[Foreign language] Happy New Year, everyone. Welcome to TSMC's fourth quarter 2016 earnings conference and conference call. This is Elizabeth Sun, TSMC's Senior Director of Corporate Communications and your host for today. Today's event is webcast live through TSMC's website at www.tsmc.com. If you are joining us via 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 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 of 2016, followed by our guidance for the first quarter of 2017. Afterwards, TSMC's Chairman, Dr. Morris Chang, will provide our key messages and conduct the Q&A session, where TSMC's two Presidents and Co-CEOs, Dr. Mark Liu and Dr. C.C. Wei, will be joining Chairman Chang and Ms. Ho in answering your questions. 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 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, for the summary of operations and current quarter guidance.
Good afternoon. You hear me okay? Thank you for joining us today. My presentation will start with financial highlights for the fourth quarter, the recap of our 2016 performance, followed by the guidance of the current quarter. Fourth quarter revenue grows 0.7% sequentially to NTD 262 billion, exceeding the high end of our guidance given in October due to a stronger demand for TSMC's 16 nanometer technology, and a more favorable foreign exchange rate than our original forecast. Gross margin increased by 1.6 percentage points quarter-over-quarter to 52.3%, mainly driven by continued cost improvement efforts. Operating margin also increased at 1.1 percentage points sequentially to 41.9%, while R&D expense continued to grow, reflecting a higher level of 7 nanometer and 10 nanometer development activities. Overall, our fourth quarter EPS reached NTD 3.86, and ROE was 30% for the quarter.
Now let's take a look at wafer revenue contribution by application. During the fourth quarter, communication and computer increased 11% and 7% from the prior quarter respectively, while consumer and industrial standard decreased by 43% and 6% respectively. On a full year basis, all four segments experienced year-over-year growth. Communication increased 16% YOY and represented 62% of our total wafer revenue. Our computer increased 15%. Consumer grew 32%. Industrial standard rose 1%. Now let's take a look at revenue by technology. Combined revenue from 16 and 20 nanometer reached 53% of wafer revenue in the fourth quarter, up from 31% in the third quarter. Our 28 nanometer contribution remained strong at 24% of revenue. Advanced technology, meaning 28 nanometer and below, accounted for 57% of total wafer revenue in the fourth quarter.
On a full year basis, the combined 16/20 nanometer contribution increased by eight percentage points and reached 28% of total wafer revenue in 2016. Advanced technologies, 28 nanometer and below, accounted for 54% of total wafer revenue, up from 48% in 2015. Moving on to the balance sheet. We ended the fourth quarter with cash and marketable securities of 632 billion NTD, an increase of 115 billion NTD. On the liability side, current liabilities increased by 61 billion NTD. On financial ratios, accounts receivable turnover days increased three days to 45 days. While our days of inventory decreased three days to 41 days, mainly driven by the reduction of working process. Now let me make a few comments on cash flow and CapEx. During the fourth quarter, we generated about 185 billion NTD cash from operations and spent 113 billion NTD in capital expenditure.
As a result, we generated free cash flow of NTD 73 billion, and our overall cash balance increased NTD 77 billion, to reach NTD 541 billion at the end of the quarter. In US dollar terms, our fourth quarter capital expenditure reached $3.5 billion, and the total year CapEx was $10.2 billion. This is a bit higher than our prior guidance of slightly above $9.5 billion, mainly due to accelerated delivery of advanced technology tools. Now, I would like to give you a recap of our performance in 2016. 2016 was a good year for TSMC, as once again, we set new records in terms of revenue and earnings. Our revenue grew 12.4% year-over-year to reach NTD 948 billion or up 10.6% to $29.4 billion in US dollars spent. As we saw wafer shipments increase across nearly all technology nodes.
Gross margin increased 1.4 percentage point to 50.1% in 2016, mainly due to continued cost reduction efforts and, to a lesser extent, a more favorable foreign exchange rate. Our operating margin also increased two percentage points to reach 39.9%. Our effective tax rate was 13.5% in 2016, same as in 2015. Full-year earnings per share was $12.89, an increase of 9% year-over-year. Excluding major one-off items, namely share disposal gains and the closure of our solar operation in 2015 and the negative impact from the earthquake in 2016, our EPS would have grown 17.4% year-over-year in 2016. On cash flow, we spent 328 billion NTD in capital expenditure, while we generated 540 billion NTD in operating cash flow and 212 billion NTD in free cash flow. We also paid 156 billion NTD in cash dividends, an increase of 33% from the 2015.
From a semiconductor industry standpoint, we estimate that the semiconductor market is likely to remain flat year-on-year. While fabless grew 5% and foundry rose 8%, the outperformance of foundry over the overall semiconductor industry was mainly driven by the strong demand from 4G plus smartphone in the China market, the replacement upgrade of gaming, and emergence of artificial intelligence. Compared to foundry's 8% growth, TSMC's 11% growth in U.S. dollars underscored our further market segment share gain from 55% in 2015 to 56% in 2016, which is mainly propelled by our leadership in 16 nanometer and 20 nanometer. We estimated fabless days of inventory ended 2016 at slightly above seasonal level by about two days. I have finished my remark on the financial summary. Now let me turn to the first quarter outlook.
Moving on to the first quarter 2017, we forecast the demand is weaker than the prior quarter.
Due to mobile product seasonality and slightly above seasonal supply chain inventory at the end of 2016. Based on our current forecast and exchange rate assumption of 1 U.S. dollar to TWD 32, we expect first quarter revenue to be between TWD 236 billion and TWD 239 billion, which represents 8.9%-10% sequential decline. Gross profit margin to be between 51.5% and 53.5%, and operating margin to be between 40.5% and 42.5%. This concludes my remarks. Now, I would like to turn the podium to Chairman for his comment.
Ladies and gentlemen, I first want to wish you a very happy, carefree, healthy, and profitable, and hopeful new year. Lora has just given her report on the last quarter, and also she gave her outlook for the first quarter. I will just spend a few minutes to talk about 2017, the whole year. For 2017, we expect smartphone unit growth of 6%, from 1.47 billion units to 1.55 billion units. Of the smartphone growth, we expect the high-end of smartphone will grow 3%, the mid-end will grow 5%, and the low-end will grow 8%. For 2017, we also expect that PC units will decline 5%. Tablets will decline 7%. PC units will decline 5%, and IoT units will grow 34%. Again, for 2017, we expect semiconductor market to grow 4%, foundry revenue to grow 7%.
We expect TSMC revenue to grow 5%-10% in U.S. dollars. TSMC revenue, we expect to grow 5%-10% in U.S. dollars. We believe that in the first half of 2017, our growth over the same period last year will be close to 10%. In the second half of 2017, our growth over the same period last year will be close to 5%. We expect to continue to increase our R&D investments, both in dollars and moderately in its % of revenue. We also reaffirm our long-term growth projection, which was made a couple of years ago. We think that in the 2015-2020 time period, TSMC revenue and operating profit will grow at a compound annual average rate of 5%-10%. TSMC will maintain an ROE of above 20%.
That concludes my outlook for 2017. I believe we are now open for-
Right
questions.
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 line. Should you wish to raise your question in Chinese, I will translate it 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, star then one. Questions will be taken in order in which they were received. If at any time you would like to remove yourself from the questioning queue, please press the pound or the hash key. We will start the Q&A session. First, we will have Deutsche Bank, Michael Chou.
Hi, Chairman. Co-CEOs, Peter and Dr. Chang. My first question is regarding the outlook in the first half. Given Q1's could be down 8%-10%, and your guidance for first half to be up 10% year-on-year. Does that imply you still expect you will see quarterly, or we can say sequential growth for Q2 or Q3 going forward?
Could you repeat the question?
Right. Michael, you are asking what is going to be our Q2 and Q3 sequential growth.
Certainly not the math.
What's the question? Repeat the question.
You are asking us about the second quarter and third quarter-over-quarter
I can't call it.
revenue growth outlook.
We expect it to be a very significant positive number.
Maybe the follow-up question for inventory. Given the Q1 outlook seems to be slightly below our estimate. Do you see the demand outlook for Q1 is weaker than you expected three months ago?
Did you hear the question, Lora? Yeah.
Michael was asking, do we see the Q1 was weaker than we anticipate three months ago?
Could you answer that?
Yes. It was weaker than we have anticipated a couple months ago.
Could you specify which segment or what kind of applications?
Basically, it's mobile product, as we have stated in the press release. All right. The question will go to Credit Suisse, Randy Abrams.
Yes. Thank you. The first question, wanted to ask, your outlook is more in line with the industry, where you're guiding 5%-10% versus foundry near similar levels. Could you talk about the factors to be more in line after gaining the last few years? Can you also address the China business? We're seeing the China foundries grow faster. SMIC is growing 20%, 30%. How does TSMC combat or defend share more on the mature nodes, where they're starting to grow faster?
Could you repeat the question?
Right. All right, Randy, your first question is made with respect to TSMC's 2017 growth outlook, where Chairman has just mentioned that 5%-10%, which is very much in line with the semiconductor industry.
Oh, I'm sorry, with the foundry industry. Yes.
He'd like to know the underlying reasons why we're only growing sort of in line with the foundry.
Why we're only in line with the foundry?
China, where the Chinese foundry appear to be growing faster than the foundry industry, and what is going to be our strategy in defending our market share or growing, competing against them?
I think that I really don't know how fast the Chinese foundries are going to grow. I do agree that their ambition is certainly to grow faster than the foundry's average. As to why I projected just now that we're only going to grow the same rate as the foundry growth. First of all, we may well be conservative. It's a little too early to say, because I said just now in answer to this gentleman's question, actually, I said when I was giving the message, that the first half year-over-year will be about 10% growth, the second half will be only 5%. The second half is farther away, and that always makes it more difficult to predict. Rather than erring on the high side, we would rather err on the low side. That's the first reason.
There is another reason. That is, this year, I think 2017, will be pretty strong in terms of technology. It will be a pretty strong 16 or 14 FinFET year. Our market share in 16, while it's quite high, it's not as high as I would like. It's actually closing 70%, or 65%-70%. That's not quite as high as our 28 nanometer, which even now, we're at almost 80%. 2017, we think will be a pretty strong year. We're running into a little air pocket here. If you call just staying par with the rest of the foundry an air pocket, if you call that an air pocket, then well, we're running into a little air pocket. The next step after, of course, is 10 and 7. We see a clear sky.
We see a clear sky anyway there. Those two reasons. First, possible a little conservatism, and second, there is a little concern that our 16 nanometer does not have as high a market share as I would like. 2017 happens to be a year that is very strong. We think a very strong 16 nanometer FinFET year. Having said all that, I will say that we certainly do not think we'll lose market share. When I said that foundry will grow at the seven, and we are going to grow at between five and 10, I would just further clarify that statement by saying that we're not going to grow less than foundry. Yeah.
Second question, a quick follow-up if you plan to introduce a 12 nanometer to enhance your 16 nanometer, if you have planned for that. The second question I had was on profitability. Margins are holding up quite well in the first quarter, where normally utilization drop, it should fall a few points. If you could talk about the factors holding it up. As we move through the year, how the expectation from that level, factoring we're ramping 10 nanometer, if you expect margin to improve from that level.
Randy, your second question actually has two separate questions. Yeah. First part of your second question is with respect to 12 nanometer. Second question of the second question is with respect to the margin profile over this year.
First question is what?
Is 12 nanometer.
12 nanometer. I think that, C.C., would you answer the first question?
I can answer. Well, let me answer the question. You asking about the 12 nanometer on what?
Yeah, if TSMC has a strategy or plan to introduce an in-between 12 nanometer to protect share on the FinFET for customers not going to 10 or seven.
Okay. In fact, our strategy is continuously to improve every node in the performance, such as a 28 nanometer. We are continually improving the 16 nanometers technology. We have some very good progress. You might call it a 12 nanometer because we improving the density, logical density, performance, and the power consumption. Yes, we have that.
The second question is the profile of the gross margin.
Gross margin profile over-
I think it will be fairly constant. I think in the second half, 10 nanometer, because the 10 nanometer will be ramping up very strongly in the second half. It will start to ramp in the first quarter, but it will ramp very strongly in the second half. The big volume for the year of the 10 nanometer is going to be in the second half. The 10 nanometer will drag down the corporate margin in the second half by about two points. If you take that into consideration, then I would say that our structural profitability will be maintained. By that, I mean something in the 50% range. Take into consideration that in the second half, 10 nanometer will drag down the corporate margin, corporate average by about two points. Yep.
Okay. Next question will be coming from UBS, Phil Lu.
Phil, yeah.
Hi. Thank you, and happy New Year. Dr. Chang gave us the outlook for the various end markets in 2017. Can you talk about your outlook for the HPC market as well in 2017? Maybe unit is not the right way, but just in terms of the opportunity for TSMC.
Our outlook for high performance computing in 2017, what is going to be the growth outlook for HPC in this year?
Mark, yeah.
Okay. The HPC, high performance compute, let me define that first.
Yeah.
It covers server, infrastructure networking, machine learning, gaming, AR, and VR. We see some other AI product to surface, but we cannot estimate it. It is just too early to estimate it. Of course, this year you have seen the gaming, AR/VR product starting to ramp. This is a very new growth area, and we see double-digit growth in the beginning. By the way, currently, we have a strong share in the networking infrastructure and gaming already. It's those other areas that are coming to the play at TSMC. You can see this year, you see a before-the-threshold kind of growth, and there's a double-digit growth only mostly in the product development phase.
If I can sneak in a question 1B, if you look at the GPU, how do you categorize it? Maybe this is for Lora. Historically, it's really for PCs, but going forward, as you use it in autos and AI, are you going to put it in computers or industrials, or how do you categorize that?
GPU for graphics, we put it in the PC. GPU for the server accelerator, we put in the HPC, high performance computing.
My second question is on capital spending. 2016 was a little bit higher than expected. What's the outlook for CapEx in 2017?
2017 CapEx will be about $10 billion.
Next question will be coming from Morgan Stanley's Charlie Chan.
Thanks for taking my question. My first question is regarding the advisory reports to the U.S. President that Friday, because in the report, the U.S. seems to want to ensure their semiconductor leadership, basically in a leading edge. What would be TSMC's strategy to fit in this policy direction and ensure your own leadership? Thanks.
You are referring to the advisory, the working group. That report was written to President Obama. There will be a new president in a few days. Whether any of that report will be adopted as policy by the new president is still unknown. Having said all that, I will say that Mr. Trump, President-elect Trump has said many times that he wants to create jobs in the United States. We highly applaud that. TSMC actually has created hundreds of thousands of jobs in the United States in the last 20, 30 years of our existence. For all practical purposes, we have created an industry, the fabless industry in the United States, and we have grown it. The fabless industry, I believe, employs hundreds of thousands of people in the United States.
We have done that by being here ourselves. I think that we will continue to create more jobs in the United States by helping the fabless industry in the United States and the IDMs that use us. There are many IDMs in the United States that use us. By using us, they have grown faster. They create jobs, too. I mean, that's actually, of course, Mr. Trump's main policy to create jobs. Of course, he has said a lot of things about other things. All right? I don't want to comment on them right now. As far as that report is concerned, yes. In fact, I think the report was preceded by the commerce secretary in the United States, Penny. Would somebody help me? Penny. Yeah. Mm-hmm.
I think that I agree with most of the things, I would agree with most of the things in the report. In fact, I agree with most of the things that Penny said. Just keep in mind that the report was written to President Obama. Mm-hmm. Actually, even if President Obama continues in office, I mean, the chance of his adopting the recommendations of a presidential task force are not very high. I mean, we have history to guide us. In fact, I will just tell you an anecdote. In July sixth, I met President Bush, then President of the United States. At that time, President Bush's task force, advisory task force on Iraq, had just submitted a report, basically recommending the U.S. withdrawing from Iraq. President Bush did not adopt the recommendation. He actually adopted the contrary.
Which was to increase the troops in Iraq. I mean, that's just an example that quickly came to my mind, when somebody talks about, "report has been written now.
Thanks, Chairman. This is very insightful. My next question is to [Lora], because it's a more maintenance financial question. Your first quarter gross margin guidance seems to be similar to your fourth quarter. Your revenue scale is down almost 5%. Why the gross margin is improving in first quarter? Can you give us some guidance about the depreciation as well?
We still have pretty big volume on 16 nanometer. The profit is improving, 28 nanometer still accounts for a big part of our revenue this year, with very good profits. We continue to do the cost reduction, we can maintain a pretty good margin in spite of decline in revenue in the first quarter. That's the reason. Your second question, you asked about depreciation for the whole year? Okay. With the $10 billion CapEx in 2017, which will be slightly front-end loaded, by the way. The depreciation year-over-year will start to grow about high teens
This year versus last year. In terms of first quarter, since last year's CapEx was very much back-end loaded, we expect a more than 10% depreciation increase quarter-over-quarter in the first quarter.
Okay, I think this is about time that we should take questions from the call. Operator, please proceed with the first caller on the line.
Certainly. The question comes from the line of Roland Hsu of Citigroup. Please go ahead.
Hi, good afternoon. Can you hear me?
Yes, I can. We can.
Okay, yeah. Thanks for taking my question. First question, I just would like to ask about recently, there is a 12-inch wafer price hike. I would like to ask, what's the impact to the overall gross margin, and how long do you think this 12-inch wafer supply and demand imbalance will last? Thank you.
I think I'll repeat Roland's question here. He's asking us about the recent raw wafer price increase. What's the impact of the raw wafer price increase on our gross margin, and how long do we expect this supply-demand imbalance in the raw wafer supply to last?
Lora?
The reason for the wafer price increase is because the imbalance supply, demand situations. It's more demand and supply. Manufacturer was trying to reduce the loss or make more profit. We anticipate some increase in raw wafer price. In terms of impact to TSMC, for this year, we expect impact will be about 0.2% of our gross margin. It's not very big, but it's not very small either. In terms of when the situation Sorry?
Yeah. Sorry. I couldn't hear that. Sorry.
Let me just answer very simply. We do buy a lot of things, and we have a lot of price every year. Every year, we have a lot of price increases, and we have also a lot of price reductions from our vendors. What we do, and I think we do it well, is to try to improve, in the past five, six, seven years, improve our structural profitability to the present point. What we're going to do from now on is at least to maintain that level of structural profitability. To manage that requires taking a lot of things into account, including price increase there, price reduction there, cost reduction here, cost reduction there, et cetera.
Understood. Yeah, I know our loss and you already made an impact to TSMC. Okay, thank you. My second question is your eight-inch revenue declined by double digits year-on-year last year, despite overall revenue increased by 11% in US dollars last year. I would like to ask, where did this eight-inch weakness come from, and how does it look for this year?
Roland is asking, last year, 2016, our overall revenue go up by 11%, but our eight-inch wafer revenue declined double digit. What causes such weaknesses? Secondly, what do you think this year is going to happen, whether or not our eight-inch weakness will persist to this year?
C.C., would you?
Yes. Thank you.
Okay.
Last year, our eight-inch wafer business decreased close to double digit, mainly because of a lot of product moving into 12-inch wafer. One of the example is the fingerprint sensor. We have customers are moving from eight-inch wafer to 12-inch wafer. This kind of a trend will continue. We expect this year will still decrease slightly. We have also do a lot of activities to increase our applications in the eight-inch wafers technology. The decreasing will be much more minimized as compared with last year.
Can we expect the utilization for eight-inch then will be much lower than previous this year?
No. Should be much higher.
The utilization will be close the same or a little bit above as compared with last year.
Okay. Understood. Okay, thank you.
All right, let's have the next caller on the line, please.
Next question's from the line of Brett Simpson from Arete Research. Please go ahead.
Thanks very much. Dr. Chang, you mentioned 10 nanometer and 7 nanometer was clear sky from a competitive perspective. Just interested, if you look at peak capacity for 7 nanometer in wafer terms, will this be a bigger node for TSMC than 28 nanometer? Thank you.
Brett, let me see if I get your question correct. You are asking us whether or not 10 nanometer and 7 nanometer will be bigger than 28 nanometer eventually. Is that your question?
In wafer terms, yeah.
In wafer counts?
Yeah.
In business, in revenue, as well as in wafer counts. You are asking too much.
Just-
We will only answer one of the two. Yes.
Only wafer terms.
Only wafer counts. Okay.
Of 10 7, is that right?
Yeah.
10 7 wafer count will be slightly less than 28. Slightly.
Okay. That's helpful.
The revenue will be much bigger.
Thank you. Just a follow-up, with regards to HPC, we are seeing a lot of specialty memory, high bandwidth memory, increasingly being packaged with your CoWoS. Does TSMC have any interest in developing stacked memories for HPC? Thank you.
Okay. Will TSMC be interested in developing specialty memory such as the high bandwidth memory ourselves to satisfy our business in CoWoS?
Well, we do not produce the special memory per se by TSMC, but we are working with the memory companies who provide the service under HBM.
Does TSMC expect
Brett.
Yeah. Would TSMC be interested in developing its own memories for HPC long term?
No.
Okay. Thanks very much.
All right. Let's come back to the floor. Next question will be coming from Goldman Sachs, Donald Lu.
Yeah. Very nice to see you, Chairman. One question is on the demand outlook for next year. Seems like everything except the smartphone is declining, and the smartphone is increasing only 6%. How do you get to a 4% semi growth? Is the smartphone content going to increase next year?
How do I get the
Right. Chairman says semiconductor growth is 4%.
Oh
Donald is saying that smartphone is 6%.
Also smartphone mix is going to the low end as well.
That's a tough question, Donald. I'm trying to think. These numbers are all compiled by our own forecaster. Again, he tells me what other forecasters are saying. It seems that all of them are saying about 4%, 3%, 4%, or even more, 5%. You're asking, where do they come from? Mark says silicon content. Answer the question loudly, please.
You're right. The silicon content of smartphone will increase from 2016 to 2017.
Even the mix is going more to the low end?
Even mix. We see the silicon content of all tiers increases, and the low end increases more. Overall, the silicon content will increase high single-digit next year.
I see. Right. Yeah. We are making our forecast based on your forecast, so it's good to know why. My second question is more on the AI, although I, like Mark commented, it's early stage, but it's pretty "hot," quote, unquote. How should we look at the opportunity for TSMC for AI? There's a big on the server side on building the model. There's also on the premise on handset or PC. Which market will be bigger, and how should we predict it in the future for TSMC? Is there any thoughts on that?
We have four world platforms. The first one is mobile, the smartphones. The second one is high-speed computing. The third one is IoT. The fourth one is automotive. We do see a lot of opportunities Or artificial intelligence in all those three platforms. Now, perhaps, Mark, would you care to say some more about the high performance computing, and perhaps CC will say something more about IoT and automotive. Yeah.
I think AI is a coming clear application for the industry, and people are excited. One of the reason is there is a tremendous progress in the algorithms. It makes the machine into much more than traditional programmable machine. This is the machine, can you think and do things. For the application side, it's immense. As many of you must have read on this, growing opportunities. All the cloud provider are in this field. Now, in AI, for us, I think it's also an important opportunity for us. Let me just talk about the computing side of AI. AI needs, for deep learning, it's collecting a lot of data, therefore, it requires very massive computation. In the past years, our technology development is collecting our pace, and we now can provide the world's most competitive technology for those artificial intelligence computing purposes. That's number one.
The people who can get into this field is across the industry using our technology. Secondly is, this AI is, the application is new. All the algorithm, all the architecture, they are all new. Hello? You hear? Okay.
You're on.
Basically, the playing field is leveled. It is not as before, where the high performance computing, you have to be a certain architecture to get into this field. This is a leveling playing field. Many players are into this field. That is where the massive innovation can come. We are very excited, and also because of this, we see the application will include many segments. We developed this advanced packaging, and that is an area that requires the AI application. That area is also, we spent several years at getting to this, and now already come to the fruit, that be it InFO or the CoWoS, and there will be multiple generations of that advanced packaging. On the technology side, we are prepared to embrace this, many people quote, "explosive growth." That is it.
Leave some time for CC too. Government through the AI.
Yeah, I think there's a lot of growth opportunities for IoT and automotive too.
Oh, yeah.
In connection with AI.
In connection with the AI, I think we need to have a lot of sensor connectivity. You know that, right? To constitute all the data format and connect with the AI, so they can do a lot of applications. We do believe that in the future, that IoT, which is a lot of sensor connectivity, put all of them together, and combine with the high performance computing, and then form the fundamental pieces of the AI. We do expect the IoT business to grow very fast because of in the future, you will see a lot of applications, even the autonomous driving. You need to look at all the background, all the environment surrounding a car, and also the car has to communicate with the cars together.
We project the future in that the whole world, you will see a lot of connectivity, a lot of sensors, and that's where we constitute our IoT business.
Okay, the next question will be coming from Daiwa's Rick Hsu.
Hi, happy New Year, Dr. Chang and all. I guess my first question goes back to your second quarter and third quarter business outlook. I think, despite the kind of a weak Q1 revenue, Mr. Chairman was talking about your second quarter revenue and Q3 revenue will grow significantly. You also mentioned about your 10 nanometer ramp-up will be also meaningful in these two quarters. What's the main demand driver behind that growth momentum in terms applications? Would that be just coming from mobile or across the board or mobile, professional gaming, et cetera?
Just repeat the question now.
Right. Rick, your question is, what's the growth driver for our second quarter and third quarter business? Is it only from mobile or from other areas?
Well, actually, the simplest answer to your question is that growth driver is always customers' demand. At our elaborate, I think we do have more concentrated customers today than we did several years ago. Those particular customers' demand fluctuation impacts our seasonal, I mean, our quarterly fluctuations a great deal. Seasonality to us doesn't mean very much anymore. Every year's seasonality is different from the next year's or last year's. Last year, we had a pretty low first quarter and a pretty strong second quarter. Then third and fourth quarters were even stronger, I think, right? This year, at least at this point, we see a somewhat different pattern. It's constantly, and of course, technology plays a part. 10 nanometer ramp-up.
Thank you so much. That's clear. The second question probably goes to Lora or maybe to Mr. Chairman as well. Can you give us more insight for color about your cash dividend this year, and kind of ballpark number?
What was the question?
He wants to know additional color on this year's dividends per share.
Lora?
Okay. I have a short answer. We will increase cash dividends in 2017. Magnitude, prefer to wait for the board meeting. February, you're going to know by then.
Thank you so much.
You won't have long to wait. I mean, February something, right? I don't remember the exact date of the board meeting.
14th.
Huh?
Mid-February.
8th or 9th. Well, I think, yeah. Of course, the final decision is made by the annual shareholders meeting, and that's in June. The board's recommendation on the dividend will be announced to the public after the board meeting.
Thank you.
All right. Next question will be coming from HSBC, Steven Pelayo.
Great. Thank you. The segments that you report to us, communications, consumer computing and other are helpful, but now two-thirds of your business is coming from computing, so we can't really divine a lot from that. We're all excited about automotive and AI and high-performance computing, IoT. If we were to use those as the segments, what percentage of revenues does TSMC get from each of those categorizations today?
Maurice?
For last year, roughly, for the mobile, it's about 55%, 56%. High-performance computing, with the definition I just said, it's about 15% roughly. PC is not a growing sector, but it's about 10%. IoT and automotive, both of them are less than 10%.
Okay. Maybe a question for Lora. I'm very impressed with your gross margins in the first quarter potentially going up even on down maybe 10% revenues. Could you talk a little bit about the nodes and the utilization rates maybe in the first quarter? Is 16 nanometer falling off significantly, yet you're making up the margin elsewhere? Help us understand maybe what's driving the margins in the first quarter if we looked at each of the individual nodes and how they're going to perform.
The 16 nanometer is not falling off. I just said that I wish our market share were higher, okay? It's not falling off.
The 16 nanometer will be flat subsequently?
I'm sorry. Other than 16 nanometer not falling off, didn't you have a question?
Steven, actually, you are asking a product mix thing, right? It's more than product mix, actually. It's improving profitability in each technology node, particularly 16 nanometer, and we have very significant amount of revenue coming from 16 nanometer in first quarter. The 28 nanometer continue to be a very strong node. Those two nodes actually supported our first quarter margin pretty well, in spite of a Lesser revenue growth or declined revenue growth. That was the main reason. Plus, as I said earlier, continual cost improvement efforts in the operations side. That also helps.
Just to follow up there, if you look at your full-year gross margins, 2016, about 50% or so. You're suggesting maybe even as much as 53% or higher in the first quarter. Are TSMC structural gross margins higher this year that we can maybe look for a couple of hundred basis points growth for full-year 2017 gross margins?
What was the question?
All right. Steven, if I understand your logic, his question is, given 50% gross margin for TSMC for the full year, 10 nanometer will take two points away from our margins in the second half, that means first half, our gross margin will be very good to get the 50% average.
Look, let's not take all the numbers I say literally. When I said we do have a narrow range. The range is not zero. When I say that we want to keep our structural profitability at about close to 50%, at least give me one or two points leeway, okay? It's not exactly. Give me one or two points leeway. Yeah. It's like 49, 50. We really can't even forecast the next quarter's gross margin that closely. We always have a range of what? Two or three points.
Yeah, maybe two or three points.
Yeah.
Can't be that precise.
I think Steven's main question is, do we increase our structural profitability this year?
Well, I think that most of the improvement has already happened. If you look back six or seven years, we have improved it by several hundred basis points, that was not easy at all. That was not easy. I think that, I would say, most of the improvement has already occurred. From this point on, our priority is to maintain it.
If I could just speak one more quick one in. When do you think 10 nanometer could be more than 10% of revenues, which quarter this year?
Which quarter 10 nanometer will exceed 10%?
Do we answer questions like that?
Sometime in the second half of this year. Yes.
If we answer questions like that, I think it should be Lora, right? You don't have to answer it, no. You don't have to.
The big volume comes on the second half. The whole year will be more than 10%.
His question is, when does it exceed the 10%, I think.
Third quarter.
Really?
Okay. Next question will be coming from Credit Lyonnais, Sebastian Hou.
Thank you, Dr. Sun. Chairman and the two CEOs.
Careful now. He means volume. He doesn't mean gross margin.
Yeah, revenue.
Revenue, yeah. Okay, good. Just want to be sure that Lora understood the question. Yeah.
Thank you. My first question is on the assumption of the growth rate that Chairman already mentioned about smartphone unit growth this year, you expect 6%, and Mark mentioned about the.
What did you say? I'm sorry. What did you say?
The growth rate. Smartphone, Chairman mentioned units.
Oh, smartphone. Okay, yeah.
6%. Silicon contents grow about high single digit, per Mark, which means we're growing about 13%-14%. Given that mobile account for 55%-56% of the total revenue this year, which means smartphone alone or mobile device alone can contribute 7%-8% of the growth of this year. Plus the growth in high-performance computing and double-digit growth, and also IoT growth of 34% of the whole market, which means it seems like easily to reach the high end of the growth, even close to 10% for this year. I'm not sure if my assumption is right or too bullish.
Assumption right, calculation may not be. I think that most of the growth this year will come from smartphones, 80% of that. That's because IoT and automotive are small. They do have looking to 20%-30% growth for those growth areas. Contribute to the whole company is still very small. Compose the rest of it. I don't calculate it up to that number. I think it's still in line with about between 5%-10%, probably in the middle. That's my calculation.
Okay. Thank you. My second question is on the President-elect, Donald Trump in the U.S., that his policy is to create more jobs. Thank you, Chairman, for mentioning about how TSMC has helped create more jobs indirectly in the U.S. I'm wondering if you already got some pressure from your customers in the U.S. that maybe when they do the calculation factoring the tax incentives, that they think that maybe having manufacturing back in the U.S. can help them to improve their profit. I'm not sure. I'm just wondering whether you think about or ever consider building a fab in the U.S. and under what kind of a scenario and circumstances?
Sebastian, your question is whether or not we have pressure from U.S.-based customers wanting us to build a manufacturing site in the U.S. so that they can capture-
Oh
incentives.
No. Okay, sorry. My second part of that question is that under what scenario and circumstances will TSMC consider building a fab in U.S.?
Will we ever consider going there and build a fab?
I do not rule it out.
Okay. Thank you.
I do not rule it out, but I see a lot of sacrifices that we and our customers will have to make if we do that. Keep in mind that we earned our business in the United States, not by lower labor costs in Taiwan. In fact, on the whole, our labor costs in Taiwan are not lower than the United States. Now, we earn our business by being good. Just to give you an example, every year, we send thousands of engineers from one location to another. We have three locations in Taiwan: Taichung, Hsinchu, and Tainan. Now, Taichung and Hsinchu are within daily commuting distance. Tainan, and you can even make a point that Tainan and Hsinchu and Tainan, Taichung are within daily commuting distance also.
Although, it's a little hard, but I know of people in New York, Manhattan, I know people that commute an hour and a half, two hours each way every day. If you are willing to do that, you can certainly commute between Tainan and Taichung or Tainan and Hsinchu also. I will even, backing up if you said you don't want to do that, it takes too long. It's still within weekly commuting distance. All right? Every year, we send literally over 1,000, 2,000 engineers from one location to another to be there for months. That allows us to ramp up things quickly and to solve problems quickly also. When one location, when one fab has got a problem, engineers from another fab can go ahead and help them out. Literally, thousands of engineer years do that every year.
If we have a plant in the U.S., won't be able to do that anymore. Second point, we have thousands of vendor people here. Well, actually they're both in Hsinchu and Tainan, and some in Taichung also. Thousands of engineers from our partners located here. These are things that, not low labor costs. These are things that we'll lose if we set up a plant in the U.S. If we lose these things, our customers will lose, too. When you say, "Are you going to" I don't rule it out. Actually, I've never ruled it out. Year after year, time and again, we consider the subject, and we have not made the decision to go there. Except, of course, early in our life, we did set up WaferTech in the United. It's still going. It has about 1,500 employees.
It's still going. Other than that was set up in 1996. After that, we consider the question every so often, and we have not made the decision to set up a plant in the United States.
Okay. With that, we need to go back to the line. Operator, can you please have the next caller on the line?
Certainly. Next question is from the line of Mehdi Hosseini from SIG. Please go ahead.
Yes. Thanks for taking my question. Dr. Chang, good to hear your voice on the call. I have a question on the 10 nanometer. If your 16 nanometer market share is about 65%, what is your view on your market share when you migrate to 10 nanometer? I have a follow-up.
Well, market share in 10 nanometer, I think I vowed about once. It must have been a year ago, because I was here last time, a year ago. I vowed then that every new node from now on, we have a market share higher than our market share on 16.
Okay. More than 65%, is that your answer?
What?
Would that be above 65%?
Very definitely.
Okay. Then when we look into 7 nanometer and some of your competitors have a different definition of seven, can you help us understand how your N7 competes and is positioned against your competitors?
Mehdi is saying that our 7 nanometer definition is different from other companies' 7 nanometer definition. How do we compete at seven?
Our 7 nanometer definition is different from somebody else. Well, I'll let Mark answer.
Okay. I wouldn't want to comment on other people's 7 nanometer. Our 7 nanometer is under qualification now, it will be qualified as according to plan around the end of first quarter. We have already more than 20 customers designing on this 7 nanometer. This year alone, we estimate will be 15-20 tape-outs already. This is our momentum built so far on 7 nanometer, no other competitor is getting to this stage of this leading-edge technology. Remember, I mentioned last time, we'll have 5 nanometer two years from now. Forget about the name, that will be a full node shrink, that will sit competing well with any technology come out at that time. Okay. Let me add some on 7 nanometer. We will maintain our 7 nanometer competitiveness just like we do on 28 and 16.
We will have a technology currently planned at 7 nanometer, but with the EUV insertion in the second year of 7 nanometer, approximately one year after. That can greatly simplify the process without increase the cost. That is if customer can take advantage of minor design, it can further increase their density, reduce their die cost, that is our plan to maintain the competitiveness of the 7 nanometer the year after. We think 7 nanometer is a well-adopted node by all the customers, we plan for the subsequent technology to shore up the demand continuously. We hope to use this technology, I mean, the second year technology, to prepare for the EUV production experience for the full-fledged EUV technology on five. Our customer can have a very, hopefully smooth, getting to from our seven to our 5 nanometer technology.
That is how we maintain our technology competitiveness.
May I ask a clarification question?
Sure.
I think if I heard you correctly, you said you will insert EUV in the second year of your 7 nanometer, which suggests to me that you may actually be able to commercialize [audio distortion] conclusion here.
Mehdi, I'm afraid that your voice was broken at some point in time. Can you please repeat? We heard you said that we will insert EUV in the second year of 7 nanometer, and then you had something that it was cut off. Can you repeat that part?
Sure. Yes. Sorry about that. I just want to make sure I understand the EUV commentary correctly. You said that you would insert EUV in the second year of 7 nanometer. That suggests to me that you may actually be able to insert EUV before competitors that have said insertion will happen at 5 nanometer. Is that a right conclusion as we compare and try to better understand your competitiveness at 7 nanometer?
Yes, we will commercialize the 7 EUV in the second year of our 7-nanometer production. I wouldn't comment on when will our competitor insert their EUV. I don't intend to do the comparison here.
Got it. Thanks so much for the details. I appreciate it.
Thank you. Now we come back to the floor. The next question will be going to Deutsche Bank's Mike Chou again.
Just two quick question. One is, what would be the tax rate in 2017?
Similar to 2016, 13%-14% effective tax rate.
Second question. If we look at the first 3 years of 7 nanometer TAM, would that be bigger than 28 nanometer for the first 3 years as well? Year 10, total addressable market.
Total available?
Total addressable market.
Addressable market. Yeah, in terms of dollar, much bigger.
Much bigger.
Okay. going to HSBC, Steven Pelayo again.
Just two quick questions from me as well. With 2016 finished now, I believe if I remember correctly from your 20F 2015, you had two customers that were 16% of revenues each. I'm curious, what was your customer concentration in 2016?
I think in general, our top five customer accounts about 50% of revenue, and top 10 about 70%. This percentage hasn't changed much over years.
Any more details on the ones you disclose that are over 10% of revenues? I can wait for the 20F, I guess.
Please wait for 20F.
Okay, my second question was, you generated over TWD 200 billion in free cash flow this year on TWD 950 million in revenues, roughly 20, 22% of revenues, free cash flow. Quite nice. If you're talking about high teens growth and depreciation this year, maybe 10% growth in your EBIT, you're probably going to add another TWD 80 billion or so, by my calculation, to your free cash flow, which could get you over 30% of revenues. Is that a realistic target to think about this year? TSMC generates more than 30% of revenues in free cash flow?
I cannot be that precise, but I think our free cash flow will grow this year versus last year. Now going to Credit Suisse, Randy Abrams.
Two quick follow-ups. The lull that you're seeing in the business now, could you talk about how broad-based that lull is? If it's across application end markets or if it's select parts. You mentioned a lull in the business that you're moving through. How broad is that lull across application or your customer base?
The lower end of?
The lull. No, you talked about like a slowdown. For that slowdown, if you could talk by application, if it's across your business, broad-based, or if it's isolated to certain areas.
First half of this year, what's the weaker area of our demand, right?
I think it's Mobile.
What was the question?
The question is, with respect to the first half of this year, what will be the area that the demand is relatively weaker?
how broad-
You mean what market sector or?
Yes.
Market.
Market. How broad-based.
Can you answer that question, Lora?
What was the question? Sorry.
The other question.
Can you tell me what was the question?
How broad-based is the slowdown, and which applications are you seeing the slowdown? Any area that's holding up stronger.
First quarter?
For first quarter or first half.
I think Mark, do you know the answer? Mark?
Let me answer the question.
Do you know the answer? The slowdown really comes from the seasonality of our major customers in the mobile areas. There will be minor inventory corrections because, as Lora Ho mentioned, exiting last year is two days above inventory, and exiting first quarter may be slightly higher still. We see some minor inventory correction going after, we don't see a major slowdown at all.
The second follow-up for the R&D expense. I think you talked about increase. If you could talk about the range you see R&D moving in an increase or as % of sales.
R&D as a % of sales, right? Mm-hmm.
Talked about rising.
Right. We said we are going to increase the R&D spending as well as the percentage. What's the percentage?
It's approximately 8%.
Percent.
8%.
Last year was 7.5%. This year will be slightly above 8%.
When I said a moderate increase in its percent, I mean several tens basis points. All right? Several tens basis points.
All right, follow-up question from Goldman's Donald Lu.
I have a question on inventory in Q4 has declined quite substantially. Will it increase substantially in Q1, and would that be potential boost to gross margins in Q1?
I don't believe so. Not in Q1. Follow-up question from Credit Lyonnais, Sebastian Hou, I think that should be our last question for this conference today.
Thank you for letting me ask. My question is there any impacts you see there from the minimum wage hike and the Yi Li Yi Xiu new regulation by the government on government on the-
What was the question?
-cost side?
Okay. What's the impact of the government's new policy on 一例一休 and the minimum wage?
Yes. Our first estimate is that it will increase our cost about 30 basis points. 30 basis points. That's right, isn't it, Lora?
That's right, 0.3% of cost.
Yeah.
It's on the cost of goods sold. My second question is on the 2016 CapEx that turns out to be actually slightly higher than the last conference revision, about like $500 million. I wonder whether that's just a cash payment pulling or any other plan change of the capacity.
It's just a pulling in of payment. There's nothing changing capacity buildup.
Thank you.
Okay. I think we have concluded our Q&A session. Before we conclude today's conference, please be advised that the replay of the conference will be accessible within three hours from now. Transcript will become available 24 hours from now.
Again, I wish you a very happy New Year.
Okay, thank you for joining us today. We hope you will join us again next quarter. Goodbye, and have a good day.