超微半導體 (AMD) 2007 Q3 法說會逐字稿

Good afternoon.

My name is Marvin, and I will be your operator for today.

At this time I would like to welcome everyone to the AMD third quarter 2007 earnings conference call.

All lines have been placed on a listen-only mode at this time.

After the speakers' remarks, you will be invited to participate in a question-and-answer session.

As a reminder, this conference is being recorded today.

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

Michael Haase, Director of Investor Relations for AMD.

Please go ahead, sir.

Thank you, and welcome to AMD's third quarter earnings conference call.

Our participants today are Hector Ruiz, our Chairman of the Board and CEO, Dirk Meyer, our President and COO, and Bob Rivet, our CFO.

This is call is a live rebroadcast and will be replayed at amd.com.

The telephone replay number is 800-642-1687.

Outside of the United States, the number is 706-645-9291.

The access code for both is 18461849.

The telephone replay will be available for the next 10 days starting this evening.

I would like to call to your attention that our Q4 2007 earnings quiet time will begin at the close of business Friday, December 14th.

Also, we're hosting our financial analyst day on Thursday, December 13th in New York City.

Additional details will be provided as we get closer to the day.

Before we begin today's call, I would like to caution everyone that we will be making forward-looking statements about management's expectations.

Investors are cautioned that our forward-looking statements are based on current beliefs, assumptions and expectations, speak only as of the current date, and involve risks and uncertainties that could cause actual results to differ materially from our current expectations as set forth in the forward-looking statements.

The semiconductor industry is generally volatile and market conditions are particularly difficult to forecast.

Because our actual results may differ materially from our plans and expectations today, I encourage you to review our filings with the SEC where we discuss in detail our business and risk factors, setting forth information that could cause actual results to differ materially from those in our forward-looking statements.

You will find detailed discussions in our most recent SEC filing, AMD's quarterly report on Form 10Q for the quarter ended June 30, 2007.

With that, I'll turn the call over to Dirk Meyer.

Thank you, Mike.

We're encouraged by the progress we made in the third quarter.

We improved our gross margins.

We cut our operating loss by more than half and improved our cash flow.

We executed well across each of our major lines of business and in all of the key components of our plan, including growing the top line, increasing internal efficiencies and continuing our transition to new technologies.

Our revenue is up 18% sequentially, fueled by a steady flow of new products and robust demand for our offerings.

We saw another quarter of great demand for AMD processors -- AMD mobile processors.

Mobile processor revenues were up 43% sequentially to an all-time record due to strong customer response to the AMD value proposition and our innovative brand strategies.

We saw continued strong performance in our desktop processor business particularly in the channel where we sold record unit volume in the quarter.

We began shipping our new Quad-Core AMD Opteron processors in the quarter with customer excitement and demand quite high.

And while our initial production ramp of Quad-Core Opterons has been slower than anticipated, we expect Quad-Core Opteron will be widely available by the middle of this quarter, and we expect to ship hundreds of thousands of Quad-Core processors this quarter into the server and desktop segments.

GPU revenues grew 29% sequentially led by robust adoption of our ATI Radeon HD 2000 family.

And we're encouraged by the widespread and growing customer response to AMD-based platforms.

A great example of this is a combination of our processors and AMD690 Chipsets and systems from HP, Lenovo, NEC, Packard Bell, Samsung and Toshiba, and we're particularly pleased with Toshiba's announcement this week to introduce commercial notebooks based on the AMD Turion 64 processors and AMD 690 Chipset.

Looking ahead, customer and end user anticipation is growing for our upcoming Spider platform.

The ultimate gaming enthusiast platform featuring AMD Phenom processors, RD 790 chip sets and RV 670 graphics cards.

The Spider platform is proof positive of AMD's commitment to deliver the ultimate visual experience and remains on track for launch in November.

In the area of increasing internal efficiencies, manufacturing performance in fab 36 in [Dresden] remains strong, exceeding targets for output, yield, and cycle time.

We remain on schedule to shut down fab 30 before the end of this year in preparation for transitioning that facility to 300 millimeters.

Our supply chain transformation efforts are starting to produce results, where we did a particularly good job matching product delivery to customer demands while, at the same time reducing inventories.

We have controlled the growth of operating expenses and will continue to focus on increasing the efficiency of all areas of the company, including R&D, sales, marketing and administration.

Finally, we continued our transition to new technologies in each of our businesses.

We completed the transition to 65 nanometer technology in fab 36 in Q2 and as I said, will be transitioning fab 30 this quarter to 300 millimeter tool -- 300 millimeter tool set and as a result, our mainstream processor business will be based on 65 nanometer, 300 millimeter technology as we leave this year.

We're looking forward to ramping 45 nanometer production in the first half of next year, and in our graphics processor business, we will launch the RV670, the world's first 55 nanometer GPU in the fourth quarter.

For the second quarter in a row, we grew our revenues and believe we gained CPU market share.

We introduced new products, we drove increased efficiencies in manufacturing and will continue to do so across the entirety of our business.

In summary, we made good strides to put AMD on course to approach profitable operating performance in the current quarter, and we will not be satisfied until we achieve sustained profitability, but were encouraged by the past quarter's improvement and by our continued strong customer relationships.

And with that, I'll turn it over to Bob.

Thank you, Dirk.

We improved our performance in the third quarter, improving the gross margin and significantly reducing our operating loss as planned.

We also reduced our operating expenses while establishing all-time records in: microprocessor unit shipments, microprocessor unit shipments into the distribution channel, notebook processor sales, and notebook processor unit shipments.

These records drove AMD's third quarter revenue to $1.632 billion, up 18% sequentially and up 23% compared to the third quarter of 2006.

We recorded a net loss of $0.71 per share in the quarter.

This loss includes charges of: one, $76 million or $0.14 per share from ATI acquisition-related integration and severance charges, which I call ARC, and two, $42 million or $0.08 per share charge associated with the impairment of our holding and the common stock of Spansion.

Excluding ARC, we had an operating loss of $148 million in the quarter, a 61% improvement from the prior quarter.

As detailed in our press release, third quarter gross margin increased to 41%.

This compares to a gross margin of 33% in the second quarter of 2007.

The eight-point increase in the prior quarter was largely driven by record microprocessor shipments, improved manufacturing efficiencies, inventory management, and a richer microprocessor and graphic product mix.

As I outlined last quarter, gross margin was negatively impact by an inventory charge of $30 million or two percentage points for older generation microprocessor material.

Therefore, operational third quarter gross margin improved by six percentage points from the last quarter.

As guided, total operating expenses which include R&D and SG&A were down $21 million from the prior quarter.

Cash flow from operation was a positive $223 million for the quarter, a significant improvement from last quarter, and adjusted EBITDA was positive at $60 million, another solid improvement over our second quarter performance.

A reconciliation to operating loss and EBITDA in accordance with GAAP is available in our press release.

Now, switching to the business segments, computing solutions revenue was $1.283 billion, up 17% from the prior quarter, primarily driven by a 19% increase in microprocessor revenue.

Microprocessor unit shipments increased 16% sequentially.

We achieved sequential double-digit revenue increases for our notebook and desktop product lines.

Notebook processor unit shipments growth remained strong, increasing 41% sequentially and 68% year-over-year.

And we significantly reduced our operating loss in the computing solutions segment to $112 million in the third quarter, $146 million less than last quarter.

Switching to the graphic segment, revenue of $252 million grew solidly compared to the second quarter.

Sales of the ATI Radeon HD 2000 series were strong, representing more than half of the graphic processor sales for the quarter.

Design win momentum continues to be strong.

Operating loss in the graphic segment was $3 million, a $47 million improvement from last quarter.

Consumer electronics segment revenue was $97 million, up $12 million from the prior quarter.

Handheld sales were higher than prior quarter and gaming console royalties increased.

Third quarter operating loss was $3 million, a $19 million improvement from the prior quarter.

Now let's turn to the balance sheet.

Cash and marketable securities balance at the end of the quarter was $1.528 billion, down $66 million from the second quarter.

This includes $118 million, representing the value of our holding of Spansion common stock, which has been reclassified for long-term investments to marketable securities, and we netted $1.48 billion from our [5.75%] convertible debt offering in the third quarter, and used the proceeds with additional cash of $200 million to repay in full the $1.7 billion outstanding balance of the term loan used to require ATI.

As previously outlined during the last earning call, we had approximately $800 million in nonoperational cash generating opportunities.

We collected approximately $200 million in the third quarter through the sale of a portion of our Spansion investment, 200 millimeter tool sales and administrative asset sales.

We expect to execute our plan to capture the remaining amount in the coming quarters.

Total inventory at the end of the third quarter was $839 million, down $53 million from the prior quarter.

Days of inventory outstanding finished at 80 days, an improvement of nine days from the second quarter.

Our inventory is fresh with the vast majority of microprocessor inventory DDR2 compatible and more than 70% of our GPU inventory comprised of the new HD 2000 family.

Now let's talk about the outlook.

AMD's outlook statement for the fourth quarter based on current expectations, the following statements are forward-looking and actual results could differ materially from current expectations.

In a seasonal fourth quarter, AMD expects revenue to increase in line with seasonality.

Operating expenses, which include R&D, SG&A, and employee stock-based compensation, are expected to be up approximately 6% compared to the third quarter, primarily driven by investments in new process technology.

ARC charges are expected to be approximately $75 million.

We expect to have a tax expense of approximately $25 million, depreciation and amortization are expected to be about $290 million, and capital expenditures are expected to be approximately $300 million.

For the year, we expect capital expenditures to be approximately $1.7 billion.

In summary, we made good progress in the third quarter and we look forward to continue to move the needle in the right direction in the fourth quarter.

And with that, I'll turn it over to Hector for some final remarks.

Thank you, Bob.

As you heard today, we have made solid progress on our plan toward a profitable model while continuing to invest in the products and technology for our long term success.

We will continue to exercise operating cost control and accelerate our efficiency thrust across the board.

We expect continued improvement in gross margins while maintaining our momentum on the development and ramp of new technologies.

We sold record units through the microprocessor channel and launched a new product cycle in each of our major business units.

And while we still have much more important work to do, our progress on each of these fronts is significant and gives us additional confidence to continue the wholesale industry transformation that I outlined in the last quarter.

Demand for the AMD value proposition is strong and customers continue to want more, more products that much their needs and value propositions, more integrated platforms and solutions and more customer centric innovation and more competition.

Our customers want choice and their decision to involve AMD in their business is very strategic.

Nearly a year after our acquisition of ATI Technologies, the result of the industry change in action are both real and demonstrable.

In the strength of our evolving relationship with key customers like Dell, HP, IBM, Sun and now Toshiba, in the growing portfolio of breakthrough products like our Radeon HD 2000 series, the AMD 690 and upcoming RD 790 Chipsets, the unprecedented performance and performance for one characteristic of our AMD Quad-Core Opteron processors, and soon to be released AMD phenom Quad-Core desktop processors.

In the form of game changing platforms like our upcoming Spider offering, our forth coming Puma platforms and the promise of fusion in an exciting new future of accelerated computing.

Every one of our business units is entering an exciting new product cycle with the added opportunity to deliver increasing value to our customers.

We are very encouraged by our progress, but dissatisfied with our financial results.

And we're working diligently to bring this company back to profitability as soon as possible.

But as I mentioned last time, we remain committed to our unique opportunity for industry leadership and to the strategic course that we have set to achieve that end.

We're making a strong progress in new products, in R&D, in our technology transitions.

And, finally, I know that you're all interested in learning about the progress we're making on our asset light strategy.

We have a bold and substantive plan and we have made significant progress in validating it.

Meanwhile, as I have told you before, for competitive and business relationship reasons, we will not be presenting details of our plan in progress until we put that plan in action.

While I can report to you today is that we have made considerable progress in our preparations and I'm personally very excited about it.

We are positioned for success and have an exciting future ahead of us.

Once again, I want to thank our customers for their support, our investors for their confidence, and most of all, our employees for their unwavering commitment to help us win.

Now, back to Mike for the Q&A.

Thanks, Hector.

Operator, can we begin the Q&A process, please?

At this time, we'll begin the Q&A session.

(OPERATOR INSTRUCTIONS) We'll pause for a moment to compile the Q&A roster.

And our first question comes from the line of Krishna Shankar with JMP Securities.

Yes, congratulations on the good progress on the quarter.

I have a couple of questions.

Can you talk about the server business?

What that did in Q2 and the outlook for growth in the server business in Q3?

You mentioned hundreds of Quad-Core, but I wanted to get a sense for how the order up business did in Q2 and the growth trajectory in Q3.

And secondly, for Bob, the trajectory for gross margins in Q4, give you had a better mix and seasonality in Q4.

Yes, I guess I'll go.

First of all, I think when you were referring to Q2, you actually meant Q3?

That's correct.

Okay.

We shipped tens of thousands of Quad-Core Opterons in Q3.

And that didn't have a material contribution to the overall server business in Q3.

So, as a result, Q2 to Q3, we were roughly flat to down a little bit.

As I said, we expected to ship hundreds of thousands of Quad-Core CPUs under both server and desktop in Q4, and we'll substantially increase the number of Quad-Core service processors shipped in Q4.

And to the question of gross margin, clear characteristics -- clearly, we're pleased with our results in the third quarter.

Fourth quarter it is up addition to the new product offerings that Dirk just kind of talked about on continuing on the backs of the GPU offering.

We saw some significant progress in the third.

And the Quad-Core offerings that will continue in the fourth quarter, we expect gross margin to improve, but it won't be at the same rate you saw the improvement from second quarter to third quarter, but directionally up.

Okay, thank you.

Our next question comes from the line of Chris Danely with JPMorgan.

Hey, thanks, guys.

Can you just I guess, sketch out a path to profitability and tell us or give us a sense of what sort of revenue level or what sort of milestones you need to hit to become profitable?

Sure.

Well, it's -- to me, kind of repeat where I was a quarter ago is clearly our goal is to make money in any given quarter.

But clearly, as I talked about there, we needed to be approaching the $2 billion revenue level and north of 40% gross margin to achieve that kind of goal since we're not going to cut our way to make the bottom line happen.

We're on a path.

We made progress -- made good progress in the third quarter.

Got above the 40% on the gross margin.

Clearly not there in the sales level.

We'll see in fourth quarter.

I mean that's our goal is to break even.

And maybe we have a shot at it.

But that's, to me, that's the directional path we're going.

It is still approaching a $2 billion top line number to drive a bottom line of break even.

Sure.

And then maybe a question for Dirk.

When do you guys expect to start shipping either at 2.4GHz or 2.5GHz Barcelona?

Yes.

The plans that we have haven't changed from what we talked about around the time frame of the Barcelona launch, which is to ship the 2.5GHz product in the middle of this quarter.

Great, thanks.

Our next question comes from the line of Joanne Feeney with FTN Midwest.

Yes, good afternoon, folks, and congratulations on a good quarter.

Couple of questions on the breakdown of sales if we could talk about that for a minute.

Could you describe, perhaps, your ASPs by desktop and notebook.

It looks like in the aggregate, they went up by about three percentage points, but did you see any difference really between the notebook and the desktop space there?

Joanne, this is Bob.

In the client space, whether it was in the notebook or desktop, we actually saw improvement in ASP in both categories.

A slight decline in the server space, but it was driven overall to be up.

Both both clients were up quarter on quarter.

Great.

So, in the notebook space, it sounds like if the price competition that Intel spoke about really sounds like it was confined to the consumer space then.

Is that right?

Well, that's the place we mostly play in.

We continue to try to expand in the commercial space, but our notebook offering sits squarely in the consumer space.

Okay.

And then, others have been concerned about double booking.

Are you folks seeing any evidence of that going into the fourth quarter?

No, not at all.

The indications from the current quarter were a lot of sell through, particularly in the distribution channel.

We're at a very low level of weeks of inventory.

And the signals we continue to get from OEMs is very bullish and strong that units are moving quite quickly.

So, we don't feel there's any overbooking at all.

Okay.

Terrific.

Then if I look at your break even path here, it does seem that given the trajectory we're seeing so far in gross margins, even if that does tail off a bit for the fourth quarter, if you guys continue at the pace you're going into terms of unit shipments and gross margin improvements, again, even if that's a little slower, it does seem if I have the math right, that it looks more likely that you'll break even in Q4 than say a quarter ago.

Is that right?

Thank you.

I can't argue with your math.

And like I said, it -- I mean, everything needs to work, it is not just one business that drives that equation.

It is all of the businesses moving at the appropriate speed.

Some of them are different in the seasonally -- in the season of the fourth quarter.

But clearly, we get to $2 billion, I think we can do it.

And then, finally, if I could, just on the graphics side, a neglected part of a business in these discussions, but it seems like you've been doing better in the mainstream.

Do you feel like going into the fourth quarter that you're getting more traction in the mainstream, and do you think you can really break in at the high-end enthusiast segment?

Yes, Joanne, it is Dirk here.

We're feeling pretty good about the graphics business, first, based on very good OEM design win momentum.

A lot of which turns into business late this year and some of it, much of it actually next year.

A lot of the business that we did this quarter was in the channel where the response has been pretty strong.

And based on the upcoming release of the RV 670, we're pretty bullish about being able to participate in a bigger and more profitable piece of the business.

Great.

Thanks so much.

Our next question comes from the line of Tim Luke with Lehman Brothers.

Thanks.

I was again, nice job on the execution, I was wondering if you could just remind us of the 45 nanometer road map in terms of movement there.

I think you sort of talked about the first half of '08.

Sure, Tim.

It's Dirk here.

First, we're on track relative to having basic yields in place in our factories on material that we're running today.

We're building 45 nanometer microprocessors as we speak and those two facts give us increasing confidence in the public statements we've been making for some time around our intent to be starting our production ramp of 45 nanometer processors in the first half of next year.

Yes.

Any sort of clarity within how you define first half and sort of with respect to the second quarter?

Nope.

Other than to just stop at the words first half.

Okay.

And then, Bob, on your inventory level, lower in the third quarter, how do you see that developing going forward and where do you want that to be?

Well, it is hard to imagine with the seasonal strength of the fourth quarter that we won't, even if we try real hard that, that inventories won't decline.

So, I expect to go out of here with less inventory that I have today.

And then try to repeat that pattern as we go through next year.

But so inventories I expect to drain in the fourth quarter.

Thank you very much.

Thanks, Tim.

Our next question comes from the line of Doug Freedman with Amtech Research.

Hi, guys.

Thanks for taking my questions.

One of them was just touched on, and that is the progress of the 45 nanometer ramp.

Is there anymore color you can offer as far as where you see those products targeted, where we should see the 45 nanometer products at first?

I think we'll provide more detail when we have the analyst conference in December.

But roughly speaking, server and desktop will be the first to benefit.

All right.

And then if I look at -- in the past, you've talked about Bobcat and Bulldozer.

Any progress or update on where those cores are at and how that progress is progressing?

It is tough to provide meaningful detail in this call.

That's probably something better left for an analyst conference as well.

All right.

Let me try one more then.

With the ATI progress, the magnitude of improvement, should we consider that what we saw in September, is that a full quarter of improvement?

Or should we expect that to continue to accelerate in the December quarter?

No, you're seeing particularly in the GPU business, you're seeing a full quarter of improvement.

We launched that new series of products in -- I'll call it mid spring, so it wasn't -- we didn't have a full second quarter impact.

We had a full third quarter impact.

And maybe you don't realize but seasonably, actually, the fourth quarter is not strong in the GPU marketplace.

So, it is actually typically flat to slightly down.

So, it will be -- we'll have to work real hard to grow sales quarter-on-quarter, which we'll do, but it is hard to make a lot of progress if you don't have it from the top line.

All right, Bob.

And in your commentaries, you stated some of the charges were related to severance and asset impairment.

Can you help us understand exactly what actions?

I know there's been some action plans put in place.

Maybe since you don't want to talk about things that are forward-looking, can you give us some idea about what actions have taken place already?

And I take that as my last one.

Sure.

Well, we've had a multiple-prong attack to address I'll call it integration issues.

Integration opportunities where we had redundancy in different categories where we needed to eliminate those resources.

And then we've been proactive in I'll call it performance management to step up and deal with some is people who weren't doing as well as we thought they should.

So, those two actions have yielded over the last two quarters, we've actually reduced some senses to allow us to upgrade the organization as we go forward or to eliminate the redundancies in the G&A areas where you just didn't need to have the same people doing the same thing because of two public companies.

So, you're seeing that in the acquisition of intangible assets and integration charges line on the P&L, which is right above operating income.

And unfortunately, in some of those cases, people actually had to leave the company.

How about, Bob, the comments on asset impairment.

On the asset impairment which is below operating income which ended up being $42 million, that's just reflecting the value of the Spansion stock, which has been more permanently impaired since it has been sitting around $8 for quite a long period of time, which was more than our current book value when we wrote it down to the appropriate level and took the noncash charge in the quarter.

Okay.

I guess one last one on just the pricing outlook.

I mean, we heard your major competitor there say that there's actually a possibility that they can't handle more upside.

They burned a ton of inventory in the quarter.

What's your outlook as far as the overall pricing environment.

What are you seeing out there?

How competitive is it for new business?

It's Dirk here.

First of all, we didn't see a qualitative change in the environment from Q2 to Q3.

Where we compete with the other guy, I would call it always competitive.

But it didn't change Q2 to Q3, and we're not expecting much of a change Q3 to Q4.

Great.

Thank you so much.

Our next question comes from the line of John Pitzer with Credit Suisse.

Yes, guys, thanks for taking my questions.

Just quickly on the top line guidance for the calendar fourth quarter.

Given how strong and how much above seasonality Q3 was, why the confidence around sort of normal seasonal for Q4?

Well, first of all, across the board, we're going into a quarter with optimism on one hand and some cautious concerns on the other.

Let me outline those for you.

First of all, the demand for products in the [pieces] phase from all of the regions around the world, particularly in the emerging markets, it is very solid, very strong.

But when we look at what our customers are telling us and none of them have changed, at least as of today, their plans, they're telling us that they also have a concern on three areas.

One, is there is a spotty places where plastic appears to be in shortage.

There are also spotty places where they see displays being difficult to obtain.

And that also applies to a concern they have over the fact that major fire in Japan where the battery factory pretty much got put out of business.

All of that has caused a little bit of uncertainty into the quarter.

But other than those three issues which are yet to be quantified by any of our customers, the optimism into the quarter is strong and particularly strong in the emerging regions.

And then, Hector, above seasonal Q3, in line seasonal Q4, any view on what that might do to the first half of next year?

Are we setting up for a below seasonal correction in Q1 or do you think there's offsets?

I think that's pretty hard to predict.

We talk to (inaudible) to our liking, but I can't tell you right now.

I think if we don't have those the three issues I talked about, consumption of product in the fourth quarter, we're going to a reasonably normal first quarter.

And then, guys, with [Penran] coming out here in mid November, how do you think that might impact sort of Barcelona adoption and volume ramps there, or do you feel like the 2.4GHz and 2.5GHz parts will sort of offset the introduction here?

Based on the input we're getting from our customers and end users, there is a lot of demand for Barcelona.

I tell you.

We're just seeing people licking their chops and ready to get their hands on the product.

Clearly, going from two cores to four cores is a big incremental increase in the Opteron line that we think will provide some benefit to the business here.

And then last question from me.

CapEx currently going to move fab 30 to 300 millimeter.

Kind of if you look at the plan for next year, how should we think about CapEx from sort of the $1.7 billion level?

Well, I don't want to get into discussing 2008 forecast at this point in time.

That's what I'll set up in the December analyst conference.

But to kind of give you at least a directional, right now, I would say it is flat to down for next year for capital expenditures.

Great.

That's helpful.

Thanks, guys.

Our next question comes from the line of David Wu with Global Crown Capital.

Yes.

Good afternoon.

Can you clarify two points for me, please?

The first one is, did you say the operating expenses would be increasing 6% quarter-on-quarter in Q4?

And if so, what happened to the 500 head count reduction savings that has occurred or should be occurring at this point?

And a second one I have is, regarding the RB 670 launch, which is I guess well anticipated to be next month, when does it really hit your revenue line since Q4?

I guess it is a little late to have much of an impact for Q4.

Let me -- this is Bob.

Let me tackle your first one.

And then Dirk can handle the second one.

And the first one, yes, you heard me correctly.

Expenses for R&D and SG&A will be up from the third quarter around 6%.

That includes stock option expense.

To your question of, so where is the savings we talked about, you saw some of it in the current quarter.

A lot of that increase in the fourth quarter as I made in my prepared remarks is for process technology development as we continue to prepare for that 45 nanometer transition that Dirk talked about.

So, it isn't people related.

I'll call it is more experiment related.

So -- and we continue to backfill those people that we had performance issues with, with better employees.

So, I mean our goal is to continue to invest appropriately in the R&D line and to prepare ourselves for the next process technology migration.

While you're answering that question, can you ask dictionary question, what on Earth is a "typical seasonal fourth quarter"?

Unfortunately, we're in an environment today.

As Bob pointed out in some of his remarks, with the GPU business as an example where it is very strong third quarter and not a strong fourth quarter in the CPU business, which tends to be actually the opposite.

You're right.

It gets tougher for us to anticipate what seasonality really means.

What we intended to say was is the market grows the percent in the fourth quarter and generally for the business in which we participate, we expect to be in that range.

Okay.

And on the RV 670 question, when we launched the product, we anticipate doing so with cards available in channels, and therefore, we will drive revenue in the quarter.

Obviously, with a mid quarter launch and on a $.25 billion revenue basis, that's not going to contribute materially to the quarter but it will provide some upside.

Yes.

Could it offset a seasonal weak fourth quarter in that kind of business?

Sure.

Thank you.

Thank you.

Our next question comes from the line of Uche Orji from UBS New York.

Hi, thank you very much.

Congratulations on a very strong performance.

Just very quickly on the graphics, a little more clarity for me.

It may seem presumptuous to conclude -- for me to conclude that you took a significant amount of market share.

Did you get a sense that you took some share and if you can just clarify for me how much confidence you have going forward that that level of performance is just a nibble and not just pent up demand for the products they launched?

Dirk here.

First of all, I'll say it is hard for us at this point to really make confident proclamations on GPU share fluctuations just because of the way the business works, in particular the way the add-in board business works as an ecosystem.

With respect to the second part of your question, a lot of our view of the forward-looking forward-looking potential in the business comes from the design win activity that we generate with our OEMs.

That's kind of a leading signal as to how competitive our products will be when they come out, and as I said in the prepared remarks, we're pretty optimistic based on those interactions.

Right.

Just another question on R&D, if you're going to be ramping 45 millimeter in the first half, how should we think of R&D spending subsequent to that ramp for the rest of 2008 and beyond?

Should we expect that to start to tail off?

Well, again, I'll provide a lot more color at the December analyst meeting, but we'll continue -- I guess I view it this way.

We're -- 45 then turns into 32.

So, you're continuing to always make investments in process technology with bumps.

So, there's that.

And we continue to enhance our capabilities in design, so it is hard to call at this point exactly where I would peg those numbers.

And I'll give you more in December.

But we don't want R&D investment to go down per se.

Right.

And just on the gross margins, the fall through sequentially was quite significant.

How much of that was 65 millimeter contributing on the cost side, or how much of that was Barcelona -- early demand from Barcelona or just service in general?

If you can just kind of explain it in terms of 65 innovator products and just help me understand how this is.

Again, back to my prepared remarks, we had record microprocessor unit shipments that helped, and that was very little affected by the Barcelona launch.

So, there's very few units Barcelona in there.

We improved the ASP quarter-on-quarter.

We had more 65 nanometer than the prior quarter.

Manufacturing continues to get higher yields, more efficiency.

We improved inventory management.

So, it really was no singular item.

It was a lot of little things.

By each of the business, whether it was the GPU business or the CPU business that yielded that, I'll call it six-point operational improvement quarter-on-quarter.

Hard to sustain, as you said.

Which I wouldn't say anybody should forecast at that kind of rate.

Right.

And just finally back again on graphics in terms of RV 670, what kind of early feel are you getting from the customer base about this product?

Customers are going to love it.

Okay.

Thank you very much on that note.

Thank you.

Our next question comes from the line of Srini Pajjuri with Merrill Lynch.

Srini, your line is open.

Hello.

Can you hear me?

Yes.

Sorry about that.

Just a clarification first.

You said you expected grow in line with in demand with Q4.

You've been gaining share notebooks and it also looks like have a couple of new products that you expect to take share in.

I'm just curious, is there anything else that you're seeing that makes you more cautious?

I mean why don't you expect to grow faster than the end demand in Q4?

I mean, let's just be totally candid here.

Why would we want to stick our neck out on a number that frankly is very difficult to predict right now in the fourth quarter.

We want to continue to increase our share participation in the mobile space.

We believe we can.

But I don't think it is prudent at all to try to pick a number, other than to tell you that it is our goal to continuously increase and improve our participation in that space.

Okay.

Fair enough.

And then one question on the CapEx.

You mentioned it will be flat to down.

I remember on the analyst day, you put out a chart that showed some capacity in terms of unit numbers.

Given the efficiency that you're achieving in manufacturing and also the demand that you're seeing, is there any change to that outlook in terms of your capacity that you expect to have in 2008?

Yes, I mean we continue -- again, I don't want to get into details of 2008, but we continue to toggle how much capability we want to have in place to produce units.

Right now, I would say probably we're less than we were a year ago at that point in time of how many units we expect to produce in 2008.

But I'll -- we'll quantify that more in December.

I guess just so I understand, as you have two more efficiencies, you expect to put that money back on to your balance sheet, not to export units.

Correct.

Instead of building inventory for the sake of building inventory, we would like to build inventory for the sake of selling it.

Clearly, we'll take the advantages we see from manufacturing and let that flow through the business whether it is the balance sheet or the P&L.

Okay, great.

And then just one clarification on the OpEx side.

Is there a steady state or normalized number for OpEx?

I mean, it's -- I was hoping it would come down a bit more.

I'm wondering after you ran the 45 nanometer, should we expect a steady state of OpEx?

That's -- no.

I guess the simple answer is no.

I mean, we continue to believe we have a business model that will gain share and grow the top line, and therefore, we'll continue to invest appropriately.

We will see the percentages of cost as a percentage of revenue to decline.

Absolute dollars is dependent on the business conditions in that moment in time.

So there's no "steady state".

This is not a steady state business.

Alright.

Thank you.

Our next question comes from the line of Sumit Dhanda with Banc of America Securities.

Yes, hi.

Question for you, Bob.

You talked about again, not to belabor the point, but talked about reducing discretionary spending, head count.

But as I look at it assuming a normal quarter here in Q4, it will be up to about $1.8 billion, $1.9 billion in revenue, give or take, but the operating expenses are still running about $100 million or so higher than where they were when you were registering comparable revenues last year.

I guess I'm unclear as to exactly what measures you've taken and how -- more importantly, how they've really impacted the operating expense structure to drive profitability.

Well, as we've said all along, we're not going to cut our way to prosperity.

It is about growing the top line and getting the appropriate margin from the manufacturing organization.

And make the appropriate investments in R&D to pull us into even higher growth rates.

So, we continue to invest in process technology development and in R&D for the multitude of products that are -- you'll see -- that you see coming out right now and you'll see coming out near term and in long term.

So, we continue to try to optimize the situation, but we're not backing off on our spending in R&D.

One further clarification, a year-on-year comparison, Q4 this year to Q4 last year, you got to be careful with because we didn't have the former ATI businesses as part of the company for all 13 weeks of the quarter in Q4, point one.

That's true.

Point two being we came out of 2006 growing pretty rapidly.

Our people and resources and capability in the CPU business and it really flattened out that growth, starting late in the first half of this year.

Right.

I wasn't comparing Q4.

My point here would be that while ATI is part of your business now.

It's really -- is a moot point whether you had it then or you didn't.

But I guess I'm a little disappointed that the cost cuts aren't coming faster than anticipated.

But we'll just let that go.

Dirk, I had a question for you.

On the Barcelona ramp you said it materialized a little slower than anticipated.

And your speed grades haven't been targeting the top right off the bat.

There has been some speculation that perhaps your yields on 65 nanometers specifically as it relates to Barcelona is the issue here.

Any comments you would want to make or different explanation you would want to offer up on why Barcelona's [silicon yields] aren't as fast as you would like?

First of all, I'll say that the basic silicon yields of Barcelona are right where we expected them to be.

They're right line on line with the previous 65 nanometer products from defectivity and overall yield perspective.

The issue is meant simply one of tuning the design to the technology so as to support a high volume ramp.

And it is that particular issue that caused us to take a few extra weeks before we turn around the high volume ramp in the middle of this quarter.

In other words, there's nothing endemic to your process technology that is problematic or which might spill over to your 45 nanometer ramp?

No.

This -- the issues -- the minor issues we've been experiencing have nothing to do with the process technology or the manufacturing capabilities.

It is all a matter of tuning the design to the technology so as to be able to ship in volume.

Anything else you could share along those lines in terms of details as what the issues have been in tuning your designs with your process technology?

I don't think in this forum.

Okay.

Thank you very much.

Our next question comes from the line of Glen Yeung with Citi.

Just to follow up on the last question, you made the point about defect density.

Is your defect density the same per unit or per square inch of silicon?

Per square inch of silicon.

Okay.

That's helpful to know.

The other question I wanted to know, and, Bobby, maybe this is for you.

And I know we're all trying to pinpoint you to a number, when you say fourth quarter is growing relative to whatever seasonality is going to be, is the current view that the demand outlook looks pretty normal versus normal seasonality, because your competitor did not guide for normal seasonal?

And I wonder if you're benchmarking to them or what one might call more normalized growth.

Well, you can't ignore the market leader of comments and statements, so to some degree, they're setting the tone and have a much broader view of what's going on in the market place because they're still the dominant player.

So, I think that's a -- they're the best indication of seasonality.

Okay.

That makes sense.

The other question I want to ask, and I'm not asking you for a number per se, but can you give us a sense as to utilization now?

Is it a high number for you net?

Do you feel like you're running your factories pretty full?

And what's the potential that you don't have enough capacity as you move into the first part of -- well, all of 2008, really?

Well, the philosophy we have in running our factories is that we have been ramping Fab 36 in capability for quite some time.

And we're not yet at the maximum capacity.

And as we would ramp, as we continue to add capability for Fab 36, we continue to utilize it.

We don't ramp it and not use it.

So, from that perspective, we use the capacity reasonably well.

But on the other side is we are -- as you heard today, is we have shut down Fab 30, we're transitioning out to Fab 38, and we're doing it in such a way that Fab 38 is going to be like a race car idling in the pitstop.

I mean we will be prepared to ramp that quickly, should we need extra capacity in that space.

Which, frankly, we hope we do.

But at this point in time, we're planning to have Fab 38 at a modest activity in 2008.

You remind me of your Ferrari laptops.

The only other thing that I understand, Hector, is that you're basically saying you feel like you have some flexibility in the way you're managing your capacity for next year, and hopefully -- obviously there's some lag between your expectation of capacity and what may actually materialize.

But you feel like you've got enough flexibility to manage.

You're not materially over or undersupplying.

That's correct.

Okay.

That's helpful.

Thank you very much.

Our next question comes from the line of Ross Seymore with Deutsche Bank.

Thanks.

Just a couple of questions on the server side of things.

You mentioned that the interest level for the new Barcelona chip is quite high.

How can we gauge the success of that and replacing existing customers?

So were they upgrade versus bringing new customers in to switch over to your platforms?

If you could just discuss kind of the success you're seeing on those two vectors, please.

Yes, first, when you use the term customer, you maybe mean something different than we mean when we say customer.

To me, the customers are the people we sell the chip to, which is the OEMs.

At this point, we've got a broad assortment of Opteron-based systems from all of the major OEMs already and Barcelona will just provide a seamless upgrade path into those systems.

Perhaps you meant interest at the end user level.

Actually, both.

Okay.

So, at the end user level, we're seeing a lot of interest both amongst installations that are already, I'll call it Opteron installations, and new opportunities.

And as I said, we've already got a broad assortment of OEM server products across blades, 2P, 4P racks and so on.

And as I think about the ASP side of that equation, how should I judge the ASP bump you may or may not get because of Barcelona coming into your server mix?

And if you want to talk about it in a relative term, if you can contrast that versus the ASP bump you got when you put Opteron out in the first place.

Yes, honestly, it is more difficult to predict the ASP effect within the server line.

It depends on the effect of the competition, the competitiveness of the competitions part, but in terms of overall effect on our ASP, the extent we drive more server business, we drive a richer mix of product overall and drive up ASPs for the microprocessor product line overall as a result, which is a good thing.

That brings over all mix, but it is still up in the air whether Barcelona's launch will increase the server segment ASPs?

Is that a summary of what you just said?

Yes, I would say that, although I will also say we already have a relatively dominant position in the highest ASP area of servers, which is four-socket servers.

And incremental opportunity for Barcelona is in the two-socket space.

Okay.

And then the last question on the ASP side, you talked about in the last quarter, both your desktop and notebook ASPs went up, which I think that's the first time that's happened in a while.

So congratulations on that.

Could you give us a little more color as to what was happening with that?

Is it less pricing pressure at that low end, or are you moving up the stack, or what dynamic is driving the ASP up?

It was reasonably good discipline on our part in terms of identifying the business we thought was good and taking it.

Modest mix upshift.

Okay.

Then the last question.

Hector, you mention that you weren't going to go into any details about the fab light strategy quite yet.

But could you give us any rough idea of when we might be expecting to hear some more details from you?

I'm going to have to tell you that I'm going to do that when we actually do it.

I'm sorry.

Okay.

Is it fair to assume we might get a little bit at the analyst day in December?

Or is it just wait and see?

If it is appropriate, will you get some information.

Okay.

Thank you.

Operator, we're going to take two more questions, please.

Okay.

Our next question comes from the line of Cody Acree with Stifel Nicolaus.

Guys, let me follow up on Ross's question on the ASP mix.

Can you look into Q4 and maybe talk about how you expect the mix with normal seasonality to impact the new products coming in, to impact the total ASPs?

Yes, I'll just be extremely qualitative.

We'll be introducing Phenom Quad-Core product into price points that we don't participate in today and we expect to sell at those price points.

On the server side, really, the opportunity is to drive more server business, and therefore, increase the richness of the mix for the overall product line.

Okay.

So, would you expect ASPs to run roughly flat, or do you think there's some room for improvement next quarter?

Across the product line, there's some room for improvement.

Right.

Okay.

Great.

But I would make sure you realize this is a client-oriented quarter, too.

Right.

I mean most of the seasonality is in the clients.

Right.

And clearly the clients are of much lesser ASP than the server.

Okay.

All right.

Not trying to diminish the Barcelona impact, but that's just the reality.

All right.

And then with thin inventories in the channel, it sounds like OEM inventories are relatively thin as well.

Your inventory coming down a bit.

We got shortages in a lot of other parts.

But how do you feel about your mix of inventory, what you have available, and is there any of -- if we have a decent seasonal period is, there any potential constraints on your part?

No.

This is Bob.

To me, we've done a good job of transitioning to have very fresh inventory.

A lot of 65 nanometer inventories.

I mean, everything is moving in the right direction from a microprocessor standpoint.

And even we're further along or even richer in the mix of all new products in the GPU business.

So, we feel like we're well-positioned to deliver what we need to in the fourth quarter and first quarter, because first quarter you're already preparing for that.

And as Hector outlined, we have flexibility to toggle up if demand continues to be at this high rate that we're currently seeing in the third quarter and fourth quarter.

All right, great.

That's all I have, and congrats, guys.

Thank you.

Our last question comes from the line of Jim Covello with Goldman Sachs.

Yes.

This is Kate Kotlarsky for Jim Covello.

Quick question on just your share gains during the quarter.

I was wondering if you could help us understand whether you've gained share across all of the market segments on the processor side, so in desktops, notebooks and servers, or was there more share gain in one segment versus the other, because it seems like your units grew relatively in line with the market.

And so I was just curious what the dynamics are between the different segments.

Thank you.

Well, we don't have all of the data to confirm by segment.

But based on the data from our competitors and our own data, we clearly gained share in mobile.

We probably lost a little share in server from a unit perspective.

And desktop is too close to call.

It probably -- we held.

So that's how I would call it at this point.

And any way to maybe comment on the magnitude of share loss in servers?

Is it one or two points, or do you think it is more than that?

It is probably in the zero to 1% kind of level.

So, it's a small -- relatively small number.

Okay.

That's helpful.

Thank you.

And again, this is all before Barcelona has much impact.

Okay.

Understood.

Thank you.

Great.

Thank you very much.

That will conclude the call.