growth. Does that require that your gross profit, structural profitability, will actually increase next year to achieve that target? Or can you sustain -- achieve that EPS growth with maintaining your current structural profitability?

Dan, we are confident we can either maintain or slightly improve our structure profitability. I know many analysts was concerned about our capital intensity being so high, how can we maintain our structural profitability? I have went through the number myself. So let me explain it to you this way. Our depreciations, there's about 85% of our depreciation that goes to COGS. If you just look at the depreciation increase, you will see, "Oh, year-over-year increased so high. Now this year versus last year is around 20%, next year will be more than 35%. So how can you sustain your structural profitability?" The thing is, we will invest in capital expenditure, actually we're adding the capability to do more business, so our quantity will also increase. So if you divide it by the unit cost with depreciation cost, actually it doesn't increase that much. Other than that, since we have migrated technology to more leading-edge, so our average ASP will go higher, as you have seen the past few years. In addition to that, with the continued productivity improvement from our operation people, we were able to drive the non-depreciation costs go down on a per annual basis. That means we can maintain the structural profitability. So make it simple, if we invest in CapEx, and those capacity get utilized, and we continue to drive the efficiency, and then we can maintain the same structure, and so does EPS will growth. I hope my explanation is clear.

All right. Next question also comes from the floor, and it would go to UBS, Jonah Cheng.

[Chinese] Very quickly for my first question, I think last time the Chairman, highlighting something about second -- Q4 will be a bit slow down. And right now, it seems the pace of inventory already were back to normal in end of this year. Can we assume in that in Q1 next year, we may see the normal seasonal pattern, like roughly flattish and start to rebound in Q2, this kind of normal pattern, or still something we need to watch?

Jonah's question is whether or not we will have a normal seasonal 1Q '14, or...

Anything different we still need to watch.

Or anything different?

What's your idea of a normal seasonal?

Okay. My idea for normal is roughly flat to down slightly Q-on-Q.

Flat to slightly down.

Well, it's a bit early to predict first quarter, but I do want to say that with the introduction of mobile products, the mobile products have a seasonality pattern that's different from what you describe -- from our traditionally normal seasonality pattern. And so I would not rely on that normal seasonal pattern. And as to what the first quarter will be, I think that it's a little too early to forecast at this point. But seasonality refers to what happens during 1 whole year, what happens quarter-to-quarter. I have said earlier that we are very optimistic about next year will be a double-digit growth year, the whole year, the full year. Now, if the first quarter comes in a bit low, you will have a surge in the second and third quarter. We'll rattle our teeth. And in fact, the mobile product's fourth quarter, well, this time in this year, was fourth quarter is a bit down, 10% down. But that's mainly because of the inventory adjustment and because, as I said, of the slowing of just a few.

High-end smartphone model.

High-end -- sorry, I just want to use the same words I used before. I know what they are, but I just want to use the same words to describe them. If you -- high-end mobile products. But that does not necessarily reflect the normal pattern of mobile products. So I don't know, I guess, my answer to you is summarized by the following sentences. You cannot rely on the normal seasonal pattern. And we are optimistic about the year as a whole.

Okay. And my second question is about the Chairman just mentioned, you are still hands-on after the June next year. So what we're trying to know is for the investment relations side, can we keep hearing from you, if in that July next year, you can still attend this meeting and then provide your views to us? That may be something we want to know.

He's making a request for Chairman to still come to the investors’ conference after relaying the CEO position.

Thank you for the invitation. And I will consider it. Thank you. Just to show how hands-on I will be.

We are expecting it.

Okay. All right. Next question comes from the floor again, it's Deutsche Bank's, Michael Chou.

So one question regarding 20-nanometer gross margin equipment pace. Given that you forecast 20-nanometer for the first year should be faster than 28-nanometer, so for the first 12 months, right? So can we expect your gross margin in 20-nanometer equipment pace will be faster than 28-nanometer?

The question was what will 20-nanometer margin look like?

You are asking what the 20-nanometer gross margin is like compared to 28-nanometer gross margin? As Chairman just talked about, the ramping profile for 20-nanometer will be more aggressive than 28-nanometer. So we expect 28-nanometer -- 20-nanometer will contribute quite significant revenue in 2014, because of our skill ramp so fast, so we can expect the margin growth will be faster, in a much faster slope than 28-nanometer if you compare to the same period versus same period.

Michael, your question really is, whether or not the gross margin will also improve faster for 20-nanometer than for 28-nanometer, the speed of gross margin improvement.

Well, I think the answer is yes. I think -- actually I said earlier that up to this point, the 20-nanometer, V0 improvement has been on a faster path than the 28-nanometer. So there's no reason to expect that, that will stop. I think it will continue, the yield improvement pace, that was the question? Yes. That was the question? I'm sorry, I don't see any -- that was the question, right? So the answer is, yes.

Second question is regarding the long-term -- or I should say, your profit before tax target of 10% growth. Given that the addressable market of the mobile devices should decline in terms of growth rate going forward, so if you continue to improve your structural profitability, the second driver could be a bigger addressable market, or you can see any new growth driver for your profitability?

Can you repeat the question for me?

The driver for structural profitability improvement, what are the...

Okay. Given that Lora just mentioned structural profitability improved, right? But the addressable market of this mobile devices should start to see the slower growth rate, so how can you address that to improve or to capture your profit before tax of 10% growth?

So, right, given the slowing growth of the smartphones and tablets, why can we still achieve PBT growth rate bigger than 10%? Given the growth drivers, which is the mobile products, the rate of growth is slowing, can we still achieve -- well, how can we still achieve PBT growth bigger than 10%?

Well, actually, you'll notice that in the past 2 years, in 2012 and 2013, our PBT growth has been higher than 10%. We had a setback in 2011, and we had to make up. So in 2012, if you show that slide again, the dotted line -- the upper dotted line is that 10% growth line, and we had a setback in 2011. And so in 2012, we had to grow faster than 10%, PBT had to grow faster than 10%. And then in 2013, it grew faster than 10% again. And now, in 2014, I'll drew a kind of hazy ellipse here to show where it might fall, and my expectation is that it's going to -- we're going to more than catch-up in 2014, and we're going to do -- by the time 2015 comes, I think we'll be above that. So now, we're going to grow double digit, I think, comfortably double digit next year, not just barely double digit, but comfortably double digit next year, okay? And it's our intention to grow our PBT proportionately. So you say how can we do that? Well, gee, that's complicated. You can't just -- it's a mistake to try to analyze our company one technology at a time, or one customer at the time. I run a company as a whole, and parts of the company may be down, but parts of them were up. And I just want to be sure, that's my hands-on, that the up parts more than balance the down parts, okay? So just to try to say that the mobile part -- the mobile part is -- mobile product growth is slowing down and that's why maybe our revenue growth next year will be a bit less than this year. This year is 17% to 18%, and next year, maybe a bit lower. But it will still be comfortably, as I said, double digit.

If I may, some comment on the smartphone growing, you are concerned about the highest smartphone growing is slowing down. But we still believe smartphone is still going to grow, at least for the 10% for this -- next year and the year after next. For overall smartphone, where we have very high penetration, on mid- to low-end, we believe the overall smartphone will grow 35% this year, and 25% next year. So our market position on the smartphone still will be a very big driver for TSMC's growth.

Okay, all right. I think it's about time that we should take our next question from the call. Operator, please proceed with the first caller in the line.

The next question comes from the line of Steven Pelayo from HSBC.

Two questions, one very near-term and one much longer-term. More color, if you could provide some more color on your guidance for the fourth quarter, revenues down about 10% or 11%. I'm curious, is that all segments declining? Can 28-nanometer maybe still grow in dollars? Will all segments be down, which one would relatively outperform? Can you give us some more color and detail on both by segments as well as technology nodes relative to your overall guidance of down about 10%, 11%?

Steven, your first question is to provide additional color on the fourth quarter's 10.5% decline, whether or not the decline is all from the decline of the 28-nanometers, that's your first question, is that right?

No. Actually my question was, can 28-nanometer still continue to grow in the fourth quarter, or alternatively, will all segments be down? Can you provide us a little bit more detail by both technology node, as well as by Communication, Computing, Consumer, will all those segments be down or any of them relatively outperform? So I'm looking for just more general color by technology and by segment on the fourth quarter guidance.

Okay, Steven, you're asking about the segment changes on the fourth quarter. Of that 10% growth -- decline for TSMC, we will see most decline will be in Consumer segment, kind of most, followed by the Computer, and followed by the Communications. In terms for the standard actually is historically down or flattish. And Steven, we do not give you any guidance on a particular node, sorry. And then what is your second part of the question?

My next question is much longer-term for the -- for Chairman there. You obviously have a much better crystal ball than all of us having seen the mobile opportunity in 2009. You just shared some forecast for smartphones and tablet growth rates falling from the 35% and 50% growth rates to the 20% to 25% next year, and I think IDC is talking about 10% to 15% in 2015. So I'm curious, as you're looking out beyond 2014, what are going to be those next big opportunities to drive growth and perhaps, just as important, will TSMC need to spend the greater than 40% CapEx of sales, that they had to over the last few years, to capture whatever those opportunities you see?

All right, so your question...

Actually, before I ask you to repeat Steven's second question, I want to add to Lora's answer to his first question, all right? His first question was whether the 28-nanometer would decline in the fourth quarter, and Lora didn't really answer that. Well, you said no, I don't mind answering it. Well, remember, we said that the -- one of the main reasons for the fourth quarter decline is the high-end mobile products, the slowing of, and 28-nanometer, our 28-nanometer, will decline just a little bit, all right? Not 10%, but just a little bit. So that's okay.

All right, the second part of Steven's question is how does Chairman see our growth drivers beyond 2014, what will be the other growth drivers beyond the mobile products for us. And then, how much CapEx we will be spending to capture those type of growth beyond 2014?

I see. Well, very difficult question. On the outlook for beyond 2015, well, look, we made a 5-year plan, and I thought that was pretty far-reaching back in 2010, and nobody even remember that until today when I brought it out again, okay? So now, I said in my autobiography that as far as semiconductors are concerned, I am an eternal optimist. I believe that people will keep finding new applications for semiconductors. Now it's the mobile products. Now the next one may well be something even more miniature and something even more convenient, now like the wearables, I like the watches and the -- but I think it could well be something even beyond that. But you -- I don't think one can foretell what the future holds for us in semiconductors. I know that it will be good, because it's a fundamental component. It's just as indispensable to the world as almost food is, I would say. Now, of course, people can never go without food, but in this information technology world -- and information technology was what drove the economic progress in the last 30 years. And in the information technology world, semiconductors are indispensable and people will keep thinking about -- keep thinking of -- thinking up new applications. So I'm optimistic about years beyond 2015, but I cannot tell you what the exact applications, where the exact applications will be. Now was there another one, another question?

Steven, thank you for your question. Now, we come back to the floor. The next question comes from the floor of Crédit Suisse, Randy Abrams. Randy Abrams - Crédit Suisse AG, Research Division: I wanted to ask, you talked about the double-digit profit before tax growth, could you talk about now your expectation on CapEx? And also, there's some talk about EUV continuing to be a bit slower. So just how that affects capital intensity and timing for EUV?

You should repeat those.

All right. Randy asked Chairman to share your view about our future CapEx, and given EUV appears to be a bit slower, what would be its impact to our CapEx and capital intensity?

EUV will be...

Slower, delayed.

EUV will be a little slower? Well, okay, we'll get to that, all right. I want to ask you where your information came from? Well, let me first talk about CapEx. Next year's CapEx will still be high, but if you want me just to give a ballpark, I would say, the ballpark is around what we -- is around $10 billion, okay? But I -- earlier, I said that this year's $9.7 billion, give or take a couple of hundred million. So if I gave you $10 billion, I have to say, give or take a couple of billion, okay? So now, it highly depends on what we see the growth opportunities of 2015 will be, all right? But I do think that the capital intensity has either peaked this year -- this year, the capital intensity is about 50%. We have $20 billion of revenue, and about $10 billion of CapEx. So I do think that 50% is about the right peak. So now, then EUV. Well, actually, I just met with our ASML partners yesterday, and they told me just a reverse of what you just said; it has slowed or something. No, I think that the EUV progress is good and we are obviously tracking it very, very carefully. And remember, we are a significant investor in ASML and we also join in their EUV R&D program, and so we're tracking it very carefully, and we are pretty sure that we're going to need it sometime in the future. Randy Abrams - Crédit Suisse AG, Research Division: To follow-up, the reason I mentioned, ASML, I think, at their investor conference, perception was throughput was improving little bit slower. So fewer tools were coming out in the next year. So it's a perception that it was just a little bit slower path in terms of improving productivity and throughput. So -- but it sounds like by 10-nanometer, do you expect EUV to start for some of your process steps?

EUV progress itself has not slowed. Well, it had a breakthrough about, I guess, early this year, and before that, it was a bit disappointing. And -- but they had a breakthrough early this year and since then, their progress has been on track.

Next question comes from the floor from Citi's Roland Shu.

First question is Chairman mentioned about we are going to have the second wave 28-nanometer product adoption. So can Chairman spend more for what are the growth driver for this second wave 28-nanometer product? Is this coming from the existing customers, or coming from the existing products, or this is a totally new customer or totally new segment?

Well, until about 3 or 4 years ago, our first wave of customers at any leading edge where graphics and baseband and the PLD and so on. But then, the first wave was taken over by the IC makers, IC suppliers for the mobile products. And the second wave now, is what used to be the first wave. Well, in the first wave, there are some users that are later than the first, first wave. For instance, the providers, suppliers to the China phone market, is later than the provider -- IC provider to the American/ European market, so...

Okay. I suspect, since the original first wave leading [ph] technology adaptor, are many for the computer for QWERTY of FPGA PLD. So this time around, I think the -- for them to become a second wave customers, I think the adapter technology [ph] world, more likely from the high-K metal gate point of view. It's not found the Poly/SiON, because I think that for them, they need more powerful devices. So that means going forward, of growth -- revenue growth for 28-nanometer probably will be even faster than the [indiscernible] growth because with this high-K metal gate adoption, we go for the higher ASP and a higher gross margin. Am I reading you right, or?

I'm sorry, would you...

All right. Roland is making a prediction that our second wave customers filling our 28-nanometer capacity will be the ones that requires high-K metal gate, instead of the Oxynitride. And therefore, our revenue growth of 28-nanometer will be faster because high-K metal gate is more value-added.

Revenue growth will what?

On 28-nanometer.

It will be faster.

Will be faster than otherwise.

Yes, because of higher ASP on the high-K metal gate.

Well, now that I think about it, you may be right.

Okay. My second question is on the 10-nanometer side. I think that Intel is pushing out 14-nanometer production, mass production by 1 quarter to maybe first quarter next year. And personally, I put this, for Intel 10-nanometer mass production probably will be 2 years after 14-nanometer, which probably will be 2016. And then we look at TSMC. According to your product roadmap, now you are going to risk production 10-nanometer from 2015. So is there any chance of -- what's the chance do you think for TSMC 10-nanometer, TSMC probably going to beat Intel in the mass production side.

Would you repeat it, please?

All right. Roland is also making a prediction on -- judging by just recently announced Intel pushing out their 14-nanometer by 1 quarter to next year. So that puts us into 2014. Mass production follows in 2 years, that's 2016. So Roland said, since we start risk production of our 10-nanometer in 2015, is there a chance that we will be ahead of Intel?

Okay. Can I rephrase my question? What's the chance? It's not, "if there is a chance?" What's the chance you think you are going to lead Intel in 10-nanometer production?

I am not going to comment on that, Roland. I'm not going to comment on that. And I never, never underestimate anybody, I would not underestimate Intel. I didn't even underestimate any better -- anybody that was lesser than Intel, then I will not underestimate Intel.

Oh, maybe I can -- in another way. Typically, now how long does it take for 10-nanometer, from risk production to mass production?

Ours?

Yes, yes.

I don't know, 10. He said 10. 10 is 2 years after what -- after '16. 10 is 2 years after '16.

I mean on your -- how long were you to take for the risk production to complete it?

Well, that can be shorter than it traditionally was, that could be shorter than it traditionally was.

Okay. We really have quite a few people on the queue.

I saw Andrew.

I know but there are several people waiting a long time.

Will Andrew comes first.

All right. Andrew comes first.

No, no, no. [indiscernible]

Okay. All right. Andrew first.

I have 2 questions. First one is regarding earlier you mentioned Q1 next year will start to roll mass production, the 20-nanometer in Q1 year 2015 mass production, this mass production is a wafer stock, is that correct?

Yes, well, wafer stock.

Is it wafer stock or...

The cycle time is a given thing, yes. So yes, wafer stocks. Well, wafer stocks...

So can we say that...

Just a minute, I think that we expect we have some revenue, right?

Around second quarter, revenue comes from second quarter, so first quarter must be wafer stock.

So we will have about 3 to 4 months wafer stock work-in-progress time, so we are going to view a lot of inventory in Q1. Can we assume...

Work-in-progress inventory.

Work-in-progress inventory, so that will share some of the depreciation costs in Q1?

Now you're looking into, deep into our accounts, okay. Sure, that's a standard accounting procedure.

The second question is, I remember 2 years ago, TSMC provide 28-nanometer per 1,000 wafer investments, it's about USD 120 million. Can we have rough color for 20-nanometer investment per 1,000 wafer, and 16-nanometer investment per 1,000 wafer?

I haven't reviewed the numbers lately. So I don't -- I can't give you an exact number. You mean, design, design, is that right? Are you talking about per design?

CapEx.

CapEx. CapEx.

Oh CapEx. So you're asking what the CapEx per...

1,000 wafer.

For what?

20-nanometer and 16-nanometer.

Okay. It's around -- approximately 20% higher than 28. But 28 has been reduced since -- you heard the number, okay? And sure, I mean, my goodness, it's a key parameter for us, capital expenditure per 1,000 wafers per month capacity, that's a very, very critical parameter for us. And it's also something that we would love to know about our competitor. So -- and we try very hard. We have standing programs, standing projects that reduce it, and so when we told you what the 28-nanometer CapEx per K per month was, what do we tell you?

20%.

What?

120 for 20.

100 what?

USD 120 million.

Oh, that's been reduced. Yes, has been reduced since then. But the 20, [ph] and I'm telling you now, the 20 [ph] is approximately 20%.

How about 16? Earlier you said it's 20 and 16% now?

We regard them almost as one node. So it's about 20%, but it's not 20% above the 120 [ph] or whatever you -- because the 120 [ph] has been reduced, okay.

Okay, the next question will come from the call. Operator, please proceed to the next caller.

The next question comes from Mehdi Hosseini from Susquehanna.

My first question for Dr. Chang. Have you seen any changes to your Q1 rolling forecast, particularly from the most of your customers that have exposure to China and specifically, post-holiday or the early October holidays?

Mehdi, your first question is, do we see any change in our rolling forecast for 1Q 2014, given our Chinese customers' business after the Golden Week? That's right, is that your question?

Yes that's the question, yes.

You want to answer that question?

It's hard to link that to a Chinese lunar new year, but of course, the 1Q forecast is constantly changing. And also, the fourth quarter is changing. So if you want to say versus fourth quarter, the number is kind of fluid. So we don't see much changes, that's the simple answer to your questions.

Mehdi, do you still have another question?

Yes. And then actually, for Lora, you were talking about the unit cost at 20-nanometer is pretty competitive to a 28, even though depreciation has gone up by 20% plus but unit cost is competitive. What about the 16-nanometer? Would the unit cost actually go down because 16-nanometer is a true shrink compared to 20-nanometer?

I think Mehdi's question is, he wants us to comment on the unit cost between 20-nanometer and 16-nanometer, unit costs of 20-nanometer and 16-nanometer.

Well, unit cost of 16-nanometer is projected to be higher than unit cost of 20-nanometer. But you're talking about cost or price?

Cost.

Cost right now is still projected to be higher than the 20-nanometer.

Okay. Let's come back to the floor. The next question I would -- I think it will come from Morgan Stanley is Bill Lu.

I'm hoping to start with a very broad question. You said earlier, when you came back as CEO, and it requires assistance strategy for TSMC. Now that we are here, in the next, say, 3 to 5 years, do you see any changes that you need to implement as far as TSMC strategy for the next several years?

Whether we will have a different strategy for the next few years.

Well, any changes...

Any change.

Yes. I think not a different direction, but I think we modify our tactics almost all the time. The general direction I think is still the same. The general direction is that this is a technology, ours is a technology business and we want to lead in technology. Now however, to allow that, we have to generate enough income from the existing products, existing technologies. And therefore, we need a good structural profitability. Now, all right, so once we have good structural profitability, we push R&D, we push R&D expenses, we push our R&D, expand our R&D capability to the limit. And of course, to realize all the benefits that technology leadership brings in, we also expand our capacity, our manufacturing capability relentlessly. That has been the general direction, all right. At the same time, in order to make sure that we have customers, large customers all the time, we, as I said, we work every day to earn the customer's trust. So I mean, all of this is, to say, integrated the whole. So that general direction does not change. Now as far as the tactics are concerned, they change quite often. How much emphasis do we put on 1 technology versus another, particularly the specialty technologies, how much do we put on one versus another? How much really can we afford to expand on the -- and how much emphasis we could put on 1 customer versus another? Those are changing all the time. But the general direction does not change. The general direction as I already described it. First, you have to have good structural profitability. Then you push, you expand, you strengthen R&D to the limit of your affordability, and then to realize the benefits that technology leadership brings you, you expand manufacturing. So I mean, just then if you have listened to what I said, you'll begin to understand why our R&D expenses have gone up, why we have emphasized so much technology leadership, and why our capital expenditures have gone up so much. Is there another question?

Yes, there is a very quick one. The Stion write-off, can you talk about whether that means a change in strategy in the solar business, or what exactly is happening there?

Stion.

Stion. So what about Stion write-off? I'm sorry that it happened, what was the question?

I'm just wondering what it means for TSMC solar business overall.

I can take that one, Chairman. Stion is the technology investment and we did in 2010, for the purpose to -- so we can license the CITS technology, and we did that. So we have take the core technology and develop ourselves, and it has been quite successful. So there's no change in that front. The write died, just because the company decide not to continue the operations, so we have to take it off from our book. But since we have greater [ph] technology, and it's perpetual, so there's no impact to our solar strategy and business.

All right, I know there are still people waiting on the queue from the call. So operator, please proceed to the next caller.

The next question comes from Brett Simpson from Arete Research.

Dr. Chang, if we go back the fourth quarter 2012, I think you said back then, after the first year of ramping 28-nanometer, that the gross margins were at or above corporate average. Can you just give us your impression, how do you see 20-nanometer playing out, should we expect we get the same gross margin levels through at or above corporate average by the end of 2014?

Well, what would be the 20-nanometer gross margin be at the end of 2014?

That's right.

Let me take that one. Our 28-nanometer, it takes 7 or 8 quarters, so the gross margin reach to the corporate average, and it has happened in the first quarter of this year. We believe the 20-nanometer will follow the same pattern.

What was she's saying is that, 28-nanometer reached...

Corporate average.

Corporate average. When?

7 to 8 quarters.

7 to 8 quarters. And she's saying that 20-nanometer will follow the same pattern. I tend to think that 20-nanometer will reach the corporate average sooner than 8 quarters, because as I said earlier, the learning curve is faster up to this point than 28-nanometers. So I think that, that will continue. And in fact, I think it's almost necessary for 20-nanometers to learn faster because the ramp-up is very fast. The ramp-up is faster than 28-nanometers. So in time, it will have to take a shorter time for 20-nanometers to reach corporate average. So I think it will -- I cannot answer you exactly what it will be by the end of next year, because the end of next year, will only be the third or fourth quarter but I do say that there will be -- the gross margin will be higher than the 28-nanometer was at its third or fourth quarter.

Okay. And just to follow-up, I think you talked earlier about unit costs, and we can all see at 20-nanometer, this is the first node you start multi-patterning and FinFET will be expensive. If you're expecting to maintain your structural profitability and the cost arising, what do you think the impact of this is going to have on the mobile value chain? Do you think the mobile market can absorb a cost increase, or do you think that chip prices are structurally going to start rising? How do you see the ripple effect from the inflation at a leading edge that's kind of?

Okay. Brett has a very profound question, saying that our 20-nanometer will be using, it will be using double-patterning, and 16-nanometer will be using FinFET, both are expensive, which would have an impact to the mobile value chain. And so how will the mobile products absorbs these higher prices or costs? What would happen to the value chain?

Well, the -- actually, I can talk for hours about this but the -- I believe that the supply chain value, the price -- the end price. The end price to consumers is a pretty elastic one. That means that the lower the price, the more you sell. And now if you are constrained by the cost, your price will have to be at a certain level and you sell less. Now, of course, everybody wants to push it to a lower price, so that they can sell more, but it start as a simple matter of being able to afford it and not being able to afford it.

All right. I think we will come back...

It is a matter of competition or? We just want to be our cost to be lower than competitors. I mean, this -- I come back to the story of 2 people in the camp and a bear, a big bear is approaching. And the first person quickly puts on his running shoes, and the second person says, "What's the use? The bear is going to out run you anyway, it's going to run faster than you anyway." And the first person then starts to run and while he departs, he said to the second person, "All I have to do is to run faster than you, not the bear." So on this price and costing, all we have to do is to run faster than the competitor.

Okay. From the floor, there are couple of media people in the back and they are raising their hands. So we'll give you the microphone but would you please identify yourself. This is the last question.

Okay. I enjoy this, but I think that the -- every good thing, every -- just as the Chinese say, every feast comes to an end.

A reporter from NewGen TV [ph]. I just have 2 questions for you. The first question is being, that you have come quite details of the TSMC outlook in the fourth quarter and the first quarter in the next year. But people are quite concerned of the current economy, and with the complication in U.S. and -- or would you like to share your wisdom on maybe, the next year's economy outlook, and if it has mistake, would it affect your expectation of TSMC in the next year?

Global economic outlook for next year.

Global economic outlook for next year. Well, my goodness. I am -- I don't know what our company official forecast is. Our company forecast that -- this year it's 2.5%, global GDP. And next year, it's 2.8% global GDP growth. That's kind of consistent with my feel. I think that things are slowly improving, I think the U.S., I think in spite of the short-term problem, which I think will probably faces a pretty quick resolution is -- the U.S. is on a recovery path. I look at the new job creation, I look at the housing prices, and I look at the -- well, those are the 2 main things I look at. New job creation and housing prices, the housing market, really. So I -- so those things tell me that U.S. is on a recovery -- continues to be on a recovery path. Europe, I think, actually is not doing any worse anyway, and Japan I think is doing better. And China, I think is going to be successful in a soft landing, you might say. So I think I'm optimistic. So intuitively, I think that the 2.5%, 2.8% forecast, which means that next year will be somewhat better than this year. Intuitively, I think it's about right.

My second question being that, of course, the Chairman Chang has just discussed with us, that when the CEO succession will be announced. But people are quite curious about who will be next in line, would that be a current operating -- Chief Operating Officers in the company, or can we expect someone else from outside of the company?

Well, the logical candidates are the 2 COOs. I think those are the logical candidates and what was the other question?

Okay, as Chairman said, all the good things must come to an end, so we will end here. But before we conclude today's conference, please be advised that this replay of conference will be accessible 3 hours from now, transcript will be available within 24 hours from now. Both of which will be available through TSMC's website at www.tsmc.com. Thank you for joining us today, we hope you will join us again next quarter. Goodbye and have a good day.