Elizabeth Sun - Director of Corporate Communications Lora Ho - SVP, CFO Mark Liu - President and Co-CEO C.C. Wei - President and Co-CEO Morris Chang - Chairman

Dan Heyler - BofA Merrill Lynch Michael Chou - Deutsche Bank Randy Abrams - Credit Suisse Donald Lu - Goldman Sachs Eric Chen - UBS Roland Shu - Citigroup Rick Hsu - Daiwa Gokul Hariharan - JPMorgan Randy Abrams - Credit Suisse Eric Chen - UBS

Welcome to TSMC's second quarter 2015 earnings conference and conference call. This is Elizabeth Sun, TSMC's Director of Corporate Communications and your host for today. Today's event is webcast live via TSMC's website at www.tsmc.com. If you are joining us through the conference call, your dial-in lines are in listen-only mode. As this conference is being viewed by investors around the world, we will conduct the 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 second quarter, followed by our guidance for the third quarter. Afterwards, TSMC's two co-CEOs, Dr. Mark Liu and Dr. C.C. Wei, and TSMC's CFO Lora Ho 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's website at www.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 statement. Please refer to the Safe Harbor Notice that appears on our press release. And now I would like to turn the podium to TSMC's CFO, Ms. Lora Ho, for the summary of operations and current quarter guidance.

Thank you, Elizabeth. Good afternoon everyone. Welcome to join us today. My presentation will start with financial highlights for the second quarter, followed by the guidance for the third quarter. So now let me summarize our second quarter financial performance. In the second quarter, we have achieved NT$205 billion in revenue, 48.5% gross margin, 37.5% operating margin, and NT$3.06 in EPS, which were all within our guidance range. On a year-over-year basis, our revenue increased 12.2%. Net income and EPS both increased 33% versus the last year same quarter. On a sequential basis, our revenue decreased by 7.5% due to customers' cautious inventory management and the less favorable exchange rate. However, we were able to largely offset the negative impacts with our continuous cost improvement efforts. Our gross margin decreased slightly to 48.5%. For the non-operating items, we have reported again of NT$21 billion. This significant one-time gain mainly came from two transactions. First, the ASML share disposal, we recognized NT$17.6 billion in the second quarter, which contributed to NT$0.60 to our EPS. Also we sold 5% of Vanguard shares, with a gain of NT$2.3 billion, which contributed to NT$0.08 of the second quarter EPS. The share disposal does not change our strategic relationship with ASML and Vanguard. We remain to be the largest shareholder of Vanguard, own 28%. You may have noticed, we incurred higher tax rate in the second quarter. The tax expense went up to 24% in the second quarter as we accrued 10% retained earnings. We expect the 2015 full year tax rate will be about 14% Let's take a look at the revenue by application. During the second quarter, all major segments showed declines. Communication, computer, consumer, and industrial standards declined 3%, 24%, 21% and 1%, respectively. In terms of revenue by technology, revenue contribution from 20 nanometer went up to 20% from 16% in the first quarter, while 28 nanometer contributed 27% of our wafer revenue this quarter. Accordingly, these two technologies, 20 nanometer and 28 nanometer, represented 47% of our second quarter wafer revenue, increased from the 46% from the first quarter. Moving to balance sheet. On the asset side, cash and marketable securities increased NT$32 billion to reach NT$550 billion at the end of the second quarter. On the liability side, current liabilities increased NT$122 billion, which was due to the accrual of NT$117 billion for cash dividend. On financial ratios, accounts receivable turnover days remained flat at 44 days. Days of inventory increased by 5 days to 62 days. So we preview certain wafers in anticipation of the capacity conversion from 20 nanometer to 16 nanometer. In addition, higher raw material and longer production cycle time for leading nodes also increased the DOI a bit. Lastly, I would like to make a few comments on cash flow and CapEx. During the second quarter we generated NT$111 billion cash from operations and invested NT$54 billion in capital expenditure. Additionally, we received NT$39 billion from disposal of ASML shares and about NT$4 billion from the sell-down of 5% Vanguard shares. On financial -- financing cash flow, we repaid NT$13 billion for short-term debt. As a result, our cash balance reached NT$529 billion at the end of the second quarter. The above are my comments on the second quarter financial performance. Now let me turn to the third quarter outlook. End-demand recovery is not as strong as we expected earlier. Customer continue to remain cautious in inventory management. Combining these factors with customers' production -- product transition, demand for our third quarter is expected to recover only modestly. Based on our current business outlook and exchange rate assumption of $1 to NT$31, we expect third quarter revenue to be between NT$207 billion and NT$210 billion, translate into 1% to 2% sequential growth; gross profit margin to be between 47% and 49%; and operating margin to be between 36.5% and 38.5%. As the remaining 20% ASML shares will be sold in the second half, we expect to recognize additional NT$4 billion disposal gain in the second half, which will translate into NT$0.12 and NT$0.01 EPS contribution to the third quarter and fourth quarter, respectively. This concludes my remarks. Let me turn the podium to co-CEO Mark Liu for his comments.

Good afternoon. I'd like to deliver the several key message to you. The first, I will cover the near-term demand and supply chain inventory. Lora reported our second quarter results were in line with guidance made in our last investor meeting, which shows a 12.2% year-to-year revenue growth. In the April investor conference, we have also noted that supply chain inventory was quite high at the end of first quarter, but it was expected to be brought down to its seasonal level at the end of Q2. However, during the second quarter, Q2, we saw demand for smartphones became weaker than we expected, due to slower demand in emerging markets and in China for mid and low-end smartphones. This weaker demand is partly due to a strong US dollar to emerging market currencies and partly due to the regional economic conditions. As a result, the excess inventory in the supply chain has only been digested about half at the end of second quarter. The recent macro economy uncertainties in many parts of the world has further dampened supply chain's confidence in end-market demand and has caused customers to become even more cautious in managing their inventory. Slow demand in emerging markets and in China, more cautious inventory management by our customers, and macroeconomic uncertainties in many parts of the world -- those three reasons are behind our modest growth outlook in the third quarter. That being said, we expect our customers' end-market demand will improve in the second half from the first half. Growth is expected to come from industrial and automotive segments, as well as from new iPhone launches and several launches of Android-based high-end phones. In addition, the continuing 4G migration in China and the demand recovery in emerging markets will further support the growth outlook of second half this year. But as the demand outlook has changed, we update our 2015 full year growth forecast as follows. Semiconductor growth revised from 4% to 3%. Foundry growth from 10% to 6%. For TSMC, we are still confident to outperform the foundry segment and still target double-digit growth rates this year. Now the message on TSMC market segment share. Thanks to our strong leadership in advanced technologies, TSMC has been able to gain foundry market segment shares in recent years. This year our market share will again be well-supported by the expansion of our advanced notes, namely 16 nanometer, 20 nanometer and 28 nanometer. For 16 nanometer, we are starting our volume shipment as we speak. The ramping of our 16 nanometer will be very steep, even steeper than our 20 nanometer. Ramping profile, similar ramping profile at similar early stage. Looking out to the future, with many more customers joining our 16 nanometer production, we are confident that we will achieve a far majority foundry share in 16 nanometer in 2016 and beyond. For 20 nanometer, we remain the only foundry capable of volume supply in the second year of its ramp-up. For 28 nanometer, we continue to strengthen our technology offerings. Following our 28LP, 28HPL, 28HPM, we have 28LPRF, 28HPC and 28HPC+. Those new offerings will enable us to protect our 28 nanometer foundry segment share. With all the above, we continue to gain market segment share in 2015. Lastly, I'd like to have some comments, key messages on advanced technology development on 10 nanometer and 7 nanometer. The recent progress of our 10 nanometer technology development is very encouraging and on track with our plan. Technology risk start qualification is targeted at the end of this year, followed by many customers' product qualifications. Our volume production is planned to start from the end of 2016. Our 10 nanometer technology is designed with excellent transistor performance spec and very aggressive chip-scaling factors. Compared with TSMC's 16 FinFET+, our 10 nanometer features has more than 15% speed gain at the same total power, or more than 35% power reduction at the same speed, and with k density of 2.2 times of that of 16 FinFET+. Many of our first-wave technology adopters have signed up for tape-outs with our 10 nanometer. So far, planned tape-outs have already include mobile application processors, network processors, and high-performance computing segments. The development activity on our 7 nanometer is also ongoing with full steam. We have a parallel team working on that program. We target 7 nanometer technology qualification in the first quarter 2017, only five quarters after 10 nanometer. With further transistor speed enhancement and chip scaling from 10 nanometer, our customer can plan their tape-outs using the latest and the greatest technology available at the time when they launch their most competitive products. For 7 nanometer, similar to our 20 nanometer and 16 nanometer relationship, we are developing 7 nanometer to be able to leverage the process tool compatibility and maturity from 10 nanometer volume message. Now I turn the podium to C.C. C.C. Wei: Thank you, Mark. I will update the technology improvement volume manufacturing and competitiveness on three technology nodes, followed by Info business update. First, on 28 nanometer. The demand outlook for TSMC's 28 nanometer remains strong. We continue to enhance our 28 nanometer technology, just like Mark just mentioned, by improving the performance. In addition to the 28LRP, 28HPM, we have introduced 28HPC last year to enable our customers' conversion to 64-bit CPU [indiscernible] market. Our 28HPC+ introduced this year can enable our customers to go after the market with multi-core which is [indiscernible] to eight to ten and advanced LTE such as LTE Category 6 to Categories 4 to 6. In addition to addressing the demand for the mid to low-end smartphone market, we have already seen demand for our 28 nanometer transceiver RF and [indiscernible] controller begin increasing over time. Based on tape-out activities, we also anticipate customer in Wi-Fi, wearable, digital TV, set-top box, and image signal process, will also start ramping next year using our 28 nanometer. Ever since we introduced 28HPC and 28HPC+, we are met with customers [indiscernible] adoption. Almost all the new tape-outs are adopting either of these two processes. And the number of new tape-outs continue to increase and reaching a record level. Our 28 nanometer [feed-in] utilization rate was in the high 80s in second quarter, which is due to customers' inventory management. We expect this [feed-in] utilization to be recovered to above 90% in third quarter, the same as we mentioned in our last conference. We intend to keep our 28 nanometer utilization rate very high by providing the best and useful technology such as our 28HPC and 28HPC+ to our customers so that their new products will grow nicely in the market, and that will translate to increasing demand and higher loading rate for us. Meanwhile, after having manufactured 28 nanometer for more than five years, we can also use our [learning] curve or cost advantage to compete with good profit margins. To summarize, we will keep our utilization rate high and we will use our technology and our cost advantage to compete effectively for 28 nanometer business. Next, 20 nanometer. Since we have begun ramping of 20 nanometer in the middle of last year, we have obtained the best progress in the reduction of defect density as compared to all previous nodes. We believe we will be able to leverage this record progress and enhance defect density reduction for our 16 nanometer. The demand for our 20 nanometer this year is still good. We expect our revenue from 20 nanometer will at least double this. And since a majority of 20 nanometer business will migrate to 16 FinFET next year, we expect our 20 nanometer business will be lower than this year. However, we still expect 20 nanometer to be a long-lived node. We are converting right now part of 20 nanometer capacity into 16FinFET to prepare for the ramp of 16 FinFET. So I move to 16 FinFET update. We have begun volume production of 16 FinFET in second quarter. Shipment has started this month. The high volume ramp in third quarter were mostly contributed to revenue in fourth quarter this year. Since 16 nanometer shares similar metal backend process with 20 nanometer, our 16 FinFET can benefit a lot from 20 nanometer's learning. We have already shipped more than half-a-million wafer 20 SoC by now, so our 16 nanometer [indiscernible] defect density has been excellent. And in fact, our 16 FinFET has set a new record for progressive [indiscernible] the defect density reduction. As for device performance, we believe our 16 FinFET class has the best transistor performance among all foundries. For low-power application, we have been developing 16FFC, which will be cost-effective and will be important for future of IoT applications as well. Last, let me update on Info business. The schedule to ramp up Info in second quarter next year remains unchanged. We have already begun moving manufacturing equipment into our new facility in Longtan which is close to be completed. Now we are also engaging with major customer for volume production in the second half next year. Our expectation of Info contributing more than $100 million quarterly by 4Q next year remains unchanged. TSMC's Info technology does not need [indiscernible] substrate and also provides a high density interconnect. This allows our solution to have benefit in form factor, that is, [indiscernible] and in electrical performance. We continue to work with major customer on developing the next-generation Info for further improvement and device performance and [indiscernible] service. I thank you for your attention. Now, turn the podium to Lora.

So I will make comments on structural profitability, CapEx and the long-term financial objectives. Let me start with structural profitability. We employ two key indices to monitor our structural profitability. The first one is expand the gross margin, referred to SGM, meaning the gross margin at a given level of utilization. The second one is expand the utilization, the utilization level that we try to achieve or exceed. Over the past five years we have seen consistent improvement if SGM while our utilization has also been at consistently high level. We plan to maintain or improve our structural profitability by maintaining or improving SGM and maintaining high utilization. For advanced technologies, we are careful in peak capacity planning. For mature mainstream technologies, we are increasing our capabilities, various special technologies to ensure all the legacy capacity can be fully utilized. After three years of operating at a high capital intensity level, about 50% CapEx intensity during the period from 2011 to 2013, our CapEx to sales has come down to about 40% last year in 2014 and is expected to be at a similar level this year. Going forward, we estimate our capital intensity ratio will be at about mid-30s level for the next few years. Regarding CapEx, there may be some adjustments but we are evaluating the capacity conversion along with better productivity improvements. We are also increasing investment in specialty technologies. So for now, we keep our CapEx budget unchanged. Let me move into long-term financial objectives. Over the five-year target, meaning from 2015 to 2019, the five-year target calls for a compound annual growth rate of 10% for both revenue and net income and a bigger or equal 20% in ROE. The long-term growth rate of the overall semiconductor industry is expected to be growth above about 4% to 6%, at a single digit -- mid-single-digit level. Foundry is expected to continue to outgrow the overall semiconductor market by a few points. We plan to achieve the 10% compound annual growth rate in revenue by carefully positioning ourselves with the right technology and build appropriate capacity to capture the business. With the right technologies and sufficient capacity, TSMC is well-positioned to continue gaining market segment share in the foundry segment. Our current five-year plan number is very close to the 10% revenue growth target. We believe we will grow double digit each year for the year 2015 and 2016, and we will continue to work on the 2017 to 2019 growth ahead of time. Thank you.

All right. Thank you. 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 their questions. Questions will be taken both from the floor and from the call. Should you wish to raise your question in Chinese, I will translate that to English before our management answers your questions. For those of you on the call, if you'd 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 hash key. Now let's begin the Q&A session. A - Elizabeth Sun: Our first question comes from the floor. It will be Bank of America Merrill Lynch's Dan Heyler.

Thanks for that. Thanks, Elizabeth. Gentlemen, good to see you again. Congratulations on the 10 nanometer progress by the way. Relative to your competitors, it's a good news. A couple of things. First, on -- I wanted to address growth for 2016. You did talk about the potential to grow over 10% next year. In light of the slowing smartphone market, if it was, say, a 5% unit growth for smartphone units per se, what type of semiconductor growth would you anticipate for the overall industry, and then secondly, for TSMC within mobile? Thanks.

Slow 2016 growth, what is the driver, is the other question?

Yes, with 5% growth in smartphones, what kind of growth can we expect from TSMC's mobile-related --

You're talking about 2016, is that right? Well, first, we expect some market share gain, as C.C. had said and as I said a year ago, we expect that 16 nanometer market share will be much greater than our competitors, our next competitors. And overall, we also see other areas where we will be expanding market share. For instance, in the IoT area where immature, more less advanced technologies are used, some of our customers are doing very well. So I expect gains in market share. And I also expect that the [indiscernible] segment of the semiconductor market will be better next year than this year. They did very well last year and so did we of course. We grew 25% last year. And we had a record fourth quarter last year. And our customers, for the most part, did very well last year also. And at the end of last year, it was, I think, last year, Christmas last year, was a Merry Christmas for a lot of people, and the expectations for this year were very high at that time. And, however, the big growth -- part of the big growth that we had last year went into supply chain inventory. And the supply chain inventory at the end of last year was indeed very high. And now this year, with a number of things happening, that Mark already mentioned, the outlook for this year has continued to deteriorate. And so in retrospect [indiscernible] averaged last year's growth rate with this year's growth rate. Last year we're 25%, and this year we're still looking for double-digit but certainly it's considered lower than last year's 25%, this year. And so I do expect next year's organic growth rate, if you will, to be better than this year. And in addition to that, as I said, we expect market share gain.

Okay. Thank you. So if I hear you correctly, mobile could certainly achieve at least double-digit growth.

Mobile. Mobile (inaudible - multiple speakers) smartphones.

Correct. So if you're growing it double-digit, would we see your -- the mobile part of your business grow, say, can it do a high single digit -- high double-digit type of growth, or should we see kind of growth (inaudible - multiple speakers) for you?

Yes.

So, would we see more growth across all of your segments be a lot closer, say, in that high double-digit range? Thank you.

Correct.

Thank you. And the second question, kind of a sub-question for this growth [indiscernible] is, as you look at processors coming out today, eight-core, ten-core processors moving to 16 nanometer, Cortex 72, that's pretty much like a PC-level processor, I'm wondering what kind of applicational opportunities are there for you to grow next year or as you look into the broader markets, with performance levels being as high as they are, we're now moving kind of out the smartphone era, much more into a compute capability --

Higher performance -- higher performance computation. Mark, maybe you should answer the question.

Yes, Dan. And indeed, the application processor of smartphones is increased -- the complexity is keep increasing, right? And that has, still, that has to do with people user experience of the smartphone today.

I think he was talking about something more than -- beyond smartphone, yes.

And therefore, [premium] content of high-end smartphone we think will keep increasing because of that. Now for IoT, I think it's [indiscernible] of IoT application because IoT devices currently open as using the smartphone as a transmitter. And -- but that computation world hasn't really flourished yet I think. That has to do with service and application associated with IoT. And that is a big market we see. The IoT will bring that in the future. And so that -- this one is not yet for the computation for the IoT world. That's my perspective. And that is the growth momentum to come.

So IoT as a portion of TSMC business today, do you guys have a rough ballpark figure of how much is IoT related to your business today and how big can it be? Relating maybe more just to the kind of the access points out of the equation. I know the processor side is difficult to calculate. Thanks.

I think Dan has got his two nickels worth with our long answers. [So let's go on with it]. Oh I'm sorry (multiple speakers).

[Indiscernible] answer to you. We see our customers bringing the connectivity function on many devices. So for us, the IoT portion is not specifically defined like other company would produce IoT product or services. So we see this trend will continue and [it proliferate] in many, many different segments. So we see that part is keep increasing. But today it's hard to [characterize] a specific IoT among so many segments that we are doing today. We just see the connectivity and the competition is increasing in many segments.

Interesting, thank you.

All right. Next we will have questions coming from Deutsche Bank, Michael Chou.

Thank you for taking my question. My first question is it seems that Intel is saying they will slow down the process migration to 2.5 years now. The cadence actually would be longer than the previous node. So will that be the case for TSMC in 7-nanometer from 10 to 7?

Well, we are looking at it right now you know. What was the time between 16 and 10? Three years. And -- it was three years. And the time between 10 and 7 is around one year right. So for those two generations for us anyway, it's three years. Now Intel says it's two years. Well, maybe we could do it sooner or maybe Intel might do it later and we might do it later too. Two and half sounds right. Did you understand my answer?

Yes. Thank you. The second question is given that you should start mass production of 10-nanometer by the end of 2016, so your customer's 10-nanometer product may hit the market in Q2 2017.

Yes. So you are trying to calculate -- triangulate the time when the customers' product will arrive at the marketplace using TSMC's 10-nanometer.

Can you answer that question? C.C. Wei: You try to narrow down exactly the time. But actually we ramp up 10-nanometer in the 4Q 2016 next year as Mark mentioned. But the real product shipment will be in 1Q 2017, yes. Exactly the new products into the market, I cannot comment.

But since Intel say that the 10-nanometer platform will be in the second half 2017 so does that mean your 10-nanometer progress could be faster than or faster than Intel. Can you say that? C.C. Wei: I only say that our product 10-nanometer will be in the first quarter of year 2017.

All right. Next we will have questions coming from Credit Suisse, Randy Abrams.

Thanks. The first question I wanted to ask about the inventory correction. It's extended for two quarters. Now are you expecting as we go in the fourth quarter, potential for restocking where we traditionally go into a low season or do we wait until after Chinese New Year? And then with 16-nanometer ramping up how much could that support fourth quarter being above seasonal? And also how do you see that 16-nanometer ramp in first half, if you expect to continue the steep ramp on 16 in the first half next year [indiscernible] more of a second half event?

Mark.

Well, the inventory picture is what you said correctly. And for the 16-nanometer, it's all our new products. So at this point we are ramping, but the inventory issue is not in the picture. We just keep the production ramp and in the second half this year and getting to the first half of next year. Now only as we know that the demand, the customer demand is very strong and that is different than the inventory level I was commenting about.

There's no supply chain inventory of 16-nanometer yet.

There's two parts to it. The first part is the core business, the existing 20 and above, if you expect restocking in fourth quarter. And then for 16 if you expect there's still a steep ramp in fourth quarter and then also a steep ramp continuing in the first half. I guess I was curious do you expect the core business to have restocking in the fourth quarter so that's above seasonal. And then also do you expect 16 to have a steep ramp continue in the first half.

We expect the core business will in -- still continue in the depletion mode until the end of next year. Now then coming to the first quarter, we might see a restocking. But as you know the first quarter is a seasonal weaker quarter. So that through competition how that reflects on our demand is yet to be seen yet.

So Randy I suppose you have another part of your question, which is with the fast, steep ramp of 16-nanometer in 4Q, will we still have a fast, steep ramp of 16-nanometer in 1Q next year. That's your second part right?

Yes. Okay, good. And if you could, the second question is on the CapEx. You mentioned capital intensity coming down. Was that a statement on next year so capital intensity reflects how we should look at 2016? And so far you're under-spending CapEx this year, so I'm just curious how you're seeing the overall CapEx levels whether it's kind of tracking similar ballpark this year and next year. And is there potential to actually start raising the dividend again next year if CapEx is not going up?

Yes, I just said the capital intensity will come down to the mid 30 level over the coming few years. With that and with the 10% growth target that we are targeting for, we are confident we are able to generate increasing free cash flow from operations. So we will consider dividend increase on a year by year basis when the time comes.

All right. I think this is about time that we will take our next question from the call. Operator, please proceed with the first caller on the line.

It's Donald Lu from Goldman Sachs, okay.

Thank you very much. My first question is on the compute opportunities. I know now TSMC seems to be narrowing the gap in the process technology and when should we see that the TSMC is really enter the notebook market? What is preventing TSMC's [Technical Difficulty] plan to enter the notebook market in large volume? Is that [Technical Difficulty]?

Elizabeth, you need to repeat the question.

Donald, I have to tell you that we did not really get your voice very clearly in here. So I'm just going to repeat your question and correct me if I didn't hear you right. Your question is what is TSMC's opportunities in computation segment. Due to ability to narrow the technology gap, when are we able to enter this market? And you are asking what is our advantage or differentiation especially if we are thinking of the market area where the ARM servers are related. Is this your question?

Yes. And also not only ARM servers but also notebook.

Also notebook, okay.

Yes. And also what is the [indiscernible] today. Is that software related or it's more of the hardware?

What will be the other factors that we need to overcome such software or architecture.

Yes.

Also notebook?

Notebook yes. Servers and notebook.

I think we are obviously very actively pursuing that market. As to what our advantage will be, it's our advantage as a foundry. We are pursuing or we're going to pursue this market, this higher performance computational market with our foundry experience and foundry record and also the degree of trust that our customers have placed on us. That's our advantage. So that's our differentiation. Now of course we need a very capable partner as well as ARM. ARM we already have. ARM is a valuable -- has been a valuable partner for us for many years now. But in addition to ARM, we also will need a capable design company. And then the partners together, the three main ones, TSMC, ARM, the design company and many others that we have always had, the equipment people etc. we think that we have a chance of becoming an important factor in that market.

Donald, are you happy with the answer?

Yes. Can you be more specific about -- Chairman, you talked about the capable partner. I believe there has been a few companies [buying] in the notebook market. Do you mean that you will be expecting a new company in this market or existing [indiscernible] companies to use TSMC as a foundry.

Again I have to apologize because of the voice transmission has not been quite clear. So let me repeat what we think we have heard from you is that you want something more specific such as when Chairman talked about the capable partners in terms of design companies. And you are asking in the notebook area, there are -- I think you were saying that there are some consolidations and what will be whether we will be able to have new customers coming in to work with TSMC as a good partner in this area. Is that what you were saying?

Specifically who the partners -- who our partners will be? Well, specifically who are the most capable design companies. And our partners will be among those.

Okay. I guess let's come back to the floor and we will see if we can fix the voice issues on the call. Let me come back to the floor. Next we'll be having questions coming from UBS, Eric Chen.

Okay, thank you. Good afternoon. My first question I would like to go to Lora. And Lora, you just mentioned the free cash flow will keep increase and let me double check. You mentioned before the free cash flow at the end of this year will double from the end of year 2013. Is that correct?

I don't have 2013 in front of me. I think based on our current outlook for the whole year, it should be right.

Okay. So that indicates probably 35%, 40% year-on-year growth. Based on the CapEx to revenue ratio, the guidance, [indiscernible] you give and also the revenue target, so can we expect the same ratio growth for your free cash flow going forward?

I cannot answer that way. You have to figure out because I just said it is revenue growth and free cash flow is going to increase. But it's hard for me to quantify the degree of increase every year. It's difficult.

Okay. Okay, how about the other way. How about depreciation expenses growth? You gave a guidance for this year around mid-teens. And based on your CapEx and the revenue target going forward, what kind of depreciation expenses growth we should expect?

I can only talk about this year, okay, because we have not fixed the CapEx for the remaining years. This year the depreciation is expected to go up by 13% year over year.

13?

13% which is lower than the number I gave you last time. I said [mid teen] last time.

Okay. And my second question I would like to go to Dr. Chang. And first on the EUV. How about the EUV, the schedule for 10-nanometer process? Some people talk about the window is closed. And how about the EUV for 7-nanometer process?

I will pass the question to Mark.

The window is not closed. We work very closely with ASML and however it's not without the challenges. And we made a very good progress on the source power as well as on the photo resist. And for mass inspection solution it appears more likely. However, the current technology's challenge is more on the mask of the EUV technologies. That is currently actively working with ASML. On the tools stability we have a criteria for the working team to reach a certain target. And that target is likely to reach by the end of this year and then we will start development using EUV tools. As you can see in our 7-nanometer development schedule that probably will not using EUV. But we are planning to exercise EUV using the 7-nanometer technology and currently we are planning to use EUV at 5-nanometer. But of course it does depend certain development criteria, milestones to be reached. And it has a good benefit from our assessment on the 7 -- on the 5-nanometer that reduce a lot of many masking layers and increase a lot of better control for the 5-nanometer.

Okay, so let me clarify. Even for the 10-nano the window for the EUV for the 10-nanometer process is not closed yet.

For 7 as I say we will exercise EUV on 7-nanometer. And we will plan our EUV on our 5-nanometer technology.

But not for 10. Okay, thank you. Okay, thank you very much.

I think Eric has his two questions. He will come back later. Next one I think is a question will come from Citigroup's Roland Shu. I'm sorry Roland before you ask questions, let me make a brief announcement. For those of you who are on the line trying to call in for questions, since we have this voice transmission problem, could you please send me your questions in email to my mailbox so that I can read out your questions. And I think most of you know my email address. That's Elizabeth_sun@tsmc.com. So if you are on the line waiting to ask questions please send me your questions through email. Thank you. So Roland now.

Thank you. Good afternoon. I think my first question [indiscernible] customer deposit to TSMC to reserve the capacity. Can we have more color on how TSMC is going to recognize the revenue for these customer deposits? Is the Q2 revenue guidance included these customer deposits? Thank you.

The customer deposits is a balance sheet item. So it not has to do with the P&L. There's a certain commitment from customers on loading. When they reach the loading we will return part of the deposit to them, so it's still a balance sheet item.

Okay, so they never go to the P&L.

They are not.

Okay, understood. Thank you. Okay, second question is for 28-nanometer, I think CC said there's a lot of second wave opportunity for 28-nanometer. And 28-nanometer last year contributed almost $8 billion in revenue to TSMC. So the question is with this contribution for 28-nanometer (multiple speakers). For next year are we going to see 28-nanometer revenue continue to expand even more than this $8 billion level?

Can you repeat this?

So Roland's question is that 20-nanometer has been a very big node for us. Last year according to Roland's estimate that we have made $8 billion revenue from 28-nanometer. So with the second wave applications coming in, his question is will we be able to expand that $8 billion revenue next year from 28-nanometer's second wave customers.

I don't know about the $8 billion. If you think it's $8 billion maybe it's $8 billion. I can't verify that. But I would say that 28-nanometer will continue to be strong for several years.

Okay, thank you. If I can -- maybe I can ask another question. For 20-nanometer and 16-nanometer, I think TSMC already said this will be the same technology node and also we are expecting much bigger market share for 16-nanometer next year. So will the combined 20-nanometer and 16-nanometer revenue to reach bigger than 28-nanometer last year?

What is the question?

For 20-nanometer and 16-nanometer revenue together next year, will this revenue bigger than 28-nanometer revenue in last year or this year?

Okay, 20 plus 16 together next year, will that be bigger than 28-nanometer last year or this year?

Again I don't know the answer to that. What I can say is that 16 next year will be much bigger than 20 this year. And as a percent of total corporate revenue, 16 next year will be higher than 20 this year. And next time Roland will probably come in and tell me how many billion dollars we are billing.

Okay. Now at this point I think we have Brett Simpson on the line and he has already sent me his questions. So operator, first please open the line for Brett Simpson and then I'm reading out his questions. And then our management can answer and then Brett may have a follow up. Brett Simpson's question number one. You talked about your outlook for 16-nanometer -- no, for 2016, double-digit revenue growth and mid 30's CapEx to sales ratio. What do you foresee FinFET to be as a percent of sales in 2016 and directionally how might 28-nanometer and 20-nanometer trend? Can they grow in 2016?

Say it again.

16-nanometer percentage of revenue next year. So whether 28 and 20-nanometer will be bigger next year than this year.

16's percentage of revenue will be what?

Yes, he's asking us how much we will get as a percent of revenue from 16-nanometer next year.

Well, I stick with my previous answer. 16 next year our revenue percentage and revenue dollars will be greater than 20 this year.

All right.

And is there another one? Is there another question?

Okay. The second part is 28-nanometer and 20-nanometer whether they will be bigger next year than this year.

28?

And 20.

And 20 will be --

His question is whether they will be bigger next year than this year.

20 -- can you maybe answer to that one?

CC and Mark was talking about expanding the 28 technology offering to more features. So with that we believe our 2016 28-nanometer revenue will continue to grow. It will be bigger than this year.

I think 20-nanometer we have already mentioned in CC's comments earlier which is a smaller level next year than this year. I think CC has already mentioned that. Okay, so Brett's second question. Consolidation among your customers. They are talking about savings with foundry partners as a result of higher scale economies. Can you give us your perspective on the impact that large scale M&A has on wafer pricing?

I don't expect any impact. I think of course when two companies combine there are synergies and there are savings. But that doesn't necessarily mean that we have to save for them.

Okay. This is a very clear answer. And Brett has another question. When you think about capacity planning for 10-nanometer, 7-nanometer how does this compare with 28 capacity and 20/16-nano nanometer. Should we assume that with lower capital intensity that 10 and 7-nanometer will be smaller nodes for TSMC in wafer capacity?

Say it again.

Whether or not 10 and 7 will be smaller than 20 and 16.

In what?

In capacity.

Look, let me just say it in an overall sense. We expect -- 28 of course was very successful and we expect 20 and 16 together to be at least as successful as 28 over it's -- over the two life -- over the life of the two. In fact, we expect 20 and 16 together to be bigger in revenue over the lives of those two technologies than 16. Now 10 and 7, our current plan is that in terms of wafers it may be -- combining the 10 and 7 together, in terms of wafers, it may be a bit lower than 20 and 16. In terms of dollars we expect it to be as significant as 20 and 16. Over -- again over the lives of the respective technologies.

Okay. Brett, I think we have addressed all three questions you sent over. And if you have further follow-up please send me an email again. Let's come back to the floor. Next we will questions will be coming from Daiwa's Rick Hsu.

Thank you for taking my question. My questions is about your Q4 outlook. Can you give us more color about that because you were talking about your demand driver is coming from iPhone -- new iPhone launches which seems to me is back-end loaded. And you're talking about your -- another your demand driver is coming from 16-nanometer FinFET ramp up. That is also kind of a back end loaded given the long cycle time. Can we fairly anticipate your Q4 total revenue would be better than Q3?

Well, we're not ready to forecast Q4 at this meeting. Now I think there was a related question that was pre-submitted. No, it wasn't pre-submitted. I saw it on a Taiwan newspaper today. The related question, I am ready to answer that. So let me just answer. The related question appeared in the commercial daily. I think it was the Commercial Times, I think it was. The question was with the inventory going down so slowly, what is the implication for the fourth quarter this year and first quarter next year. So you know I thought it was a pretty good question, so I got ready to answer that one. So maybe answering that one will also at least partially answer yours. Clearly you know the slowly decreasing inventory is not a very good omen for the fourth quarter. The inventory is being depleted more slowly than we expected. However -- however, the impact on the first quarter is an entirely different story. We do expect that the fourth quarter, by the end of the fourth quarter, the inventory will be back to the so-called seasonal level or even lower. And that bodes very well for the first quarter. But still I, having said all that, I still repeat to you another thing I said. For the entire year we still think that we will have double-digit growth.

Okay, thank you so much. My second question is about maybe in 2016 or 2017, how do you see the rise of the Samsung in-sourcing strategy.

How do we see the what?

The increase of the Samsung in-sourcing strategy.

Samsung's using more of its internal products than go to foundry.

It reduces the foundry market. We count that as non-foundry market.

Thank you so much.

Okay, all right. Now since several analysts have sent me questions, I need to read out the next one. And operator, please open the line for Mehdi Hosseini. Mehdi has two questions. One, how will consolidation among China based smartphone OEM's impact TSMC's revenue CAGR of 10%? China smartphone OEMs consolidation impact TSMC's 10% CAGR.

CC, you're going to answer.

China's OEM consolidation. Smartphone OEMs.

Smartphone OEMs consolidation. I don't think any of us is really qualified to answer.

Sorry Mehdi, we cannot answer. Second question.

I think [it gets] to the consolidation of some of our customers' customers in China. I think that's what you ask.

Right.

So it doesn't appear that we can answer that.

Second question is still related to our CAGR of 10%. What are the key assumptions for wafer shipment and ASP trends in our revenue CAGR guide of 10%?

I don't know. All I want to do is to make at least 10%. And it's a bottom up thing you know. It's not something that -- it's not top down. Actually 99% of our plans and forecasts are bottom up. And so when you say, what are the assumptions and so on, I have to talk to, ask 50 people what their assumptions are. And I just or I should say we, we just tend to trust them. Actually on the whole, for years or maybe even decades on the whole we have done a really good job. We've done a much better job in forecasting our planning in this bottom up way than any top down kind of thing. Top down you make assumptions and so on and your assumptions could be all wrong. Bottom up, of course many people make assumptions. And I -- we, I will have to ask 50 people or maybe even 100 people and I certainly am not going to do that. But anyway it's -- that's our 10%, that 10% is the bottom up kind of a forecast.

Okay. Since there are still quite a few questions posted online, I'm going to the next one online which is from Steven Pelayo at HSBC. Steven has a couple of questions. I'm just reading it out. Clarification of full year guidance. You said full year to grow double digit. But you also recently said second half 2015 will be greater than first half.

I said that. I said that in June of -- June 9 something like that, shareholders meeting. Well, that forecast has become more marginal now.

Okay, so then.

Well I said two things, I am not meeting first half and second half will be better than the first half, second whole year will be opposite the second part still holds the first part may still be true but as I said has become more marginal now.

Chairman has already answered this, so I won't get Steven's follow-up questions, which is all the minute details of that first half, second half thing. But then he indeed had a question about third quarter guidance. He said, 3Q guidance of 1% to 2% growth, what node will grow and which will decline or all stable. Unidentified Company Representative^ Say it again.

All right. He wants to have colors on third quarter node by node sequential growth.

For third quarter, node by node sequential growth.

Growth or decline?

Can we answer that Lora or C.C. or Mark? Do you want to answer that?

Let me try. I will not answer node by node revenue growth, but I can give you some color on the segment growth. The third quarter we believe communication will come down, computer will go up, consumer will go up and industrial and others will go up.

So, okay, with that, let`s come back to the floor. Michael Chou from Deutsche Bank has follow up questions and then afterwards we'll go to JPMorgan.

Thank you.

Michael is with Deutsche Bank? Were you the one that asked that I thought I said was a pretty smart question, in the commercial --

In 10 nanometer actually given what would seem 16 nanometer for your market share, do you think your first year in 10 nanometer --

You didn't answer my question. Were you the one that asked that question about how the inventory decrease will impact, were you the one? No? Randy? I thought it was Deutsche Bank. Randy, you're not Deutsche Bank are you? Okay, never mind, you ask your question then.

Okay. Do you expect your 10 nanometer market share in the first year will be quite dominant?

What's that?

Your 10 nanometer market share in the first year.

10 nanometer?

Market share in the first year going forward will be quite big. Given what we've seen in 16 nanometer --

What will 10 nanometer market share be?

Yes. For the first year.

Do you want to answer that? Well actually the answer is, we don't know. For first year, 10 nanometer will be, really first year will be year after next, all right, 2017. I mean we'll start end of next year, start means you know, not very many vapors. And 2017 it will become quite significant. All right, I will try to answer you. I think if you -- and the first year 16 nanometer market share.

That's a very safe one. Okay, now we're going to JPMorgan, Gokul Hariharan.

Yes, thank you. Chang, just wanted to get your updated view on what you think about locating 12 inch fab in China given that [indiscernible] seems to be developing very quickly. Any updates or any new thoughts on also what's happening in terms of the Chinese [indiscernible] ecosystem buildout.

All right Gokul's question is, can we talk about whether we have a plan to build a 12 inch wafer fab in China? Can we comment on our competitors building out ecosystems in China?

We are only considering it and we have no comment on the competitors doing it.

Okay. A follow up on the new areas of growth that you mentioned like IOT, automotive, industrial going from next year onwards. There is a perception that a lot of that growth could be happening in mature nodes and not so much in advanced nodes. Could you address what areas would that growth be coming from because in the past a lot of the new markets have been demanding advanced process as well than the mature processes.

All right, so Gokul's second question is with respect to the new growth area, since we have mentioned the new growth area of IOTs, automobiles, industrial, he thought that these are all more related to the mature nodes and not necessarily the leading edge or the advance nodes. So he likes to have us --

I missed the last few words.

The [indiscernible] applications tend to use only the mature node and not the advanced nodes. So what will be the growth opportunities that would probably use the advanced nodes, right.

As long as there is a new application is growth, I mean yes, they use mature nodes, but generally, they don't use those, just the mature equipment. We have to upgrade the equipment. We have to make couple of investments. In fact, we have all along been making considerable capital investments in the mature nodes, in some years more than other years. I think we have been telling you most years how much of our capital CapEx is in the leading nodes and so on. And there is a pretty significant portion that's not leading edge [as the] mature nodes. And usually there is considerable investment, additional investment we need to make to adapt to the equipment to -- it's not really changing the equipment, its buying new equipment in order to do the specialty IOT kind of products. So even if it's a mature node, it means growth.

Okay, I think Donald Lu over some follow-up questions, I will read it out right here. Donald's question, do you still think that global foundry, I mean it's not that company in New York, its worldwide foundry 14 nanometer and 16 nanometer capacity is rational, i.e. whether or not the world's overall 14 and 16 nanometer capacity will that be rational? Second question, does TSMC still expect its smartphone content to increase year-over-year in 2015 and 2016?

First question?

Whether the total capacity will be rational? Whether they will be rational?

To 16 nanometer?

Yes.

I think the 16 nanometer global capacity will be rational. And the second question, well look, there are many companies in the 16 nanometer. So where there are just a few players, I expect the capacity to be more rational than if they are lot of players. And another thing is its very expensive, well anyway more expensive than the 10 and a lot more expensive than the [indiscernible], more so than the 20. And they are a lot more expensive than the 28 on top of the 16 now. So those two reasons, few players and very expensive capacity, make me think that the 16 capacity will be rational. And the second question?

Do we expect the smartphone content to increase for us?

The answer is yes. And do you have a more quantitative answer? You do? Well, good.

The overall content for the smartphone for us will increase next year versus this year in general slightly, but especially on the high end --

I think he means how many dollars we have --

Yes, that's what I am talking about. Especially high end smartphone, we expect the content dollars will go up by $1 next year from this year.

And mid to lower end also?

Yes, mid, low and also increase slightly.

[Integrated] Our TSMC dollars in it.

Okay, now we are coming back to the floor. This question will be coming from Dan Heyler, Bank of America.

Just a quick follow up on Lora's numbers there, could you remind us from what -- you said $1, so what's the absolute number for 2016 in terms of content at the high end and the middle to low end?

Okay for high end the dollar will be from $18.7 to $19.7 for TSMC.

Right. Okay. And then on the market share opportunities that you highlighted in mobile, Dr. Chang you mentioned that next year within the mobile market you anticipate the potential to increase share. Does that require significant share gains across kind of -- will that be share gains across all the customers or do you anticipate some share loss at say one of the major customers in those assumptions, because it looks as though competition is actually increasing in mobile, so I wanted to get more color.

Share gain, well every customer, well that is -- that can only be dreamed of. There are always some share gains, some share losses, but on the whole I expect a share gain.

Follow up question from Credit Suisse, Randy Abrams.

I just wanted to follow up on last quarter you mentioned FinFET would be better than your prior plan, implying like low teens growth by fourth quarter. Could you give an update on the pace of that, like if you'll get some revenue in third quarter and you still expect over 10% fourth quarter and the 20 node in the second quarter got close to 20%, so if we could get toward that by first half next year?

Will you repeat the question?

Sorry, I was speaking with my BlackBerry.

Okay, so I was just asking if FinFET, if you still expect to have some revenue in third quarter over 10% in fourth quarter and then to match 20% to get about 20% of sales by first half.

Do I expect in the third quarter 10% did you say?

In third quarter to get revenue and then fourth quarter 10% or more revenue and then first half that start approaching 20%.

I don't think there is anything above 10%.

That's right.

I said 10% for the whole year.

I think last conference the comment was --

Randy, I think we remember, last conference one analyst was saying that he predict we will get 12% revenue from 16 nanometer in the fourth quarter. But the management did not agree to that number. The management did not accept that 12% number.

Okay.

That was last time.

Was that your question?

Yes my question was if you could comment like how it will grow as a percent of sales over the next couple of quarters?

How will the 16 nanometer grow as a percent of our revenue in the next few quarters?

Want to talk about that? Well, do you want to talk about -- I know that we don't want to talk about fourth quarter, but do you want to talk about third quarter?

Third quarter will be very small single digit revenue.

16 nanometer, it's very little.

Okay. Unidentified Company Representative^ I mean we started ramping, when, three months, four months ago, last month. And the deposit cycle is very long, Randy.

Okay.

Its months before we get the first finished wafer out. And some customers want us to do the bumping and all that stuff, and that takes longer and so it's a long time between for starting to ramp and the time you get even the first wafer [out]. So third quarter will be very little.

Okay. It sounds like you don't want to mention on fourth quarter, I guess the other question I was going to ask on structural profitability where you were talking in the prepared remarks, do you still expect structural profitability to be in line to better, like you've kind of maintained that track record the last few years, if as you look out over next 12 months, you still have that trajectory?

Yes, the answer is yes.

Okay, thank you.

Good, all right. There is a question coming to my BlackBerry from a [indiscernible], his question is this, how does the strategy of quicker node transitions and the reuse of tools as we have seen in 20 nanometer to 16 nanometer affect the depreciation for new process nodes? And does this tool transition allow you to offer better ASPs to your customers? So quicker nodes transition, reuse a tools whether this translates to better ASP to customers and how that affects depreciation?

Well, I can answer in a very general manner. That is something making one generation's tools compatible with the next generation or vice versa. That has been a prime effort of TSMC [indiscernible] and TSMC operations in the last three or four years. That of course started because the complexity of the prices has increased and because we have adopted the faster cadence of the technology. Now I think the prime effort that this very big focus on reducing what we call internally, we call it the conversion loss. Our effort in reducing the conversion loss has borne another developed fruit. And now in the ideal case, all the tools are compatible, but that's impossible now because when you advance, go from 16 to 10, you do need new equipment, that's even now still being developed by our equipment partners such as supply, materials and so on. However we have been able to reduce the conversion loss to a very low level now. And I think from that you can start calculating depreciation and what not. You want to add anything to what I said?

I can add some comments on this equipment migrations, just like Chairman said, the operation R&D people made a tremendous effort trying to increase the migration rate. As far as we can see now from 20 to 16, 16 to 10, we can manage migration more than 95% of the tool can be reusable.

Okay, now there is another follow up question sent over by Steven Pelayo at HSBC and his question is this. Since TSMC learned a lot at 20 nanometer, will a steeper ramp of 16 nanometers this year not impact gross margin too much, i.e. not much of a dilution or perhaps the faster learning will allow for higher stocking 16 nanometer gross margin compared to 20 nanometer, whether 16 nanometer gross margin will be higher than 20 nanometers at the starting stage.

Yes, I think the answer is, we have learned a lot from 20 nanometer to benefit 16 nanometer and therefore as a result we expect 16 to be more profitable than 20 and we expect the yield learning curve to actually be a part of the 2016 total learning curve from -- and I think that's actually one of the advantages of a faster cadence.

Okay, now we come back to the floor. The question will be coming from UBS, Eric Chen.

Okay, thank you, my question, [indiscernible] have one question and I'll go through the other way. You talk about a 28 nanometer process and then you talk about HPC. And I would like to get a rough idea for your 28 nanometer process business, how many percent go through the HPC and how about HPC Plus and going forward, what kind of exploitation we should keep.

Actually I cannot give you the exact number, but I can give you the [indiscernible]. Our 28 nanometers are tape out right now is higher than any quarter previously we announced and most of the tape out give 28 HPC and HPC Plus, because that's give you a hint of future business what is the percentages is there going to be.

Okay, got it. So I don't get it, I mean for the HPC minus spending is lower than average of a 28 nanometer process business and given the lower I just mentioned, for your low-end smartphone I see the content value will increase slightly. It doesn't mentioned if HPC the revenue can increase, I assume the low end smartphone, I see the content value should be cut. Thank you.

Let me be more precise on the content. I was talking about the high end smartphone. TSMC's revenue will go up. Actually the number is, this year is $17.80, next year $19.7, so its $2 up, that's the high end. The mid-end, $6 this year, $6.4 next year, slightly up. The low-end keep the same, $3.40 and next year will be the same.

Okay, so that's good. And even other HPC the percentage will increase but will still keep at the low end, probably at the low end the ASP, right?

Actually it's not only the smartphone using the HPC or HPC Plus, it's the second wave a lots of other products, they are using HPC and HPC Plus.

Got it, so [indiscernible] about the other foundry makers said a second tier the foundry maker, they are more aggressive with the wafer prices strategy at the 28 nanometer process, we would like to keep the overall the [indiscernible] ASP.

Well Eric, maybe I can do a little bit clarification here. When Lora talk about mid end, low end smartphone content, it is not just from one geometry, it has many silicons inside and that's an average price. HPC or HPC Plus is only adopted by certain applications in those boxes and they are not the whole thing.

So that's more like product mix, probably 14 to 28 nanometer.

The overall silicon content of the smartphones doesn't matter it's a low end or mid end or high end, those contents are increasing because of complexity and functionality.

Well at least ours.

Ours is increasing.

Content is increasing yes.

Thank you very much Lora, [indiscernible] Wei and [indiscernible] Chang.

Thank you. So I think it's about time that we should end our conference here and thank you very much for coming over. Before we conclude the conference, please be advised that the replay will be accessible within three hours from now, transcript will be available 24 hours from now. And those are all available through TSMC's website at www.tsmc.com. So thank you for coming. Hope to join us again next quarter.