Welcome to TSMC's fourth quarter 2013 earnings conference and conference call. This is Elizabeth Sun, TSMC's Director of Corporate Communications and your host for today. Before we begin, please let me extend our warmest wishes to all of you for a very happy New Year of the horse. Today's event is webcast live via TSMC's website at www.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's Senior Vice President and CFO, Ms. Lora Ho, will summarize our operations in the fourth quarter and full year of 2013, followed by our guidance for the current quarter. Afterwards, TSMC's Chairman, Dr. Morris Chang, and TSMC's two co-CEOs, Dr. Mark Liu and Dr. C.C.
Wei, will jointly provide a couple of key messages. We will open both the floor and the lines for the Q&A. For 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. As usual, I would 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. Now, I would like to turn the podium to TSMC CFO, Ms. Lora Ho.
Thank you, Elizabeth. Good afternoon, and happy New Year to everyone. Thank you for joining us today. I will start my presentation with financial highlights for the fourth quarter and a recap of 2013 financial performance, followed by the guidance for the first quarter of 2014. In the fourth quarter, TSMC's business was negatively impacted by semiconductor supply chain inventory management. On a sequential basis, fourth quarter revenue declined 10.3% to TWD 146 billion. Gross margin was 44.5%, down four percentage points from the third quarter, mainly due to lower capacity utilization. Total operating expenses decreased 11% as we took several expense-cutting measurements during the quarter. This led to an operating margin of 32.8%, down 3.9 percentage points from the third quarter. Non-operating items was a gain of TWD 2.6 billion. We have reversed the TWD 1.2 billion impairment loss due to strong business recovery from one of our invested companies.
In addition, we received the last payment of litigation settlement from SMIC. Overall, EPS was TWD 1.73 in the fourth quarter. ROE was 21.7%. Let's take a look at revenue by application. During the fourth quarter, inventory correction continued across all major segments of the semiconductor supply chain. Our communication, computer, and consumer-related revenue declined 13%, 7%, and 20% respectively, while industrial and standard product remained flat during the quarter. On a full year basis, communication increased 29% and represented 54% of our revenue of 2013. This reflected the strong demand for application processor, baseband, display driver, and other peripheral ICs used in mobile computing devices. Computer decreased 10%, with PC graphics declining the most. Consumer increased 43% year-over-year, reflecting the strong demand in the next-generation game consoles.
Industrial and standard increased 14% as mobile devices use various ICs enabled by our specialty technologies, such as Touch controller IC and Power management IC, et cetera. Our technology, 28 nanometer, contributed 34% of our total wafer revenue in the fourth quarter, up from 32% in the third quarter. On a full year basis, 28 nanometer tripled and contributed about 30% of our total wafer revenue, a big increase from the 12% in 2012. Let's move on to the balance sheet. Cash and marketable securities was TWD 245 billion at the end of the fourth quarter. Current liability increased by TWD 42 billion due to higher accounts payable to equipment suppliers. Accounts receivable turnover days increased to 48 days. Days of inventory remained at 45 days. Let me make a few comments on the cash flow.
During the fourth quarter, we generated TWD 103 billion from operations, invested TWD 74 billion in capital expenditure, and repaid TWD 2 billion short-term loans. At the end of the fourth quarter, our cash balance increased TWD 26 billion to TWD 243 billion. In the fourth quarter, free cash flow was an inflow of TWD 29 billion. I would like to give you a recap of our performance in the year 2013. 2013 was another record year for TSMC. Thanks to the technology leadership and excellent manufacturing execution in 28 nanometer, we were able to capture the strong growth in mobile computing market. From the financial point of view, revenue increased 18% year-over-year to reach TWD 597 billion, or in the US dollar term, about $20 billion. On the profitability side, gross margin declined 1.1 percentage point to 47.1% due to lower capacity utilization in 2013, while our structured profitability continued to improve.
On cash flow, we spent TWD 288 billion, or $9.7 billion in CapEx, up 17% from 2012. We were able to grow operating cash flow at a faster rate of 22% to reach TWD 347 billion. Free cash flow increased by 54% year-over-year to reach TWD 60 billion. On earnings, income before tax increased by 19% versus 2012. Meanwhile, we took a hit by the higher effective tax rate. The effective tax rate went up from 9% in 2012 to 13% in 2013. EPS increased at a slower pace of 13% to reach TWD 7.26, set a new record. Overall, our ROE was 24% for the whole year, meet our long-term financial objective of bigger than 20% ROE. I have finished my report on financial part. Let me provide you a first quarter outlook.
Based on the current business expectation and the forecast exchange rate of 30, we expect our revenue to be between TWD 136 billion and TWD 138 billion. This would translate into 6% Q-over-Q decline. On the margin side, we expect the first quarter gross margin to be between 44.5%-46.5%, and operating margin to be between 32%-34%. First quarter gross margin is expected to be better than the fourth quarter despite a lower revenue. Let me explain why. You all know our business has certain seasonality. First quarter is normally a slow quarter. If we can start a wafer earlier in the slower quarter, we will be able to complete more wafers in the following quarters when the business pace picks up.
This way, we're able to complete and ship more wafers to support a surge demand in the peak quarter without having to build to a peak capacity for that quarter. Meanwhile, we can also benefit from better utilization rate in the slower quarter. It improve our profitability. We began to implement this practice few months ago and expect it to benefit our gross margin by two to three percentage points in the first quarter. Another thing I want to highlight is the tax rate for 2014. We expect our whole year effective tax rate to be about 13.3%, which is about the same as 2013's 13%. However, from a quarterly perspective, second quarter would carry about 40% the total tax burden for the whole year, due to the need to accrue the 10% return on distributed return earnings. This concludes my remarks.
Let me turn the podium to our Chairman, Dr. Morris Chang.
Good afternoon, ladies and gentlemen. Today, our comments are scheduled as on the slide on your left. First, I am very glad to have the opportunity to introduce our new top management team. I'll first start with Lora, although I think everyone knows Lora well. Lora has a bachelor's degree from Tunghai University, a master's degree from National Taiwan University, both degrees in finance. She worked for Cyanamid, Wyse, Thomas & Betts, and TI Asia before she joined TSMC in 1999, and she has been TSMC CFO since 2003. Next, Dr. C.C. Wei. C.C. has a bachelor's degree from Chiao Tung University and a PhD from Yale University, both in electrical engineering. C.C. worked for TI, SGS, Chartered, before joining TSMC in 1998. C.C. has been Senior VP of Operations, Senior VP of Business Development, co-COO, and in the co-COO job, C.C. was successively responsible for R&D and operations.
C.C. is President and co-CEO. C.C. is 60 years old. I should add that Lora is 57 years old. Mark Liu. Mark has a BS from National Taiwan University and a PhD from Berkeley, both in electrical engineering and computer science. Mark worked for Intel, Bell Telephone Labs, before joining TSMC in 1993. At TSMC, he has been Senior VP of Operations, and he was also a co-COO, and all the time he was co-COO, he was responsible for our sales, marketing, and planning. Now, Mark and C.C. are Presidents and co-CEOs of the company. Mark is 59 years old. I'm not going to introduce myself. I think you all know me. Let me talk about Q4. I will just give a few comments, a few words on the year 2013. I will say more words on this year.
Q4, I think Lora Ho already discussed it. It's pretty much as we guided it 3 months ago. Our revenue was affected by the supply chain's inventory management across almost all major segments, particularly in communication-related applications. The supply chain DOI declined from one day above seasonal in the third quarter to one day below seasonal in the fourth quarter. However, our structural profitability remained intact. The lower utilization caused by lower revenue resulted in a lower gross margin and operating profit % than Q3. On 2013, overall, we are pleased by the results. It was another consecutive year of record performance. As Lora Ho pointed out, revenue grew by 18% and profit before tax grew by 19%.
Looking at 2014, we expect still another consecutive year of double-digit growth in revenue. I will talk a little more about profit in a few minutes. 1Q is a seasonally weak quarter for IC companies, including us. Supply chain continued to manage inventory conservatively, even when the DOI has already reached below seasonal level. We expect the supply chain DOI to be two days below seasonal in 1Q, at the end of 1Q. Our structural profitability in 1Q still remains intact. In fact, it will probably improve slightly, as you can tell from the guided profit margins. For full year 2014, we forecast the following industry numbers. For worldwide semiconductor industry, we forecast 5% growth. For fabless industry, we forecast 8% growth. For foundry industry, we forecast 10% growth.
For TSMC, as I already did, we are forecasting revenue growth surpassing the growth of the foundry industry. 2014 capital budget is estimated to be between TWD 9.5 billion and TWD 10 billion, similar to that of last year. About 95% of the capital expenditure is for advanced technologies, and we mean 20 nanometer, 16 nanometer, more 28 nanometer, R&D, and our mask shop portion. Depreciation expense is expected to increase by about 35% year-to-year. The ratio of depreciation expense to revenue is expected to rise by a few percentage points in 2014. However, we will have a higher blended ASP by several percentage points resulting from our new technologies, principally the High-K metal gate 28 nanometer and the 20SoC. We also plan to have better operating efficiencies. The higher operating efficiency would include lower variable cost to wafer, higher manufacturing productivity, et cetera.
Between those two, principally the higher blended average price, which results from the richer product mix, principally from that and also partly from the better operating efficiencies, we intend to offset the increase of depreciation as it is semi-revenue. We believe we can maintain and perhaps even improve our structural profitability in 2014 versus last year. From second quarter on, we'll see strong growth. Where are our growth engines? In terms of market segments, mobile products, smartphones, and tablets will be the growth engine. In terms of technology, High-K metal gate 28 nanometer and 20SoC are our growth engines. I will say a few words about mobile products now. C.C. Wei will, a little later, talk about the technologies, the new technologies. Smartphones are expected to grow by 25% to 1.246 billion units, and tablets are expected to grow by 21% to 307 million units.
To be a little finer grained, the high-end smartphones are expected to grow 8% year-over-year to 325 million units. The middle end, 22% year-over-year to 449 million units. The low end, 45% year-over-year to 472 million. Due to the continued performance improvements, which means multi-mode and 64-bit, et cetera, also due to feature enhancements such as fingerprint, MEMS, NFC, and new standards in Wi-Fi and Bluetooth, we expect silicon content for the mobile devices to continue to rise. Silicon content in the mobile devices to continue to rise. TSMC's share in the non-memory silicon content of these devices is expected to increase from 45% last year to 47% this year. As a result, TSMC revenue from mobile products are expected to grow more than 35%.
Just looking a little further ahead, further than this year, we expect to continue benefiting from the growth of mobile products in 2015, expect to see emerging devices such as wearables and others to join the line for the mobile products. Let me ask C.C. to speak to the technology aspects of our growth engine. C.C. ?
Thank you, Chairman. Good afternoon, everybody. I'm C.C. Wei, I'll give you the update of our 28 nanometer High-K metal gate version. Let me recap the history. We start the 28 nanometer volume production in 2011, mainly on the 28 LPE, the oxynitride version. Since then, the business continued to grow. Last year, we had tripled the 28 nanometer of business versus 2012. In this year, 2014, the business 28 nanometer will continue to grow at least by another 20%. All the increase are coming from the 28 nanometer, the High-K metal gate version, which is we named it 28 HPM. Let me add more color to it. We expect we are going to have about more than 100 tape-outs from about 60 customers that in this year, in 28 HPM.
You may ask why there are so many products will be designed on this technology. One of the main reason I can give it to you is the performance, the superior performance. For example, 28 HPM compared with the 28 LPE, that will gain another 30% of the speed at the same kind of power consumption. Or you can say that at the same speed, it will consume 15% less power. Everybody knows that the power consumption in the mobile device is very important. That's why we think we have a very high growth business on the 28 HPM. Furthermore, after the 28 HPM, we also offer 28 HPC, which is a low-cost version of the 28 HPM. The 28 HPC is developed to meet the customer's demand to compete in the mid to low-end smartphone market.
We expect that this 28 HPC will have a very strong demand in the next 2 years. That's what we have. Okay. Let me give you some information on the competition to explain why we are so confident on this 28 nanometers High-K metal gate business. If you still remember that long time ago, we mentioned about gate first and gate last. Still remember that terminology? All right. Simply to say that gate-last version will give you better performance and better process control. As a result, all our customers will enjoy using the gate-last version of technology to have a higher or better performance than other products which are designed on the different approach. In addition to that, we are saying that because of the better process control and TSMC's manufacturing excellence, we have a much better yield than our competitors.
So that our customers will enjoy the lower cost. That's what we have. That probably explains our confidence that on the 28 nanometers business, continue a very good business for us. Now let me switch the gear to 20SoC . That's another exciting news that we have, I want to share with you. 20SoC is the technology that we developed to enable TSMC's customers to dig in the mobile device market. This technology, we are believing this year, next year, will have a very good business to capture. What is the status now of the 20SoC ? We have 2 fabs, fab 12 and fab 14, that complete the core of 20SoC . As a matter of fact, we started production. We are in the volume production as we speak right now.
It's in the high volume production as we are speaking right now. Let me add more information to that. First, there are more than TWD 10 billion has been committed to build the capacity. Second, we have more than 2,500 engineers and 1,500 operators right now in manufacturing doing the 20SoC volume production. The ramp rate will be the fastest one in TSMC history. Using the ramp rate, you can get a hint of the business, how big the business is. Another fact to share with you, at the end of this year, we have more than dozens of tape- out from about a dozen customers, that they are producing the 20SoC product. Okay. You might ask, "Good business. How about the competition?" If you have a very strong competition, you cannot have too much confidence on the future. Let me talk about the competition.
I'm very confident that our 20SoC is the highest gate density in volume production at the 20 nanometers node. Okay, please remember that. Highest gate density and high volume production. I don't think any competitor today can claim on this kind of production and with this kind of gate density at this time. Nobody. Most of our competitors, to be frank with you, they're not even into this game yet. We are confident we will have a good business that will contribute to TSMC's revenue, wafer revenue, by probably around 10% this year. That will conclude my presentation. Thank you for your attention.
Thank you, CC. Looking into next year now, 2015, and perhaps even 2016, I think next year this time, I will be telling you that our growth engine in technology next year, a year from now, is going to be 16 FinFET. Recently, Intel published some data which showed our 16 FinFET technology to be behind. We think that the data is highly misleading. I now would ask Mark Liu to speak to TSMC's competitiveness versus Intel and Samsung.
Good afternoon, ladies and gentlemen. I will start this topic by update you our recent development status of our 16 FinFET technology. 16 FinFET technology has been a very fast-paced development work in TSMC, and we have achieved the risk production milestone of 16 FinFET in November 13, November last year. This month, we should have the 1,000 hours, so-called the technology qualification. The technology is ready for customer product tape out. Our 16 FinFET yield improvement has been ahead of our plan. This is because we have been leveraging the yield learning of 20 SoC. Currently, 16 FinFET SRAM yield is already close to 20 SoC. With this status, we are developing an enhanced transistor version of 16 FinFET+ with 15% performance improvement. It will be the highest performance technology among all available 16 and 14 nanometer technology in 2014.
The above progress status is well ahead of schedule. Let me comment on Intel's recent graph shown in their investor meetings, showing on the screen. We usually do not comment on other companies' technology, but because this has been talking about TSMC technology, and as chairman said, has been misleading. To me, it's erroneous, based on outdated data. I like to make the following rebuttal. On this view graph, the vertical axis is the chip area on a log scale. Basically, this is to compare the chip area reduction. On the horizontal axis, it shows the four different technologies, 32, 28, 22, 20, 14, 16 FinFET, and 10 nanometer. 32 is Intel technology and 28 is TSMC technology. Is the following three nodes. The smaller number, 20, around 14 FinFET is Intel, 16 FinFET is TSMC.
On the view graph shown at Intel investor meeting, it is with the gray plot showing here. The gray plot shows the 32 and the 20 nanometer. TSMC was ahead of the area scaling. However, with 16, the gray data shows a little bit uptick, and following the same slope, go down to the 10 nanometer. What's the correct data we show on the red line? That's our current TSMC data. The 16, we have been volume production on 20 nanometer, as CC just mentioned. This is the highest density technology in production today. We took the approach of significantly using the FinFET transistor to improve the transistor performance on top of the similar back-end technology of our 20 nanometer. Therefore, we leverage the volume experience in the volume production this year to be able to immediately go down to 16 volume production next year, within one year.
This transistor performance and innovative layout methodology can improve the chip size by about 15%. This is because the driving of the transistor is much stronger so that you don't need such a big area to deliver the same driving for the circuitries. For the 10 nanometer, we haven't announced it yet, but we did communicate with many of our customers that that will be the aggressive scaling of technology we're doing. In the summary, our 10 FinFET technology will be qualified by the end of 2015. 10 FinFET transistor will be our third generation FinFET transistor. This technology will come with industry-leading performance and density. I want to leave this slide by 16 FinFET scaling is much better than Intel said, but still a little bit behind Intel. However, the real competition is between our customers' product and Intel's product or Samsung's product.
TSMC's Grand Alliance, that is the alliance of us, our customers, EDA, IP, community, and the supplier, is the largest and the only open technology platform with the widest range of product innovation in the industry today. As for the tape outs of our 16 FinFET, more than 20 customer product tape outs of 16 FinFET technology scheduled this year already. They include wide range of applications: baseband, Application processor SoCs, graphics, networking, hard disk drive, field programmable gate array, CPUs, and servers. Our 16 FinFET technology captured the vast portion of products in the semiconductor industry. We've been actively working with our customer designer on this. It's been since last year. TSMC's speed and productization of the customer's product and our ability to execute for a short time to market for our customer are far superior than Intel and Samsung.
Lastly, I would comment on the mobile products. With this 16 FinFET technology and the innovations of processor architecture and various IP from our customers, we are confident that this planned 16 FinFET mobile product which is going to take off to us, will be better than Samsung's 14 nanometer and better than Intel's 14nm SoC. Thank you very much.
Thank you, Mark. Summary. In summary, I want to say the following. First, in 2014, we expect double-digit revenue growth, and we expect to maintain or slightly improve our structural profitability. As a result, we expect our profit growth to be close to our revenue growth. In 2014, the market segment that most strongly fuels our growth is the smartphone and tablet mobile segment. The technologies that fuel our growth are the 20SoC and the 28 High-K metal gate, in both of which we have strong market share. In 2015, our strong technology growth will be 16 FinFET. We believe our Grand Alliance will out-compete both Intel and Samsung. Out-compete. Thank you very much.
All right. 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 that to English before our management answers your questions. For those of you on the call, if you would like to ask a question, please press the star then one 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 hash key. Now, let's begin the Q&A session. First question comes from the floor. It will be from Roland Shu of Citi Securities.
Happy New Year, Chairman, Dr. Liu, Dr. Wei, and Lora. I'm very happy to see Chairman, who has energy today. My first question is on the 16 nanometer FinFET and thanks for Dr. Liu, in Intel for your technology trends on the 16 nanometer FinFET. However, most of our competitors are still talking about 14 nanometer versus 16 nanometer. I know we have a 16 FinFET+. My question is on that. Whether this 16 FinFET+ is improving from the design win side or this is just for the performance enhancement? Are we going to consider to change our 16 FinFET+ to even to the better than 14 nanometer? This is a question for Chairman. Thank you.
Mark, would you answer that? If you don't mind if Mark answers it.
16 FinFET+ is a transistor enhancement. For the design, back-end design rules are similar to 16 FinFET. Designer can design on 16 FinFET and re-characterize, upgrade their product performance. This transistor, as I mentioned, also can reduce the cell size, standard cell size and with the enhanced performance transistor. That's the way to reduce the chip size. We do not intend to change the naming. This is the name that we chose earlier, based on the physically consistent number, we do not intend to change the name.
Okay, thank you. 16 FinFET+ or 16nm actually is same available as the 10T and with the status of at least a similar performance as the 14 nanometer. This is what I mean by I don't know for the question, I think. I am actually very happy to hear you are so confident about your 16 FinFET and 16 FinFET+ half the margin. Okay. My second question actually is for Lora Ho, actually for the cash dividend. Since with a very good bottom line last year and this year, and also with the improving free cash flow. Question is on that. Under what kind of a criteria TSMC is considering to increase cash dividend? Thank you.
We look at the free cash flow. When the free cash flow sufficiently covered our current TWD 3 per share dividend level, that is the time we will seriously consider to increase cash dividends.
In my model, actually, I estimate this year 2014, free cash flow probably reach about TWD 3.5, TWD 3.6 per share. This is the only one consideration. When the free cash flow to about TWD 3, you will increase the cash dividend.
That's the major element. I will not comment on your model. Other than the free cash flow consideration, another thing is we look at the capital intensity. Our capital intensity in the past four, five years has been very close to 50%. That number will come down this year, 2014. It will even come down further in 2015. That is another element for us to seriously consider cash dividend increases.
Okay. Thank you.
I guess he's asking when we raise the cash dividend. I don't think that matters to say that, certainly 2014 is perhaps the first time. 2014 dividend payable in 2015. I think it's not out of the question. Okay?
Okay. Our next question comes from the floor. That will be from Bank of America Merrill Lynch, Dan Heyler.
Thanks. I came with a lot of questions. You answered a lot of them earlier on. I've got two. I first want to comment to congratulate Dr. Chang, and your team as well, for your achieving your 3-5-year target of turning around technology and turning around growth in the company. It's been remarkable to see. Also, congratulations on handing off the transition here to the management team, too.
Thank you.
I've got two questions. First, I guess as we look at, maybe perhaps for C.C. Wei on the technology side. As we look at EUV, do you agree that most likely implementation of EUV occurs to be for foundry, number 1, over kind of the one to two layers at 10 nanometer? That you can kind of achieve those X number wafers per hour to hit a rollout of that technology. What exactly wafers per hour do you need, and at what time will you achieve those?
Mark, will you answer that, please?
I think he was asking whether we are considering using EUV in 10 nanometer, maybe one or two layers. That's the question, right?
Yes.
The current process of record, our current execution plan is on 10 nanometer, do not have EUV. It's all in immersion layers. However, we have been working with ASML very closely, we set a target for their EUV's throughput. Together we have the ambition to improve the 10 nanometer cost if the EUV development in ASML can reach our target, the target we gave them. Okay. These two parties have been working very closely. EUV is a cost reduction opportunity for us on 10 nanometer.
Okay, great. A quick follow-up on that. If we were to draw that curve that you put up earlier, would that look fairly linear based on your current model? On your cost curve previously that you showed on-
That curve is without EUV.
Exactly. With EUV we imagine a dotted line, do we go through a bit of a penalty there and then we're back on trend? Do you think it's even-
It's the same. With EUV, our design would be the same. We're not going to wait for that. We're looking at EUV just as a cost reduction.
It's a cost reduction opportunity.
Excellent. Thank you very much. I was going to ask you previously on managing these swings in utilization that we've seen over the last couple of years. You're addressing that through production during the cycles. Are there other things that you can do going forward to smooth out these big swings in utilization? I guess part of the answer, I'd like to know what you think the industry dynamics will be on these swings, and how should we think of it? Is it pretty much business as usual? From an order standpoint the last three years-
Well.
Could we see the orders move? Thank you.
I think that Lora just explained what we have started to do, which is really what the IDMs do all the time. That is, they build some inventory when the utilization is slow. They build some inventory. That's what we are doing, too. Of course, in our case, we have to bet right. I do believe that we have organized ourselves. I wanted to start doing it as early as August of last year, because even at that time, we saw that the fourth quarter and the first quarter would be rather slow. Well, we had to organize. Organization, people have to bet, if not their job, but their bonus is on the line. That this inventory we are building will not be written off. Also, we have to decide how much reserve to take, et cetera, et cetera.
Now we didn't get completely organized. We really didn't get going until sometime around the middle of fourth quarter. The effect of inventory building on the fourth quarter was minimal. In the first quarter, as Lora has already pointed out, it does have an effect. I think she said two or three percentage points of gross margin.
Is that two to three percentage points something that you're modeling through the course of the year? Is that to see what?
I don't think so
through the course of the cycle, or is that just one quarter?
Yeah, I don't think you can model it as for the whole year, because we have to look at the lower season and off-season. They are meant to be due in the lower season. Only the low quarters will start to produce some inventory.
Okay. I misspoke there. I kind of meant through the cycle. Is that kind of your expectations for as you go through the cycle? At each of the troughs, could we kind of see this as something you can repeat on the next correction, a couple percentage points?
The answer is no. When we see that we need to use it, we'll do it.
Thank you very much.
We don't model it way in advance, okay?
Is there anything else you can do to affect the volatility and improvement in margins? Are there other things in the toolbox there?
Yeah. For instance, we very carefully control the capacity build.
All right. We will now take our next question from the call. Operator, please proceed with the first caller on the line. Your first question from the phone comes from Randy Abrams from Credit Suisse. Please ask the question.
Hey. Thank you. My first question on the management structure now with the Co-COOs promoted to Co-CEO. If you could talk about how the responsibilities will change with their promotion to Co-CEO. For yourself, Dr. Chang, how will your activities change versus before this move? If you could talk about the roles for each of the different Co-CEOs and yourself now.
Yes. We started the present Co-CEO in November, and it has been now two months. If you ask me now, has my life changed in the last two months? My answer is no, it has not changed. I think that my effort, my time has been spent more on the coaching aspect. I do believe that I do more coaching. If I spent 100 hours, I now perhaps spend 20 hours of the 100 hours on coaching, whereas in the past, I probably spent only five or 10 hours of the 100 hours on coaching. Actually, this is an overseas call. Is this correct? Yeah. Let me just explain very briefly what the Taiwan law and customs are in relation to a chairman's authority and responsibility. Basically, by both law and custom, the chairman of a company has the ultimate authority and responsibility.
However, he may delegate this authority and responsibility to the President. He may also take it back anytime. He may delegate any and all of the responsibilities to the President. These two gentlemen, their title is President and Co-CEO. President comes first. They are, in a very legal sense, Presidents. The Co-CEO is basically a Western term, in the U.S., a CEO usually bears the final ultimate responsibility and authority as a Chairman in Taiwan does. In the U.S., it's the CEO. My goal in the future is really to convert these two gentlemen from the Taiwan sense President to the U.S. sense CEO, and it will be a gradual process. Do you think I've explained this clearly? Yes?
Randy, do you think Chairman has explained your questions clearly?
No, that was clear. I could ask a second question. It goes back to the last conference. There was the talk about the battle raging on 16 nanometer in 2015. I'd want, I guess, an update on how, given the technology progress you've made, how you're looking at 2015, those battles playing out.
On what?
Your thoughts on your market share position, and also if you see any challenge, whether it's market share or pricing with two new-
with Samsung and Intel both more aggressively at least pursuing foundry.
Could you repeat the question? I think it's about 16-
16 nanometer battle in 2015.
Oh, 16 nanometer.
Right now, we may intend to maintain our position in the foundry business on 16 FinFET, as we did on the 28. As I mentioned, our 16 FinFET, we have over 20 tape out already working with our customer. Of course, these products will be competing with Intel's product in 2015. That is that we are so closely working with our customers together. The technology is ready to tape out. We need to work with our customer closely so that their design and the time to market can be catched as early as possible for their 2015 ramp.
Well, we live on technology just as Intel thinks or says, maybe we do that too. It's extremely important to us that on 16 FinFET, we need to provide good enough technology so that the grand alliance of customers, our key equipment vendors, the EDA, Arm, and third-party IP developers, all those are in the grand alliance. Our role is to provide good technology so that the grand alliance together can out-compete Intel and Samsung. We feel pretty confident that we'll do that. If you talk about market share, we feel that we'll win a pretty big share of the market. I don't know how much yet, but I think that it'll be pretty big.
All right. Now we are coming back to the floor. Next question will be coming from Michael Chou of Deutsche Bank.
Actually, it's a follow-up question on 16 nanometer, because you mentioned you should take the lead.
We should comment regarding the competition. What's your power consumption efficiency versus competitors 14 nanometer FinFET? Would you be still the leader in power consumption efficiency at the same speed versus competitors of 14 nanometer FinFET normal devices in 2015?
In their transistor design, the speed and power are convertible. From our intelligence, our 16 FinFET+ technology, with 15% improvement on top of 16 FinFET, is about the same as Intel's transistors. That is what we are targeting at, and customer can convert that speed to power consumption.
Okay. Second question is, could you update your gross margin expectation for 20 nanometer? The guide for high single-digit growth on 20 nanometer.
Didn't hear that. We expect.
Mark.
I think it will be down.
I thought I already said around 10% wafer revenue.
For the year.
For the year.
For the whole year.
Yes.
It will, by the fourth quarter, will be much higher, yeah.
How about the full 2016?
Negligible. Q4 will be.
Q4 this year is very high-
Q4 this year-
Q4 2014 will be very high. That's what I said earlier. Q4 last year was negligible.
With 10% for the total year of wafer revenue, we expect full quarter along, 20-nanometer will contribute more than 20% of our wafer revenue.
Thank you.
Okay, next question will come from the floor again. It will be Barclays, Andrew Lu.
Dr. Chang, Dr. Liu, Dr. Wei, and Lora. I have two questions here. First one is regarding the capacity design for 20-nanometer and 16-nanometer FinFET. Some of you indicate you are going to prepare 60Ks a month for 20 and 60Ks a month for 16-nanometer FinFET. Last time, remember, Dr. Chang mentioned the 16-nanometer FinFET capacity is quite similar, 90% switchable. can we-
More than.
Do that for next year, CapEx itself and the CapEx number, probably even lower than this year?
Oh, I don't want to say that. We may-
Maybe equal.
We may capture even more market share than we now think we will. If that happens, we'll spend money. You're right about the 20 to 16 capacity, or 20 capacity being quickly convertible to 16, at a loss that's a lot less than the 10% that you cite. A lot less.
What do you mean?
Well, you said 90%, I think.
Yeah. I'm saying it's more than 90%.
Switchable.
Switchable.
We might see a kind of different case. Depends on subject change on customer demand. We might see the next year, if 16 nanometer FinFET is quite packed, we might not see 50/50% 20 and 16.
No. We are not, by the way, we are not building 60K per month capacity of 20 and then 60K per month capacity of 16. They are not additive.
Okay, thank you.
Yeah.
Kind of a floating change based on the customer demand.
I think it floats one way, from 20 to 16. It will not float the other way. Yeah.
My second question is, since the 14 and 16 nanometer FinFET argument among the top six players so much, why we choose 16 FinFET at first place?
I think it's your strategy thinking back to two, three years back.
Well, the best laid plan by men is sometimes undone by God, okay. Let me tell you a story. Well, it's not God, it's customers, okay. We actually first got committed to the 20. We got committed. Suddenly, we realized that there was a need for FinFET, for a faster FinFET. This meant one of the gorillas in the business announced that they were going to do FinFET, okay. Now, if we hadn't committed already to 20, we might have skipped 20. I think we would have skipped 20. We were committed already. That was a good thing, getting committed, well, the customer was committed too. That was a good thing. That's why we gained a great deal, I think, on the 20. We had to hasten up to do the 16 now.
I think the team in TSMC, the R&D team, did a tremendous job in now getting the 16 to where it is now. That's the brief history. Did that answer your question? Oh, okay. Not exactly. What was your question exactly?
Okay. Let me ask another way. Two years ago, when you considered 20 and 16 investments, do you consider switchable as an important part?
Yes. Always. For instance, when we consider 10 nanometer, we consider the 16 to 10 switchability. We keep saying switchability. Actually, it's just convertibility. One-way convertibility. I don't see it being converted back. That is an important factor when we consider.
If we want to know the answer, if we invest our fortune in 10, 15, what equipment can use for 14 nanometer versus 16.
If we invest in 14, we're not investing in 14. I'm sorry. Did I misunderstand you? You said if we invest in 14.
No, if earlier decision is 20 and 14, what percentage-
Well, earlier decision-
of 20 nanometer can be used in 14 nanometer?
My goodness. The earlier decision was not 20 and 14. Earlier decision was 20 and 16. The reason it was Well, actually, 16 and 14 is the same. What the heck?
Okay. Since there's so many hands being raised so many times, I have to go to HSBC, Steven Pelayo.
Great. When you're talking about 20 nanometer, you talked about it being, I think, as much as 30% faster ramp than once you had at 28 nanometer, 20% of revenue by the fourth quarter. That's all the top line. I'm curious about the margin impact to it. Obviously, you get a good ASP premium, but I think it took about, I don't know, maybe five quarters to get 28 nanometer corporate average margin. Could you talk a little bit about the pace, given the accelerated ramp to get 20 nanometer for corporate average margin?
Well, actually, five is pretty normal. I mean, yeah, 28, it took, I guess five, right? Was it five?
I guess.
Yeah. Mm-hmm.
7 or 8 quarter for 28 Nanometer. I can tell you.
Eight
20 SoC will be the same as 28 Nanometer, will reach corporate average at 7, 8 quarters, like 28 Nanometer.
If I could just get a clarification on the smoothing effect to the first quarter to try to take in some of the potential ramp in the second quarter? Can you measure how much is that? Does that then impact the potential growth in the second quarter? Are we taking 2 or 3 percentage points of growth away from the second quarter by pulling that into the first quarter? Do you think it doesn't really impact the fact there will be enough demand to continue to-
No. The so-called smoothing, it helps the margin. It does not affect the revenue.
If I can add one more point. The real purpose, not to lose capacity. When you have a surge demand and your capacity is not enough, and you lose the demand, you lose the business. By switching and smooth our among quarters, we don't lose the product and don't lose the market share. That's the main purpose. We don't have to build so much capacity, especially such peak capacity probably only exist for a few quarters, which is bad for the company. You understand that? Okay.
The next question will come from JPMorgan's Gokul Hariharan.
Hi. Thanks for taking my question. Dr. Chang, Dr. Wei and C.C. Wei and Lora Ho. I had a quick question, just switching gears a little on the back-end side. We hear a lot about TSMC working on the back-end side, especially on 2.5D, 3D, as well as die-level InFO. You also had given a kind of a target of about a TWD 1 billion revenue sometime in 2015 or 2016. Can you update us on what your plans are on the back-end side? Secondly, also, how does that work when you work with the back-end partners? Are you going to be some budgets going into the just die-level stuff, or is it going to die-level stuff as well? Thank you.
Excuse me. Do I need to repeat the question? Okay. To answer your question on the revenue of reaching the TWD 1 billion for the CoWoS and 3D IC. Since we announced the CoWoS technology and we work on the 3D IC, we're using a lot of key technologies and develop a lot of derivative applications. Today we estimate probably not go up to TWD 1 billion, but it's around TWD 800 million in 2016. That's our estimate today. All right. What's your second question, by the way?
Yeah. Second is, the back-end activity, is it going to be limited only to the very high-end, 2.5D, 3D related products? Or is it also going to be whatever is going to be wafer level package, like Wafer-Level Chip-Scale Packaging, those kind of things?
Okay. The correct answer is that we are not shooting for some kind of technology by ourselves. We're always working with the customer to meet their demand. All right? To answer your question is that whatever or whenever that our customer ask us to integrate and to give them better service, we do it. All right. It's not limited to the 2.5D, 3D, or whatever it is. All right.
Next question will come from the floor again. It will be from Goldman Sachs, Donald Lu.
First question is I want to ask the Chairman, are you satisfied with the transition so far? Also, how the two presidents will share their work? Are they still rotating or not?
Are they still?
The three Co-CEOs were rotating, probably not now. Maybe give us some details about how the company is run, I will have a follow-up question on competition.
All right. I am quite satisfied with the transition. These two gentlemen, Mark is now responsible for sales, marketing, strategic planning, business development, information technology and materials management. All those. C.C. is responsible for operations, all the operations. He's also responsible for specialty technology, R&D. Specialty technology, incidentally, accounts for 25% of our total business. Donald, your other question is whether they're going to rotate. My plan currently is I don't plan that way right now. However, I view it as a pretty flexible thing. Tomorrow, I may take one part of Mark's, give it to C.C., or vice versa. I'm not considering rotation, per se. Yeah. Does that answer your first question?
Yes. Thank you. My second question is, this is as far as I can remember, the first time TSMC comments on competitors, namely Intel and Samsung.
That's because you picked the ones.
I understand. That is a very good reason. Okay. Why don't give us more color? 16 nanometer, for example, are we saying that in terms of die size, performance, our product would be very similar to Intel 14 nanometer FinFET? Also, Mark commented that for the FinFET pickup, specifically with the CPU and server chip. Can we say that TSMC CPU and server chip will have the similar physical performance as Intel product today?
Well, I think, Donald, we have already given everybody enough information on our 16nm. I think that if we keep giving more, we would be helping our competitors who have picked us. Now, we do stand on what we said. Our Grand Alliance will out-compete Intel and Samsung. Our Grand Alliance on the 16 FinFET will out-compete. By that, I don't mean that we'll compete to exclude them. No, we can't, we won't be able to do that. Our Grand Alliance, with us as the foundry supplier, will capture a large share of the 16 market. Do you agree with that, Mark Liu?
Yes. I think if you have a product in mind today, want to get to 16 FinFET, I think that your job will be more secure coming to TSMC, and make sure your product will be able to deliver in the market.
Okay. The next question will come from Morgan Stanley's Bill Lu.
Smartphone what?
Smartphone
Yeah, right. Yeah, I said that just now. No, I'm no expert on that. There are a lot of other non-semiconductor stuff in the handset. Maybe those people are getting squeezed, or maybe the handset maker is getting squeezed, or maybe the carrier, the subsidizers is getting squeezed. You're asking me a general question, which I'm not expert enough to answer.
Just to make sure I understand. You kind of categorized it into high-end, mid-end, and low-end. Do you see that the semiconductor is going up for each one of these segments, or what's the-
Do we have data on that, Lora?
The high-end content definitely will go up, but the growth for high-end will be slower than the mid and low-end. The mid and low-end stay constant, but because mid and low-end has higher volume, which to TSMC advantage.
Well, I guess I had a microphone already. I actually look more at our value in the smartphone. It's gone up. This is TSMC, right? It's gone up from average meaning averaging high-end, mid-end, and low-end. Average of TWD 6 in 2012 to TWD 8-- This one? To TWD 8 last year, to TWD 11 this year. Six, eight, eleven.
That's very helpful. Thank you.
You're welcome.
Second question is for Dr. Wei. I think the expectation is that you are going to try to switch customers from poly to High-K metal gate. If you look at the market this year for 28, what do you think is a split between High-K metal gate and poly? Is it 70/30? Is it 80/20? Can you also talk about your expected market share for TSMC for High-K metal gate and also for poly? Thanks.
In terms of TSMC, I would say greater than 80% of the 28 nanometers built will be in High-K metal gate.
How about for the industry?
I cannot comment on that.
Thank you.
Thanks.
Okay. The next question, I think we should go back to the line. Operator, could you please proceed with the next caller on the line? Thank you. The next question from the line comes from Mehdi Hosseini from SIG. Please ask the question.
Thank you for taking my question. I have two. The first one is for Dr. Liu and Dr. Wei. You offered some metrics comparing HPM with LP, talking about speed improving by 30% and power by 15%. Would you be able to offer similar targets or similar metrics, comparing '28 to '20 and '20 to '16? I have a follow-up.
I think it sounds like. Did you hear?
Well, the question is that they ask from the 28 nanometer HPM compared with the 28 LP.
Go ahead and answer that, if you know what the question is.
All right. If I repeat the question correctly, they ask about from 28 HPM, compared with the 28 LP, there will be 30% speed improvement. Now, how about the 16 FinFET? 16 FinFET actually improve much more compared with the 28 HPM.
Okay.
Does that answer the question?
What about 28 to-
Does that answer the question?
Yes, how about 20?
I cannot hear you clearly.
Mehdi, are you still on the line?
Yes. Can you hear me?
Mehdi, are you still on the line? Okay, Mehdi, I think your question was asking Dr. Wei to explain what's the power and speed improvement from 28 to 20 and from 20 to 16.
Yes.
From 28 to 20, we got about 20% improvement in the performance. From the 20 to 16, specifically talk about 20 SoC to 16 FinFET. The 16 FinFET will have a higher than 30% improvement in the speed. Does that answer the question, again?
Yes. Yes, it does.
Okay. Thank you.
Yes. Let me move on to the second question that I have for Dr. Chang. Over the past two years, we have seen a significant improvement in revenues from Q1 to Q2. This was followed by seasonal or below seasonal trends in the second half. How do you see this year evolving? Do you see this year being any different than the trend over the past two years?
Mehdi, your question is about seasonality of this year, whether we will have a strong rebound in the second quarter from the first quarter,
Yes
Whether we will have a stronger second half than the first half.
Our current outlook is.
Yes, correct.
Our current outlook is-
Yes, correct.
Our current outlook is very anti normal seasonal. Our current outlook is that the first quarter is the lowest, second quarter is stronger, third quarter is stronger yet, even fourth quarter is a bit stronger than the third quarter. I certainly would not call that anything seasonal at all. I think that that's pretty unique for this year. Mainly because of the ramp-up of 20 SoC.
Thank you. That's very helpful.
Thank you. We come back to the floor. The next question comes from Daiwa Securities' Eric Chen.
Hi, Dr. Chang, Dr. Liu, Dr. C.C. Wei, and Lora, Happy New Year. Let me know if I get it wrong. You just mentioned your smartphone IC, the ASP getting higher. Right now, it's like $11. I believe, probably because of the, what do we call about a smartphone from a single core to dual core to quad core. My point is that, probably we will now have the 16 core, and 16 core FinFET probably will be the next launch. I try to say is that we probably will see the, what do we really call the low-end smartphone, the [ramp] will really catch up. That probably will give you some ASP impact. Actually, my question is your strategy in terms of the low-end smartphone IC.
Once I look at the China smartphone IC maker, your market share probably, if I'm saying this, again, let me know if I'm wrong. Your market share as the China smartphone IC maker, probably lower than your market share as the global smartphone IC maker. What is your strategy to deal with the wafer fab competition?
We intend to be a significant factor in the China market. We intend to be a significant factor in the medium to low-end, middle to low-end
Kind of smartphones.
Okay, let me say more exactly. The data where you talk about the 28, the HPC, that's mainly for the lower-end mobile applications. For the client, why I choose the 28 HPC, why don't I choose the 28, the poly/SiON, what kind of advantage can we talk about? The ending thing, then finally, the technology.
28, yeah.
Okay. 28 HPC is still the High-K metal gate version.
Yes.
The performance will be much better than 28 LP. It is a low-cost version. You can trade off the performance and the die size sometimes. You can get your optimal position that depends on your product. From our point of view, and we work with the customer, that we offer this solution for them to meet the challenge of the market, of mid, low-end smartphone. It's a low-end, you expect the pricing to be lower. For us, we offer the technology to compensate that.
How many % is performance improvement and how many % the price is lower than the gap from the poly/SiON and how about the trade?
From LP to HP.
From HP.
Yeah. From LP to HPC. I'm not going to tell you how many % of the price or something like that. I'll tell you the
You realize that it depends on the kind of designs too.
Yeah.
As far as the intrinsic stuff is concerned.
You are mentioning about the performance or the speed. The speed, actually, 28 HPC, they're exactly the same as the 28 HPM.
How many % better?
30% better.
30% better.
TSMC, the customer, can trade off that 30% the speed again, performance, and then trading with the die size, so that you can get the die size.
Right.
If you don't enjoy the 30% performance, you can enjoy the die size shrinking.
Okay, I see.
That's the idea.
I assume the wafer price is probably like 10% higher than the poly/SiON.
We're not going to talk about wafer price.
Okay. Thank you.
We will tell our customers wafer price.
Okay.
You're not our customer, we're not going to tell you that.
Thank you. Basically, the same question as the founder, the question for Lora. I remember that Dr. Chang just mentioned that all the ASPs here were down by several points. My question, when I look at.
No, that's not what I said. I said our blended ASP will rise. It will not jump, it will rise.
Sorry.
Our blended ASP will rise by several points.
The rise also quite positive. The rise of several points. When I look at the TSMC, the revenue growth is 18% year-on-year, but utilization rate is down. The competitor probably grows another 14% currently year-on-year. My point is that TSMC's success last year is because one of the main reasons because of blended ASP, and I believe the blended ASP last year probably rose by probably 5% more. Either Morris Chang or Lora Ho, how you think about this year? You mentioned the 20 nanometer process in Q4 probably will reach the 10% of the total revenue. Sorry, 20 will reach the 10% of the total revenue. In that case, can we give the same expectations that the TSMC, the blended ASP, and will rise by over 5% year-on-year?
Actually, chairman just gave you the answer. He said the blended ASP with the product mix richer, it will go up several points. He has already answered your question. What's your true question?
Eric Chen wants to know-
I would like to know compared to the last year-
He wants to know.
We have a stronger confidence on the ASP.
Well, Eric assumed that 2013, our ASP went up by 5%, which we will not comment. He wants to know whether 2014 we can go up above 5% or below 5%, which we will not comment. Okay, next question will come from the floor, and it will be UBS, Jonah Cheng.
Just one small question from me. Recently, we know the China government, they try to launch some stimulus programs to the local semi industry. We know it's very difficult for them to do the semiconductor, but maybe, you think there is an opportunity with the fabless or something from equipment side or how can TSMC position to get this kind of opportunity or business? That's my question. Thank you.
No, we have not considered that. We understand the impact that only Chinese fabless companies will benefit from. Not companies located in China, but truly Chinese-registered companies, that only they can benefit from.
This subsidy. Funding all fabless.
I think, Jonah, your question is, can TSMC benefit from the benefit of the Chinese companies?
Well, I don't know about that. Does anyone have an idea? We do have a lot of business in China from Chinese fabless companies. Our business has been rapidly growing. I somehow doubt that in giving us their business, they are benefiting from the subsidy. I somehow doubt that, but I don't know for sure.
May I add a point?
Yeah.
I think we will benefit.
Oh, we will?
The fabless company in China are very aggressive approaching leading-edge technologies. Just to tell you, on 16 FinFET this year, already some of the fabless companies will be using it in tape outs. I think all those fabless companies, subsidy will propel them into the leading-edge technology more.
I don't see how those two are related. The subsidy is not for them to use leading edge, is it?
Our leading-edge technology is not yet available in China.
Yeah, that I know. How is that related to the subsidies again?
Their business will be aggressive moving into leading edge, which will benefit our competitive advantage.
They will be generally better off after receiving the subsidy. That allows them to move faster into leading edge. That's the point. That's just like saying that somebody will sell to us. To me, will be better off if I make more money. What happens? Or anything, never mind. Okay.
All right. Well, I think with the strong prospect in China, we will be very happy to end our quarterly conference call here. Before we conclude the conference, please be advised that the replay of the conference will be accessible within 3 hours from now. Transcript will become available in 24 hours from now. Thank you for joining us today. We hope you will join us again next quarter. Goodbye and have a good day.