Welcome to TSMC's fourth quarter 2012 earnings conference and conference call. This is Elizabeth Sun, TSMC's Director of Corporate Communications and your host for today. The event is webcast live via tsmc.com. If you are joining us through the conference call, your dialing lines are in listen-only mode. As this conference is being viewed by investors around the world, we will conduct this event in English only. The format for today's event will be as follows. First, TSMC's Senior Vice President and CFO, Ms. Lora Ho, will summarize our operations in the fourth quarter and for the full year 2012, followed by our guidance for the first quarter 2013. Afterwards, TSMC's Chairman and CEO, Dr. Morris Chang, will provide his general remark and a couple of key messages. We will open the floor to 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.com. Please download the summary slides in relation to today's earnings conference presentation. Before we begin, I would like to remind everybody that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the safe harbor notice that appears in our press release. Now I would like to turn the podium to TSMC CFO, Ms. Lora Ho.
Thank you, Elizabeth. Good afternoon, everyone. Thank you for your participation today. My presentation will start with financial highlights for the fourth quarter and a recap of 2012 financial performance, followed by the guidance for the first quarter. In the fourth quarter, despite inventory correction in the IC supply chain, demand for our products was higher than we expected three months ago, resulting in above-guidance revenue and profit margins. On a sequential basis, our fourth quarter revenue decreased 7% to TWD 131 billion. Gross margin was 47.2%, down 1.6 percentage points from the third quarter. This is mainly due to lower capacity utilization, while cost improvements and favorable inventory valuation adjustment offset some of the decline. Operating margin was 35.2%, down two percentage points from the third quarter. Operating expense as a percent of revenue increased 0.4 percentage point on a smaller revenue base.
Non-operating items were small losses of TWD 7 million in the fourth quarter, as we recorded an impairment charge for certain invested companies. Overall, our fourth quarter EPS was TWD 1.61. ROE was 23.8%. In terms of revenue by application, as I just mentioned, inventory correction in the IC supply chain has affected overall demand for TSMC's wafer. However, demand for mobile computing devices remained firm, making communication the only growing segment in the fourth quarter, while demand for computer, consumer, and industrial-related products all declined by double-digit. As a result, revenue contribution from communication-related applications further increased from 49% in the third quarter to 54% in the fourth quarter. On a full year basis, communication increased 23% and it represented 50% of our wafer revenue. The major contributing segment include application processor, baseband, CMOS image sensor, and wireless LAN, reflecting the strong demand for mobile computing devices.
Another fast-growing application is industrial and standard, which grew 42% year-over-year. The growth was mainly contributed by the increasing usage of power IC, data converter, touch controller, and flash controller within mobile devices. This also reflected our success in developing specialty technology businesses. If we look at the revenue by technology, thanks to customer strong demand for TSMC's 28-nanometer technology and the excellent execution by the operation team, revenue contribution from this node jumped from 13% in the prior quarter to 22% in the fourth quarter. We see 28-nanometer as a very successful node for TSMC, which already accounts for 12% of our full year revenue in 2012. Looking forward, we remain confident 28 revenue contribution will exceed 30% for the whole year in 2013, and margin will be at corporate level starting from this quarter. Taking a look at the balance sheet.
On the asset side, cash and marketable securities ended the quarter at TWD 151 billion, slightly increased from last quarter. Long-term investment increased to TWD 66 billion, mainly due to the TWD 31.5 billion investment in ASML shares. This strategic investment will be recognized as available for sale in our balance sheet. On the liability side, current liability increased TWD 23 billion, mainly due to increase in payables to contractors and equipment suppliers, as well as increase in short-term loans for hedging purposes. Long-term debt increased by TWD 4.4 billion to TWD 82 billion, as we issued more corporate bonds in the fourth quarter. Our days of inventory increased by 6 days to 50 days, mainly due to higher working process for 28-nanometer products in response to a strong market demand and an increase of raw material.
On the cash flow side, we generated TWD 85 billion from operations in the fourth quarter and invested TWD 60 billion in capital expenditure. Overall, our cash balance increased TWD 5 billion to TWD 143 billion at the end of the fourth quarter. Due to higher operating cash flow and the lower capital expenditure, free cash flow improved from negative USD 2 billion in the third quarter to positive USD 26 billion in the fourth quarter. Now let's take a look at the capital expenditure. We spent USD 2 billion on capital expenditure in the fourth quarter. As a result, full year CapEx ended at USD 8.3 billion, in line with our prior guidance. Let me make some comments on our capacity plans. As we continued adding capacity for 28-nanometer process technology, our total capacity increased 4% to around 4 million 8-inch equivalent wafers in the fourth quarter.
For the full year, our 12-inch capacity had increased by 21%, and the total annual capacity increased by 14% to reach about 15 million wafers. We expect the total capacity to decrease slightly by 1.2% in the first quarter of 2013 due to fewer working days and scheduled maintenance. Now, I would like to give you a recap of our performance in the year of 2012. 2012 is a record year for TSMC. Our strength in technology and the manufacturing made us well-positioned in the mobile computing market. We are happy to mark a year with record sales and profitability, despite difficult macroeconomic environment. Compared with 2011, our revenue increased 18.5% year-over-year to reach TWD 506 billion. We continue to outperform the semiconductor and foundry industry. For profitability, although the rising depreciation and fast ramp of 28-nanometer have indeed put pressure on our gross margin.
Nevertheless, our gross margin has increased to 48%, a 2.7 percentage points increase from 2011, thanks to higher utilization of the capacity we invested in. Earnings per share increased 23.8% to TWD 6.41. ROE for the whole year was 24.6%, compared to the 22.3% in 2011. Before I dive in the first quarter guidance, I would like to brief you about the changes in our effective tax rate in 2013. TSMC's tax rate is projected to be 14% in year 2013, which is significantly higher than 8.7% in 2012. The higher tax rates are due to the following reasons. First, the introduction of capital gains tax in Taiwan, coupled with the increase of the AMT rate from 10% to 12%, will cause our effective tax rate to increase by two percentage points.
Second, higher tax penalty on appropriate retained earnings, as we are making more net income by keeping the same dividend. This will add 1.7 percentage points to our effective tax rates. Third, a 0.9 percentage points increase caused by the expiration of certain tax exemptions. Lastly, the 0.7 percentage point is attributed to the absence of certain subsidiaries loss carry-forward. Under the current tax environment, we expect the tax rate for 2013 and 2014 will be around 14%. I have finished my financial report. Now let me provide you our first quarter guidance. Based on our current business expectation and the forecast foreign exchange rate of 28.90, we expect our revenue to be between TWD 127 billion and TWD 129 billion. In terms of margins, we expect the first quarter gross margin to be between 43.5%-45.5%, and operating margin to be between 31.5%-33.5%. This concludes my remarks.
Let me turn the podium to our Chairman and CEO, Dr. Morris Chang.
Good afternoon, ladies and gentlemen. I'll make a few comments on last year's achievements, on this year as a whole, and specifically on first quarter. I'll also comment on 28 nanometer technology and on 20 nm and the 16nm FinFET, then finally on this year's CapEx. Lora has already reported the financials of last year to you. Basically, last year was a year of achievements for us. Revenue grew 18% to reach $17.1, and EPS grew 24% to reach TWD 6.41 per share. 28 nanometer technology was a resounding success. The production in 2012 increased more than 30-fold over 2011. We have enjoyed throughout the year, in spite of a lot of attempt of the competition, we've enjoyed throughout the year close to 100% foundry market share in 28 nanometer technology. We have also, in the year 2012, further strengthened our R&D.
Our R&D expenditure increased from TWD 33.8 billion in 2011 to TWD 40.4 billion in 2012. Our R&D people increased from 3,400 at the end of 2011 to 3,900 at the end of 2012. A few words on 2013 and the first quarter of 2013. For the full year 2013, we are forecasting a global GDP growth of 2.6%, which is a bit higher than last year's global GDP growth for 2.4%. We're forecasting 2.6% for this year. We are forecasting a world semiconductor market growth of 3%. We are forecasting a fabless company growth of 9%. We're forecasting a foundry industry growth of 7%. We are forecasting a TSMC revenue growth much higher than 7%. Those are our forecasts for the full year 2013. For Q1 2013, I have some comments on supply chain inventory to make.
Three months ago, in the last investors conference in October, we expected the supply chain inventory to decline from seven days from above seasonal in Q4 to one day below seasonal in Q1. That was our expectation three months ago, that the supply chain inventory would decline from seven days above seasonal in Q4 to one day below seasonal in Q1. All together, an eight days decline. Because many mobile product manufacturers have accelerated their new product launch this year. Instead of late in the year, they have now pulled ahead to earlier in the year. Therefore, they need IC supplies, IC inventories earlier. The supply chain inventory now is forecast to decline only slightly from Q4 to Q1. Instead of seven days above normal to one day below normal, we're now forecasting that the inventory will decline from six days above seasonal.
Well, instead of seven days. Three months ago, we thought the inventory was going to be seven days above seasonal. We have better information, we think it's six days above seasonal in Q4. The big change is that we are forecasting it to decline to only four days above seasonal in Q1, which is only equivalent to a two-day decline, instead of the eight-day decline that we had forecasted three months ago. All this resulted in a higher Q1 than we thought three months ago. We now expect the Q1 revenue to decline. Three months ago, we expect it to decline. The Q1 revenue, we expect it to decline from Q4. We now expect the Q1 revenue to be essentially flat in US dollars from Q4. Essentially flat from Q4 in US dollars. A few more words on 28 nanometer technology.
After accomplishing a 30-fold increase in production, 28 nanometer capacity and output continue to ramp up aggressively this year. Production of 28 nanometer wafers in 2013 will triple that of 2012. I think the newcomers may ask, do we have customers? The old customers, I think, know us well enough not to ask that. Yeah, we have customers. High-k Metal Gate will surpass oxynitride. That's in the 28 nanometer. We have, as you know, actually four types. Three of those are High-k Metal Gate, and the earliest type that we introduced was the oxynitride, and indeed, last year, the majority of the production was the oxynitride. The more advanced version, High-k Metal Gate, will surpass oxynitride in Q3 this year.
In Q4, it will even surpass oxynitride even more. Gross margin percentage of 20 nm in 1Q13, which is this quarter, will be slightly higher than corporate average and is expected to remain so in 2013. It's neither dragging the corporate average down, nor is it pulling it up. A few words on 20 nm and 16nm FinFET. Both technologies are in progress in R&D. Both represent state-of-the-art, leading-edge technology, not just in foundry, but in the whole semiconductor industry. Enough discussions have taken place with enough customers with large requirements to lead us to believe that in both its first and second year of production, in both the first and second year production of 20 SOC, first year will be next year, 2014. Second year will be the year after that, 2015.
In both those years of 20 SOC production, the volume of 20 SOC will be larger than 28-nanometer in its first and second year of production, which were last year and this year. That's a long sentence, but let me repeat it in slightly different words. We think that our volume of 20 SOC next year, 2014, will be greater than the volume of 28-nanometer last year. We think the volume of 20 SOC in 2015 will be greater than the volume of 28-nanometer this year. A few more words about CapEx. This year, it will be about $9 billion, give or take a few hundred million, $9 billion. 88% of it will be for 28-nanometer, 20 nm, 16-nanometer building facility equipment. Basically, 88% will be for 28, 20, 16, everything. We're building new buildings now, in fact.
We have been building new buildings for quite a while. Facilities and equipment. 5% will be for R&D equipment. That's basically 10 nm and beyond. CoWoS and whatever. 2% for specialty technology equipment. That is power, embedded flash, microcontroller, imaging, and so on. 2% of capital. 1% is for the piece of land that we just acquired in Zhunan. I think I told you about it last time, did I not?
Yes.
I think I told the conference about it. We bought a piece of land in Zhunan, which is about, I understand, 15 or 20 minutes driving from our Hsinchu headquarters. Those are the comments that I've prepared.
All right. This concludes our prepared statements. Before we begin the Q&A session, I would like to remind everybody to limit your questions to two at a time, to allow all participants an opportunity to ask questions. Questions will be taken both from the floor and from the call. Should you wish to raise your question in Chinese, I will translate it to English before our CEO or CFO answers your question. 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 the order in which they were received. If at any time you would like to remove yourself from the questioning queue, please press the pound or the hash key. Now let's begin the Q&A session. First question goes to Deutsche Bank, Michael Chou.
Comes from Michael.
Hi, Chairman. Could you give some update for your LED and solar business? Thank you.
Yeah. On both LED and solar, our strategy is first to achieve a technological distinction. Either better performance than competitors, than current competitors, or better cost than current competitors. Anyway, a technological distinction before we start large-scale production, before we put a lot of capital money into it, et cetera. That's the philosophy. That is the strategy, philosophy for both solar and LED. At this point, I think we are perhaps a little further ahead in LED than in solar. I think the environment is a bit more friendly to the LED also. In LED, we are forging our way ahead. In fact, in the fourth quarter, last quarter, we already had some commercial revenue. By commercial revenue, I mean more than samples, but really real product for real money. We already had some revenue, and we expect the revenue to increase pretty dramatically this year.
I don't know whether it's going to be 30-fold or not, but it will increase pretty dramatically this year. On solar, as I said, the environment is not as friendly because the world has a, at least silicon solar capacity glut, and major countries are putting up incentives for more production. However, we think that our thin film, our CIGS, C-I-G-S technology, is a pretty promising one. We are spending a lot of R&D money to try to achieve a distinction with the CIGS technology. We are not in a commercial production, commercial revenue stage yet.
Thank you. A follow-up question is, would your LED be based on silicon substrate or traditional sapphire substrate?
I guess I think we said that earlier.
We actually have both.
Oh, you answered the question. You answered the question.
Yes, I think we have both.
Thank you. Second question is.
I'm not sure if her answer is right, but okay. We'll go into that, yeah.
I will verify this and let you know.
Yeah.
Okay. Thank you. My second question is, what is your total capacity increase in 2013? Would you increase some capacity for 20 nanometers as well? Thank you.
With the TWD 9 billion CapEx, we expect to increase the total capacity this year versus last year about 10%. We are putting money on 20 nm.
We are building it.
Yeah. We have a small pilot line for 20 nm.
Why do you think I composed a long sentence where I said we are planning our capacity accordingly? That means we are spending money on 20 nm. We will be spending money on 20 nm capacity.
Thank you.
Next question comes from Citi, Roland Shu.
Hi, Chairman and Lora. My first question is to Lora. Under your TWD 9 billion CapEx spending this year, how is the depreciation increase this year?
With the TWD 9 billion, which is very much front-end loaded, about two-third front-end loaded in the first half and one-third in the second half, we expect depreciation will go up around 23%-24% year-over-year. It's very similar to this year. This year, depreciation went up by 22%.
Okay. This is a little bit bigger than 20% you guided us last quarter. What is the reason?
As I just said, it will be front-end loaded, depreciation will happen earlier for the whole year.
How about the first quarter depreciation?
First quarter depreciation will be only slightly higher than fourth quarter. Starting from second quarter, it will increase faster.
I think my follow-up question for this one is, under IFRS, actually, the depreciation years actually can be revisited every year by the company. Are you considering changing your depreciation year on your 20 nm, or even 16 nanometer investment since most of this investment on this back-end online equipment for this metal interconnection actually can be shared. That means the investment on the back-end online for this technology, the lifetime definitely will be longer than previous technology. Are you considering revisiting the depreciation years for the investment on this node?
Actually, our economic life of our technology has always been longer than the depreciation year, which we use five years. We have no plan to make it longer. For certain tools, actually, the economic life might be lower than five years. For those parts, we may consider to use lower depreciation years.
Okay.
That's the other way. I think he's suggesting that since a lot of equipment can be used for 20 nm, blah, blah, blah, and 16 and so on, we can make the life longer. Well, thank you for the thought. We will think about it.
Okay, thank you. My second question actually is to Chairman. Again, Chairman is commenting this year's TSMC growth will be much greater than 7% of the foundry. Can you comment about what's the growth driver to drive such great growth? Do you see any momentum change for the IDM outsourcing in this year? Thank you.
Well, first of all, let me say that I'm not trying to be coy when I say that our growth will be much higher than 7% without specifying a number. I'm not trying to be coy. It's just that if I predict more specifically, I think that we'll get a call from the Taiwan SEC that will want it. That's the reason. Now you asked about the driver. The driver is the 28-nanometer. Just a quick back of the envelope calculation will show you that the growth of 28-nanometer this year will actually be greater than the total growth of the company, even if the total growth of the company is much higher than 7%. I was talking about what? This year's production of 28-nanometer.
Triple.
Triple, yeah. That growth is I think your question was originally on what customers, and I'm not going to comment on customers, but I will say that the driver is 28-nanometer technology.
How about the IDM company's contribution this year?
The what? The what?
IDM, especially for Japanese IDM company.
Oh. You know the answer, or shall I answer it? Shall I answer it?
I also know the answer.
I think, yeah, I do expect the IDMs in Japan to contribute more. Compared to the contribution that the 28-nanometer technology will make, the Japanese IDMs increase outsourcing to us. It will be relatively small. Yeah.
All right. I think somehow this time, there are quite a few analysts who used to attend our conference in person here are actually overseas this time, and they are on the call. I think we will now take our next question from the call. Operator, please proceed with the first caller.
The first question on the line today comes from the line of Daniel Heyler from Merrill Lynch. Daniel, please go ahead.
Thanks for that. Thank you, Elizabeth. I'm sorry I couldn't be there this time. A quick question. I guess two. First relates to Dr. C. L. Lei's growth outlook. As you mentioned, foundry growth up 7%, fabless up 9% in 2013, TSMC much bigger than that, and you did mention 20 nm. I want to ask, maybe talk a little bit more in terms of some of the end markets, because I did notice that your computer and consumer businesses were pretty good in 2012, showing a pretty strong growth. I'm wondering, in the non-communications business, what your prognosis is for your businesses in 2013?
I didn't get it. You could repeat?
If we exclude the communication-related applications, what would be the growth for TSMC, both non-communication-related business growth?
I think that almost all the 28-nanometer is communications related, right?
There are some-
There are some computers, too.
There are some computers.
Yes.
What would you say? I'm here roughly calculating with you, Dan. I think that maybe one-third is computers. One-third of the 28-nanometer growth is computers. That leaves two-thirds communications related. With that growth, then we try to answer your question now. I think the rest of it is almost the same. It doesn't change very much.
Okay, great. Fair to say, you don't need to see significant
You are talking too loud, but not very clear. If you maybe slow down a little bit then, and pronounce each word more clearly, maybe it will help. Yeah. Mm-hmm.
Thank you. Thank you. I'm trying to save time for my colleagues. Okay, fair to say that the growth this year that you're bullish about, you really don't need to see necessarily a big recovery in consumer or computer to achieve your bullish forecast, that you can pretty much generate that kind of growth singularly from communications. Is that a fair summary?
Yeah, that's a fair summary, I think really a more basic answer is that we intend to generate our growth with new technology. Mm-hmm.
Great. Okay.
Yeah.
Great. My second question relates to the cost of 28-nanometer is, many companies, fabless included, have obviously, again, concerns that perhaps 20 nm cost per transistor would see a pretty significant uptick. You have a pretty bullish forecast as well, and outlook for that. Do we need to see pricing increase either by chip companies and others to pay for this? Or do you think Moore's Law is pretty much intact at 20 nano?
Moore's Law is what?
I think Dan's question is some customers are saying that when we get to 20 nm, the cost per transistor actually is higher. Will TSMC pass on the cost to our customers?
I'm not going to answer that. I will just repeat what I said earlier in my statement, that enough discussions have taken place with enough customers who have large requirements to lead us to believe that the volume will be very large.
No, that's fine. I'm actually wondering if some of the chip companies have said they may need to increase their prices, and they're willing to try to do that. The expectations are that costs will be higher generally, I think, from your customers. I just wanted to get your perspective on that.
Again, you're asking me to say something about price, and I really am not going to say anything about price. I would just say that enough discussions have taken place with enough customers. Of course, those discussions involve price discussions too, Dan. They were not just technical discussions.
Yeah.
Yeah. Mm-hmm.
Yeah, not so much about price, just as the economics. Obviously the design ecosystem is getting very good and manufacturing is getting better. I just wondered how things are progressing. Thanks.
Thank you.
Okay. We will continue on the call. Next question will also be coming from the call. Operator, please proceed to the next caller.
The next question on the phone line today comes from the line of Randy Abrams from Credit Suisse. Randy, please go ahead.
Okay. Thanks so much. My question first is on competition. You mentioned in the prepared remarks, near 100% share on 28 nanometer last year. Could you talk about how you're viewing the effective capacity and competition on 28 nanometer as it matures this year? Whether you're seeing competitors ramp up, and if you could give an early view on your market share in the early stage of 20 nm.
I don't know what the 28 nanometer capacity is. Incidentally, when I say 28, I meant 28. I don't include the 32 nm. If you include 32 nm, which is not 28 nanometer, by the way, but if you include that, then you of course, we have a lower market share than almost 100%. The capacity will be correspondingly a lot bigger.
Okay, could you say if you're seeing more competitors on, say, 28/32 as the node matures this year? As you look at the tape-out activity on 20, do you view the same-
We don't group them together. 32 is not 28. Randy, isn't it?
Yeah. Randy.
No, I know. Okay.
Yeah. Can I say something about competitors in general, the advanced technology competitors? Well, you know who they are as well as I do.
I think every one of them is a formidable competitor. I also think that we are ready to tackle, to fight every one of them, as we have always been ready to fight every new competitor. I remember all the grilling that I went through when we had UMC, we had SMIC, then we had GlobalFoundries, and now we have Samsung, and we have Intel. Well, every one of those, I think everyone was a formidable competitor at the time. Everyone was. We take everyone very, very seriously. We also fought with everyone in the past and with the current competitors that have just appeared on the scene. We have always felt that we have a lot of strengths.
I think in the past, I have gone over our strengths before, and if you want me to repeat them, I'll be very happy to repeat them, but I'm just afraid of boring you.
Okay.
Yeah.
Okay. No, they're good. If I could follow up on the R&D investment, where you had significant growth in 2012, could you talk about whether you expect the same growth in R&D and OpEx as a % of sales, or as you accelerate your sales growth rate, do you think there's a bit of operating leverage?
We expect a great deal from our increased R&D spending, from our strengthening of the R&D, more and more strengthening. First of all, I said a little earlier that the growth engine in the next few years is going to be technology. Our growth engine is going to be technology. We have proven it now already with 28 nanometer. We are going to prove it with 20, with 16, and then a little later with the 10. Why have we so dramatically increased our R&D effort in the last few years? Number one, all right, because technology is going to be our growth engine. I have said more than once, more than two or three times, in fact, that as far as Moore's Law is concerned, if anyone is going to pursue Moore's Law to the end, we will be there.
A large part of the money is being spent on the further pursuit of Moore's Law. That, of course, also includes EUV, the investment that we have made in ASML, and it also includes the more pathfinding type of R&D to improve transistor performance. Now, a smaller part will be spent on CoWoS. Then another part, which is perhaps a bit greater than what's spent on CoWoS, will be spent on specialty technologies, embedded flash, imaging, power management. All those specialty technologies that we expect to fill our more mature capacities. See, we grow on new technology, and we stay profitable, both with the new technology and with the backfilling of the capacity by specialty technologies. I guess those are about the main uses of our R&D fund, R&D money. Those are the biggest reasons for us to expand R&D.
First, relentless pursuit of Moore's Law, all facets of it, performance as well as cost, as well as density, including EUV and transistor performance improvement. Second, CoWoS and specialty technologies. Those are the main uses.
Chairman, I think Randy's question also includes with all these increases in R&D, whether R&D as a % of revenue will go up.
No. Well, I'm not going to predict for the very distant future, but for this year, my plan now is to keep it at about 8%. As I said, the next few years will be either growth years or strong growth years. Each one of the next few years will be either a growth year or a strong growth year. Well, by strong, I mean double-digit. By growth year, well, I don't consider a 1% growth a growth year. 1-10 is a growth year, and a strong growth year is double-digits. The next few years, I expect that every one of them will be either growth or strong growth. R&D expenditure will go up, even though the percentage as a % of revenue is not going to change very much. I don't think it's going to decrease.
Okay. Thank you, Dr. Chiang.
I think we'll come back to the floor now. Our next question comes from Donald Lu from Goldman Sachs.
My first question is, if I hear correctly, Chairman, you said fabless growth is 9%, foundry is only 7%. Why foundry usually grows even faster than that?
I think it must have to do with the inventory change and all that stuff.
So maybe-
Yeah. I saw the numbers. I really had the same question. It wasn't important for me to ask. I didn't feel it was important enough to ask the guy that gave me the numbers. The guy that gave me numbers is our veteran forecaster, who is sometimes right and sometimes wrong.
Okay. It must be not very important.
I think it has to do with the inventory change, yeah.
Okay. Just follow up on that. Communication in terms of demand is growing really strong in the last few quarters. How much is your revenue from smartphone and tablet today?
How much from smartphone and tablets? Hmm?
32.
Why don't you say it?
It's about 32%.
32. Okay. My next question is more on.
You mean our total revenue, 32% is from smartphones and tablets.
Global computing devices.
Not just for the whole year.
Yes.
Last year.
2013.
In 2013, 32% were from smartphones and tablets.
Tablets.
That's including touch controller.
Right
Components. Okay. My next question is on structural profitability.
Yeah. My favorite subject.
Oh, great. The first part is the currency is moving in an adverse kind of direction. What's the impact of the currency on gross margin in Q4 and Q1, so we have an idea on an apple-to-apple basis?
You want-
Every 1% of FX change, we have a 0.4 percentage point of our margin impact.
Okay. For the whole year, would we expect structural profitability to trending in either direction?
Actually, our structural profitability this year will be slightly better. About one percentage point better than last year. 100 basis points better than last year. On the other hand, we have a couple of uncertainties, one quite large and one not so large. The quite large one is the loading of the maturity, all the technologies except 28. 28-nanometer, we are quite confident will have a utilization rate of around 100%. All the other technology lines may have a lower utilization than last year. Now, keep in mind that last year was a tough comparison. Last year, everything was loaded close to 100%. I'm not saying at all that this year our utilization will be bad, but I'm saying that last year was a tough comparison. Even though our definition of our structural profitability leaves utilization to one side.
Our structural profitability this year will be better than last year. A little less than one percentage point better than last year. If you start to talk about actual profit margin, then you have to take into account the utilization. As I said, the large uncertainty is utilization. Utilization, we have quoted numbers before.
Yes, it's 1% utilization translating to about 0.35-0.4 percentage point.
Almost like exchange rate. Every percentage less utilization means 0.35 to 0.4 percentage point less gross margin. Anyway, utilization is the greater uncertainty. Then exchange rate, of course, is uncertainty too. Here, I say it's perhaps not as big an uncertainty because, my goodness, I sure hope that the exchange rate doesn't vary all over the map. Okay.
Okay. Next question will also stay with the floor and will come from Morgan Stanley's Bill Lu.
Hi, Dr. C. Hi, Lora. Dr. C, if I hear your comments today, it seems like you are positive for the next multiple years and not just this year, whether it's the CapEx comments, whether it's saying that 20 nm is going to be bigger than 28, it all points towards several good years ahead of us. For that to happen, I think, obviously technology, you need to migrate. I think we can all feel pretty good about that. The end market also has to be growing at a much bigger, faster pace than before as well. I'm wondering if you could talk about what gives you the confidence in multiple years out, because a very different comment than just saying customer forecast are good for a couple of quarters.
Then secondly, what signs are you looking for that things aren't going to be that good?
I think, Bill, your question is what would be the growth drivers for the end market-
Yeah
right? End market.
In multiple years, I mean four years. All right? Last year, I meant five years, and now I mean four years. Okay? It's not multiple, multiple years, okay? It's four years, okay? You are concerned that the world may end in the next four years? Anyway, the economy may collapse. No, of course, we're not predicting anything like that. We are just following the conventional, consensual economic scenario, which is that the developed countries, U.S. and Europe, will grow very slowly. Well, actually, U.S. will grow sort of medium slowly, 2.5% perhaps, Europe will do worse. Japan also will do worse than the U.S. Now, of course, one never knows what the new prime minister is going to do, or one knows what he has already done, and what he has already done seems to be a bit on the positive side.
Anyway, we're assuming that, and we are also assuming that the developing economies will do quite well. China, principally, but also a lot of other countries that China now exports to. Now, all these smartphones and some tablets that China makes are not for Chinese consumption. They go to other countries. They go to other developing countries. We believe that those developing economies will do quite well. That's our macroeconomic scenario. As far as the market is concerned, the application market is concerned, we predict that the mobile products will be the dominant new products on the computer scene. I think that they will continue to eat into the traditional notebook PC market.
I think these are pretty conventional assumptions, and those are the assumptions that we're making when we say that we are going to grow quite strongly in the next 4 years now, 2013, 2014, 2015, 2016. Yeah.
I'm also wondering if there is anything that you're looking at that might change your mind.
Oh, wow. A lot of things could change my mind. A lot of things. Yeah. What do you have in mind that would change my mind? I think a lot of things could change my mind. If the U.S. economy suddenly falls apart, that would be a very big change, a pretty significant change. It won't change everything, but it would be a significant change.
Second question is maybe for Lora. We're now in, I guess, what? The fifth or sixth quarter of 28-nanometer production, and typically, you would expect to see pricing start to come down a little bit. This year, there's also a move from poly to High-k/Metal Gate. Can you help me with how I should think about pricing for 28 this year, versus how I forecasted for the previous nodes?
We look at the structural profitability other than just look at the pricing. It has to work both on the pricing and also for the cost end. I think Chairman and I just said that from this quarter, the 28-nanometer with high utilization will track to corporate level margin, and going forward, we'll keep on maintaining like that. I think on the longer term, that's what we also intend to do, to maintain a similar level or even slightly higher standard gross margin.
Phil, as you know, that we cannot talk about specific prices. All right. Thank you for the understanding. We'll go back to the call. Operator, please proceed with the next caller.
Your next question on the phone comes from the line of Andrew Lu from Barcap. Andrew, please go ahead.
Thank you. Dr. Chen and Lora, thank you for taking my question. My first question is regarding earlier, Dr. Chen mentioned the year 2014, the 20 nm volume will be higher than 12, and the 15 will be higher than 13. Is that including 16nm FinFET in year 2015 as well?
No, it does not.
Only 20 High-k/Metal Gate?
Right.
Okay. Thank you.
If you go to 2016, then I will tell you something different. Let's not go there for the time being.
Do you suggest there will be no 16nm FinFET production in year 2015?
I think it'll be relatively small.
Thank you. My second question is, the CoWoS, the 2.5D packaging. If you are charging customer TWD 5,000 per wafer, if you are doing a turnkey for customer on CoWoS, what additional value you can get from this customer basis?
Well, I think the advantage that a customer will get is that we will be in charge of everything, we will be responsible for the whole thing, till it's all packaged. If he uses another OSAT, who is going to be responsible for the yields? I think it'll be a more difficult situation. Our advice to our customer is to let us handle the whole thing. Of course, we have to be competitive in price and all that, of course, we're willing to do all that. At the end, I think that it will be advantageous for the customer to have us handle the whole thing.
Yes. Can we make some revenue back for us? For example, if we sell the customer $5,000 wafer, can we get additional $2,000, $3,000 on the whole thing back in?
Well, things are getting a little expensive, aren't they? Applications still grow, and I think that, of course, in free market economy, every price finds its equilibrium, so there needs to be enough demand and enough supply for there to be a price. That's how it works. We feel quite confident that we can find this equilibrium, and it will make it worthwhile for both the customer and for us to do this thing. Otherwise, it won't be a business, and we think it will be a business. We think it will not be a significant business until 2015, 2016. As I said earlier, even a year ago, I said that it would not be a TWD 1 billion or more business until 2015 or 2016, and I still feel that way.
Thank you.
We will continue with the call. Operator, please proceed with the next caller on the line.
The next question on the phone comes from the line of Mehdi Hosseini from SIG. Mehdi, please go ahead.
Yes. Thanks for taking my question. I have one question for Dr. Chiang and one for Laura. It was interesting to hear Intel as part of your competitors. Dr. Chiang, can you please elaborate at what technology node and what kind of end market application or segments you view Intel as becoming a more fierce competitor?
Mehdi, your question is, according to TSMC's view, at which technology node will Intel becomes a more fierce competitor to us? Is that right?
Yes.
Okay.
Did he use the word fierce?
Yes.
Intel is very fierce to me now, already, whether they are a competitor or not. At what node? Well, my goodness, I think that they are a competitor to us. Don't they have at least two foundry customers already? Is it two? Is it more? Two, yeah. They are a competitor. I do believe that Well, I think Intel said it themselves, and I believe them, that they will be very selective and that they will not go out in a general way. They are not going to be like a PSMC. I don't know whether they said it complementarily or deprecatingly. I imagine it's the latter, actually. Yeah. Anyway, I believe it. Yeah. They're not going to be like a PSMC, which means that they will be selective, and they will not be a general competitor.
As a selective competitor, I think they're already a competitor, and I think they intend to be a bigger competitor in the future. They can do that anytime they become qualified.
Got it. My follow-up has more to do with the P&L. If R&D is going to stay around 8%, how should we think about SG&A for 2015? Is that also going to stay in the 4%-5% range?
SG&A is in the 4%, roughly 4% range.
4%. Thank you.
If revenue gets higher growth, it can be maybe 3.5%-4% in that range.
Okay. One final question. Is there any way you can elaborate the number of tape-outs at 20 nm, or if you can elaborate or quantify how many customers or anything that would give us a sense how the early demand looks like?
Lora, you want to answer that question, or you don't want to answer that question?
I don't.
Up to you. I said it's up to you whether to answer the question or don't answer the question.
I will not answer this question.
He's asking how much of-
Tape-out, 20 nm.
Tape-out.
How much of CapEx?
Tape-out.
Tape-out.
Tape-out.
For 20. Okay. For 20 nm, no. I guess I would just limit myself to the comment that I already made, that we have had enough discussions with enough customers who have large requirements that lead us to believe that the volume required in 2014 and in 2015 will be bigger than the 28-nanometer volume in 2012 and 2013, respectively. Did you get that this time?
Yes.
Thank you.
Okay.
Thank you.
Now we are coming back to the floor. Any question? Okay. Next question comes from J.P. Morgan, Rick Hsu.
Yeah. Happy New Year, Chairman.
Likewise to you.
Thank you. Just one question from me. You mentioned earlier about your customer inventories. It's going to be still four days above seasonal in Q1.
Supply chain inventory.
All right. Supply chain inventory. Yeah, rather than one day below seasonal in Q1. Would that negatively affect your second quarter loading? That means your customer may exit Q1 still with excess inventory.
Oh, that's pretty smart.
Rick.
Rick, yeah. Yes. I think it means that the second quarter growth may not be quite as strong because first quarter is higher than we thought. Our scenario three months ago, six months ago, was that first quarter will be low, and second quarter, I think six months ago, I even used the word strong rebound. Did I not, Rick?
Yes.
Yeah. Three months ago, I deliberately omitted the word strong, but I said there'll be a rebound. Now I can maintain my words that I used three months ago. There will be a rebound. It would not be a strong rebound as we thought it was going to be six months ago. Six months ago, we thought first quarter would be quite low. In fact, six months ago, we thought the fourth quarter and first quarter will be quite low. It has now turned out that the fourth quarter was quite a lot stronger than we anticipated six months ago. The first quarter is also stronger than we anticipated six months ago. Second quarter, the strong rebound that I referred to six months ago in the second quarter will not be quite as strong as we thought six months ago.
All this came about because I said too much, I guess. Maybe the thing to do is not to say very much, is keep my mouth shut. I think it'll be a less interesting conference.
Thank you.
Yeah. Mm-hmm.
If there's no more questions for Oh, there is. Okay. Next one. The next question comes from HSBC, Steven Pelayo.
Just two quick questions. One, some of the longer-term concerns to follow up Bill's question. I think we are in the steepest part of this transition now to the smartphone revolution. I think maybe it's even majority smartphones are shipping today than feature phones. It's the glory days, I guess you could say. We were talking even a year or two ago about how much your dollar content increased with the silicon content increase. This is a billion unit market, maybe a two billion unit market even in a couple of years. Clearly, we are passing that threshold that's kind of now 50%. We do worry two years from now, three years from now, where maybe smartphones are-
You said a billion, two billion dollar market.
No, billion unit.
Unit.
Units.
Yeah.
We do worry, what's that next billion-plus unit market that's going to have the same kind of silicon content increase that we've enjoyed in the smartphone revolution?
I wish I knew. I wish I knew. Does that answer your question?
I don't know as well. The one follow-up I will say is that, I think the math says that your 20 nm revenues now are bigger than all of UMC. Congratulations. That formidable competitor is a distant image in the rear view mirror.
I think it's a lot bigger than yours.
I think it is a lot, actually. I do want to talk a little bit about the surrounding nodes, I guess. First going up to 40 nm, because a lot of guys are still there, and it is still more than 20% of revenues for you. Is there any more intense pricing environment or intense competition? Are people picking up second source in 40 nm here?
I think that our 40 nm market share has dropped. I know it has dropped. We anticipated that, we certainly were not going to lower our price below what we intended to lower it to. We had our price strategy, we're not going to do anything different on price just in order to keep the market share. We do try to let it drop in a controlled manner. The way we do fight to make the drop a sort of controllable one is to backfill the capacity with specialty technologies. Specialty technologies are actually migrating upwards. They are just two or three or four generations behind the latest graphics, the latest APs, and so on. Images are still at, which generation?
90. 90 nanometer.
90, yeah. They will be migrating.
Five. 65.
65, yeah. They had migrated.
0.1
0.13 earlier. That's just one example. Another example is all these microcontrollers and so on, they're still at, what?
0.18. 0.18.
0.18. They'll be migrating into 0.13, and so on. The power stuff is.
0.25, 0.18
0.25, yeah. They've been migrating into 0.18, and so on. That's the picture. That is the strategy. As I said earlier, a small part of our R&D money is being spent, and a small part is really not that small. It's quite a large group. It's of course much smaller than the Moore's Law pursuit. It's important to us to develop these, to let them migrate into the older technologies. Very important to us.
If I could just do two quick follow-ups to that. I really appreciate you guys commenting on 28-nanometer mix for the full year and corporate average gross margin. Could you just help us think if we were trying to forecast the bottoms-up gross margins for you by node? What is a 40 nm relative to your corporate average? What is a mature node relative to your corporate average gross margins? You obviously don't have to be specific, but help me understand if I'm trying to think about bottoms-up gross margin forecasting.
I will give you a general philosophy, all right? It was a philosophy that I started 27 years ago when TSMC was started. My philosophy then and my philosophy now is I want all the products we make to be within a pretty narrow gross margin range. That remains true till today. The reason I started that philosophy was because I hated to have to pick because of the gross margin. Capacity becomes tight, you make more of this because it's high margin, and that was such an unnecessary burden on the manager's mind. I didn't want that. Now I still don't want that. I don't want to have to pick between the various opportunities just because of gross margin. A simple answer to your question is that we try to keep the margins on all products within a pretty narrow range.
I would say it's a 10-point range. Well, not even that big. I would say it's a seven, eight-point range.
My last question was just on some of the expanding your addressable market into things like LED, solar, packaging. If I try to corner you and ask you, when do you think those would be 5% of revenues, would you give me a year?
Solar, did you say?
Your new market, solar, LED, and packaging. When do you think they could be 5% of total TSMC?
Oh, I'm not even counting on solar and LED. Those we have already spun out. Packaging, as I said, will be TWD 1 billion in 2015, 2016. TWD 1 billion by then is only a few %.
Thank you.
Yeah.
All right. I see there's still one caller on the line who has been waiting for a long time, operator, let's open the line to the next caller.
Your next question on the line comes from the line of Brett Simpson from Arete Research. Brett, please go ahead.
Yeah, thanks very much. Dr. Chang, you mentioned this year, wafer capacity for 28 nanometer will triple for TSMC. How do you assess the total industry supply at 28 nanometer this year for foundry? What level of increase do you see overall? What level of increase do you see overall, and what market share do you think might be achievable for TSMC at 28 nanometer this year?
All right, Brett, your question is, what is our view on the overall supply of 28 nanometers this year, and what TSMC's market share is going to be on 28 nanometer this year. Is that correct?
That's right, thanks.
Overall supply, demand?
Capacity of 28.
What is the number?
Yeah. Like, since we will be triple, what will be other people?
I don't think we give
No.
I was going to give the capacity, our CFO told me that we don't do that. Market share. Since our CFO didn't tell me not to do it, I'll venture forth and say that I said that the last year was almost close to 100%. I do expect that it will drop slightly from there. I think that this year, maybe we'll be close to 90%. Okay.
Okay, that's helpful. Can you talk a bit about the 20 nm ramp schedule? When specifically might we start to see first revenues for TSMC at 20 nm?
When will we see 20 nm revenue? When?
Yes.
When?
This year.
We are going to see part of it in 2014. Didn't I say that 2014 will be bigger than 2012?
That's right.
Yeah. 2014 will be bigger than 2012 at 28. 2012 for 28, yeah.
Okay, got it. Then maybe just final question on inventory. Dr. Chang, you mentioned this, if inventories, supply chain inventories, aren't falling as aggressively in Q1 as you first thought, how do you assess the situation beyond Q1? Do you think that the industry's inventory levels gets back to normal levels in Q2, or do you think it will take longer before this equilibrium takes place?
Well, you're asking me about second quarter, and really we aren't ready to give our second quarter guidance yet. I will say that our current view of second quarter inventory, supply chain inventory in the second quarter, is also going to be higher than our previous view. Yeah. In other words, the change from our previous view of inventory is that the previous view forecasted a pretty steep decline of inventory in the fourth quarter, in the first quarter, and in the second quarter. Our current view is that the inventory will stay a lot flatter.
Great. Thanks very much.
Okay, we are coming back to the floor. Next question comes from Goldman Sachs, Donald Lu.
My first question is on tax. We have a new tax rate this year, which is quite taxing in a way.
You are right.
There are various ways you might be able to reduce that. For example, going to a less taxing jurisdiction, like Intel always does that. For example, Intel built a fab in Israel, Ireland, Dalian, et cetera. Would TSMC consider that given the change in tax rate?
I think we may consider that not purely on the tax reason.
Okay.
Beauty of fact, outside of Taiwan, it's a big thing. You have to look into the efficiency, the cost, and the local environment, et cetera. Currently, we are not thinking just for the tax reason to go offshore. Maybe for the overall growth reason and any other reason, bigger picture reason, we may consider that.
I think Donald is thinking about Cayman Islands and so on.
No
he's not thinking of moving our factory offshore, I don't think. Right?
He's talking about that.
I think both. I mean, whatever works. That in the future years, maybe we can model the tax rate might decline again at some point.
I'm not very optimistic about that. Yeah, we are paying 14% this year. I think you just said that you expect to pay 14% next year also, right?
Yes.
Basically, the Taiwan corporate income tax rate is still not very high. I mean, 17%. The U.S. has a 35% income tax rate. Now, actually, I looked up, Morris Chang, you looked up Intel's effective tax rate for me, right?
What was it? Do you remember? Yeah.
It's like 27%. Yeah.
No. You looked up for me. It was a lot lower than that, I think. Well, nobody knows. Yeah.
27.
Was it 27% for the year?
Yeah. Good looking number.
I thought it was lower than that, but yeah.
20.
I thought it was lower, at least in the first, second. Well, anyway. It was all right. The United States has a corporate income tax rate of 35%. No. Anyway, we do have a lot of things to consider, yeah. I agree with you that the 14%, a jump of five point some points.
Six
between last year and this year, I think that's certainly not a good thing for us. Yeah.
Okay. Second question is on the treatment for ASML investment. You just said you are putting that all in the investment ready for sale.
There are two portion of that. One is for the stock, one is for R&D investment. The later part, I mean, what can you sell? You mean the part you will give ASML to help with the R&D.
Oh, that R&D support will be expenses, will be our R&D expense when it comes.
That would be expensed over certain years or?
Five years.
Five years.
Five years.
Starting this year, right?
Yeah.
Okay, that's part of the R&D expense already.
Correct.
Okay, great. Thank you.
All right. If there's no further questions, I think we'll conclude this quarter's investors conference. Thank you very much for joining us, and I hope we will see you next quarter