Welcome to TSMC's fourth quarter 2014 earnings conference and conference call. This is Elizabeth Sun, TSMC's Director of Corporate Communications and your host for today. Before we begin, let me wish you a very happy and prosperous New Year. Today's event is webcast live via tsmc.com. If you are joining us through the conference call, your dialing lines are in listen-only mode. 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 Senior Vice President and CFO, Ms. Lora Ho, will summarize our operations in the fourth quarter and full year 2014, followed by our guidance for the current quarter. Afterwards, CFO Lora and TSMC's two Co-CEOs, Dr. Mark Liu and Dr. C.C. Wei, will jointly provide our key messages.
After that, TSMC's Chairman, Dr. Morris Chang, will host the Q&A session. For those participants on the call, if you do not yet have a copy of the press release, you may download it from tsmc.com. Please also download the summary slides in relation to today's earnings conference presentation. As usual, I would like to remind everybody that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the safe harbor notice as this appears on our press release. Now, I would like to turn the podium to TSMC CFO, Ms. Lora Ho, for summary of operations and current quarter guidance.
Thank you, Elizabeth. Good afternoon, and happy New Year to everyone. Thank you for joining us today. My presentation will start with the financial highlights for the fourth quarter and a recap of our 2014 performance, followed by the guidance for the current quarter. Fourth quarter was another record-breaking quarter for TSMC, with revenue, earnings per share, and cash balance all reached historical high levels. Despite the moderate impact from supply chain inventory corrections, our revenue increased 6.4% sequentially to reach TWD 222.5 billion. This was mainly due to the strong demand for our 20 nanometer technologies. On the profitability side, gross margin was 49.7%, down 0.8 percentage point from the third quarter. This was attributed to 20 nanometer margin dilution and a lower capacity utilization, while cost improvements and favorable foreign exchange rate offset some of the decline.
Operating margin was 39.6%, also down 0.8 percentage point from the third quarter. Overall, our fourth quarter EPS was TWD 3.08, increased 4.8% sequentially and 78.5% year-over-year. Let's take a look at revenue by application. During the fourth quarter, the strong 20 nanometer ramp was mainly driven by communication-related applications. Result, communication grew 18% sequentially, and the revenue contribution increased from 59% in the third quarter to 65% in the fourth quarter. For other applications, computer grew 7%, while consumer and industrial declined 21% and 11% respectively. On a full year basis, communication increased 39% and it represented 59% of our revenue. The major contributing segments included baseband, application processors, image processors, and display drivers. Another fast-growing application in 2014 was industrial and standard, which grew 30% year-over-year.
The growth was mainly driven by increasing usage of power management ICs, near-field communication, and audio codec within the mobile devices. By technology, 20 nanometer revenue contribution started with a very small number in the second quarter, jumped to 9% in the third quarter, and it reached 21% in the fourth quarter. Such unprecedented ramp cannot be achieved without seamless teamwork with our customer, the R&D and operation people in TSMC. On a full year basis, 20 nanometer accounted for about 9% of our full year wafer revenue. Looking forward, we are confident that 20 nanometer will continue its momentum to contribute 20% of the revenue for the whole year 2015. Meanwhile, customer demand for our 28 nanometer wafer remains strong. Accordingly, these 2 advanced technologies, 20 nanometer plus 28 nanometer, represented 51% for our fourth quarter total wafer revenue, a big increase from the 43% in the third quarter.
Now let me move on to the balance sheet. On the asset side, cash and marketable securities increased TWD 147 billion to reach TWD 437 billion at the end of the fourth quarter, mainly due to higher free cash flow generated from the fourth quarter, and the receipt of TWD 30 billion guaranteed deposit. Total liabilities increased by TWD 56 billion, mainly due to increase in guaranteed deposit, increase in the tax payable, the employee profit sharing. On financial ratios, accounts receivable turnover days was 47 days. Days of inventory increased by 2 days to 58 days, reflecting longer production cycle time for leading nodes. Now let me make a few comments on cash flow and CapEx. During the fourth quarter, we generated about TWD 153 billion cash from operations and invested TWD 52 billion in capital expenditure. As a result, we generated free cash flow of TWD 101 billion in this quarter.
Overall, our cash balance increased TWD 132 billion to reach TWD 358 billion at the end of the quarter. In U.S. dollar terms, our fourth quarter capital expenditure was $1.7 billion. This adds to the total of $9.5 billion of capital expenditure for 2014. Now, I would like to give you a recap of our total performance in 2014. TSMC set records in terms of revenue and earnings in 2014. Our revenue grew 27.8% year-over-year to reach TWD 763 billion or $25 billion in U.S. dollar terms. On profitability, although the rising depreciation and fast ramp of 20 nanometer has indeed put pressure on our margins, our gross margin actually improved 2.4 percentage point to reach 49.5%. This is because the capacity we invested were fully utilized, and we continue the productivity and cost improvement, and to a lesser degree, a favorable foreign exchange rate environment.
Our operating margin increased 3.7 percentage point to reach 38.8%. This demonstrated our ability to drive higher operating efficiency. The operating expenses as a percentage of revenue decreased from 12% in 2013 to 10.6% in 2014. As a result, our full year earning per share increased by 40% to reach the historical high level of TWD 10.18 per share. On cash flow, we spent TWD 289 billion in capital expenditure, which is about the same level as 2013. Meanwhile, our operating cash flow increased 21% to reach TWD 422 billion. Accordingly, our free cash flow more than doubled in 2014. Overall, our ROE increased by 3.9 percentage point from last year to reach 27.9% in 2014, exceeded our long-term financial goal of equal or bigger than 20%. I have finished my report on the financial part. Now let me turn to the first quarter outlook.
We expect a slightly weaker demand in the first quarter due to seasonality. However, we also anticipate that a more favorable foreign exchange rate will moderate the seasonal weakness. Based on current business expectation and the forecast exchange rate of 31.80, we expect our first quarter revenue to be between NT$221 billion and NT$224 billion, representing a flattish quarter. In terms of margins, we expect the first quarter gross margin to be between 48.5% and 50.5%. We expect operating margin to be between 38.5% and 40.5%. This concludes my remarks. Thank you very much. Now our executives will deliver the key messages. The messages will be offered by our CFO, as well as by the two presidents and co-CEOs. We will start with Lora. I will make a few comments. I will start with the capital expenditure for this year.
As we continue to expand our business in advanced technologies, we estimate our 2015 capital expenditure to be between $11.5 billion-$12 billion, which is about a 20%-25% year-over-year increase. In addition to the investments for 16 nanometer capacity, we also spend for 10 nanometer tools and facilities to be ready for customer product tape-out by end of this year. More than 80% of the planned CapEx is budgeted for advanced technologies, while eight-inch capacity tools for specialty technologies and the back-end capacity investment constitute the rest of the 20% of 2015 budget. I would also like to make some comment on the solid-state lighting selling.
As you know, last Friday, January 9th, upon TSMC's board of directors approval, we have signed a contract with Epistar to sell TSMC's entire holding shares, which is 94%, of TSMC Solid State Lighting to Epistar, and we will exit the LED industry. Despite several years of dedication and hard work, as a late entrant to the LED industry, TSMC Solid State Lighting faced difficulties overcoming patent obstacles and sales channels. Not seeing how the company will be able to reach profitability due to the oversupply exhibited by massive expansions of LED companies worldwide, we have decided to transfer the ownership to Epistar, which is the world's largest manufacturer of LED epitaxial wafer and dies. The share transfer is valued at NT$1.46 per share, with a total proceed of NT$825 million to TSMC.
We have took NT$740 million impairment loss in the first quarter last year, with a minimal impact on EPS by about NT$0.03. The most important part of this deal is that no TSMC Solid State Lighting employees loses his job. Everybody has a job. My last comment is about the ASML stock sale. As you know, in August 2012, we acquired about 21 million shares of ASML under its customer co-investment program. The purchase price was €39.91 per share for a total of €838 million. There was a lock-up period of 2.5 years. In the last two years, TSMC has entered several hedging contracts that fully covered our position with an average hedge price of €62.59 per share, resulting in a locking profit of €483.5 million.
As the lock-up period is to be expired in April this year, we will be able to book a total profit of about TWD 21 billion in 2015. This one-time non-op gain is expected to increase our EPS by TWD 0.61 in Q2 and TWD 0.13 in Q3, and for the full year will be about TWD 0.75. That concludes my remark. Let me turn the podium to Mark.
Okay. I will follow to give you key messages on the near-term demand. We have just concluded a strong 2014 with a 27.8% revenue growth. In particular, the strong demand of our 20SoC overcome the normal inventory adjustment pattern and enable a 6.4% quarter-to-quarter growth in the fourth quarter 2014. The fabless company exited 2014, based on our estimates, with the days of inventory two days below seasonal. It was from a four days above in the third quarter. Now we see such inventory adjustment should come to a close. We estimate the days of inventory at the end of the first quarter 2015 should be one day below seasonal level. We see our near-term demand is quite healthy. Since our fourth quarter last year set a high base, we guide a good quarter for the first quarter 2015.
We are essentially flat from fourth quarter 2014, clearly better than our seasonal again. Looking forward to 2015, it should be another upbeat year. We forecast the semiconductor industry revenue growth to be 5%. The foundry revenue growth is 12%. For TSMC, we are confident we can outperform the foundry revenue growth by several percentage points in 2015. Now I give you a few words on 10 nanometer development update. Our 10 nanometer technology development is progressing, and our qualification schedule at the end of 2015, end of this year, remains the same. We are now working with customers for their product tape-outs. We expect its volume production in 2017. On the new technology development in TSMC, I will begin with beyond 10 nanometer I just talked about. We are now working on our future generation platform technology development with separate dedicated R&D development teams.
These technologies will be offered in the 2017 to 2019 period. We are committed to push forward our technology envelope along the silicon scaling path. In addition to the silicon device scaling, we are also working on the system scaling through advanced packaging to increase system bandwidth, to decrease power consumption and device form factors. Our first generation InFO technology has been qualified. Currently, we are qualifying customer InFO products with 16 nanometer technology, it will be ready for volume ramp next year, 2016. We are now working on our second generation InFO technology to supplement the silicon scaling of 10 nanometer generation. On the other side, in addition to the recently announced 55 ULP, Ultra-Low Power technology, 40 ULP, 28 ULP technologies for Ultra-Low Power application, such as wearable and IoT. We are also working on 16 ULP technology development.
This 16 ULP design kits will be available in June this year. It will be just suitable for both high performance and Ultra-Low Power or ultra-low voltage, less than 0.6 volts applications. Now I turn the microphone to C.C.
Thank you, Mark. Good afternoon, ladies and gentlemen. I will update you 28 nanometer, 20 nanometer, 16 nanometer status and the InFO business. First, on 28 nanometer. Since year 2011, we started to ramp up 28-nanometer production. Up to now, we have enjoyed a big success in terms of a good manufacturing result, and most importantly, the strong demand from our customer. This year, we expect the success will continue. Let me give a little bit more detail. First, on the demand side. The demand continue to grow, which are driven by the strong growth of mid and low-end 4G smartphone, as well as the technology migration from some second wave segments, such as radio frequency, hard disk drive, flash controller, connectivity, and digital consumers. Second, on the technology improvement, we continue our effort to enhance 28-nanometer technology by improving the speed performance while reducing the power consumption.
28 HPC, 28 Ultra-Low Power technology are some examples. To conclude the 28-nanometer status, we believe we can defend our segment share well because of excellent performance and performance cost ratio and our superior defect density result. Next, I'll talk about the 20SoC business status. After successfully ramp up in high volume last year, we expect to grow 20 nanometer business more than double this year due to high-end mobile device demand which was generated by our customers' very competitive products. Our forecast of the 20 nanometer business, as Lora just pointed out, will contribute 20% of the total wafer revenue. That remain unchanged. Now on 16 nanometer ramp-up. We expect to have more than 50 product tape-outs this year on 16 nanometer. High volume production will start in Q3, with a meaningful revenue contribution starting in Q4 this year.
I would like to stress again that our Chairman already mentioned that combining 20 nanometer and 16 nanometer, we expect to enjoy overwhelming market segment share. Last, our update on the InFO business. The traction on InFO is strong. We have engaged with many customers, and a few lead customer are expect to ramp up in Q2 next year. Right now, we are building a small pilot line in a new site to prepare for high volume production next year. Also, we expect this InFO technology will contribute sizable revenue in 2016. Thank you for your attention.
Okay, this concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to limit your questions to two 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 questions in Chinese, I will translate it to English before our management answers your question. For those of you on the call, if you would like to ask a question, please press the star, then 1 on your telephone keypad now. Questions will be taken in the order in which they were received. If at any time you would like to remove yourself from the questioning queue, please press the pound or the hash key. Now let's begin the Q&A session. First, we will invite Bank of America Merrill Lynch, Dan Heyler.
Thank you. Good afternoon, congratulations on a fantastic 2014 and guidance. I had two questions. First, I wanted to talk a bit about the ASP situation. You did talk about blended ASP. You did talk about growth in semis being 5% this year, foundry being 12%, and TSMC growing several percentage points beyond that. Should we expect that your blended ASPs should rise again this year and continue to rise in future years? Thank you.
The answer is yes. We do expect the blended ASP will continue to grow in 2015.
In order of magnitude relative to last couple of years?
We expect the trend will continue to the year after next year as well.
Okay.
Order of magnitude, he said.
Oof. A low single-digit range year-over-year.
Okay. The specific drivers of that, I know you're going to say mix, but could we talk a little bit more detail on the drivers of the ASP improvement as my follow-up?
I think it's just 20 will be bigger this year, percentage-wise, than last year. 16 will start, 16 will be much bigger next year than this year, et cetera. 10. In the past several years, our ASP has been increasing because the mix the advanced technologies keep coming up. That's the driving force.
Thank you. The second part, I guess as we look at your pie chart on your slide with communication and computer being amazingly only 9% of your revenue. Say 10 years ago, that chart looked much, much different with compute being the biggest. As we look at computer opportunities going forward, I think to some extent, there's maybe a sense of a little bit of disappointment in that we don't see ARM necessarily in PCs yet. We haven't really necessarily seen that ecosystem come through in the server business. Big data being such an important trend going forward with compute growing about 15% per year. I'm wondering what TSMC is doing or what your view of that opportunity will be in the future as a potential growth driver. Thank you.
Servers is one of them. Mark can Well, there's IoT actually also. Just don't forget that mobile actually, we think, has a few more years to run yet. Really, the TSMC silicon content in the average phone is actually increasing, which is something that is not recognized by a lot of people because everybody says that the weight, the gravity is shifting to the middle level, lower level priced phones. According to our data, and we have kept track of it for quite a long time, the average, our TSMC silicon content in the average phone is actually increasing. Look, we still look for, I think the number we have is that by 2019, there'll be 2 billion phones manufactured. I think last year it was what? 1.3 billion, I think. Yeah, 1.3 billion. 1.3 to 2 billion.
The average TSMC silicon content per phone is increasing, and the number of phones is going up. That's by no means I mean, it's still there. It's still a growth engine. IoT, I think we talked about IoT before, and now we are certainly not oblivious to the server possibility. Why don't I ask Mark to talk more about the server, and maybe CC will talk a little about the IoT.
Okay, Dan, I just respond to you on the server part. Chairman talked about the area we mostly focused on phone today, that would give us growth momentum in the next several years. On server, we work with the product innovators around the world, such a field definitely will not lose in our radar screen and theirs. The TSMC has, over the years, developed our technology to suit for high-power computing. From 65, 40, 28 to 16 nanometer, we continuously improve our transistor performance. Today, we believe our 16 FinFET Plus transistor performance probably is one of the top of the world that well suitable, well capable doing the computing tasks. Actually, before server, there are several supercomputer around the world in U.S. and in Japan, already powered by our technology, doing the weather forecasting, whether the geo-exploration applications today.
On the server, on ARM particular, we have very close partnership with ARM in recent years. ARM is a very innovative company. They produce CPU core and the new architecture every year. We bridge our leading-edge technology very early with ARM, to design their leading-edge CPU cores. That will continue, several our customer are taking advantage of that. Yes, in the past, it's been getting into slower as expected. That's because the software ecosystem is slower to come, but actually, a lot of service company and system company is continually investing in this ecosystem. Linux-based ecosystem is coming very strong, too. I think the trend will continue, we will, with our customer, get into these segments in the near future. Yes.
Okay. CC, say a few words about IoT.
Okay. For the IoT, that will be a big topic right now in the whole industry. All I want to say is that we are happy to share with you that long time ago, we already focused on our specialty technology, which are CMOS image sensor, MEMS, embedded flash, all those kind of thing. Today, we add another new technology, Ultra-Low Power into it, that will be the basis for the IoT technology necessary in the future. We believe that when the times comes and IoT business become big, TSMC are in a very good position that to capture most of that business. That's I share with you. Thank you.
Thank you very much.
Next, we will invite Credit Suisse, Randy Abrams.
Thank you. The first question, actually wanted to ask about the CapEx increase, where it's moving up for this year. Could you talk about the allocation, how much you plan for the 10 nanometer, how much for the 16, and also if you still see growth out of 28, and the eight-inch, if there's plans to expand eight-inch. If you could give more flavor on the CapEx.
Yeah, Randy, in my remark, I was talking about 80%. 80% of the CapEx goes to leading-edge technology. That actually cover a very big part of 16 FinFET capacity, and also the 10 nanometer for the engineering line and R&D expenditure. The altogether leading-edge technology will be 80%. We are also increasing our investment in the back end and also in eight-inch. These two things together will be about 10%, and the rest are the smaller items.
How big is the 10 nanometer?
We do not disclose specific numbers.
Okay. Sorry. Okay. The follow-up question on profitability, if you could give a flavor on structural profitability for 2015 and some of the flavor for 20 nm, how quick that may get to corporate margins, and for 2016, because it's an extension, whether that could be near corporate margins as that comes up. If you could give a comment on the inventory at current levels, if that'll stay at these higher levels from the WIP you've been building, or that may come back down to a different level.
Okay. Randy, you have a multiple question. I recall you asked about a structural profitability. That's your first question, right? Okay. From what we can see now, we are quite confident we can maintain equal or slightly better structural profitability standard gross margin versus 2014. Okay. For the 20 nanometer and 16 nanometer ramping, how would that affect the corporate margin? I have said in last July, it usually takes seven or eight quarters for any new leading-edge technology get close to the corporate average. For 20 nanometer, it will take eight quarters, we believe. 20 nanometer start to sell in second quarter 2014. We expect by first quarter 2016, that's eight quarters, it will be at corporate average level. For 16, we are going to mass produce this product. It will follow the similar trend.
A 16 nanometer will be based on the feature of 20 nanometer, so the margin will start to be higher. But it will also follow the similar trend. It takes seven quarter to reach to corporate average. So say, we plan to mass produce 16 FinFET in third quarter 2015, so by first quarter 2017, it will get close to corporate average. So before that, there will be still small dilutions. For this year, the dilution will be two to three percentage point, and last year, the second half will be three to four percentage point, and very low in 2016. Okay.
Okay. Just the one follow-up on inventory, is there any impact from the inventory at these current higher levels? Do you expect, because of the longer cycle time, inventory would stay at these inventory day levels, or that would come back down to the historical levels or closer to history?
I think Mark has talked about the inventory level. We have went through the inventory depletion period. Now we see inventory was four days above seasonal in third quarter last year.
For TSMC's own inventory.
Yeah, TSMC's inventory level.
Oh, TSMC, you're talking about the 50 days I was talking about? Okay. Normally, in the past you have seen around 45 to 50, maybe 45 day in average. This increase of inventory is mainly because of 20 nanometer ramp. It has much longer cycle time, both in the wafer fab, also in the back end. That's the main reason inventory is going up. It will not continue to go up further. It will probably stay at a similar level. It will not go down either. Okay.
All right. Thank you.
All right. Next, we invite Citibank 's Roland Shu.
Hi. Good afternoon, Chairman Mark, C.C., and Lora. I think first question, I would like to learn from Chairman, your view about the effective capacity of 40 nanometer. Since I think back to 40 nanometer days, you precisely foreseen the effective 14 nanometer definitely will be much smaller than the capacity, 14 nanometer capacity will be much smaller than the building 40 nanometer capacity. What's your view for 14 nanometer this time? Thank you.
Sure. Yeah. Roland, your question is with respect to effective capacity, where Chairman has defined the capacity to be effective when you have a useful technology.
Yeah.
Your question is wanting to hear from Chairman about the situation of 14/16 nanometer, what kind of effective capacity scenario we are facing today.
I define effective capacity as what?
A capacity with a useful technology.
Yeah, I think that is in effect to 14 nanometer days. Chairman guided us on the effective 40 nanometer capacity. That means it's going to be much smaller than the overall build capacity. Just want to really learn from you, what's your view for 14 nanometer effective capacity these days?
Yes. That is capacity that is going to be used. Right?
Used, and also.
That is going to make a profit for us, right? We are just in the middle of building up our 16 nanometer effective capacity strongly.
I think Roland's question, you probably are asking us about
It's the ballpark overall. Yeah.
Overall industry, whether or not other players are building effective capacity as well.
I think I have pointed out many times in the past that some companies, some foundries, build capacity on speculation. Just like builders build houses or condos on speculation. They haven't sold them yet. The speculation is that after they build the apartments or houses, they'll be sold. That doesn't always happen, of course. We, however, are different. We build capacity when we know that it's already sold.
Maybe I ask in other way. Compared to 14 nanometer days, and now we are looking for the 14 capacity, do you think the effective capacity for 14 nanometer industry will be bigger or smaller than the effective capacity when you see at the 14 nanometer days?
Do you understand the question?
Sort of.
40. He's talking about 40. Well, tell me.
He's talking about-
Yeah, tell me.
He's asking Chairman to compare 14 nanometer today versus 40 nanometer a few years ago in terms of the capacity.
What today?
14/16.
Yeah.
You should use 16.
Yeah.
16 nanometer versus 40 nanometer a few years ago.
In what respect?
Whether the industry has oversupply in capacity, whether the oversupply capacity is effective.
I don't know yet. From the data we have now, I don't think You're talking about 14, right?
Yeah.
14, right? From the data we have now, I don't think that they are building too much 14 capacity. Am I correct? From the data we have now, I don't think they are building too much capacity yet. You want me to compare with 40? 40, yes. I think at about the same point in time in the 40 nanometer cycle, we Well, my memory is a little hazy now. That was five, six years ago.
Too long ago. Yeah.
Okay. Thank you very much. I think my second question is, I think most investors are very happy to hear, Chairman, you have ranked return of shareholder as the top priority when you are running TSMC. I think some of the customers probably are very upset to hear about the TSMC put customer on the much lower priority when you are running your business.
No. Customers have a very high priority in our company. In fact, very, very high. As I think we have said many times that we really have three major strengths. One is technology, second is manufacturing, and third is customer's trust. This has been our model ever since we started the company almost 30 years ago. No. Now, if you are talking about We do, of course, place shareholders also in a very high priority, very high position. I think that's quite common. I think that's the way it should be. Obviously, you need very good customers. You need customers that trust us. We need customers that trust us, that work with us in order to satisfy our shareholders. It's a part of the same equation.
Yeah. I think maybe I should rephrase my question again. I think in the past, TSMC, because your technology-
Why do you have to rephrase your question all the time? Okay. Go ahead. Yeah.
My question in simple is just for the profitability and also customer relationship. I think most of the time, I think this is conflict. I think in the past, due to very good profitability for TSMC, probably some of the customer actually was not happy. Going forward, how TSMC to balance its profitability and the customer relationship going forward?
Roland, your point is that because we have very high profitability, therefore our customers are unhappy. That is not the right logic.
We think we earn our profit. We think everyone has to earn his profit. We think our customer has to earn his profit too. I think our customers do think they earn their profit, just as we think we earn our profit. There are always people who think that you're making too much profit. Some of our customers' customers think that our customers are making too much profit too. Okay. Thank you. We think we earn our profit. If any customer is unhappy with us, he sees me, he comes to see me. He comes to see us, okay? We try to correct the situation. We try to improve the situation. Yeah.
Okay, good. Next, we invite Goldman Sachs, Donald Lu.
Congratulations on a very good 2014 result, also very strong 2015 CapEx guidance. I think this year maybe TSMC will top the world in logic.
Oh, really?
Maybe. Okay. Now, two questions. One is, Chairman, about six months ago, you gave us a comment on your estimate on TSMC's market share in FinFET in 2015, 2016, 2017. Has that changed? Second question is, Lora, you commented that someone paid you TWD 30 billion capacity guarantee. Is this something new? I don't remember that TSMC take.
What did you say? I didn't hear the last one.
The second question is about capacity guarantee of TWD 30 billion.
Yeah. Customer deposit.
Deposit. Customer deposit.
It's a guaranteed deposit. Yeah. You're asking if this is something new?
Yeah.
Actually, I think more than 10 years ago, maybe 15 years ago, TSMC have done this with several customers. It's not something really new to us.
You mean 15 years ago?
Late '90s.
Late '90s.
Forgive me, I wasn't-
After that, I guess we didn't have-
We hadn't.
Yeah. Mm-hmm.
Okay.
I thought it was pretty good in the late '90s, so we started again. Customer likes it too.
Maybe you can explain a little bit what it guarantees, TSMC's obligation on that, and how long it will be in place.
Are we still answering the second question? I want to answer the first question. Donald's question was Actually, I looked up my statement at that time, July 16th of last year. I said on the subject of 16 and 20. 16 nanometer and 20 nanometer technology. I actually made three statements. The first statement was that because we started the 16 a little late, our market share in 2015, our 16 nanometer market share in 2015 will be smaller than our major largest competitors. The second statement I made was that we started 16 late because we wanted to do 20. If you combine 20 and 16, our major competitor who will be slightly ahead of us this year on the 16, he has very little 20, almost no 20 at all.
If we combine 20 and 16, our combined share in this year will be much higher than that competitor's. The third statement I made is that in 2016, we will have much larger share in just 16 nanometer than that competitor. All right? First, I want to say that I, at this time, stand on those statements. In fact, I now will add a couple of statements. The statements I will add are. That's fourth statement now, okay? When we have a larger share on just 16 alone in 2016, the 16 market will also be much larger than this year, 2015. Yeah, we are slightly behind. We have a smaller market share in 2015 in a smaller market. Next year, we'll have a larger share, in fact, much larger share, in a much larger market, 16.
Another statement I want to make is that I'm, at this point, very, very comfortable with all those statements that I have made on July 16th last year and the statements I have added today. I'm very comfortable. I don't know whether that answered your question or not, Donald.
Yes. How about 2017 if-
The what?
20-
2017.
Huh?
2017.
What? Well, 2017. 2017, the trend is going to continue. We're not going to lose the leadership on 16 market share once we recapture that in 2016. It's going to continue 2017, 2018. Also, both 20 and 16 are going to live longer than you might think now. 28, for that matter, also will live longer than you think.
Okay. Next we will invite Deutsche Bank's Michael Chou.
Thank you. Chairman, do you see the 16 nanometer FinFET Plus PPA is better than tier 2 foundries 14 nanometer at this moment? Given that you mentioned you expect TSMC's 16 nanometer market share should be higher than your major competitor in 2016. Based on-
I'm sorry, what?
Yeah.
Based on current R&D progress or any product design progress, do you think your PPA of the 16 nanometer FinFET Plus is better than your competitor's PPA, 14 nanometer?
PPK?
PPA. PPA.
Power performance.
PPA.
Yeah, PPA.
Okay. Michael's question is, if we look at the definition of the technology in terms of performance, power, and area, is our 16 nanometer better than our competitor's?
Better?
PCU.
Yeah. Mm-hmm. Well, why don't you be the judge?
As a follow-up, do you think that most customers will stay in your 16 nanometer rather than shift to tier 2 foundries? I mean, for over the next 18 months.
Will stay in?
Stay in your 16 FinFET Plus rather than move to your competitor, 14 nanometers.
Oh, you mean after we have ramped 16 FinFET Plus?
Yes.
Will our customers shift to our competition's offer?
Oh, I thought the question has been answered already. Once we capture that larger share, we stay there for many years.
Let me put it another way. Can we say your 16 nanometer market share in 2016 will be quite similar to your dominance in 28 nanometer, given that your 20 nanometer is the only provider? The apple-to-apple comparison should be 28 to 16 nanometer.
Market share in 16 nanometer in 2016, will that be the same as our market share at 28 nanometer, I would say, back in 2013, 2014?
Yes. Oh, no. I don't think so because 28 of course we were virtually sole source. 16, we already know there's at least one major competitor, and then there's another one that's just kind of eager to get in.
Okay.
I don't mean the first competitor's accessory. I mean another one.
Okay. Second question is regarding the InFO. Do you expect the gross margin InFO will have a negative impact to your overall gross margin in 2016 or beyond? Given that you mentioned it could be sizable revenue in 2016.
Your question is whether or not InFO business in 2016 will impact our margins.
Yes.
Whether InFO will impact our margins.
No, probably not. The back-end business actually is a low margin, but the turnover is faster. Put two together, it's comparable. InFO will have lower margin than our wafers business, but it will actually have higher return on invested capital than our wafer business.
Okay. I think we really should go to the line and invite questions there. Operator, could you please invite the first caller on the line?
Certainly. First question comes from the line, Brett Simpson from Arete. Please ask question.
Thank you very much. My question is around 28 nanometer. You're running a large capacity at 28 nanometer at the moment. Can you share with us what the capacity plan is for '28? As you migrate more business to 28 nanometer and below over the next couple of years, do you intend to convert 28 nanometer capacity to lower nodes, or do you think you can keep the existing 28 nanometer capacity running full going forward?
All right. Let me repeat Brett's question so that people here can hear it better. Brett's question is, TSMC's 28 nanometer capacity is very large. As our technology migrate to more advanced nodes, such as 20 and 16 in the next few years, what will be our plan on capacity of the 28 nanometer? Will we still have large demand to utilize those capacity, or we need to do some changes?
In every generation, we worry a lot about the conversion loss we will suffer when we convert the capacity of that generation to the capacity of the next generation. We do two things. First, we try to minimize that conversion loss. Since we've been living with the problem for so long now, I think we're getting to be pretty good at it. The conversion loss from one generation to another is normally in the low single digit, low middle single digit. Now, the second thing we try to do is. I think that we actually have been doing it perhaps even more successfully than the first thing. The first thing was to try to minimize the conversion loss. The second thing we try to do is we try to prolong the life of each generation.
I was just saying just five minutes ago that I think that the life of 28 nanometer may be longer than a lot of people think. I mean it. Actually, we're still making half micron stuff. We try to prolong the life of every generation as we continue to migrate to advanced technologies. 28 is certainly a generation that we want to prolong the life of.
All right. Brett, do you have a second question?
Yeah. My follow-up question is around China, maybe you can share what's your plan for 28 nano production in China. Would you expect this to happen over the next 12 or 24 months? What's the size of the potential capacity you might be looking at in China?
The question is.
Whether or not we'll build a 28 nanometer capacity in China in the next 12 to 24 months, how large will that be?
We are seriously considering the possibility. In fact, we are gathering data and making contacts, et cetera. There are obviously both pluses and minuses. We're seriously considering the proposition of making 28 nanometer in China. There are also barriers. As I said, at this stage, we're exploring, we're seriously considering, we're exploring, we're gathering data and making contacts.
All right. Let's go back to the floor. Now it will be from Barclays, Andrew Lu.
Test. Dr. Chang, Dr. Liu, and Dr. C.C. Wei, and CFO. [Foreign language] First one is regarding the revenue outlook for the next few quarters. Are you expecting any single quarter for the next few quarters, Q2, Q3, Q4, revenue below first quarter?
Now we what?
Andrew is asking us to give him a guidance, whether or not our Q2, Q3, Q4 revenue will be lower than Q1 level.
Will Q2, Q3, Q4 be lower than Q1?
Any single quarter in your internal forecast saying will it be lower than the first quarter?
What is the Q1 times four?
That would be.
I'll work out the answer here, right here.
That's the following one I am going to calculate. Based on the estimate, Q1 is quite similar to Q4 from the revenue, from EPS point of view, from OP margin guidance, gross margin guidance. If we times four, revenue is up 16% year-over-year, the EPS up 20% year-over-year.
That will be consistent with what Mark said. He said that it will outperform the foundry growth.
Yes.
Which is what? 12%? He said we'll outperform it by several points.
This is based on flattish environment. We've got no quotes for the next few quarters. That's why I'm asking you any downside risk or.
All right. Let me just tell you what I think. I think we have upside, okay?
Okay.
In the next year.
That's enough.
In this year, yeah.
Andrew, I also have to remind you, the foundry numbers are based in US dollars, but the fourth quarter or the first quarter revenues are based in New Taiwan dollars.
Yes. Second question is, are we still planning to raise our cash dividend?
Well. Look. Yeah, we're seriously considering it. Obviously, I can't answer the question because the board has to approve it, of course, the shareholders' meeting has to approve it.
What's in Dr. Chang's mind?
Huh?
What number in your mind?
Oh, I'm not going to go there. I'm not going to go into that.
Thank you.
Yeah.
Okay, now we will invite HSBC's Steven Pelayo.
Thank you. It seems like, especially last week, there's been two or three key concerns people are talking about relative to TSMC. The first one is smartphone growth slowing down. I think your guidance for the foundry market growing 12%, you growing several points faster, kind of answers that growth. The next two concerns are really about competition, customer concentration. I wonder if I just ask, you look at your top three customers in 2015, do you expect them each to grow year-on-year? And do they grow that several points above 12% foundry market?
Steven's question is regarding our top three customers, whether or not their business with TSMC year-over-year growth rate will be at least in line with the foundry's 12% rate of growth.
If you just limit it to three, it is getting too specific, because you almost know who the three are. Let's say 20, okay? Our top 20 customers.
Yeah.
I expect the vast majority of them to grow every year. Yeah.
Well, maybe as a follow-up, I think we're all kind of dancing around the same general questions. I asked this of you, I think, last quarter. At the 28 nanometer node, you had seven quarters of sequential growth, absolute TWD. 20 nanometer ramped up so significantly because you've had some significant customer wins there. When you look at it on a quarterly basis, do you expect every quarter of 20 nanometer to be higher than the prior quarter in TWD as you go through 2015?
You're asking 20 or 28?
20 now.
20. Okay. 20 nanometer every quarter in the following quarters, whether they will be higher than the prior quarter.
20 nanometer every quarter, whether it will be higher than the prior quarter.
This year?
This year.
I think the answer is yes.
Yeah.
The answer is yes. By the way, going back to the last question, were you just asking about this year or every year from now on?
I think the customer concentration concerns are primarily to this year, trying to offset such huge gains last year, and it appears as though you are absorbing those.
Well, my answer is still the same. Of the top 20, I expect the vast majority of them will grow, each will grow this year. Mm-hmm. Yeah.
Thank you very much.
All right. Next questions will be coming from Morgan Stanley's, Bill Lu.
Hi there. Thanks very much, also let me add my congratulations on spectacular 2014.
Thank you.
My first question is on 28 nanometers. If I look at your capacity this year versus 2014, how much is the increase in capacity?
28?
28, yeah.
C.C., you want to answer the question?
High double digit.
High double digit.
You mean high teens or high double digit?
High teens.
High teens, actually.
High teens. I'm sorry. Okay, great. Do you think revenues can grow? In other words, do you think ASP decline could be less than that unit growth?
We are not supposed to comment on a single node's price. I'm sorry. Our legal advice is not to comment on single node's price.
Okay, great. My second question is on your China strategy. I think you talked about potentially looking at 28 nanometers. Correct me if I'm wrong, but my understanding was that Taiwanese companies cannot do 28 in China. Can you talk a little bit more about that?
What?
28 nanometer
Oh, sorry. Yeah.
28 nanometer manufacturing in China. Your question is, most of the players cannot do 28 nanometer properly in China right now.
I thought by law Taiwanese companies cannot do 28.
Oh, Taiwan.
Yeah.
Whether Taiwan-
Taiwan, no. Actually, Taiwan, I think, has a rule now that says you still have to apply in every instance. The general rule is that the N-1 technology is allowed. That's Taiwan. You still have to apply in each case.
Okay. Next, we will be having questions from UBS, Eric Chen.
Hi. Very quick, my first question regarding to your China investment. I would like to know why you picked the 28 nanometer process. We know your China client already, and do very good business with your TSMC and in Taiwan. What's the point for you to build out a 28 nanometer process in China? What is the trigger and what's the benefit? My first question.
Yeah. Eric's question is, since most of our Chinese customers already do 28 nanometer with us in Taiwan, why do we need to go to China to capture the 28 nanometers there? What's the plus and the minus?
Well, because they're telling us that, yeah, they will continue to do 28 with us, but it'd be better if they do 28 with us if we're in China, you know.
Okay.
You have to realize that there are companies, there are foundries in China that are also going to do 28 nanometer. Our customers may prefer to buy from the Chinese foundries when their 28 becomes available.
Okay. Can we assume from the profitability point of view, it's no big change, no big difference between the manufacturing in Taiwan and the manufacturing in China?
What's that?
Whether or not the profitability.
Well, that's why I said there are pluses and minuses. Basically, I think the cost. We have had the experience for more than 10 years now of operating an 8-inch factory in China. Okay? The cost is higher. All right? Let's say minus. If you lose business.
That's not good.
That's even worse. All right?
Okay. My second question regarding the CapEx. We raised all the CapEx higher than the market expectation. Can we expect, can we assume all the equipment, all capacity for the 16 nano, the FinFET probably will move ahead of all the earlier schedule? In terms of the 16 nano FinFET, the equipment schedule, will that move ahead?
Eric, your question is, since our CapEx guidance is higher than market expectation, whether or not we are moving the equipments earlier or ahead of our original schedule?
Yes. For the 16.
For 16 nanometer?
Yes. Thank you.
Well, I don't know what the market expectations are. We don't benchmark ourselves against market expectations.
Yeah.
We benchmark ourselves against needs.
Yeah, you are right. I remember.
Yeah.
Thank you. Dr. Chang, I remember probably six months ago, you talked about, probably three months ago, six months ago, in conference, probably Lora mentioned the CapEx for this year, probably slightly higher the year 2014. I assume we get more aggressive at the CapEx. Am I right?
I think we have always been reasonably aggressive in CapEx without speculating at all. That's our standard. All right? I don't know what your question is anyway. Are you asking whether we are moving in or how soon we are buying, we are setting up the capacity? Is that what he's asking?
I think, Eric, you are really trying to see if we are becoming more confident, convinced of the demand-
You are right
We are pulling in the equipment sooner, right?
Yeah.
I said earlier that we don't build capacity on speculation.
That's what we need.
Yeah. Good. Thank you.
Translation.
All right.
Next, I think we will, because JPMorgan's Gokul was already raising his hand, we'll give the microphone to him. Thank you.
Congrats on a good 2014. Thanks for taking my questions. First, I had a question on, there's been a lot of controversy about cost per transistor, whether the economics of Moore's Law is slowing down. Your competitor, Intel, has put out a very emphatic statement saying that until 7 nm, they're seeing that continuing at the same pace as before, while there has been a lot of noise from the fabless community in the last couple of years that at 20 nanometer or at 16 nanometer, there is a potential slowdown. Could we have TSMC's version now that you're pretty much ready to start 10 nanometer and thinking already about seven? That's my first question.
All right. Let me repeat. Gokul, your question is mainly on the comments of cost per transistor. As some of the other player, I think you are referring to Intel
Yep
who has made comments that they do see the cost per transistor to continue into 7 nm, they can handle the economics of the Moore's Law. Whereas, on the other hand, fabless companies begin to complain about not seeing enough economics starting with 20 nanometer. What is TSMC's statement regarding this economics issue?
Okay, let me answer this question. The cost of transistor continues to go down, and by scaling mostly is everybody knows, nobody I think refuse that statement. We see the cost of transistor continue going down in the constant rate and in going forward, the cost of transistor going down probably at slightly slower rate. That's the argument. It really depends on companies, and for some company simply do not have the technology capabilities. Today, further going down the Moore's Law technology development, just a few. As far as whether those cost can get enough returns, and of course, that has to do with how much that technology bring the values to the product where the command the price. Today, we see certain segments will continue to need it, that type of system performance, to get enough return.
This is the reason we committed to push the system scaling.
Can we say that for customers who can afford it is still going to go down basically? Even at 10 nanometer, for customers who can afford it, afford the development cost, and have the volume, the cost is still going to be going down substantially.
Oh, of course. It will go down very significantly. Yes.
I had the second question, just a clarification on the 16 nanometer ramp-up. I think last conference, C.C. Mentioned that 16 nanometer ramp-up is likely to be at or even faster than the 20 nanometer ramp-up that we saw last year, with a five-quarter delay. basically meaning that first quarter 16 nanometer revenues could be even higher than what 20 nanometer revenues were last quarter. Is that still the expectation for the 16 nanometer ramp-up, in the next few quarters?
What was the question?
The question is whether or not the speed of the ramp-up of 16 nanometer will be faster than the speed of the 20 nanometer ramp-up in the first three quarters.
In the first three quarter, our ramping up speed very similar, but maybe a little bit faster, but very similar. I'm sorry.
Okay. Thank you.
All right. Do you-
Don't we have any more questions from the-
Yes. There are people raising hands here.
Yeah.
Okay. There is Daiwa's Rick Hsu.
No, I mean, overseas.
Go ahead, Rick. Thank you.
Yeah. Hi. Sure. I'll do this quick. Yeah, this is Rick from Daiwa. Just got one question here. I remember in the last four years, post the financial crisis, I think TSMC tended to build about two fab shells per year for expansion a year ahead. If I look at this year, correct me if I'm wrong, it seems to me that you don't have any new fab shell under constructions. Does that mean you guys are turning a bit more conservative in 2016 or 2017?
Rick is asking us whether or not we will be building new fab shells this year at the same speed as we did in the past, which is two shells per year.
We will. We continue this trend. Two shells a year, huh? Two shells almost for one generation.
Thank you.
Okay. Randy has a follow-up question.
Thank you. My first question on the guidance you gave for first quarter is holding up pretty well flat. Looking at the last four to five years, it's also been much better than it used to be at the beginning of the year in first quarter. If you could talk about if you're seeing seasonal patterns shifting more, customers getting more aggressive first half, and if you see the same type of scenario where you have second half slowdown again. If you see a different pattern of seasonality.
Seasonality, Randy's observation was that in the past, he saw our customers to be optimistic in the first half, and then going through an inventory correction in the second half.
This rate change is equivalent to 0.4% point of our margin. I think you're asking whether that still holds true, is that right?
Whether if the currency stays at that level, if that's a permanent benefit, if we were to stay at 32 or over time, you share some with your customers.
You are saying that since, if TWD remains this low, whether or not we will share the exchange rate benefit, at least part of that, with our customer.
Whether we will be sharing the exchange rate benefit with our customers, i.e., whether or not we are willing to take a lower U.S. dollar price.
Well, they didn't share the exchange rate loss with us.
Thank you.
Okay. All right. Follow-up question from Roland. Citi's Roland Shu.
Thanks. Just a 10 nanometer question to C.C. Since, C.C., you said we are expecting to volume production 10 nanometer in 2017. I remember in the past two quarters, actually, your outcome was to pulling in 10 nanometer mass production by end of 2016. Are we pushing out the 10 nanometer mass production schedule a little bit or not?
Let me explain that, because 10 nanometers, the masking layer is about 70 to 80. You got to start in 2016 to have output in 2017. What I'm talking about is 2017, you still start to have revenue.
Okay, thanks. Wafer start schedule definitely does not change.
No, I cannot say more than that.
Okay. Thank you.
Okay. All right. Andrew Lu also has follow-up question.
I remember last investor conference, C.C. Wei mentioned 16 FinFET revenue have a high single digit by Q4 this year, and maybe few percentage in by Q3. Is that number unchanged?
Unchanged.
He sounds less confident.
The more I say, the more the information from the customer will be released as well.
Okay.
Steve.
Understood.
All right. Andrew, you are done, right?
Yes.
Okay. We are going to Dan Heyler.
Yeah, thanks. I had a question on the more the mature nodes situation. Still a nice chunk of your revenue there on the mature 12-inch nodes. As we move into IoT, there's a lot of interesting products that are coming out, Ultra-Low Power for one. I'm wondering is, as you look at the 40/65 nodes, what's happening on device complexity? Is device complexity there increasing? We hear about device complexity maybe on the mature node may be increasing. I wonder if you have a view on that. Sorry, it's more design related. What I'm getting at there is the ASP trends. I think there's a traditional view of mature technologies as being low margin business.
I'll let C.C. Answer the question. In general, yeah, the device complexity on mature nodes is increasing. That's how we are prolonging the life of the mature nodes, though.
Yeah, usually we develop the pure logic technology into the derivative technologies, which is more complicated. One good example is from the logic to embedded flash. You add quite a few steps, and become a CMOS image sensor or those kind of things. All that's more complex. Yes.
Implications there for, I would presume pricing and market share then would be quite favorable. When device complexity goes up, does it hold blended ASPs flat or does it increase ASP in general terms?
What's the question again?
Because of the complexity of the mature technology is increasing, whether we will benefit from ASP.
Whether we will what?
That our ASP will benefit.
We'll benefit?
No, does market ASP.
Well, actually, I would say that our profitability has remained pretty constant. In the early stage, the profitability of a node is often low. As Laura pointed out, it takes about eight quarters for the margin to get to the corporate level. After that, it stays pretty constant. Or it increases a little bit, in fact. Increases, particularly in the last few years, I think we have pushed up the structural profitability. Yeah, I think that the added complexity or the Actually, a lot of new things are happening on the mature nodes. So the mature nodes today are nothing like Well, not nothing like, but only about 50%-60% like what they were when they were first introduced. That's about right? Yeah.
Okay, thank you. Great. The second question is on the 20. Do you think 20 revenue will grow this year?
Of course.
You expect it will, okay. If that's the case, does the mature technology overall, everything else, say 40-90, is that able to stay flat or does that go down? Because there's still a lot of migration taking place to 20. 20 is a very attractive node. I'm just wondering what's happening on the 40, 65, and 90. Can that hold flat or does that decline?
I think it depends on nodes. Dan's question is, if we are growing our 20 nanometer revenue, we are growing our 28 nanometer revenue, whether or not our 40, 65, et cetera, those older nodes' revenue will be growing as well.
Yeah.
If I can make some comment on your questions. Actually, we have very strong demand on those specialty technology. As you know, that's 0.15, 0.18. It is very high demand. We are also increasing the technology offering for 40 nanometer, 65. From what I can see now, I believe that 2015, those mature technologies revenue will be bigger than 2014.
Excellent. Thank you.
All right. I think really, we will just allow two hands, okay? It will first go to Michael, and then it will be Steven, and then that'll be it.
Sorry. Just the capacity increase this year. What will be your forecast?
Okay. With the $ 11.5 billion-$ 12 billion CapEx, we expect to increase our capacity by about 11%-12%.
Thank you.
Thank you, Steven.
Of course, we're all going to try to reconcile that with the full year. Andrew's question with the full year guidance. I guess, could we talk just a little bit about the second quarter, but talk about it relative to, I think you mentioned last quarter, and I think you did it last year, you do some pre-building in the first quarter ahead of the second quarter. How much of that is impacting your first quarter guidance?
Lora?
Very minimal. As we are going through the inventory depletions, this pre-build gets less and less. I think you know the purpose, trying to utilize the capacity without any waste. It does help a little bit for the utilization, but it's getting lower and lower now.
Okay, thank you.
All right. I think in the interest of time, we will end our conference here. Thank you for coming, and I hope we will see you next quarter. Goodbye. Have a good day.