聯華電子 (UMC) 2007 Q1 法說會逐字稿

Welcome to UMC's first-quarter 2007 earnings results conference call.

This is [Shiquana], the moderator for today's conference.

Please be aware that each of your lines is in a listen-only mode.

At the conclusion of the presentation, there will be time for a question-and-answer session.

At that time, instructions will be given about the procedure to follow if you would like to ask a question.

For your information, this conference call is being broadcast live over the Internet.

A replay of the call will be available at www.UMC.com, under the Investor Relations Investor/Events section through Friday, June 29, 2007.

Also, a telephone replay of the call will be available from 10 AM Eastern and will run until midnight on May 3, 2007.

To access the replay, please call 1-888-286-8010, or 617-801-6888 if you are calling from outside of the U.S.

The access code will be 60438802 any time during this period.

I would now like to turn the conference over to Mr.

Chitung Liu, CFO of UMC.

Please proceed.

Thank you, operator.

Welcome, everyone, and thank you for joining our first-quarter earnings conference call.

We are hosting this conference call from Taipei.

With me here are Dr.

Jackson Hu, Chairman and CEO of UMC, and Mr.

Bowen Huang, Senior IR Manager.

During today's conference call, we will first review our financial results for the first quarter of 2007, and then Dr.

Jackson Hu will provide an update for our business and forward-looking guidance, followed lastly with a Q&A session.

Before I get started, I would like to remind you of UMC's Safe Harbor policy.

That is, certain statements made during the course of our discussion today may constitute forward-looking statements which are based on management's current expectations at beliefs -- are subject to a number of risks and uncertainties that could cause actual results to differ materially, including risks that may beyond the Company's control.

For these risks, please refer to UMC's filings with the SEC in the U.S.

and the ROC security authorities.

For the first quarter in 2007, revenue was NT$23.03 billion, representing an 11.8% quarter-over-quarter decrease from NT$26.11 billion in Q4 2006 and a 5.6% year-over-year decrease from NT24.38 billion in 2006.

Gross profit for the quarter was NT$3.68 billion or 16% of revenue, compared to NT5.45 billion or 20.9% of Q4 '06 revenue.

Operating profit was NT$18 billion or 0.1% of revenue.

This is compared to NT1.37 billion or 5.3% of Q4 '06 revenue.

Net income was NT1.46 billion in Q1 2007, compared to a net income of NT5.69 billion in Q4 '06.

EPS for the quarter was nt$0.08, and the earnings per ADS were US$0.012.

Above is a short summary for the results in Q1 2007.

Please refer to our press release today for more detailed financial data.

Now, Dr.

Hu will provide you with an update for our business and the guidance for the second quarter of 2007.

Thank you, Chitung.

Hello, everyone, and welcome again to our conference call.

Thank you for joining us today.

As always, we appreciate your interest in UMC.

First of all, I would like to touch on the highlights and then provide further details on our performance and the key activities impacting our business.

The first quarter of 2007 was clearly a challenging quarter for UMC.

Our results reflected the anticipated soft demand caused by the industry-wide inventory correction.

For Q2, we have noticed that the customers' inventory levels have been continuously going down.

Starting from June, demand for communication applications, such as the cell phone, and the consumer applications, including digital TV and (inaudible) flat-panel driver ICs, will pick up strongly.

The demand for the computer segment is also starting to recover, although it is not a strong as the other two segments yet.

Given that demand continues to remain as strong as we are seeing now, Q3 revenue with be very strong, which will significantly improve the overall bottom line.

As far as new technology development is concerned, multiple 65 programs, 65-nanometer programs are moving forward smoothly.

Yield improvement is very encouraging.

Besides the existing 65-nanometer products that have been in production, additional business contribution from this technology (inaudible) start to come in later this year.

Furthermore, we are working with multiple customers to prepare IP for 45-nanometer technologies.

We expect to have multiple product tapeouts in Q3 this year and to have working prototypes in Q4 2007.

Particularly noteworthy is the effect that more and more IDMs are becoming fab-lite or fabless.

This trend will help extend the business for the foundries.

Specifically, it will help UMC diversify into other product areas, such as CPU and memory.

Currently, we're working with many customers to explore mutually beneficial opportunities in this area.

This trend will also allow us to collaborate with our IDM customers more closely in developing future technologies such as 45 nanometer and 32 nanometer.

As usual, our industry is full of challenges and opportunities every day.

We are optimistic about the foundry industry's future as well as UMC's own business outlook.

Now, let me provide you with the guidance for the second quarter of 2007.

Wafer shipments are expected to increase by 6 to 8%.

Wafer ASP in U.S.

dollars is expected to be flat from the previous quarter.

The capacity utilization rate will be approximately 75%.

Please notice that our overall capacity will increase by 2.1% from Q1 to Q2.

As for profitability, there will be a slight improvement from the previous quarter.

The percentage of revenues from 90-nanometer and below will be approximately 20%.

The Communication segment is expected to be the strongest, followed by the Consumer and the Computer segment.

Our 2007 CapEx budget will be within the range of US$1 billion to US$1.2 billion.

That is all for my report.

Now, we can begin the questions and answers.

(OPERATOR INSTRUCTIONS).

Randy Abrams, Credit Suisse.

I wanted to see if you could talk about what is driving confidence, perhaps by application or customer type, into a sustained rebound into the third quarter.

You mean besides the strong demand side or you are (multiple speakers)?

Yes.

When you're talking about strong demand, where you are seeing it?

And extending into third quarter, is it across all applications?

Then is it more IDM-led coming back or is the fabless continuing to be strong?

Yes, this industry is interesting.

When you see a strong rebound, it usually happens across the board, and now the situation is similar.

We see stronger demand from both IDM and fabless.

Okay, maybe talk about the Asian fabless.

It has rebounded significantly the past two quarters.

Do you still see room into the next quarter or two for that to continue to grow as a percent of sales?

Actually, in the last two quarters, the demand for consumer products was low, right, in Q1.

So I do not know exactly what you're referring to, and so can you maybe reformulate your question or clarify it?

Yes.

The question -- your revenue from the geographic region from Asia has gone up 27% to about 43% of revenue, so I wanted to see if that potential, that region will drive more of the revenue growth in the coming quarter or two.

Okay, I see what you mean.

Yes, relatively speaking, actually, we have large cell phone customers and in Q1, the seasonal, low season -- it's typically the low season.

That is why I think, relatively, the sales from North America is lower, and relatively from Asia is stronger.

You know, to the Chinese New Year is actually was in February, right, so the consumer demand and some of the cell phone demand continued.

Okay, just one final question -- could you talk about the factors for profitability being up only slightly in the second quarter with shipments rebounding and ASPs holding flat?

Yes.

I think I mentioned it briefly -- or implied that due to the capacity increase, the utilization is at 75%.

So as the utilization continuously improves, then the profitability will have a more significant improvement.

Okay, thanks a lot, guys.

William Dong, UBS.

Good evening, two quick questions.

One is on the memory side.

I know that, with working closer with the IDMs, there is potential on the memory side, but as we know that the equipment mix is a different between memory and logic.

So how do we plan to address that or is there any insight you can provide to help us understand that?

Okay, very briefly, DRAM does use different equipment than logic.

However, other memory, including SRAM, NOR, NAND, flash memories, the compatibility on the equipment side is much higher.

Okay.

So one final question then is, in terms of sort of looking at the trend before versus today, I mean, we've seen sort of the breakeven point, utilization point has gone up to about 70, 75%.

I was wondering, going forward, is that going to be more the norm or are there things we can do to actually lower that breakeven point to kind of back to more the historical norm?

Yes, I think it's a matter of technology mix.

If we have more products on the high-end technology side, then that tends to lower the breakeven utilization point.

That's the direction that we should drive for.

Okay, so maybe just to clarify, so for the 65-nanometer, could you -- just a quick update in terms of number of customers or products that's currently running or taping out at the 65-nanometer side?

Yes, we have grown many customers and many programs ongoing.

Just a second, please.

So for example, we have at least ten engaged customers, and two of them have been in production since last year, and we have more than 16 products tapeouts by the end of 2006, and that number is increasing.

Thank you.

Shailesh Jaitly, Nomura Securities.

I actually wanted to just understand a bit more on your ASP guidance.

You have mentioned the ASPs to be flattish and your product mix doesn't seem to be improving.

You mentioned that 90-nanometer is going to be 20%, so I guess flattish to down.

So obviously there are certain areas where you are seeing the ASP improvements, which are offsetting the declines, so if you could run through (inaudible) which are the areas where you are seeing ASP strength and what is driving it?

Yes.

I think the pricing pressure is more on the mature technology [nodes], and we did see increasing demand on the (inaudible) nodes, specifically 1.3-micron and 90-nanometer.

So that's how the balancing comes in as how we can keep the ASPs flat.

Given that your 90-nanometer is declining from 21% of the revenue to 20%, are you expecting a massive pick-up on the 0.13-micron which is offsetting the decline from the mature?

If you could also quantify as to what proportion of your revenue is going to be 0.13-micron?

0.13-micron in Q1 is 16%, so definitely, for Q2, it will increase.

Yes, I don't have the exact percentage estimate for you yet.

Okay.

Also, in your prepared remarks, you talked about new areas of engagements going forward, potential engagements; that is memory and CPU.

If you could help understand the progress you have made on the sepia CPU front, and also if you could provide a time line as to when do you expect the CPU part of the business to start kicking in?

We are discussing with multiple customers in the CPU and memory area, and at this moment, we cannot comment on the specific time line and who the partners, potential partners are.

At the proper time later, we may disclose those information.

Just to clarify, is it a MPU customer or is it CPU customer?

I don't know the difference as to what you are referring to.

That is the processors for the PCs, or the processors for the other consoles.

(multiple speakers)

I consider both of them as one category, as CPU.

I was CPU designer myself when I was young, so I consider them all in the same category.

Fine.

One last question on your capacity -- if you could help provide the details as to what you expect year-end capacity to be?

Would all the capacity additions be at the leading edge or are you also wrapping up some of your mature processors' capacity?

Capacity for the 2007 should be 7% higher than 2006.

The (inaudible) equivalent, it will be about 4.3 million [pieces].

In terms of breakdown, we believe the 12-inch -- or the breakdown by geometry, that 36% to 40% by year-end will be below -- will be 0.13 and below.

Thanks, Chitung.

Thanks, Jackson.

Steven Pelayo, HSBC.

Good evening.

A couple of quick questions on the composition of your cost of goods sold -- it was a little bit of a surprise to me in the March quarter, both on the depreciation coming down as well as the cash COGS, if you will, both components.

First on the depreciation, I guess, based on some earlier feedback, I understand you guys are guiding second-quarter depreciation to increase slightly here but the full-year to still see a decline of about 10%.

For that to happen, then it means you're looking at a pretty big pickup, then, in depreciation in the second half of the year and the corresponding impacts to your margins.

So, any comments on your second-half depreciation quarter-over-quarter trend?

Then my last question is really just relative to the cash COGS component.

Could you talked a little bit about the drivers of that?

I expected it to be a little bit more variable.

As a percentage of revenue, that was 51% of revenue.

I think that was the highest level I've seen in many years.

Can you just talk a little bit about the drivers of that and your outlook for that component of your cost of goods sold?

Okay, yes, I think the numbers you're quoting are correct, that we are seeing this quarter to be the bottom of the recent depreciation level.

For the following quarter, Q2 2007, we expect to see a 3% increase in depreciation quarter-over-quarter.

For the whole year, we are expecting to see a 10% decrease over 2006.

According to our internal numbers, the sequential growth in depreciation should be fairly mild and a linear trend for us throughout the year.

So, that does suggest than about a 10%, maybe 10 to 15% sequential increase in either Q3 or Q4, so that's going to impact, I don't know, a couple of hundred basis points to my margins assumptions in the second half.

So is that what we should be thinking about?

There is a 10 to 15% sequential increase in Q3 or Q4 for depreciation and COGS?

It shouldn't be the number.

I think it's quite different from our internal numbers, so you may want to recheck your models.

But if you have any questions, just e-mail us.

We can help you with that.

Okay, then they other question, on the cash COGS component of margin?

That could be similar to your first question.

Again, our internal number doesn't really show that, so if you can send us more specific questions, we can help you to clarify that.

Okay, well, I mean, just looking at the numbers from the direct press release here, you had roughly 11.7 billion on the non-depreciation COGS.

That's about 51% of revenues.

That's the highest level it's been in many years.

I was just trying to understand the drivers behind that.

Just the other side of the depreciation cost is at the lowest point ever.

Okay.

I will take that one off-line.

My last question then will just be your guidance for a slight improvement for profitability, is that the bottom-line guidance?

Is that an operating-income guidance?

Just what you general thoughts are for the nonoperating income line investment sales, asset disposals for second quarter and the remainder of the year.

That's it.

Okay.

The slight improvement in profitability refers to both gross margin as well as the operating profit margin.

As for the guidance for nonoperating income, up to now, there is about NT$2 billion also disposal gains from the exchange of our [AUO] exchangeable bonds.

That will be the [least] number for the second quarter and yet we don't really have any concrete trend for major disposal in the second quarter for our noncore assets.

Great, thank you very much.

Bhavin Shah, JPMorgan.

Chitung, if I can just pickup on the last question, last year, the depreciation in costs was about 38.5 billion.

If you start from the first quarter number you reported and just take your comment of, I think you mentioned (inaudible) increase linearly throughout the year.

Even if I assume 5% increase throughout the year, we end up with about 14% decline year-on-year in COGS depreciation.

So it just seems like there has to be a meaningful jump in the second half, otherwise your year-on-year depreciation would decline a lot more.

Could you just help us with that?

Again, it's hard to pinpoint the percentage increase, but I think what I can give you is a 10% decrease year-over-year comparison for the whole year depreciation, plus a linear pick-up in depreciation throughout the rest of the year.

The 3% increase is as far as I can give you for the second quarter.

We should be -- I mean, the number we have been calculating (inaudible) several times.

It should be correct.

Okay.

Can you just give us some guidance on how to project the investment income for this year?

I don't have any guidance for the whole year, but like I mentioned, for the second quarter, so far we have realized about NT$2 billion from the AUO sales through the bond conversion.

Other than that, we don't have any concrete, major plan yet for the second quarter.

Right.

Let me just go to the comment on (inaudible) opportunities in CPU and memory.

Obviously, memory is a pretty tricky business, as you know from the past experience.

Could you just elaborate on exactly what sort of opportunities in the memory area would you be interested in?

SRAM is clearly one, and flash memory, which have high compatibility from the equipment point of view, is another area.

DRAM is not in the consideration.

Okay, thank you.

Pranab Sarmah, Daiwa Securities.

Good afternoon.

This is Pranab from Daiwa.

I have a couple of questions.

The first one -- could you give us some idea what was your -- how your utilization rate was on that 12-inch fab in the first quarter and how you think that 12-inch fab will grow by the third quarter this year, utilization rate?

Well, first quarter, as we reported, if you included the annual maintenance in the base, the reported average utilization rate was 74%.

If you factor any shutdowns, then the actual loading was only around 70%.

For the breakdown range in Q1, it was above average, and 12-inch was below average in Q1.

According to the second quarter, we don't have the annual shutdown factor any more.

The 2% increase in overall capacity, which increased wafers, with lead the utilization rate to roughly 75%.

In the second quarter, it's actually quite average between 8-inch and 12-inch in terms of loading.

We have (technical difficulty) to see the utilization for the (inaudible) technology node to continue to improve quarter-over-quarter.

So in third quarter, could we expect like 12-inch will take over on the utilization rate over 8-inch?

(multiple speakers)

Quite possible.

That means ASP improvement could be quite significant on the third quarter?

We certainly hope so.

Okay.

Very good.

The second one is on -- do you have any timeline about your stake in Hsun Chieh, like when you will be reapplying to the Taiwan government and what would be your strategy for next -- this year on the Hsun Chieh stake?

Are you referring to capital reduction or --?

No, no, Hsun Chieh stake, that is -- which is a non-accrual account, and you are not allowed to take that stake because of the Taiwan government regulations.

Do you have any timeline that you're going to reapply with the Taiwan government again this year sometime?

I do not know whether it is a matter of re-apply.

We did ask for government guidance for instance, on the timeline, for taking the 15% of Hsun Chieh.

So far, we have not received a very clear response yet.

Since now 0.18-micron is allowed to go through China, we may formulate our approach to the government again.

Any timeline for [it] and whether this quarter or next quarter you might approach the government again?

Yes, we can do it anytime, yes.

Okay, that's good.

Could you also tell us about the top five customers, how many percentage of your revenue accounted by the top five customers in the first quarter, and how many customers were above 10%?

I think our top 10% accounted for 62% of revenue.

Top ten is 62%.

How many are above 10%?

How many above 10%?

Only one.

Only one, okay.

The last one is on -- when you define about the third quarter that it is very strong, what is your definition of very strong?

Historically, if you see like quarter-on-quarter and a very strong quarter, [we have] grown [about] 20% sequential growth.

So far, we don't really have any guidance for the third quarter yet.

I think that strong was referring to the momentum of the leading-edge revenue picking up.

So the whole quarter, we have to wait until the next earnings conference call.

That's true.

My last question is on the 65-nanometer.

What is the percentage of revenue that came from the 65-nanometer in the first quarter, and how you think that will do in the second quarter?

Well, 65-nanometer, it was 1% in first quarter.

And second quarter, what is your expectation on that?

Between 1 to 2%.

Mehdi Hosseini, FBR.

Yes, thanks for taking my question.

I have a couple of questions.

Going back to your comment that you will see a significant pick-up in demand by the June timeframe, I want to know if that pick-up has already been seen in your wafer starts.

In other words, are your wafer starts beginning to increase as we speak?

If not, are you referring to the start picking up in the June timeframe?

No, we are referring to the wafer out for the quarter otherwise we can't -- we will not be able to account for revenue.

So that increase in wafer shipments has already been seen in the wafer starts?

The wafer shipment in Q2 has been accounted for.

Okay.

Then you talk about mix of 65-nanometer as a percentage of revenue.

How should I think about the mix as a percentage of actual capacity?

Just a second.

By the end of the year, I think the capacity for 65-nanometer will be around 3%, 3 to 5%.

3 to 5%.

Okay, so your utilization rate is about 75%.

If the second-half seasonality comes in, I will see your utilization rate to get up to at least 90% sometime in the second half.

So as we go into Q3 and you begin to plan for capacity in 2008, including the required R&D for 45-nanometer, would that imply that your CapEx would at least be flat for next year?

It's a little too early to talk about that and we will have to -- we need more time to gather the market demand situation and also the R&D requirement.

Fair, but would this year require you to start planning earlier than previous years, given the fact that we have already hit the [low end] in utilization rate?

Well, what we have done is we have announced a second phase of our 12 [a] fab in (technical difficulty), and that's a total $5 billion project.

Once the share is done, then that gives us plenty of flexibility in terms of equipment moving.

Remind me again, has the construction already begun?

Yes, we have made a press release.

Yes, it says that Phase III and the Phase IV of the current (inaudible) and the construction of that fab is expected to finish by the end of this year.

(multiple speakers)

Okay.

Then one question regarding the macro big picture -- how do you see the overall health of the semiconductor industry this year in terms of the growth?

To that extent, would you expect foundries as an industry to underperform?

I think [in] more and more markets, the lead companies are predicting a milder growth year for the entire semiconductor industry, maybe around 8%.

Historically, foundries have outperformed the other sectors, except last year, I think the memory was very, very strong.

So this year, there's still a chance that the second half is very strong for the foundry to outperform the average.

Now, is that driven because someone or a few of your largest customers want to outsource more, or is that more of an organic growth?

It's more like a market demand kind of thing, you know, instead of one or two specific customers.

You know, we have a very broad customer basis.

Just using cellphone as an example, we have pretty much all the leading players (multiple speakers) using us.

Sure.

If I may do just one quick follow-up?

With one of your largest customers increasingly talking about a higher mix of outsourcing and actually outsourcing the R&D portion of the process development, would that help you with the higher growth maybe next year or in 2009?

Yes.

As I mentioned earlier, the trend that more and more IDMs are becoming fab-lite or even fabless on the advanced technology node is a very positive thing for the foundry industry.

We expect that more business from the IDM customers.

Also, we expect it to collaborate more with them in technology development.

The reasoning is very simple -- because in the past, they developed their technology internally, and (inaudible) the products are working, then they asked us to match their products.

Now, they don't do development any more, so they have to work with us more closely in defining the process of spec needed for their products.

Wouldn't that require you to increase CapEx intensity or even R&D expenses?

No, not necessarily.

Okay.

All right, thank you.

Daniel Heyler, Merrill Lynch.

Good evening.

I had a question, Jackson, on your product diversification.

That was one of the key areas that you focused on in the past, in terms of trying to improve UMC's profitability and diminishing its volatility through the cycle.

Here, you talked about the digital TV product, which is a new product.

Would you elaborate a little bit in terms of your progress here in diversifying your product base and how that's progressing, and maybe quantify, if you could, how many new products you think we can see in the second half that are meaningful and high-volume in nature?

Okay, I wouldn't think of [HDTV] as a new area.

Actually, we have multiple customers shipping product (inaudible) in this category in last year and also this year.

So, but as you pointed out, diversifying our product portfolio and the customer portfolio is something that we continue to work on, and including the new areas that I mentioned today, the CPU and the memory.

So, it may take some time.

I cannot give you a definite timeline for when those products will be available and start to contribute to the revenue.

Typically, if we engage with someone, it may take a year and a half to two years to see significant revenue contribution.

That can give you a (inaudible).

Okay, I was referring to more of the core business, the digital logic part of the pie, which is relatively close to what you've been doing in the past.

You've had strength in graphics and core logic previously, and you've talked it about as an opportunity for a couple years now.

Is there any progress in the graphics and core logic business where that can become a more meaningful part of your business?

We started to ship more revenue from those areas that you referred to.

As the demand comes back, we expect that trend to continue.

Then we have new programs following the 90-nanometer programs, and those will continue as well.

Okay.

Then you referred to -- actually you said flash memory.

There's obviously two types.

Would you consider NAND or would you focus more on the NOR market initially?

We would look at both because, from an equipment compatibility point of view, they are very close.

Okay.

Since that's a relatively commoditized product, increasingly commoditized, are you looking for mainstream -- maybe I guess explain the logic there.

Is this going to be a way to ramp your advanced fabs or are you looking for commodity products, or are these relatively customized embedded-memory products with System-on-a-Chip capabilities?

I think it would be helpful for investors to understand whether you plan to go into the mass market or whether this is more of a foundry model.

Okay, all of the above.

We have been doing embedded flash memory in the past, and we are making more progress over there.

For commodity memory products, flash memory products, we are also exploring because the flash memory is growing rapidly, and we simply feel that we cannot be -- we cannot ignore it, and we have to be present.

Okay.

Then finally, on the DRAM side, you are considering DRAM.

You've been in and out of that business in the past and it hasn't been a successful business for you.

What has changed there?

Why are you considering the DRAM market again?

Oh, I said that we would not consider the DRAM.

Oh, would not, I'm sorry.

I thought you said under consider -- okay, great, that's good news.

Then finally, on CPUs, there are a number of segments here obviously competing in the high MHz products segment.

It's rather difficult.

Is this an area where you would be focusing perhaps on more of an integrated or power-sensitive product segment suited for the (inaudible) high-volume segment (multiple speakers) performance markets?

Initially, we would probably focus in on Box CMOS, and I did mention in the past that we are continuously developing SOI technology and we will disclose that at the proper time.

Therefore, when that technology is available, we would explore business opportunities as well.

Okay, great.

Then at 32-nanometer, will anything change in terms of your partnership strategy?

You've been relatively close to a number of key partners, both fabless and IDMs, in terms of developing your advanced technology.

Given the difficulties and complexity we're hearing about at 32, would we see anything potential change, potential different alliances or possibly even consider IBM as a partner at that [node]?

Yes, joint -- a similar question was brought up in afternoon local conferences.

You know, someone mentioned that technology development is not a team sport.

You may have heard of this in the past.

Joint technology development is not an easy thing to do, and UMC has engaged in GDP with multiple partners in the past, right?

However, we noticed this, that sometimes each partner may have their own wish list or their own priority and how to make everybody happy or with limited resources and manage the engineering resources.

It's not a trivial matter.

So for joint development with many partners, it's something we have been very, very cautious about because we got burned in the past.

So everyone knows we have been successfully developed 90-nanometer and 65-nanometer, and moving onto 45.

So at this point of view, we would consider to continue our path.

I got it.

Thank you.

Mark Fitzgerald, Banc of America.

Thank you.

You guys have announced a share buyback.

When I look at the share count, it still kind of creeps up here, so I was curious.

Can you give us an update where you stand on that and the plans here?

Okay, it's not a share buyback; it's actually a capital reduction proposal, which is subject to the shareholders meeting approval on June 11.

Once we've got the approval, the whole process may take up to four months before the shareholders receive NT$3 per share back.

So the share count increase a little bit probably due to the share employee option exercised.

Okay.

I notice that our long-term debt came down and the short-term debt went up.

Is that in preparation of this capital adjustment that's coming up?

No, it should be a routine payback, redemption of our debt.

The magnitude should be quite small.

Okay, thank you.

Sunil Gupta, Morgan Stanley.

Thank you.

Good evening, Jackson, Chitung and Bowen.

I just wanted to follow up on questions earlier about memory and CPU.

In terms of your memory businesses, how large do you think it could be as a percentage of your revenue by, let's say, end of this year and end of '08?

It's too early to talk about that and definitely, it would not show up this year.

As far as percentage-wise how much we wanted to engage, that is a question that we have to discuss with the potential customers.

Yes, one thing, if I may add, is really that we see the trend of IDM outsourcing becoming a fab-lite or even become a fabless company.

In the past, CPU and memory always excluded from the addressable market, from a foundry point of view, but we see accelerating outsourcing trends, even CPU and system memory now becoming addressable on the foundry operator.

So that's why we are quite excited and start to restore to this upcoming opportunity more aggressively.

Okay, so from your internal planning point of view, if memories aren't going to be big this year, what's the internal thinking or constraint that you might have or limit that you might have for your memory business, say, by the end of '08?

I understand that it will depend on how the market plays out.

But what is on your wish list?

Well, I cannot give you an exact number, a percentage of revenue for memory.

It's too early to estimate.

But you know, think of this as a diversification of our business, diversification from the traditional logic-only business into other areas, which is, I believe, healthy to the Company over a long time and it will help [validate] the fluctuations, the seasonality fluctuations and other factors.

Okay.

When you are thinking about these two memory initiatives, you said SRAM and flash, based on the way the market is playing out right now, what do you think could be the related margin structure of that business compared to what you're doing for your silent logic and mixed-signal?

What kind of return on your assets could that generate and how does it compare?

That is also a little bit early to talk about, yes.

I hope that, at the proper time later, we can disclose this information.

Okay.

Could I just pursue also your comments earlier about CPUs?

You said that your (inaudible) multiple -- potential customers there, and it seems like you are going to tackle this first (inaudible) CMOS and some point in time you might complement that with SOI.

Would you have an SOI process ready and working by, say, the end of 2008 to go after the CPU market?

Actually, we are developing 65-nanometer SOI, which may be available by the end of this year.

But would that still be relevant to the CPU market if you are already on 65-nanometer only by the end of this year and I guess (inaudible) may have moved onto the next process?

Yes, we hope there is still opportunity, but you know, through this development experience, we would solidify our SOI technology.

Okay.

Would you expect your CPU business, whether it's about CMOS or SOI, to be more than 2 to 3% (technical difficulty) say by the end of 2008 (inaudible) visible in your business?

I think, again, it's too early to estimate.

All right.

Okay, thank you very much.

Ivan Goh, Dresdner Kleinwort.

Good evening, a couple of questions.

The first one -- not too long ago, I remember you guys talking about the SoC market and how many opportunities there were there.

Today, you're talking about CPU and memory.

I just wanted to backtrack a bid and ask is that shift in focus due to your view that the SoC market is a bit -- has been exhausted?

SoC is definitely trend.

When we mentioned about SoC, it was at least two years ago, if not even earlier.

Today, you have already noticed there are single chips, over a single chip cellphone that has baseband and RF (inaudible) integrated it.

You can clearly see similar products in the Wireless LAN area, Bluetooth area and so on and so forth.

So, that is definitely the trend, and more and more products will follow that trend.

One question on the CPU -- may I just find out how long have you been in discussion?

It appears that (inaudible) came out fairly recently.

I just wanted to see when you got through the door with the discussion, and at what stage the discussion is and if customers have given a firm commitment that they are definitely going the outsourcing route or are they still in the exploratory stage on whether it is doable?

Yes, the discussions have been ongoing for a while, yes.

Some commitment was firm and others are still being developed.

Okay.

Some housekeeping question to Chitung -- just now that you mentioned that Q2 depreciation will rise by 3% Q-on-Q?

Yes.

Okay, and then one question on tax.

What kind of tax rate can we expect for Q2 and for the whole year?

It seems to be 16% in the first quarter.

Yes, 15% in Q1 was a bit misleading due to the big component of nonoperating income.

For the full year, we're still expecting to see 5 to 10% of the tax rate, plus Taiwan has adopt this minimum tax of 10%.

One final question -- I got just now that you're talking about the utilization at the leading-edge increasing into the third and the fourth quarter.

You also mentioned that there is not much of a gap between utilization of 8-inch fabs and 12-inch capacity in the second quarter.

If you can, would you review what the utilization rate of 12-inch fabs as you exit the second quarter?

We don't have that kind of data.

It's a dynamic, moving target.

Okay, thank you very much.

Shailesh Jaitly, Nomura Securities.

Yes, thanks.

For the PC segment, what proportion of your revenue in the second quarter was graphics-related?

Just a second, we are getting that data.

Actually, a significant portion.

Yes, PC was about 18% of our (multiple speakers) revenue in Q1.

I would say about 20% to 25% in the PC segment is graphics-related.

Thanks.

Of your older revenue, how the LCD driver market revenue has panned out -- you mentioned that market is recovering, so you alluded to that comment for second quarter, or you have already seen growth in the current quarter?

Or pardon, the first quarter?

Depending on which customer you are referring to, we see some demand picking up in the beginning of the Q2 and some of them start to increase the demand in June.

Okay.

As an aggregate, if I recall correctly, it used to be almost a quarter of your business in terms of shipments.

Has that proportion changed materially?

Can you repeat your question again?

I recall -- correct me if I'm wrong -- but LCD drivers used to account for most 25%, a quarter of your business in terms of shipments, not revenue.

I was wondering.

Has that number changed significantly?

(technical difficulty) much.

I mean, the shipment and revenue are pretty much similar.

Similar to that.

Thank you very much.

Steven Pelayo, HSBC.

Yes, if I could rephrase my earlier question, I think Bob (inaudible) tried to follow-up on it as well, in order for your depreciation to decline about 10% this year, it implies you exit the in fourth quarter of '07 with a quarterly depreciation in COGS of perhaps as much as 30% higher than what you just showed in the first quarter, roughly high NT9 billion to NT10 billion.

A, does that sound right to you?

B, if it is true, this then implies that if we flat-line that kind of run rate, that your depreciation in 2008 could beat up as much as 10% or more.

Can you help me understand that?

Just for clarifying that first of all your assumption was not correct, and second quarter, we are seeing 3% increase.

The third quarter, we are seeing 5% sequential increase, and fourth quarter, we are seeing roughly 2 or 3% increase.

So hopefully that clarifies your question.

Yes, then that doesn't add up to the down 10% full-year.

All right, we will take the rest of that off-line.

My second question was on the R&D.

It was down about mid-teens or so, quarter-over-quarter, but still about 10% of revenues.

I'm curious about your full-year thoughts for R&D.

For the whole year, it would be similar to or close to 10%.

Great.

Thank you very much.

Pranab Sarmah, Daiwa Securities.

Thank you, I have a question on your R&D strategy for memory business.

As Jackson has mentioned about going to flash memory business, and probably R&D technology development card with the flash memory is quite different from the logic side.

How are you going to look at resources on these two areas from your R&D point of view, and (multiple speakers)?

The model would be foundry model, so similar to the model that we used for our IDM logic products.

So basically, we will be doing the process of matching [reporting] for them.

Again, still that means you have to probably look at quite a sizable amount of fab to one customer in that area.

Well, it depends how much business we are willing to take, right?

Okay, got it.

But again, in logic -- I mean, sorry, the flash memory side, your technology card is probably one year, you migrate to next technology node.

Logic is one and half year.

Would it be easier for you to do both together?

For your information, we have been doing process matching and [reporting] for many, many years and for many, many products.

Every year, we have many programs ongoing.

So you know, we have accrued significant experiences, and we have processes and procedures that we follow and can assure a first (inaudible) success.

So we have been doing that extremely well.

Okay, thank you.

At this time, I would like to turn the call over to Chitung Liu for closing remarks.

So, that concludes our call today.

Thank you again for joining us.

You are more than welcome to contact us directly if you have any further questions.

So now, over to you, operator.

Thank you.

A link to the replay of this call will be available until the end of Friday, June 29, 2007 on the Investor Relations section of our Web site.

A dial-in version of the replay can be accessed by calling 888-286-8010 if you're in the U.S.

or at 617-801-6888 for international calls.

The dial-in replay will be available until midnight on May 3, 2007.

The access code is 60438802.

If you have any additional questions, please feel free to contact us directly.

Thank you all again and have a good day.