Taiwan Semiconductor Manufacturing Company Limited (TSM) Q1 2012 Earnings Report, Transcript and Summary

[Call Starts Abruptly]… For those participants who 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 quarterly review presentation. I would like to remind all listeners that the following 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. Information as to those factors that could cause actual results to differ materially from TSMC’s forward-looking statements may be found in TSMC’s Annual Report on Form 20-F filed with the United States Securities and Exchange Commission on April 13, 2012 and such other documents as TSMC may file with or submit to the SEC from time to time. Except as required by law, we undertake no obligation to update any forward-looking statements, whether as a result of new information, future events or otherwise. And now, I would like to turn the call over to Lora.

Thank you, Elizabeth. Good morning and good evening to everyone. Thank you for joining us in the first quarter earnings conference call. Tonight I will start with the financial results for the first quarter and give you the outlook for the second quarter followed by some comments on the supply chain inventory. The first quarter 2012 was the stronger than seasonal quarter for TSMC. Compared to a normal seasonal decline, our revenue increased 0.8% to NT$105 billion. In the U.S. dollar terms, the increase was 2.7% over last quarter, which was a slightly better than our guidance. During our last earnings conference call, we had expected the beginning of inventory replenishment in certain applications to drive first quarter demand. On top of that, the better technology also contribute to the first quarter trends. On the margin side, first quarter gross margin was 47.7%, or 3 percentage points higher than that in the fourth quarter of 2011. This is also about 3 points higher than our guidance. Thus the first quarter utilization rate was a lot higher than we had expected, reflecting a stronger demand in the first quarter and second quarter. The first quarter gross margin also include the negative impact from NT dollar’s appreciation and the temporary margin dilution effect from 28-nanometer which takes place during the initial ramping stage of every technology node. Operating margin was 33.6%, up 2.2 percentage points. Operating expense increased about NT$1 billion, mainly due to higher operating expense for Fab 15 in preparation for 28-nanometer ramp-up, as well as the increased R&D investments for 20-nanometer technology. Overall our first quarter EPS was NT$1.29. ROE for the first quarter was 20.8%. Let’s move on to revenue analysis. For applications, computer, consumer and industrial related revenue benefit from the customers’ inventory replenishment and increased by…

Hi ladies and gentlemen. Good evening or good morning, if that is the case maybe. The news in the last quarter has ranged from encouraging to exiting. Encouraging was the macro-economic news from the United States and China which are our largest end markets, both economies seen to have averted a worse possible scenario and are playing out being a medium, good scenario. Encouraging was also the world semiconductor market trend and now are looked to exceed our earlier forecast of plus 2% growth this year. Exciting was that after two quarters of sequential decline, TSMC’s business points to a solid growth year this year. In the first quarter, three months ago, we had estimated that first quarter would be flat or a little down. It turned out to be a little up. Not only that, the incoming orders were strong enough that allowed us to guide a very solid growth second quarter as Lora has just done. Not only that, the orders forecast which is our forecast of incoming orders in the next few months is also very strong. And that means that third quarter in all likelihood will be a continued growth quarter. Most exciting is that our 28-nanometer which is being ramped up just now. In fact, first quarter was the first ramp-up quarter. 28-nanometer is now turning out to be a roaring success. The success has exceeded the expectations of our partners who have designed with our 28-nanometer technology and the success, of course, has exceeded our expectations too. Let me talk more about 28-nanometer. Its performance and cost are both attractive. We are the only attractive foundry supplier. We are quite confident that we will remain the only effective foundry supplier for quite some time to come. And we are also confident that in the…

All right. This concludes our prepared statements. Operator, please open the floor to questions.

(Operator Instructions) Your first question comes from the line of Daniel Heyler of Bank of America Merrill Lynch.

I had a follow-up on the FinFET question, Dr. Chang. Today you had -- in just your opening comments you had mentioned that FinFET would be introduced after the 20-nanometer planar version. I don’t know if I am reading into this too much or not, it sounded like maybe a little bit of a different tone than previously, I believe last time you guys said that FinFET would be on mostly likely on 14-nanometer. So just to clarify that, would FinFET possibly be on a later version of 20-nanometer or is it still embedded in the 14-nanometer ramp?

Well, I am not saying that, I am not saying either Dan. I think it could still be the 14-nanometer but it can also be a later version of 20-nanometer. And we -- at this point I am not going to predict or commit either way.

And then with regard to the huge growth that you are seeing in 2Q and it seems like pretty good visibility as you said likely growth into the third quarter for this year. Maybe walk us through the sustainability of that. There is clearly a lot of smartphones getting introduced by handset companies across the world then and tablets. So beyond the product introduction in 2Q and 3Q, what are your thoughts on sustainability of growth from the fourth quarter and into next year?

I think that there may be some cyclicality in the one year, two year term. But as far as the long term is concerned, I am talking about three, four years, I am quite confident that our three or four year growth is sustainable because we are not just talking about the handheld products, the mobile products sector, although we do depend quite a bit on it. But in addition to the strong growth of the mobile product sector, which benefits us, we are also talking about leadership – technology leadership, capacity leadership in the entire global logic IC industry. Of course, I am talking about leadership of -- leadership as a foundry. But that’s getting to be more important in the whole scheme of things also now. So I am banking on leadership technology and the capacity leadership plus trust from our customers which we have had for a long time. And we plan to maintain in the future. I am banking on our leadership in those areas plus the strong growth of the mobile products.

And if I may ask just one more and then I will get back in the queue. So today you had mentioned that it’s quite difficult to predict the mix of 20-nanometer which is exceeding your expectations for the fourth quarter, I think, over 20% of revenue by fourth quarter which is a very steep ramp. But predicting the mix between the bulk version or so called high end version and then the higher performance low power version and the high-k/metal gate version, predicting the mix between those is difficult and I understand that’s somewhat customer driven. Once the high-k/metal gate kicks in, are there ASP implications for that? Is that a higher – command a higher ASP variety of the 28?

Yes, I would say it does command a higher price than the LP version.

Is it a meaningful price difference? I mean given the performance in the power envelope is I believe materially better. I am just wondering whether you would be capturing a meaningful ASP list from the high demand --

Yeah, I think I will call it a meaningful, yeah.

Your next comes from the line of Mehdi Hosseini with Susquehanna International.

Going to Dan’s question, Dr. Chang, can you please help us better understand or evaluate the 28-nanometer high-k/metal gate coming out of Taiwan versus competitors gate-last versus gate-first? We do know that gate-last is sustainable but is there anything from your end that you can offer to better help us understand the differences and how it is going to help you keep the leadership? And I have a follow-up.

Now as far as gate-last versus gate-first is concerned, that battle is over. And gate-last has won. Now I think two or three years ago when I first resumed the CEO responsibilities that was a very controversial issue. And at that time, I think I took some time to explain the investors what the technical differences were. And basically you know both have the advantages, gate-last and gate-first advantage. And both also have their disadvantages. And just to put it very simply, gate-first appeared to be easier to implement. However in the actual implementation you could run into serious difficulties. And we anticipated that perhaps better than the people who decided to use the gate-first approach. So we felt that ultimately gate-last would be the correct route. And all right. So I say the battle is over because even though some companies have continued to pursue gate-first on the 28 and 32-nanometer but they have also decided to switch to the gate-last route in their 20-nanometer. So I don’t think that is an issue any more.

Got it. And then one follow-up on the operating margin. Last time it peaked at 42% in Q4 of ’05. Given the improving blended ASPs and the challenges with increasing utility costs and everything, do you think that you can meet or exceed that 42% operating margin from Q4 of ’05?

I may ask our CFO Lora to answer that one.

Q4 of ’05 actually was a long time ago.

I am trying to understand how better blended ASP is going to help offset some of the higher costs associated with like electricity and would you be able to capitalize on it and actually be able to continue expand operating margin going forward?

Let me comment your – to Q4 2005 first and I will go to the utility cost impact to our margin. Before 2008 actually the profit sharing expense in Taiwan was not expense.

Bonuses, yeah.

Employee bonus. After 2008, it was expense. That actually has quite significant impact to our profitability. Now we have allocated about 13.5% of our net income to employee bonus. If I remember correctly, the impact to operating margin is about 5% each point. So that’s the major difference. Nowadays, utility costs will go up. In the second quarter, we anticipate the utility costs for TSMC will go up by 38% in average. That will cause the 0.5% operating margin decline in the second quarter and going forward to second half, the impact will be bigger, will be about 1 percentage point. So other than that, we also have foreign exchange impact. NT dollar has appreciated pretty much in the recent few years. If I – you remember, we have -- keep a rough estimation for every percent of exchange rate change, the impact to our gross margin was about 0.4 percentage point. So that’s another factor that’s beyond our control.

But given the trend in 28-nanometer ramp and better ASP contribution, is that – would that be enough to offset some of these headwinds on the margin front?

Well, let me take up the question now. I think you understand that you hear Lora’s explanation of the employee profit sharing which was not expensed until ’08. So the period that you mentioned, the last quarter of ’05, did you say?

Yeah.

Last quarter of ’05, the employee bonus, the employee profit sharing was not expensed. And so if you normalize the profit margin, the operating profit margin at that time to the present day situation, you would lose 5 points right away there. So you said 42%, did you not?

Yes.

So you lose 5 points by that, so it’s 37%. And Lora then went on to explain the power cost and the exchange rate and so on. So now if you are asking whether one day we’ll get back to 37% again -- operating profit 37% operating profit, my answer is yes, yes. And I would like to see that day come. Yeah, and I see no intrinsic reason why we can’t.

The next question comes from the line of Randy Abrams of Credit Suisse.

Wanted to see if I could ask on what are the factors that prompted the decision to pull in $700 million to accelerate 20-nanometer. And if you could talk about with the pull-in, would this pull the volume inflection, so it comes back to the two year cadence so that second half ’13 it would ramp up rather than early 2014.

I got the first part of the question. Let me – two year cadence, second half ’13, yeah. Well, look let me answer the first part of your question first. What were the factors that caused us to pull in the 20-nanometer, let’s say, pilot line? Actually the motivation was quite tied to the experience we have – we are having on 28-nanometers. Now in the 28-nanometer node, 28-nanometer is the first when the mobile product IC customers played a big role in the ramp-up. And they are very big users. So that right now we are having to ramp up very fast, faster than we ever have previously because previously we did not have big users that needed a lot of chip wafers at the initial stage of the ramp up. Now we anticipate that in the 20-nanometer we will have big users just as we have done now in 28-nanometer. In 20-nanometer we will have to ramp up very fast, even faster than we are doing, I think we will be faster than we are doing in the 28-nanometer. So it will be much to our advantage to shorten the learning cycle. And the $700 million capital expenditure was just pull in, we were originally going to do it sort of a bit later in next year. And now we feel that it will be to our advantage to learn faster sooner and that’s why we want to pull in this $700 million capital expenditure into this year. Now, the second, the two-year cadence – what does that mean?

Yeah, what I was implying by that is 65-nanometer ramp up second half 2007, 40-nanometer second half 2009, so 28, instead of the second half 2011, you’re getting a much faster ramp up first half of ’12. So wanted to see if that gets you back so that second half of ‘2013 we should start the real volume inflection on 20-nanometer node?

No, I don’t see the real volume of 20-nanometer in the second half of ’13 yet. I see the real volume in the first half of ’14, perhaps starting from the first quarter of ’14. So as far as the cadence is concerned, the two-year cadence or the most low 18 months, I really think that it may slow down a little bit, yeah.

Okay. And maybe the following question, wanted to see because capacity will take a few quarters to ramp up and I think you mentioned first quarter 2013 you will be fully caught up. Is the shortage opening the door of that to some multi-sourcing or more than you might have wanted? But I guess just wanted to see what you are seeing from customers that they are trying to do that or and keep up with most of the business?

Yes, it has. But at the end, we believe that those or he who has the best and the most will get to be the primarily supplier.

The next question comes from the line of Steven Pelayo of HSBC.

Question for Lora first. Just a little explanation on the March quarter results especially the gross margin upside. Revenue grew about NT$800 million, the gross profit grew about NT$3.5 billion. So a huge incremental margin there. I am struggling to understand what were the drivers behind that. It was quite the step-function improvement in gross margin than what I was looking for.

You are talking about our first quarter?

Correct.

So can you repeat your question again?

I am struggling to understand why the incremental gross margin in the first quarter was so strong, yet a little bit of revenue growth was very big gross profit growth. What were really the drivers behind that?

Okay. The driver is the utilization for the first quarter being associated with the strong second quarter because we have seen the demand were coming very strong for the second quarter. So we started to build the work-in-process in the first quarter. That helped to offset (ph) the overall utilization. So we gained utilization from that point.

I guess I am a little clear with that but wouldn’t the cost raise up with building up that WIP?

Yeah, the WIP costs go up and I explained the DOI went up by four days mainly because of the WIP.

And then a question on production, I asked earlier today about the capital intensity ratio that’s kind of 40% to 50% of revenue. It looks like you’re going to extend for extend for three years here. And I asked is that the new norm and you said, you didn’t say it necessarily that was a norm, it’s part of the cycle. But I guess I am thinking out just over the next two three years, FinFET is going to be pretty expensive and I assume there’s going to be a 450 millimeter fab investments some time in the next few years as well. Isn’t there a risk that we actually do have to sustain this kind of high level of capital intensity?

Well, look, I really can’t answer your question in one or two sentences. I mean it all comes back to having to integrate all the financial factors into the business equation. All I can do is to say that before we commit the kind of capital we must have the potential returns, meaningful in potential profit and the potential return on investment in pretty good grasp before we commit the capital. So in a strong growth period, I do not think that the 40%, 50% capital intensity per se bothers me but in the long run, I don’t really think that it’s permanently sustainable. I think that’s – you have to run huge gross margins in order to justify a 40%, 50% capital intensity. But in a short burst, and I mean, by short burst I mean a three, four year burst, I really think that as long as we have the potential returns and grasp, I really think that just 40%, 50% capital intensity per se does not bother me. And we have history to prove that. We have experience to prove that. We had experience to prove that in periods of several years we could sustain 40%, 50%, and then for again a pretty long period of time, we had only 25%, 20% capital intensity. And those happen to be low growth years too -- the 20% to 25% capital intensity years happened low growth years too. So it’s a question – I think it’s a question of seizing an opportunity, a growth opportunity when it presents itself.

And I could just sneak in two housekeeping questions here. Your tax rate is a little bit lower than I expected to lower, I am curious what the outlook is your tax rate. And then secondarily, I didn’t see that you gave any guidance on where you thought the segment’s growth rates were going to be, what was going to be the relative outperformers or underperformers in the second quarter? If I could help with those two last questions, that’s it for me.

So the tax rate – the tax rate like last year is about 7% to our net income before tax and this year, we expect it’s going to be around 8%. The reason being it’s lower than the corporate tax rate of Taiwan is because we have some tax credits. And we’re also grandfathered by the council of taxing pandits (ph), we’re doing through that and we will continue thought that for several years. So before that happens, our tax rate will not be as high as 17%. So it will be like 8% and it will really go up to 10%, 11% in the coming two years. Okay. Your second question is about segment growth for the second quarter. In the second quarter we see across the board growth for all segments. The three segments will grow above seasonal, they are industrial others, communication and the computer. But for consumer we expect the growth will be lower than the seasonal.

The next question comes from the line of Michael Chou of Deutsche Bank.

Dr. Chang, my question is that do you expect continued shortage for 28-nanometer high-k/metal gate in 2013 given the financial challenges for K2 foundries? Thank you.

20-nanometer?

28.

Do I anticipate a 28-nanometer high-k/metal gate shortage in 2013? Is that the question?

Yes.

No, I don’t.

The next question comes from the line of Satya Kumar of Credit Suisse. Satya Kumar – Credit Suisse: Dr. Chang, I was wondering if you could comment on your current process development plans for 14-nanometers. Specifically I was wondering if you could comment on whether you are starting this development with multiple patterning lithography or with EUV and by when you will sort of make up your mind on what lithography process to use?

Yes, on the 14-nanometer, it’s still a choice, two possible routes, EUV or immersion. And our equipment supplier is working very actively on both immersion and EUV. And it’s not certain yet which one will be better. It’s always a question of throughput. And immersion throughput is, there are plans to increase it pretty dramatically. Now EUV throughput has improved quite a bit just in the last six months but it’s still pretty far from being acceptable in production. Now so we’re going to wait and see. We actually – I think that the next, let’s say, 12 months or 14 months, I am thinking of the middle of 2013, the next 14 months will be quite critical in our making a choice, making a decision of which way to go. Faster immersion, or faster EUV. Satya Kumar – Credit Suisse: And a quick follow-up, you mentioned that you’re adding some extra capacity at 28-nanometer. I was wondering if you could let us know what your original plan was in wafer starts capacity at 28-nanometer and what the new plan will be at the end of this year. And you also mentioned that you expect 28-nanometer may be the biggest node that you have had. So I was wondering from a high level perspective do you expect that the rate of capacity additions at 28-nanometer to be comparable next year, compared to this year or this year is the peak year for capacity additions at 28-nanometer? Thank you.

Sorry, I missed the last sentence – missed the last sentence of your question. Do I expect the latest addition of 28-nanometer will be less than – Satya Kumar – Credit Suisse: The question was what were you previously expecting to ramp 28-nanometer capacity to and what do you now plan to ramp 28-nanometer capacity to by the end of this year? And the second part was you said 28-nanometer would be the biggest node for TSMC. So I was wondering if the amount of capacity additions on 28-nanometer is going to be higher, lower or comparable next year compared to this year?

Lora?

I will take the first question. The capacity was originally built in, and with recurring incremental capital expenditures, that will give us 10K more – and 10K per month, I mean in terms of overall capacity for 28-nanometer for this year. Your second question is talked about, will we have enough capacity for 2013? Satya Kumar – Credit Suisse: Yeah, 2013 do you expect that the capacity additions will be higher in 28 or would it be similar compared to 2012?

We will be adding a lot more 28.

Yeah we will continue adding 2013 – the 2013 will be continued growth in the demand over 28-nanometer overall.

Well, look when I said that it will be the biggest node, I am measuring it in terms of the peak or the level of the saturation output per month. And if you look at the past, we had records and it’s the best node, the highest node so far was the 65-nanometer. And I believe if my memory serves me correctly, the mountain top output was in the order of 120,000 or 130,000 per month. And when I say that 28 will be the biggest node, I mean that I expect 28 will top that.

All right. In the interest of time, I think we will only accommodate two more callers’ questions. The next one please.

The next question comes from the line of Dan Malkoun of Viking Global. Dan Malkoun – Viking Global: Hi, just a quick question on one that I saw on the capacity slide, you guys have, it looks like for Fab 14, you got wafer capacity declining in the second quarter and then ramping back up again in 3Q and 4Q. Can you just explain what’s going on there? I assume that’s just a transition to different process node?

Okay. The reason for Fab 14 capacity goes down for next quarter is because we are migrating the 65-nanometer to 28-nanometer. So we have the more 65 which can be used -- most of the equipment can be used for 28. So we are doing a migration. Dan Malkoun – Viking Global: And then just on the last comment, from March about 20-nanometer being a bigger peak than the 65-nanometer peak, what’s the timeframe when you expect to hit that? Is that this year or you expect that would happen some time next year?

I think it will be – it will possibly be 2014, I think. Dan Malkoun – Viking Global: Okay. And when you surpass it. And have you said what you guys think –

When we reach the peak up, yeah. When we surpass it. Dan Malkoun – Viking Global: When you see the 120,000 to 130,000 that you were –

Now when we surpass the 120,000, 130,000, you are right, yeah. Dan Malkoun – Viking Global: Okay, have you said what you expect your wafer capacity on 28-nanometer be in total this year, by the end of this year?

It will be ranging from 350K to 400K for the whole year. Dan Malkoun – Viking Global: For 28?

Yes. Dan Malkoun – Viking Global: 350K to 400K for total for the year. Thank you very much. Appreciate it.

Last question now.

Yes, your last question comes from the line of Szeho Ng of BNP.

For 20-nano, would TSMC still offer both 40 times (ph) and high-k/metal gate solution or just the high-k/metal gate?

For 20-nano just high-k/metal gate.

And second one, what percentage of the 28-nano equipment will be upgradable to 20-nano? Just I want to get a thought from you?

About 70% of 28 equipment is upgradable to 20-nanometer.

That’s all about it right.

70%, yes.

Okay. So thank you. This concludes our Q&A session. Thank you for joining us today. We hope you will join us again next quarter. And good bye.

Before we conclude TSMC's first quarter 2012 results webcast conference call today, please be advised that the replay of this conference call will be only be accessible through TSMC's website at www.tsmc.com. Thank you all. You may all disconnect.