台積電 ADR (TSM) 2003 Q4 法說會逐字稿

Welcome to the Taiwan Semiconductor Manufacturing Company's fourth-quarter 2003 results webcast conference call.

Today's event is chaired by Ms. Lora Ho, Vice President and Chief Financial Officer, and Dr. Rick Tsai, President and Chief Operating Officer.

Today's conference call is being webcast alive via the TSMC's website at www.TSMC.com and only in audio mode.

Your dial-ins (ph) are also in listen-only mode.

At the conclusion of the management presentation, we will be opening the floor for your questions.

At that time, further instructions will be provided as to the procedure to follow if you would like to ask any questions.

Please be advised for those participants who do not yet have a copy of the press release, you may download it from TSMC's Web site at www.TSMC.com.

Please also download the summary slides in relation to today's quarterly review presentation.

Once again, the URL is www.TSMC.com.

I would now like to turn the conference over to Dr. Elizabeth Son, TSMC's Head of Investor Relations, for the cautionary statement, before the main presentation by Ms. Ho and Dr. Tsai.

Good morning and good evening to all participants.

This is Elizabeth Son, Head of Investor Relations for TSMC.

Before we begin, I would like to state that the management's comments about TSMC's current expectations made during this conference call are forward-looking statements subject to significant risks and uncertainties, and that actual results may 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 June 23, 2003;

TSMC's registration statement on Form F3 filed with the SEC on October 21, 2003; 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 statement, whether as a result of new information, future events, or otherwise.

And now I would like to turn the conference call over to Ms. Lora Ho, our Vice President and Chief Financial Officer.

First of all, Happy Chinese New Year to everyone, and wish you many happy returns in the year of monkey.

Welcome to TSMC's fourth-quarter conference call.

We have prepared some slides for you to look at as you listen to the call.

These slides are available through the TSMC Web site.

During the call, I will refer to these slides by their individual numbers.

To start with, I would like to share with you some of the highlights of our fourth quarter 2003 operating results.

These figures are shown on slide number 4.

As announced on January 8, our fourth-quarter total revenue amounted to NT$57.8 billion dollars, which is a record high for TSMC.

Net income after-tax totaled NT$16 billion, also the highest since the fourth quarter of year 2000.

Earnings-per-share reached NT$0.79.

TSMC ended the fourth quarter with NT$110.8 billion in cash and short-term investments.

Our total long-term debt has remained unchanged at NT$35 billion.

During the fourth quarter, we shipped a total of 1,427,000 8-inch equivalent wafers.

Overall utilization rate, which included TSMC Tech (ph), WaferTech, and SSMC was roughly 101 percent in the fourth quarter.

I am also glad to tell you that our fourth-quarter ROE has reached 19.9 percent, which is very close to our long-term target of 20 percent.

Now let's go to slide number 5.

Compared with the third quarter 2003, revenue increased by 5 percent sequentially.

This was a combined result of a 13.5 percent increase in wafer shipments, a 5.5 percent decrease in ASP, and a slightly stronger NT dollar against the U.S. dollar during the quarter.

Overall utilization increased from 98 percent in quarter three 2000 to 101 percent in the fourth quarter.

As a result, gross margin improved slightly from 39.1 percent to 39.3 percent.

Operating expenses as a percentage of our net sales was 11 percent for the quarter, up from 9 percent in previous quarters.

This is (indiscernible) to a onetime non-recurring R&D expense associated with the 65 nanometers evolvement.

Next, nonoperating expenses increased from NT$0.14 billion to NT$0.19 billion.

Investment income, on the other hand, saw a significant improvement from NT$0.36 billion in third quarter to NT$1.14 billion in the fourth quarter.

The reasons were firstly, capital gains realized from selling certain equity shares from our other investment were higher than those recorded in previous quarters.

Secondly, operating results from SSMC, WaferTech, and Vanguard continues to improve, with WaferTech operating income more than double during the fourth quarter.

Income tax expenses increased from NT$1.54 billion in the third quarter to NT$1.57 billion in the fourth quarter, the result of a higher pretax income.

Net income for the quarter was NT$16 billion.

Net margin remained at 28 percent, the same level as the third quarter, demonstrating yet another quarter of TSMC's strong performance.

Let us now turn to slide number 6.

On a year-over-year basis, our revenue increased by 40 percent, primarily due to a 65 percent wafer shipment increase, partially weighted down by ASP declines of 8 percent.

Gross profit was NT$22.7 billion, a 113 percent increase over fourth quarter last year.

Meanwhile, our gross margin improved from 26 percent to 39 percent.

Total operating expenses, although increased 21 percent from the same quarter last year, took out only 11 percent of the total sales, down from 12 percent a year ago.

Our nonoperating expenses for the quarter totaled NT$0.2 billion, substantially lower than the NT$0.4 billion in the same quarter last year, mainly as a result of low interest expense after TSMC repaid NT$9 billion quarterly bonds.

TSMC had a net investment of NT$1.1 billion versus a net investment loss of NT$2.2 billion a year ago.

This is mainly due to a significantly improved operating result from WaferTech, Vanguard, and SSMC, and the realization of substantial capital gains from investment.

Pretax income and net income were NT$17.6 billion and NT$16 billion separately for the quarter, representing a year-over-year increase of 471 percent and 527 percent respectively.

Earnings-per-share was NT$0.79 cents, more than six times the EPS achieved from the same period last year.

If you see our 2002 and 2003 full-year summary in slide number 7, on a full-year basis, our revenue increased by 25 percent, the wafer shipment increased by 38 percent, and ASP declined by 7 percent.

As a result, gross profit was NT$72.9 billion, a 40 percent increase over the NT$52 billion reached a year ago.

Subsequently, gross margin percentage improved from 32 percent to 36 percent.

Total operating expenses, although increased 14 percent from last year, represented only 10 percent of the total sales, down from 11 percent a year ago.

As for nonoperating expenses, our nonoperating expenses totaled $NT2.4 billion, higher than the 1.4 billion in the last year, mainly due to a write-off of certain (indiscernible) facilities.

That amounted to NT$1.2 billion, partially offset by savings made from lower interest payments from repaying of NT$9 million of bonds.

As to the net investment income, TSMC's (indiscernible) net investment income of NT$0.8 billion versus an investment loss of NT$5.7 billion a year ago.

This is mainly due to all WaferTech, Vanguard and SSMC turning profitable in 2003 and the realization of substantial capital gains from the investments.

Pretax income and net income were NT$51 billion and NT$47 billion (ph) separately, representing a year-over-year increase of 88 percent and 119 percent respectively.

Consequently, I am pleased to report EPS for year 2003 has come to NT$2.23, more than double the EPS of 2002 of NT$1.05.

Slide number 8 will provide you with the best (ph) shot of TSMC's balance sheet at the end of fourth quarter and some key financial ratios.

TSMC continued to maintain a very healthy balance sheet, as we ended the quarter with cash and short-term investments of NT$110.8 billion, representing a little over one-quarter of our total assets.

Our accounts receivable and our inventories all remained very healthy levels.

Despite increase in total sales, accounts receivable turnover days improved to 40 days, down from 41 days in the previous quarter, while total accounts receivable increased slightly to NT$25.8 (ph) billion due to higher sales in the quarter.

As for inventory, TSM ended the quarter with NT$10.9 billion in inventory, lower than that of the third quarter.

The entry (ph) turnover days improved to 36 days.

TSMC has remained virtually debt free as our total cash and cash equivalents far exceeded our total liabilities.

Slide number 9 summarizes TSMC's cash flow for the quarter.

As you can see, TSMC continues to generate strong cash flow from operations.

During our fourth quarter, cash flow generated from operating activities totaled NT$31.5 billion and cash used in divestment activities was NT$15.9 billion during the same quarter.

This includes capital expenditure during the quarter that totaled NT$12.8 billion and cash used to purchase short-term investments totaled NT$0.9 billion, mainly in government bonds and money market funds.

As a sidenote here, CAPEX for the full year 2003 was US$1.1 billion.

Now let us look at some business analysis.

Slide number 10 provides a quick look at our revenue breakdown by technology.

The top portion of the bar represents revenue percentage from 0.13 Micron products, which for this quarter accounted for roughly 18 percent of our total wafer sales, maintained a similar weighting in quarter three, with revenue actually exceeding that of quarter three.

Although demand for advanced technology has increased in fourth quarter, the increased demand for mature technology surpassed the increase for the advanced technology.

As a result, revenue from advanced technologies, which includes 0.13, 0.15, and 0.18 microns, accounted for approximately 63 percent of our total wafer sales, down from 66 percent in the previous quarter.

Slide number 11 provides a view by application.

For the fourth quarter, revenue from communications segment represented 42 percent of our total sales, a 6 percent increase from the last quarter.

Revenue from computers accounted for 35 percent of our total sales, slightly down from 36 percent in the previous quarter.

Revenue from consumer segment represented 17 percent of our total sales, also down from 21 percent in the third quarter.

We saw robust growth in the communications sector during this quarter, mainly from some strong end-market demand for mobile phones and wireless LAN, reflecting increasing demand and recovery of the economy.

Slide number 12 shows the geographical breakdown of our revenue.

Revenue from North America, Asia-Pacific, Europe, and Japan now accounts for 75 percent, 12 percent, 5 percent and 8 percent of our total revenue of the fourth quarter respectively, in line with the trend of previous quarters.

Slide number 13 provides a revenue breakdown by customer type.

In the fourth quarter, revenue generated from fabless customers accounted for 71 percent of total sales, 2 percent down from quarter three, (indiscernible) IDM customers account for 29 percent of our total sales, 3 percent higher than that of Q3 levels.

This is very much in line with our normal pattern, which is roughly 70 percent fabless and 30 percent IDM.

Now let's take a look at slide number 14.

This slide shows our fab utilization and ASP trends.

Due to much higher demand for mature technologies, our ASP declined by 5.5 percent from the previous quarter, while our utilization improved significantly to 101 percent from 90 percent in third quarter.

Let's turn to slide number 15.

This slide summarizes our in-store capacity by fab.

Our total capacity during the fourth quarter was 1,067,000 8-inch equivalent wafers, 5 percent higher than 1,012,000 8-inch equivalent wafers in the third quarter and in line with our third quarter guidance.

Currently, we expect our capacity for the first quarter of 2004 to be roughly about the same.

We will continue to increase our capacity for advanced technologies in order to meet our customers' demands.

Meanwhile, some of our 8-inch fabs, such as Fab 2, Fab 5 and Fab 8, will go through their annual maintenance.

Therefore, there will be a small reduction in the in-store capacity in first quarter 2004.

Overall, our capacity for the whole year 2004 is expected to be about 15 percent higher than that we had in (indiscernible) 2003.

The last slide, number 16, provides a recap of the major events during the fourth quarter.

I will leave that for your own reference.

Finally, let's briefly go over our guidance for the first quarter 2004.

Coming off a very strong fourth quarter, our first quarter 2004 guidance is as follows.

Wafer shipments to increase by a low single digit percentage point sequentially.

ASP to decline slightly, but magnitude is less than that of the previous quarter.

Overall utilization rate to be about 100 percent or slightly higher.

Gross profit margin percentage to remain essentially unchanged sequentially.

Demand will improve in the communications segment, remain flat for the consumer segment, but experience seasonal decline in the computer segment. 2004 CAPEX expected to be about US$2 billion.

This ends our presentation today, but before we open for Q&A, I would like to mention that TSMC on December 19, 2003 has filed a lawsuit against SMIC.

The purpose of filing this lawsuit is to protect our intellectual property and our trade secrets.

Currently, the lawsuit has entered litigation stage.

Therefore, we will not make any comments on this matter at all in today's conference call.

I'm sure you understand our position and respect our decision.

Thank you.

Thank you.

At this time, we will open the floor for questions. (OPERATOR INSTRUCTIONS) Kalpesh Kapadia.

Our next question comes from Matt Gable (ph).

Calypso (ph) Capital.

Congratulations on a nice quarter.

Two quick questions.

What are wafer starts doing right now?

Are they showing a slightly increasing trend, a flat trend?

Question number two, how were 0.13 ASPs sequentially in the fourth quarter of 2003?

What was the sequential performance there?

Wafer starts continue to increase because the loading (ph) picture is quite good.

As to the -- we don't -- as to the ASPs for 0.13 micron technology, sorry, but we do not comment on the ASP of a specific technology.

We do give you the overall ASPs.

Okay, thank you very much.

Ali Irani.

Congratulations on a great quarter and outlook.

I was hoping you could give us an idea of your technology roadmap between the 200 and 300 millimeter fabs, and in particular am wondering if you will maintain retooling at the old (indiscernible) down into the 65 nanometer node.

And also in the same line of thinking, if you could give us an update on the timing of your China fab?

Good question about the technology roadmap for 200 millimeter and 300 millimeter.

Basically, for 200 millimeter 8-inch wafer fab, I think our technology will go to down to 0.13 micron, but pretty much stay there.

We will do our 90-nanometer technology all in our 12-inch wafer fabs and beyond.

Of course 12-inch we will also do in 0.13 and 0.11.

So it sounds like some of the new exotic technologies that are being offered to extend 200 millimeter fabs down to as far as 65 nanometer, you don't feel that that is the most cost-effective way to reach your customers' needs?

Yes, we agree with that comment.

Basically, I think it would be quite difficult actually to get even a good supply of 8-inch equipment, for 65-nanometer technology generation.

I think 90-nanometer, there may be (indiscernible) of supply from 8-inch, but 65 will be unlikely.

As to our China fab, your second question, I think we are basically moving according to our announced schedule before.

The construction is moving well.

We plan to move in equipment sometime by midyear and to have early production by the end of 2004.

Great.

Thank you very much and congratulations on your strong results.

Michael McConnell.

Pacific Crest Securities.

Looking at your gross margins, both fourth quarter and your guidance for Q1, quite impressive that with the pricing declines you're still able to exceed gross margins and look like you're guiding for flat gross margins for Q1.

Should we think of this as a function of higher utilizations as well as some yield improvements in 0.13 micron?

Is that the reason for the gross margin strength?

Yes, I think definitely the new utilization play a major role on the margin improvements.

So you can see our fourth quarter has reached to over 100 percent.

That is a very important factor.

And one other question.

You usually provide some strength by products.

Could you talk about where you saw strength in communications and PCs by product, please, as well as consumer electronics?

Very quickly, wireless mobile phones remained the strongest.

However, wireless LAN is also doing quite well from the wireless domain.

Wireline, I think the ethernet (indiscernible), the ADSL, Blue Tooth, (indiscernible) very well actually in the wireless space.

And has the wireless strength continued into Q1?

I think basically communication will continue -- as we said in the guidance -- communication will continue improving in the first quarter.

I think both segments, wireless and wireline, are going to do well.

Okay, thank you very much.

Prudential.

Good evening and good morning.

Two questions.

We are looking at the gross margin number and TSMC without (ph) affiliate is doing gross margin of 44 percent and in aggregate gross margin is 39 percent.

Should we be looking that as the pricing comes back that your gross margin should go north of here?

Then I have one more.

Yes, definitely.

And the second question for Rick, I believe Morris on the afternoon conference call talked about maybe semiconductor industry growth of 25 percent or so and foundry industry likely to grow higher.

So assuming a 30 percent '04 growth for you and given you're adding capacity by 15 percent, then let's assume that utilization rates are somewhere like 103 percent or whatnot.

That would suggest that actually ASPs have to go up somewhere like 6 to 8 percent range.

Is that consistent or will that be your view overall for the year?

We cannot really comment on the overall growth for the whole year for (indiscernible).

We do of course maintain our statement that the ASP environment has stabilized and will improve starting second quarter and our product mix (ph) will continue improving also throughout the year.

So we look forward to a good 2004.

Thank you.

Mark FitzGerald.

Banc of America Securities.

I had two questions.

The first was could you give us a quick explanation of the R&D charge, the one-time 65-nanometer, and dollar value for that so we could adjust forward for the first quarter here?

Okay, we had a one-time non-recurring R&D expense in the fourth quarter of 2003, and maybe it was around the 25 million range.

Okay, and that basically would come out of any sort of first-quarter number?

A better way to ask it, do you have guidance for R&D for the first quarter?

We don't guide R&D for the first quarter, but what I can tell you is when markets start to grow, normally our R&D expense in terms of absolute dollar value will also grow slightly.

Okay, but --

Our Chairman said in the afternoon that we manage our R&D expense to be around 50 (ph) percent of our sales in the past, and we believe that ratio will probably be somewhat lower going forward.

Yes, that's right.

Okay, just a quick follow-on.

In the past conference call, you commented that the 0.13 micron designs had not been really coming out of your customers very aggressively.

Have you seen any pickup in those releases of designs to you at this point?

That 0.13 design tapeout is doing quite well.

Some of the numbers that we mentioned in the afternoon was (indiscernible) the customers would -- it's actually more than what -- the number of 0.13 micron users are around 50 plus, but also it's around 20, 20 plus of them are having the major volume.

So we are doing quite well in 0.13 micron because the people -- the 0.13 micron is getting quite mature from a manufacturing point of view in TSMC, and the customers starting (indiscernible) enjoy the cost benefits from 0.13 compared to 0.18.

So we are quite positive and optimistic about the 0.13 micron of the TSMC offering.

Would that suggest as a percentage of the mix that the trend would reverse here in terms of it declining as a percent of the mix in the first quarter?

I mean the advanced technology (indiscernible) will improve.

Okay, thank you.

Ben Lynch.

This is Ben Lynch from Deutsche Bank.

I have two questions please.

The fact that the lagging edge proportion took up bigger share of the mix in Q4, is not indicative of end demand or just maybe you have less leading-edge technology availability?

Had you had sufficient leading-edge technology do you think the mix would have still been the same?

By the way we don't call them lagging. (multiple speakers) Basically the end market demand for those mature technology products is very strong.

I don't think we are alone in observing that trend for the short term.

I think we are growing in both advanced technology and mature technology in volume, but the percentagewise mature technology demand is growing faster during the past quarter.

So it is not just that you did not have sufficient leading-edge technology available?

No, that is not the case.

Just the follow up question, do you think that maybe that tight utilization rate you had perhaps led some business to go elsewhere?

And if so, how much business could you have generated had you had a lower utilization rate than the 101 percent?

I think so far we have managed to increase our capacity both advanced and I don't know whether we made that clear that we had also increased our capacity in the mature area.

TSMC's objectives always are of course to satisfy our customer's requirement and to keep our market share in the up cycle also.

So you don't feel you are maybe foregoing some business to the people with lower utilization rates at the moment?

No, we are not seeing that.

Thank you.

Timothy Arcuri.

This is actually Dan Barramont (ph) for Tim from Deutsche Bank.

Follow up on utilization actually, what are your plans for that and how long would you be able to sustain it at greater than 100 percent?

Are there any plans to attempt to bring that down to normalized levels, are you happy keeping it up there?

We expect to do a (indiscernible) in the utilization in 2004 and we have high confidence in our manufacturing operations' capability to do better than 100 percent utilization if the demand is there.

You talked about 130 nanometer tapeout activity being good.

Can you maybe give us some historical perspective on that?

How did the number of tapeouts received for the 130 nanometer, which as you mentioned is becoming mature manufacturing technology for you, how does that relate to previous up cycles?

And you seeing more designs?

Fewer designs for the mature technology?

I have not really looked at the data in detail myself, but just off the top of my head I would say it is fairly close to 0.18 micron generation. 0.15 has been kind of a shrink of 0.18, did not have as many tapeouts.

And so 0.13 I would say follows a similar pattern as that of the 0.18, and we expect that to continue to be strong for quite a long time, because we're still seeing strong demand for our 0.18 micron technology even now.

Great, thank you.

John Pitzer.

Credit Suisse First Boston.

Just a quick question.

When you look at the planned capacity expansion for this year, what percent is coming at 200 versus 300?

And I guess as a follow up, can you help me understand how the design activity at 300 has been?

When do you think most customers will reach a cost crossover and what kind of tapeout or design activity have you seen at 300 to date?

In terms of the capacity expansion, as I just mentioned, the overall year-over-year is the 15 percent increase.

Of that number, I think the 12-inch, the capacity will exceeding 130 percent -- more than double.

For the 8-inch, we expect will be like 7 or 8 percent increase in 2004. (multiple speakers) the crossover items --

And design activity that you are seeing at 300 right now?

We will see the crossover for 200 millimeter and 300 millimeter will happen in our late second quarter this year, in terms of die cost (ph).

(indiscernible) for 300mm, I all the major customers are designing into 300 millimeters.

And the new tapeout for 0.13 micron are also -- the majority of the new tapeouts for 0.13 micron also go to 300 millimeter.

Great, thank you.

Christina Osmena.

Needham & Company.

A couple of questions here.

I was hoping you might reconcile something for me.

You said your ASPs had stabilized and that the mix shift in the first quarter should improve and shift back towards the more advanced technologies, yet the ASPs are declining.

So maybe if you could explain to me what is causing the decline if it is not a mix shift?

I think what we meant is the ASP environment is stabilizing in the first quarter of this year, but somewhat the ripple effect on last year's continued into first quarter.

That is why we're seeing still some slight ASP decline in first quarter.

So our comment, I think, probably can be applied to times after first quarter.

Okay, so we should expect first quarter's mix to include a higher percentage of advanced technologies, then?

Yes, I think you can say that.

A couple of questions on your CAPEX, then.

Is that 2 billion going to go primarily into equipment or is there some spending on land and buildings?

And also, if you could give us what the mix is between 90-nanometer and 0.13?

Almost 2 billion CAPEX.

I think close to 90 percent were associated with a capacity increase, of which 75 percent goes to 12-inch and the rest, around 10 percent, goes to 8-inch, and the other 10, 20 percent -- 15 percent is related to supporting area.

Okay, and the mix between 130 and 90?

We usually don't separate CAPEX that way.

Sometimes the machine is combined, used for both technologies.

Okay, and also the CAPEX for your JVs, do you or do you not include that in the CAPEX numbers that you give us?

And if you do not, I understand WaferTech is doing better, SSMC has some increased plans.

Could you give us what the JV CAPEXes are?

This 2 billion includes our JVs.

WaferTech around 100 million range.

Could you give us an update on how your mask shop is doing?

I know you don't break it out, but maybe a sense of how it might have done quarter-over-quarter and what you are expecting for Q1?

The (indiscernible) businesses to support our wafer business actually fluctuate quarter by quarter.

No, I cannot give you an absolute projection on that one.

Okay.

And finally, could you make some comments on when you plan on rolling out SOI, or if that is on your roadmap?

It is (indiscernible) we have an active program on SOI, and we -- I think at the 65-nanometer node, we definitely will have something.

But before that, we will have to depend on the market situation.

Okay, thank you.

Ashish Kumar (ph).

You do comment on book-to-bill.

Can you tell us how it has been sort of behaving over the last few months?

Well, we have not shown the data specifically, but for the last few months it has been quite good, definitely above one, with good margin, shall we say.

Sure.

And my second question is that when Mr. Morris Chang spoke about 26 percent (indiscernible) growth rate in '04, any assumption on U.S. dollar versus some machinery (ph) that you assume?

Can you repeat your question?

You know the 2003 of saving invested growth rate benefited meaningfully from U.S. dollar depreciation against the various major currencies.

There is a lot of fear of positive (ph) position of U.S. dollar.

Mr. Morris Change, I think (indiscernible) did say that the 26 percent saving in (indiscernible) revenue growth is in U.S. dollars.

I was wondering whether somebody had an assumption of such positive depreciation or you have just taken the current foreign exchange rate?

I think the 26 percent primarily based on U.S. dollars.

However, of course, we are not the economists, so we cannot really judge whether the U.S. dollar will continue declining in the year to come.

But I am sure with the current U.S. dollar exchange rate you can make your own assumptions and get your own numbers pretty easily.

So if I understood correctly, your 26 percent forecast has used the current rate, so there is no benefit for the depreciation of the U.S. dollar in there?

I think you can say that.

Okay, thank you.

Amri Tamise (ph).

I was just calling to learn about your depreciation forecast for next year.

Depreciation for the year 2003 is approximately NT$60 billion.

We expect the depreciation will only be increased slightly in 2004, as we had increasing CAPEX because a lot of our (indiscernible) had expired depreciation.

Okay, and also what is your planned 0.11 micron ramp for 2004?

At what rate do you plan to exit the year?

We plan to have production output starting second quarter of 2004.

I do not think we give year-end specific technology output in the conference call, but we do have, of course, high hopes for and high confidence in having a good production resulting from our technology.

Okay, thank you.

Michael McConnell.

Pacific Crest Securities.

A quick follow-up.

There has been a couple instances in the past couple quarters where you have seen some of your customers shift products to your competitor, given the fact that you do not offer a current SOI process.

Could you give us an update with your plans regarding SOI and if you are in development right now of SOI?

Yes, the Chairman actually elaborated in the afternoon that we -- it is true that we have not engaged in very high-speed microprocessor equipment of the semiconductor units we have (indiscernible).

Basically the market size is quite limited and we have been focusing on much higher volume technology note (ph), so that we can have that much better return for our shareholders.

However, we also recognize that the performance requirement for the advanced technology (indiscernible) is getting higher.

So -- and as I said earlier that we do have active programs on SOI and we definitely expect to do something by the 65-nanometer node.

I think with microprocessor equipment, for the very high-end microprocessor, we need to really evaluate the market size and the profitability.

However, for the more -- (indiscernible) computer type of microprocessor will be a major player in that, with our current roadmap and our current transistor performance improvements (ph).

So you believe SOI will be coming potentially forward or you'll be offering it with 65 nanometer, is that correct?

Did I understand you right?

It will be, I would say, a profitable technology by that time.

However, I am not -- we don't believe that will be a mainstream field (ph).

Thank you.

Bhavin Shah.

J.P. Morgan.

My first question is on capital spending.

Could you put a range on that in the sense that what would be (indiscernible) in which you'll exceed that guidance or maybe come below that?

Right now, we don't have a range.

However, if you also understand that -- on the CAPEX, especially on the CAPEX (indiscernible) of the year, will impact mostly the (indiscernible) for next year.

So really, the important factor for how we look at next year's semiconductor industry growth.

As the Chairman said earlier today, that we are looking at it right now, currently, which we believe will change as we go ahead.

But right now we're looking at a 10 percent growth rate for the semiconductor industry next year and our CAPEX scenario I think reflects that belief for now.

I see, okay.

I had a question on the ASP difference between 300 millimeter and 200 millimeter for like-to-like technology.

What is the difference and is that a factor to suppress ASP going forward?

No, Bhavin, I don't think so.

We definitely have priced our 300 millimeter wafers in such a way that we can have similar or better margins compared to our 8-inch wafers.

However, with that, we can also share -- give our customers a better cost benefit.

So I don't think that we're going to have a depressing effect for ASPs.

I believe in the Taiwan meeting, Morris Chang might have commented that 0.13 could be as high as one-third of revenues by end of the year.

I was wondering if you can confirm that?

And secondly, if you were -- do you (indiscernible) any indication whatever of where 90 nanometer and 110 nanometer nodes could be by the end of the year, or when we would expect meaningful revenues from those nodes -- meaningful meaning, let's say, 1 percent of your revenues.

I think Chairman's comment on one-third, I think that we have a pretty good -- very high confidence in achieving.

Of course, that 0.13, I might mention, a 0.13 micron also includes 0.11.

We kind of put them together, 0.13/0.11 micron.

I think that how you should look at that number.

Okay.

And any thoughts on when you can expect meaningful revenues from 90 nanometer?

By the end of this year -- is that the question?

We will see early production, but I do not expect to have a major impact in our revenue.

Right, okay.

A couple of last questions.

One is, if the decision has been made on the dividends?

We are seriously considered to have some cash dividend we would propose to the Board meeting, but it is up to the Board to decide.

We have not have had Board meeting yet, so we cannot comment on how much will be our cash dividend.

Thanks, Laura.

My last question is on any change in customer behavior you might have noticed following the legal activity in December, either on the side of customers holding back some business to you or maybe expressing their reactions, because effectively you are seeking to stop shipment of their products made elsewhere.

Bhavin, neither Laura nor Rick will answer that question.

I got that, thank you.

Ivan Goh.

Dresdner, Kleinwort, Wasserstein.

Just have two questions here.

This one is regards to your 200 millimeter expansion plans in 2004.

Can you just give an idea of where the capacity is coming up?

That is my first question.

Mostly will be from 8-inch fabs.

We are continuing buying some equipment to have incremental increase in the 0.18, 0.15 technology, so we are doing that in several fabs -- WaferTech at (indiscernible) Fab 6 and Fab 8.

Does any of the CAPEX actually include portions for your Chinese fab?

And how should we think of this?

Were (ph) you going to move all equipment through China, will that free up space in fabs in Taiwan, then you can add in more equipment?

The Chinese fab we will ship the used (ph) equipment from Taiwan as a starting point, so there will be no major additional CAPEX for our Chinese fabs.

But would that free up space for future expansion of 8-inch capacity in Taiwan?

You are right.

We will have some free space; if we believe the demand is strong, we may consider it.

For right now the current plan is to (indiscernible).

And the second question is regarding the consumer area.

Can you provide some idea what is going to carry the consumer segment revenues into the first quarter?

I see that you have it for flattish trends.

Perhaps an idea of what is strong in that area?

Thank you.

As usual, DVD players will be a major product, along with the digital cameras, games (ph) and also game consoles.

I think those are probably the top revenue for us from the consumer segment.

Thank you very much.

David Wong.

Morgan Stanley.

I have a quick question regarding mix of high-end products.

I heard on the call before that we may expect this to (indiscernible) this quarter.

I was just wondering what magnitude it will increase (indiscernible) fourth quarter.

And under such a scenario, (indiscernible) previously mentioned ASP declines (technical difficulty)

David, your voice breaks.

Can you repeat your questions?

Sure, is this better?

Yes, that's much better.

Okay, I was just wondering, I heard on the call before that we may expect increased mix of high-end products in the first quarter.

I was wondering to what magnitude we expect this increase to be versus fourth quarter.

And then also under such a scenario, if this is going to help in terms of keeping ASPs flat or maybe even boosting it?

(indiscernible) in the first quarter we will increase just small scale, but after second quarter the increase will be more significant.

Okay.

That is actually the only question I have.

Thank you.

Daniel Heyler.

Merrill Lynch.

Wanted to get a sense of how lead times were developing from, say, December, because November was down and December was strong.

Where have lead times gone from December now into January?

I don't have that number with me.

I don't think it is getting shorter, obviously, but on the other hand I don't think it is getting too much longer either.

I think we're managing that pretty well.

Our fabs have done actually a wonderful job in maintaining and sometimes even continually approving the cycle time, even under this kind of loading situation.

(multiple speakers) before loading, is it roughly averaging two to three weeks or is it stretching to more weeks?

You mean two to three or stretching to what?

Has it stretched beyond two weeks?

Two weeks sounds about right, but it's definitely not four weeks, for instance.

Okay, great.

And from the standpoint, I guess, how previous upturns have happened, are you sensing that some customers are having to start to expedite some orders, that you are seeing some rush orders?

One large IDM indicated today that they had to significantly expedite some orders to their foundry partners.

I'm wondering if you're starting to see more of that?

We are seeing quite a few -- large or small, IDM or fabless, but we are rushing some orders.

But I tell you, Dan, the customers feel -- I think in the bad time we have many customers rushing orders also.

I think that is a habit -- now is habit.

Right, okay.

And then you had indicated 8-inch capacity roughly -- I think Laura had indicated 7 to 8 percent increase for 2004.

Is that right?

Yes, for the whole year.

For the whole year, right?

That is not December to December, that is the full year over the full year, right?

Yes.

Where is that capacity going to be added?

Which fabs?

We just mentioned it -- Fab 6, Fab 3, and Fab 8, most of the 8 inch fabs will have -- WaferTech of course.

And those fabs have different technology nodes, so is it fair to assume that you're adding across various nodes, or will it be 0.18 and 0.13 mostly?

The 8-inch mostly at 0.18, 0.15.

And so not a whole lot on Fab 6.

So is fair to assume you're moving some of the 0.13 guys from Fab 6 into Fab 12?

We are, because Fab 6 is getting very full.

We are also -- we are squeezing a few pieces of equipment into Fab 6 also to also have some more capacity in Fab 6 for 0.13.

Okay, great.

Given the massive inventory swings that you have been through, Morris has talked about this to some extent, where customers have given really bullish forecasts.

What do you think you have learned going into this upturn?

Would you tend to take forecasts with more of a grain of salt or would you be more inclined to try to build to what customers are telling you?

And if anything, has the ability to scrutinize customer base improved at all or are we still dependent on customer forecast?

How do you think about that now?

Dan, as you know, it is a complicated process.

Of course we look at customers' demand forecasts seriously.

On the other hand, we have a major marketing function which looks into (indiscernible) from the end-market demand point of view, look at the system units and the component units.

And we try to do some sanity (ph) check, and so we come up with certain demand, shall we say demand portfolio from various customers, and then we work with them accordingly.

So it is not purely from their demand forecasts.

Thanks.

Ashish Kumar.

Could you throw some light that when you ship customers (indiscernible) and there is cost advantage, and assuming stable yields, there is about 50 (ph) percent cost advantage broadly.

How do you share that with your customers?

I don't think that is a question that we can give you a specific answer, but as Chairman said earlier, we do share some of those gains with our customers (indiscernible) and we need to our margin or better margin from all of those great efforts and the capital investment.

And once yields have (ph) stabilized, would the margin on the 12-inch be meaningfully higher than margin on the 8-inch?

I think it is fair to say that we try to maintain a similar or better margin at 12-inch compared to 8.

And one more question.

As capacity utilization rates are so (indiscernible) and demand has been good and you are the top tier foundry, logistically I was thinking how do you shut down some capacity and start shipping into China?

How are you thinking about this issue?

We have been planning on this for a period of time, since I think early to mid last year.

We have been working with our customers through the time so the customers knew that coming quite some time ago, and they have been working with us for either to move their product to other fabs or to have the last (indiscernible).

So we believe -- and of course, we try to suspend the occurring fab operations as long as we can, so that the customers' demand can be met and we can make as much money as possible.

And if I may ask, is it possible that the quarter you start shipping there could be one quarter in '04 where sequentially there would not be any capacity increase despite our building capacity in Taiwan, because some capacity in that quarter gets taken out and will finally, eventually come up there in China?

It won't be a one-shot transfer of equipment.

It is a gradual process.

So gradually slow down, not one shot.

And the fab from which we will take the capacity to China, total -- one of our smaller, actually the smallest fab, so the capacity reduction is not diluted (ph).

We definitely will not see a capacity reduction from quarter-to-quarter because of that.

Thank you.

A quick follow-up question.

When you look at your CAPEX budget for 2004, lead times on tools are beginning to stretch out.

Are there any concerns on availability of equipment?

I guess another way to ask the question, when you look at the $2 billion, what percent is equipment that you have already taken physical delivery of and have not paid for, and what percent have you already ordered with your suppliers?

Thank you.

I don't think we have the breakup of points in CAPEX by quarter at hand, but we (indiscernible) of course.

As we said last time that we're going to start shipping -- moving equipment for our Fab 14 in the March/April time this year, which is two months away.

So obviously, we have all those tools booked and confirmed in their delivery.

And that is a big chunk of our capital expenditure.

I think for Fab 12 equipment, that should probably be (ph) coming longer -- in the second quarter and early third quarter time also. (indiscernible) has been -- we've been working with our suppliers early and they have been very cooperative with us.

And we are getting our share of tools, I can assure you of that.

So no major concern that tool availability is going to dampen the ramp in your capacity plans?

Not for us.

Thank you.

Shekhar Pramanick.

Again, the question on the gross margin side, particularly affiliate.

What do you think needs to happen that, assuming that gross margins go north of 40 percent, what is really possible there?

You're asking whether WaferTech all margins were exceeding 40 percent?

That is correct.

Without affiliate, WaferTech and SSMC, you are already achieving gross margins of 44 percent.

So you're asking, even with the (indiscernible), will we be exceeding 40 percent?

That was your question.

That is the question.

You see now, with one billion we are at 31 -- 39.3, and we're looking for a very good 2004.

So even under that condition, I think it is not so difficult for us to get higher than 40 percent in 2004.

That is an aggregate, but is it possible for WaferTech and SSMC standalone gross margins to grow also close to 40 percent?

WaferTech has regularly (ph) higher cost structure than the fab in Taiwan, so their own margin is lower than TSMC's own fab.

I can only comment.

And their margin will not directly reflect into our margin.

We will count on their revenues only, not their own margin.

And a big chunk of their performance we reflect in our investment income.

Okay, thank you.

Christina Osmena.

Just a quick follow-up.

Somebody had asked this question but I wanted to get more detail.

May I please have your cycle times as they are currently at the end of January from start to out by technology node?

We do not have that number at hand.

Actually, we publish that to our customers.

I believe we're running towards roughly 1.8 (indiscernible) there, something like that, plus/minus.

So it depends on how many (indiscernible) the customer has.

Okay, I will try to get it off-line.

Thank you.

Mark FitzGerald.

Can you give us a little guidance on share count for the first quarter, the March quarter?

The guidance of what?

The count in shares.

(indiscernible) is 20.2 billion.

When do the bonuses get paid where that bumps the share count?

Early in the second quarter.

Okay, thank you.

Ms. Ho, there are no additional questions at this time.

Please continue.

If there are no further questions, we thank you for joining us for this morning and evening.

Thank you very much.

Before we conclude TSMC's fourth-quarter 2003 results webcast conference call today, please be advised that the replay of the conference call will only be accessible through TSMC's webcast at www.TSMC.com.

Thank you all.