超微半導體 (AMD) 2005 Q3 法說會逐字稿

Ladies and gentlemen, thank you for standing by.

And welcome to AMD's third quarter 2005 earnings conference. (OPERATOR INSTRUCTIONS).

As a reminder, this conference is being recorded.

I would now like to turn the conference over to the Director of Investor Relations, Mike Haase.

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

Our participants today are Hector Ruiz, our Chairman of the Board, President and CEO;

Bob Rivet, our CFO; and Henri Richard, our Chief Sales and Marketing Officer.

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

The telephone replay number is 800-475-6701.

Outside of the United States the number is 320-365-3844.

The access code for both is 790990.

A telephone replay will be available for the next ten days starting at 7 PM Pacific time tonight.

For your planning purposes I want to call to your attention an upcoming AMD event.

We will be hosting our Analyst Day in Sunnyvale, California the morning of November 15.

Invitations will be coming out in the near future.

In addition, I would like to call to your attention that our Q4 2005 earnings quite time will begin at the close of business Friday, December 9.

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 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'll find detailed discussions in our most recent SEC filings, including AMD's annual report on Form 10-K for the year ended December 26, 2004, and AMD's quarterly report on Form 10-Q for the quarter ended June 26, 2005.

With that I will turn the call over to Hector Ruiz.

The third quarter of 2005 was a strong indication of both the growth and earnings potential of this new AMD.

Solid execution across the board led to a quarter of records in a number of important areas, including total revenue, microprocessor units and revenues, microprocessor gross margin and operating income, as well as Flash unit sales.

We are pleased with the trajectory demonstrated by our Spansion Flash Memory business, which produced both revenue and operating performance improvement.

Strong customer adoption led to record MirrorBit sales in the period, up 35% over the prior quarter, and 49% over the prior year.

In the meantime, our Spansion team maintained their technical leadership by demonstrating working silicon on a 1 gigabit ORNAND Flash memory device based on 90 nanometer MirrorBit technology.

In our microprocessor business this was our ninth consecutive quarter of 20 plus percent year-on-year growth.

Once again represents the largest expansion since the launch of the AMD64 platform, up 44% over the same period a year ago.

With the help of our expanding set of world-class partners, we continue to gain momentum steadily across product lines and around the world.

In the enterprise we now claim more than 85 of the companies on the Forbes Global 100 as AMD customers, and more than 25% of the global 2000 are now buying AMD-based solutions.

The past quarter was highlighted by a Company record for both mobile processor shipments and revenues, a sign of growing worldwide demand for our AMD Turion 64 Mobile Platform.

AMD Turion 64 revenues were up 72% over the prior quarter.

And our special edition HP Live Strong laptop is proving to be one of the most successful co-marketing collaborations in our history with HP.

Less than five months after launching our initial dual-core server offerings we introduced three new models in the quarter, again, delivering the world's highest performance processors for 1-way to 8-way x86 service and workstations, while maintaining our clear performance for (indiscernible) leadership.

But technical leadership is only relevant if your customers can truly exploit it.

We were excited to see that according to the latest 4P/2P (ph) benchmarks the HP ProLiant BL585 and BL385 servers both based on AMD Opteron processors are the world's highest performing x86 4P and 2P database servers.

We were also pleased to help Sun Microsystems introduced their new Galaxy family of servers and workstations powered by AMD Opteron multicore processors.

According to Sun this new 4-way servers consume about one-third the power, produce 1.5 times the performance, and cost half as much as comparably configured Intel-based servers from Dell.

And we're particularly proud to have been named Supplier of the Year by Sun for extraordinary contributions to Sun's record of delivering top-quality technology and service to their customers.

We were also honored by the editors of Popular Mechanics who recognized the AMD Athlon 64 X2 Dual-Core Processor as one of only ten product winners for their 2005 Popular Mechanics Breakthrough Awards.

Finally, our Dresden team in Germany continues its excellent performance record, both in terms of the continued operation of our world-class 530, as well as in ramping our new state-of-the-art 300 mm Fab 36 on schedule and on budget, for which we are hosting a grand opening ceremony on Friday this week.

In summary, the third quarter was a very good one, characterized by strong customer adoption of our expanding AMD64 and MirrorBit product families, continued recognition by both industry partners and objective observers, flawless manufacturing execution against our MirrorBit and AMD64 road map, and improving financial performance in our memory business, and continued record-setting performance in our microprocessor business.

Also, we're confident the quarter represents a very healthy foundation for continued success.

I would like now to turn this to Bob to review the results of the quarter as well as the outlook.

AMD had an outstanding third quarter.

We had strong execution across the board with record-breaking CPG performance and significant improvement in our Memory Group's business.

Once again, our microprocessor business established all-time quarterly records in total sales and units shipped, server, mobile and desktop processor sales, gross margin, and operating income.

Total sales were $1.523 billion for the quarter, up 23% compared to the third quarter of last year, and up 21% compared to the second quarter of 2005.

We recorded net income for the quarter of $76 million, amounting to a fully diluted EPS of $0.18.

Third quarter gross margin was 41% compared to 39% in the second quarter of 2005.

This improvement was driven by record CPG gross margin and a near doubling of our Memory Group gross margin.

Third quarter total operating expenses, SG&A plus R&D, increased by 9%, while sales were up by 21%.

As a percentage of sales both SG&A and R&D expenses were down compared to the second quarter.

Cash flow from operations was $429 million for the quarter, and EBITDA was $387 million.

Now I will switch to the business segments.

CPG sales were $969 million, another record, and a 44% increase over the same period a year ago.

Compared with the second quarter of this year we increased CPG sales by 26%.

To highlight our growing momentum through the first nine months of this year our processor business is up 38% over the same period in the prior year.

Unit sales in the quarter increased by 41% over the same period a year ago, led by 100% increase in mobile units, and 144% increase in server units.

Total unit sales were up 27% compared to the second quarter of 2005.

Geographically we saw especially strong sales in Russia, India and Greater China.

And we continue to gain momentum in the enterprise market driven by AMD Opteron and AMD Turion 64 penetration on a global basis.

Despite the higher seasonal mix of mobile and desktop sales, we maintain a comparable ASP with the second quarter.

CPG's gross margin increased in the second quarter to a record high of 60%.

CPG's operating income of $209 million established a new high watermark in the quarter, up again from the record level of the second quarter of this year.

And with an operating margin 21.5%, this was the first quarter our operating margin broke through the 20% of sales level.

Now switching to the Memory Group business.

Flash memory sales were $516 million, down 4% from the third quarter of 2004, but up 12% from the second quarter of this year.

We shipped 15% more units compared to the second quarter, and ASPs have stabilized.

MirrorBit sales increased 35% from the second quarter on improved unit volumes and ASPs.

Much of the increased demand for MirrorBit Flash solutions came from large wireless customers.

We recorded a $50 million operating loss in our Memory Group in the quarter, a 40 million improvement from the second quarter.

Memory Group's gross margin nearly doubled in the quarter to 13%, reflecting record unit shipments, a richer mix, and better factory utilization.

Turning to the balance sheet.

Cash balances ended the third quarter at $1.342 billion, a $122 million increase from the prior quarter.

Overall, our cash conversion cycle improved to 50 days in the quarter, down from 71 days in the second quarter of 2005, and down significantly from 83 days in the third quarter of last year.

DSOs improved to 52 days in the quarter, down from 53 days in the second quarter, and down from from 58 days in the third quarter of 2004.

Weeks of inventory improved in the quarter declining 2 weeks to 13.5 weeks from the second quarter.

Now let's talk outlook.

AMD's outlook statements for the fourth quarter of 2005 are based on current expectations.

The following statements are forward-looking, and actual results could differ materially depending on the market conditions.

AMD expects fourth quarter microprocessor sales to grow between 7 and 13% sequentially, which is a 42 to 50% increase compared to the fourth quarter of 2004.

Because of Spansion SEC Form S1 filing, AMD is not providing guidance for the Flash Memory business.

Based on the continuing development ramp of Fab 36 and the seasonal fourth quarter marketing investments, we anticipate total AMD fourth quarter operating expenses, SG&A plus R&D, to increase by approximately 8% from the third quarter.

In summary, we are very pleased with our progress and look forward to a strong holiday season in the fourth quarter.

And with that I will turn it back to Hector.

I would like to close our prepared comments by offering a perspective on our business, and what we're achieving as we take AMD forward.

As you look at this Company I would encourage you to note that we are operating in the context of a few very simple guiding principles.

First, we never forget that our stakeholders are looking for sustainable growth.

We are aware that particularly today, great growth prospects are not easy to find.

And we know that not only all growth brings the same value.

There is opportunistic growth, the kind that can be produced relatively quickly at the risk of being difficult to reproduce.

And there is sustainable growth, the reversible advances with blue-chip partners, customers and users, most notably in our industry the enterprise.

At AMD we are committed to constantly improving both the depth and breadth of our customer and end-user base.

And as the quality of our customers will attest, we believe we're making significant strides to building a sustainable growth engine for the foreseeable future.

Second, we know that sustainable growth is the product of true innovation.

We believe that in the long-term our industry places a strong value on a steady stream of innovation.

Innovations like AMD64, MirrorBit and our ATM technology, which allows us to be able to turn on a dime in our world-class manufacturing operation.

Innovations built to create value, not just for ourselves, but for our customers.

We are proud of the fact that almost 14% of our microprocessor revenue in the third quarter came from our dual-core technology introduced in the latter half of the second quarter.

We're also proud to have introduced over 20 new processors in just this past quarter.

We know that the only way to keep our position as an innovation leader is to keep pushing ourselves and our technology as far end and as fast as we can.

And that investing in new, true innovation should prompt spending on building barriers to other success every time.

Finally, we believe that true innovation can come only from a customer centric approach.

If you look at every one of our recent product successes, they stem from our relentless commitment to customer centricity.

It is our customers who keep us focused on relevant innovation.

While those of you outside the Company see and evaluate us on the basis of the products and technologies we produce, what you cannot always see are the hundreds of changes going on across our Company and around the world to further our sales to continue to be the customer centric innovation leader.

The results today have been great.

And we're confident that they will only get better.

Once again I want to thank each and every AMD employee for remaining laser focused on the task of creating the world's most innovative semiconductor products and technologies.

In particular, I want to acknowledge the hard work and perseverance of our Spansion team in positioning that business for success.

Like our revenue, our progress in attracting world-class talent is accelerating.

And this is another reason why we're confident about the future ahead of us.

We remain highly motivated about the prospects of helping to create a marketplace that is better for our customers, partners and end-users around the world.

And that we're encouraged that our industry is showing signs of their willingness to break free from the artificial competitive barriers that have constrained all of us to date.

Thank you, and I would like to now turn it back to Mike Haase for the question and answer period.

We are now going to start the Q&A process.

As a reminder, because Spansion is still in registration, we will only address historical questions regarding our Flash memory business.

Back to you, operator.

(OPERATOR INSTRUCTIONS).

Tim Luke with Lehman Brothers.

Congratulations on the numbers.

With the operating margin moving up to these new levels of 21%, I was wondering whether you could give us a sense on the processor side of what some of the key factors were driving that improvement, and how you see the sustainability of that level.

Or whether, Bob, maybe you can give us some new metrics for looking at how your operating margins might move in terms of the upside for that going forward.

And secondly, I was wondering, as you just launched on Friday Fab 36, whether you could give us any color on how you see capacity growing there?

I think in the past you had talked about 15 million going to 75 million by the end of next year -- 50 million this year run rate.

And how quickly does it get to that number?

Thank you.

Lots of questions.

We're going to try to make sure that we don't forget them.

So let me start first of all by saying that quite some time back we committed to use as our goal was to operate in a business model, having a successful company, being able to return value to stakeholders by operating in the 20 plus percent operating margin as a microprocessor Company.

As we achieve that, of course, the key for us is to continue to significantly improve the top growth.

We believe that that is going to give us the opportunity to then invest the proper amount of money, while seeing -- still being able to give a very decent return to our stakeholders.

We are -- I will let Bob comment a little more about that particular business model, but we're very pleased with having reached some of those aspects already.

And we now have to focus on continuing to grow the top line so we can continue to spend the right amount of money to invest.

As far as the Fab 35 and the manufacturing capacity, you're right.

We are going out of the year at approximately a 50 million unit clip.

We are going to start the production in Fab 35 here in the first quarter.

The official opening is Friday.

And we will ramp it as fast as we possibly can.

We believe that puts us in position to significantly improve our unit making capability.

And as you stated, we think it could be 75 million, but it could be plus or minus depending on the market.

The next year it could be 100 million, plus or minus depending on the market.

And our desire is to, while ramping it as fast as it is physically possible, to make sure our customer engagement and marketing programs are aligned so that we can sell all of it.

And that is our plan.

That is our desire.

And that is, at this point in time, how we're marching to that.

And, Bob, would you like to answer?

Sure.

On a long-term basis, I will call it a steady-state basis, as Hector said, the goal we have to run the business in microprocessors is in that 20 to 25% zone, probably erring more on the side of the 25%.

Talking about -- in particular in gross margins of 55 to 60% of range.

Clearly we're at the top end of that range at this moment in time.

But the part I would caution a little bit as we transition into the new Fab 36 factory there will be some incremental costs associated with depreciation that will come online.

And so we will be challenged a little bit of maintaining the high 60% level as we go forward from a cost perspective.

Clearly though as we move through time and the units start matching the cost perspective, we will get back into the model of the high 20% kind of operating level and pushing the 60% envelope.

A critical piece, as you questioned, in drivers.

Drivers continue to have the right mix.

The right mix of products, of participating in all the segments, in particular in the server and thin and light mobile base, and then particularly in the segment I will call enterprise, of having both clients and servers in the enterprise space.

So those are kind of the key issues there.

As you see, as I have stated, we poked through the 20% barrier.

That is the first time we have ever accomplished that.

So we are well on our way to that game plan.

The first half of next year, well I will call it, be a little more challenged from a cost structure, because we will bring on the factory from that perspective.

But clearly for the fourth quarter if you have the sequential increase in revenue you would expect that the operating margin structure to -- and processes to move upwards again.

That's right.

Mark Edelstone with Morgan Stanley.

First off congratulations on the really strong results here.

Bob, can you just follow on there on the Fab 36?

When do you expect the depreciation actually to hit the P&L?

Is it still kind of middle of the quarter in Q1, or do you have a bit better handle on the exact timing?

It will probably -- and thank you for the compliment -- it will start more early on in the quarter than later on in the quarter.

Right now we're planning to start manufacturing fairly quickly in the quarter, with outfit (ph) toward the end of the quarter, beginning of the second.

So depreciation will start early in the process.

Kind of from a modeling perspective, I am kind of going to give everybody a little bit of guidance here, because it has been a big number.

So this will comprehend the incremental depreciation associated with next year for Fab 36, and the decreasing depreciation in Fab 30.

So I'll kind of give you the all-in net number for the microprocessor business, including the back end manufacturing operation.

Think of a number of about 150 million of incremental depreciation net for next year compared to this year for the microprocessor business.

Obviously, it will be heavier loaded in the back half of the year than the front half of the year, as we continue to deploy more tools to run manufacturing.

But hopefully that will help a little bit more.

That is great.

Can you state -- I know you were looking at whether you would stick with the five-year depreciation schedule in Fab 36?

Right now that is still modeled under a five-year scenario, though we continue to look at if we will make that change.

I would like to hold off, and I will actually announce exactly what we're doing at the analyst conference coming up in November.

Right now the numbers are I just quoted would be in the five-year vernacular.

Could you also on related to Fab 36 -- just give us a sense for what the total amount of expenses in R&D for Fab 36 were in the third quarter, and what the incremental increase was Q2 to Q3?

Most of the increase -- I'm not going to give you the totality of the number.

We haven't disclosed that.

But all the increase -- I would say 90% of the increase is associated with Fab 36.

So the increase we continue to see on a quarter to quarter basis.

Some of it is in people, of hiring new designers, but the bulk of the increase is associated with running engineering material in Fab 36.

I guess -- as that fab turns into production we're going to see those costs puddle in the COGS though.

So can you give us a sense as to how we should think about that in Q1?

Q1, there will be some proportion, I would call it probably less than 25% of the number will move up.

So it is a relatively small number.

The biggest number will be depreciation.

Okay great.

Thanks a lot guys.

And I will give you more granularity when we get to the analyst conference.

Michael Masdea with CSFB.

Just a quick question on the ASP.

You made a comment that was kind of flattish overall despite more Turion and service.

Can give us a little color behind that?

Sure.

As you noticed, we had a significant increase in both the desktop and mobile units.

And yet our ASP remained almost identical.

So that talks to the strength also of the server business in a quarter where typically we have a lot of demand on the client side because of the consumer aspect of our business.

I don't want to give you more granularity, but it was stable as a total, as an aggregate ASP, and it was stable also on a line by line basis.

There has been some commentary that the -- some of the OEM are less aggressive in the low end.

Maybe the clones are taking back from market share.

Any color on that or thoughts on that?

I'm sorry, I didn't get the first part of your question.

Yes, there's some commentary that the OEMs are less aggressive on the low end ASP boxes out there, and the clones are taking some market share back potentially.

Are you seeing that at all, or are you seeing any difference?

No, I haven't seen anything that would validate this assumption.

Great.

The last question is just emerging markets.

It seems like there is more and more growth coming from there.

Your competitor has pretty big wallets they are putting towards that effort.

Tell us about what you're doing to compete against that in the emerging markets?

First, those high-growth markets are slowly but surely consolidating under key players.

If you take the example of China, you know have really three or four local players that have the lion's share of the business.

Our strategy in those high-growth markets is obviously to partner with the key players in each of these markets, and also to partner with our multinational partners, who are important players, particularly in the enterprise segment.

So there is no special recipe, it is just being focused, being customer centric, being close to those markets, and continue to deploy our resources accordingly.

Adam Parker with Sanford Bernstein.

A little bit more color on the servers please.

Were your server ASPs up sequentially?

Our server ASPs were stable.

Even though your blended ASPs in server are at a premium to Intel, it seems if you compare apple for apple, 2-way to 2-way, 4-way to 4-way, etc., you are still selling them at a discount to Intel.

I guess what I'm asking should we think there's any reason to think there is pricing pressure in server, or is there still ample headroom for stability in your server ASPs over the next quarter or two?

I think that we have a superior product offering.

And I actually don't subscribe to your notion of a discount.

I think our competitor has a blended ASP that takes into consideration some high-end products.

As far as the 2P and 4P x86 space, I think at the processor level we're very comparable.

There may be pressure at the system level, but that is essentially because some of our competitors' customers have to discount heavily to sell the inferior technology that they offer.

So it is not true that you have a better mix within server than they do?

Clearly our leadership in the scalability of the architecture gives us a favorable mix of 4P versus 2P.

So when you look at the blended ASP for our server business, of course that helps because we have a very healthy 4P business.

Secondly, another question is about memory.

Hector, your attitude about Spansion seems to have, I think, improved over the last six months.

Not that you were ever negative about it.

But do you think is this is because of AMD Spansion specifically, or more broadly the Flash market?

And given there are whispers about that business being quite positive in the fourth quarter, is there any reason to think that a book (ph) way above 1 isn't achievable for AMD in the fourth quarter in Flash?

First of all, I just want to underline one point you made.

Is I had never been negative on Flash.

I tried to say that, but I meant you seemed incrementally more positive from a positive starting point.

Here's what we see happening as we talk to the leaders in Flash today.

Let me speak from the market perspective that the use of NAND has continued to grow.

And a lot of that is driven by new products that didn't exist before, for example, iPods who are heavy users of NAND technology, as one example.

And so that particular side of the business continues to grow, but it has not really had a significant impact on the outlook of NOR in the flash markets because what is happening is, as one example, when the year began you might remember that most of the consensus around the number of units built by cellular handset manufacturers were somewhere around 650 million units for the year.

And as we talk to these people, the leaders in this industry, what they believe now is there is very likely, very possible, that they will exceed 800 million for the year.

Which means the outlook for them improved significantly as the year began to develop, and these are all heavy-duty users of NOR Flash technology.

And also we pointed out in the written remarks, the introduction of the sampling of our first ORNAND product also indicates that our business is in a position to also be able to benefit from some of those additional market opportunities.

The market seems to have improved from what the users of technologies felt at the beginning of the year.

And I think that is what is giving people a little bit of a positive outlook into the balance of the year.

And I would just response -- in Q1 I think you said something on the conference call like the memory business -- the losses made you want to puke, I think is what you said.

So I was reacting to being more positive relative to that.

I am just trying to figure out how much was the market sentiment versus kind of AMD specific progress.

Okay.

It sounds like you think it is both.

Yes.

Joe Osha with Merrill Lynch.

Congratulations on the numbers.

A couple of questions.

Bob, you gave what I thought was a very helpful comment in terms of incremental depreciation associated with the processor business in 2006.

Could you perhaps take us to the same place in terms of what the incremental operating costs will be '06 over '05 for that business?

I'm not prepared to do that on this call, but I will at the analyst conference on November 15.

Would it be fair to say that as is the case with sort of the depreciation elements of the equation that perhaps, as you bring on cost to associate -- to support Fab 36 initially that you maybe drop below your target, and then come back to it as the year progresses?

Like I said, definitely challenged in the first half of the year versus second, as we try to match costs to incremental capacity.

But kind of as a framework, I will just talk in math and general numbers.

Because we would expect dilution in our R&D category and our SG&A category as time goes on proportional to sales.

Got it.

So you absorb more cost as the revenues go up.

That's right.

That makes sense.

To me -- we are on -- obviously with the kind of numbers I talked about and we have just reported, we are on a very fast growth rate.

And clearly we will not add costs at the same growth rate of sales.

That makes sense. (multiple speakers).

If I look at Fab 36 my own kind of sniffing around had suggested that that fab could be running around 2,000 wafers a week by the fourth quarter, which would imply actually that you're almost double in terms of square feet of silicon where you are right now.

But it sounds like I'm a little off there.

Hector, can you maybe tell me where I am making a mistake?

That is not out of the question.

And I just would like -- when you said fourth quarter, you meant fourth quarter next year I'm sure?

Right, 2,000 times 2.5.

And I know you are running around 25 K. at fab 30 right now.

So that basically that doubles you year-on-year.

That is not entirely out of the question.

And we are going to do everything we can, as I said, to ramp as fast as is physically possible.

Super.

And then last question, back to you, Bob, just in terms of the -- I have beaten up Hector about personal connectivity losses, so I'll just maybe ask again.

But then also you've got another $66 million loss in sort of other products.

Can you tell me what the story with that is?

As I think we have stated before, but I will just restate it.

The other products represent costs that we do not allocate to the product lines.

In particular, for AMD, that is mostly centered on long-term and short-term bonuses and profit-sharing associated with how well the Corporation or individual people or business units do.

And so with the large improvements of profitability in some of our businesses has dictated that we actually need to accrue quite a bit of bonuses.

Now does this recur, and is it attached to the processor business or the Spansion business?

The honest answer, of course, most of it is coming from the processor businesses because the size of the operating income.

We don't pay many bonuses for losing money.

But as we continue to take the 200 to 400 to 500 to whatever the numbers are, bonuses would move proportionally to that.

So I need to think about the $66 million as being attached in part to the processor business, and as recurring as you continue to be more successful in that business?

Correct.

And then of course like all bonuses, I'm sure just like yours, they get reset every year.

No, mine never goes up.

All right, thank you very much.

Ben Lynch with Deutsche Bank.

Just joining the crowd to congratulate you for what it's worth.

First question, maybe this is for Henri.

And I know the final data is not out, but I'm sure you have a rough feel for this.

Your estimate for you 2P and 4P server share for the -- currently -- define that as you like, currently.

And I'm assuming 4P is a lot higher than 2P.

What do you attribute that to, I guess, specifically just the scalability advantages?

And is there -- do you have specific plans to boost the 2P share significantly in the next year or so?

That's the first question.

First, I am not going to make guesses.

We will wait to see -- as the dust settles on the third quarter and where the shares are.

I keep my horizons focused at about six to nine months.

So on my pipeline -- my pipeline and opportunities is very strong and continues to grow.

So I'm comfortable with the velocity of our server business.

With regard your specific comment on 2P versus 4P, clearly our scalability advantage on the 4P space makes Opteron a particularly compelling solution in that space.

Obviously, the 2P space is more of a commodity.

The 1P space is actually a total commodity.

And we have a lot of opportunities to grow in those two segments.

Frankly, the opportunities there are available to us as we continue to augment the number of platforms available from our Tier 1 partners, and as we develop our offering in the SMB space.

I think that the more you move towards the commodity market, the more it is about market coverage.

Clearly in the 4P space you are confronted with very, very smart and informed buyers who can understand what is the best value proposition in the marketplace.

But I'm confident that we will accelerate in the 2P space and then in the 1P space as we continue to augment the number of platforms available.

It will be less -- if I hear you properly it is maybe less the technical arguments, which clearly there are as well, but may be a bit less emphasis on just going in and penetrating your customers.

Absolutely.

Okay.

(multiple speakers) market coverage.

And the second question I had was going into the quarter on your Q2 comments AMD was talking a little bit about supply constraints, and also how the higher consumer mix in the second half might be somewhat of a drug on margins.

And clearly there are constraints which are facing third-party chipset vendors and Intel, etc.

It would seem that they would apply to you as well because you have the same sort of partners.

And yet you had this phenomenal quarter.

I'm just trying to understand how those supply constraints that seem to be forming just didn't seem to matter?

First, we haven't experienced any service sort of supply constraint during the quarter.

I want to dispel any notion that that is the case.

With regards to the fact that our results are in the upper end of the range that we could expect -- or you can expect-- it is really a statement of the brand positioning that we have done over the course of the last year and a half.

The Athlon 64 brand is really doing very well.

The Athlon 64 X2 is doing exceptionally well.

And Turion 64 is doing very well.

And so the establishment of those brands and how they are anchoring themselves in the market and the price tags are obviously helping us in getting those phenomenal revenue growth.

Glen Yeung with Citigroup Investments.

Just a question about your fourth quarter guidance.

It looks pretty much in line with normal seasonality.

I'm wondering if you could explain to us, given that you have been obviously taking share now for some time, if you believe you are no longer taking share, or if you believe that we may in fact have a lower than normal seasonal fourth quarter?

It is an interesting question for the following reasons.

I believe that we -- it is very tough and challenging for -- at our side frankly to be as predictable in terms of the total market.

There are probably other people with a significantly larger share able to see that and have more visibility on to that.

We believe we will outperform the market in general.

But we also believe the fourth quarter seasonality numbers are beginning to be difficult to predict, because we find ourselves in a different pattern that we see the rest of the competition.

And so it is hard to blend it.

From our perspective it was best for us to indicate to you that the very strong year-on-year growth that we have had that we expect to continue.

To add to your point, I guess that year-on-year growth you are expecting in the fourth quarter, actually exceeds the year-on-year growth you have had in the first nine months.

So maybe some conservatism in the way you thought about the quarter?

No, I think is reasonable.

Okay.

(inaudible).

Fair enough.

And the other question I had, maybe for Bob, was in the second quarter you had seen some inventory build and made the point at the time that that was -- given that you are looking forward to what was a good third quarter, and you obviously gave us that, that inventory build was justified.

When we look specifically at the inventory of Opteron, whether that is (indiscernible) or finished goods, does your Opteron inventory come down in the third quarter amidst a 2% increase in overall inventory?

You've got to remember the inventory report is for both businesses, both Spansion and (indiscernible).

And total I actually did not make comment on the specifics, but we -- hopefully we continue to show you -- we demonstrate strategically building the right inventory and then moving it appropriately.

Opteron continues to set records every quarter.

Just because we don't talk about it as much in any particular tone, doesn't mean every quarter is not a new record.

We're very happy with our inventory levels.

Of going into the fourth quarter, as we have stated in our forecast, we expect decent growth in the fourth quarter again.

And so it will continue from that standpoint.

I am sorry, Hector, just to go back to the question I was asking before.

Just to be specific on this particular fourth quarter, given we have so many macro issues that are out there, is your sense that -- are you trying to factor in the concerns we have on a macro basis?

Is that part of your thinking in this calendar Q4?

No, no.

It doesn't mean we don't care about the macro issues.

We are aware of them, but frankly they are beyond our ability to forecast, and so we are not taking those into account.

Christopher Danely with JP Morgan.

Great quarter.

Can you just talk about the relative sizes and the growth rates of the three different processor segments, server, desktop and laptop?

Not more than the color we gave you in the statement, which is the growth was led by Turion 64 this quarter.

But that is not to say that the other segments didn't grow very well as well.

They were all in the double-digit mode.

And again a spectacular performance from Turion 64.

And then how about going forward, can you give us a sense of do you expect the laptop segment to drive growth going forward, or could we see something else pop up?

If you look at our three segments, our largest market share is in the desktop.

So obviously although we are performing very well, we have more opportunities to grow in the server space than in the mobile space in relative basis.

Although, as Bob was pointing out, there's a huge opportunity for us as we continue to develop our business in the client space for the enterprise.

But overall I would say we would continue to expect record after record on both mobile and server, with the desktop doing very well, but probably as a second -- as a second growth rate.

Was there any key metrics that caused the switch flipping in the laptop segment?

Well you know we introduced the Turion really at the end of the first quarter.

Our customers ramped production of about 60 different platforms in the second quarter.

By the time these hit the shelf for back-to-school for the consumer space and started to be evaluated by Fortune 500 companies in the commercial space, if you look at the cycle, it is totally logical.

But it also shows that the market needs an alternative in the mobile space, and that Turion 64 provides both our customers and end-user with a very high-quality alternative to the competition.

And then last question.

I know you guys can't talk about Flash going forward, but can you just address the ORNAND market a little bit?

Doesn't Samsung have a similar type of product, and does anybody else have a similar type of product?

There might be some attempt to make it looks similar, but I'm not aware of anything that looks at all like what ORNAND is.

And just to remind you a little bit is the NOR technology, which is based on this thing called MirrorBit, which is very unique, and only AMD has the IP portfolio to be able to do it, it is a derivative of that leads to this technology called ORNAND.

And there is no such thing as an ORNAND market, because there is -- only AMD owns that product.

But it is a NAND-like application based on technology that is a derivative of NOR.

And that, as perhaps we stated before, brings with it a very cost competitive space, as well as a very high quality in terms of data retention.

In some instances it will take away some of the market that NAND might be approaching.

In other cases it will be a lower cost alternative for some not so demanding NOR solutions.

That makes sense.

Thanks a lot guys.

Great quarter.

Jim Covello with Goldman Sachs.

I am trying to get a little more granularity on the fourth quarter gross margins.

Bob, you had given us terrific color for the out year margins.

But relative to the fourth quarter, maybe if I could start, is there an update to your full year depreciation and amortization guidance?

No, pretty much the same as we have had from that perspective, not materially different from the prior annual guidance.

If I'm not mistaken that was kind of 1.2 billion.

So you would actually expect depreciation to come down a decent bit in the fourth quarter?

Depreciation will trail in that direction -- constant the trail.

Again you also remember depreciation is a factor of currency rate.

Sure.

And could you give us any other Caller on the factors that might affect the fourth quarter gross margin margins specifically?

As Henri stated, seasonally fourth quarter is a client-based quarter.

So clearly clients will be a strong piece of the business from a growth perspective, and a percentage of the business.

But as he also said too, we're continuing -- we have a healthy pipeline of servers, and those will continue to kick on on a continuous basis.

So we see no issue in ASP.

ASP will continue to move as it has historically, trending north.

That clearly helps gross margin.

And we have, as I said, we will have cost increases associated mainly with Fab 36, and merchandising the product for the fourth quarter.

So we should (multiple speakers).

So when you net that all out, I think the gross margin drop through was about 50% or so this quarter.

Should we expect that to remain about the same or go up or go down for the fourth quarter?

I think it is in the relative zone.

And then second question, relative to the emerging market impact on the business, do you think that has any change on the seasonality that we may see?

A little less pronounced positive in the fourth quarter, a little less pronounced negative in the first quarter of the year, or the first half of the year, given the differences in the seasonal patterns in the emerging markets?

Yes, I think that actually it is going to help us overall.

Because if you look at the cycles, they are different in those markets.

Both vacations and holidays and actually important dates.

As our business in the past was heavily dependent on both the U.S. and Europe, we have the seasonality that was impacted by really the retail cycles in those two markets.

Clearly the high-growth markets are representing an increasing percentage of our total revenue.

And that should smooth out the curve.

That is partially why Hector indicated that we're moving a little bit away from the traditional seasonality that has been ours for the last few years.

Because there is a significant change in both the geography mix of our business, but also our consumer to commercial ratio.

David Wong with A.G. Edwards.

A couple of points -- quick sort of administrative things.

Can you give us some idea of what tax rate would be for AMD if it didn't have Spansion with it?

Are there any net operating losses that would go away with Spansion that would force the tax rate up?

Actually, no, Spansion today is a partnership.

And actually in the partnership all the losses accumulated to date stay with AMD.

The benefit accrues to us.

Along the same vein, legal costs, would there be any step up going forward, or does this current quarter's SG&A and your guidance for next quarter sort of map out what we would expect into next year?

Or is there any point where things step out, or alternatively step down?

I haven't given the granularity on next year, but as I stated in the last conference call, I always will give you what we think it's going to grow quarter on quarter.

The 8% again growth rate from third to fourth is what I expect.

And I will give more color and detail and specific at the November analyst conference.

My final question.

I think Hector had said there were no shortages of AMD parts.

At this point in time what does your capacity utilization look like?

Do you have -- is it relatively full utilization?

Yes, the question is -- the simple answer is yes, but I have got to modify it.

Let me explain to that that we continue to be ramping.

The capacity of Fab 30, even though has fixed number of wafers, has continued to go up because we have improved the node of technology from 130 to 90 nanometers, and we have been able to add productivity changes which have improved also the capacity.

So that is this discontinued increment of capacity that has been occurring for the last several quarters.

Of course, that will continue not only in Fab 30, but it will continue even at a faster pace as we bring Fab 36 on board.

And our plan plans are, with Henri sales and marketing team around the world, is to sell it all-out.

And therefore at every quarter in time, even though we improve capacity, is to be able to be as close as we can to full utilization, where we are right now.

The rate at which your computational revenues have grown, for example, is roughly equal to the rate at which the market has grown.

Then you haven't been production limited in terms of what you can sell?

No, we have not.

Allan Mishan with CIBC World Markets.

Nice quarter.

A couple of questions.

Bob mentioned before 150 million in depreciation incremental.

Can you just split that first half versus second half, so we understand how that hits the P&L?

I'm not prepared to do that right now.

I will do that at the analyst call.

Fine.

And then do you guys have any anecdotal evidence or any other evidence that you might have benefited from the Intel chipset shortage this quarter?

And if so, do you worry that that share goes back to Intel at some point?

Our understanding is that our competitor is challenged from a manufacturing perspective in certain areas.

In the case of chipset it is the low end of their offering.

As I pointed out, we're doing very well on Turion.

And mainly the growth of our mobile business was driven by Turion.

So whether we had some benefit in our Sempron mobile offering of Intel's failure to execute remains to be seen.

But I'm not worried about their overall ability to fight back.

Because again, in the space where they haven't announced any sort of shortage or issues is where we performed the best.

Chris Passal with Friedman Billings Ramsey.

I just wonder if you could give some more color regarding the Turion ramp?

And could you give us some sense of what -- where Turion sits as a percentage of your mobile business right now, and where you think that is likely to be as you go into '06?

I really would prefer to stay away from that level of granularity, but obviously Turion 64 is the only 64-bit platform available.

We intend to drive as big a percentage of the total mobile business as we can.

But we also recognize that, and particularly in a lot of those high-growth markets, there's a need for an entry-level solution.

And that is why, again, we did very well with Turion, but I don't want to have you leave with a false impression that our mobile Sempron business didn't do well as well.

Let me follow it with something else then.

Regarding just on follow up to an earlier question that you guys talked about the capacity utilization, the plan to basically sell out all your capacity.

As things go forward with your competitor being tight, I guess is it right to say that you guys are making some decisions in terms of what business you're actually going to take in the quarter, given the capacity constraints?

Or is -- there's some other things that you guys could think of such as outsourcing?

Maybe just give some color on that?

You know what, the only thing I would like to comment on outsourcing is we have announced plans to use Charter Semiconductor on some foundry.

We are very happy with that.

We expect that to complete in the second half next year.

That will be part of our desire to aggressively grow in the market and use the capacity fully.

Other than -- I will let Henri comment on the other question you had relative to whether we actually are selective on the businesses that we take or not.

But we right now have a strategy from the beginning to serve the enterprise market with servers.

We aggressively are moving to the client to continue to do well in the consumer side, where we have been very strong, and also begin the path to significant penetration on the commercial space.

That has been demonstrated over the last several quarters that we are making progress on and where some of the growth has come.

That will continue.

And, Henri, anything you would like to add to that?

No, just Hector that like in everything we do we are being customer centric in how we look at the portfolio of products that our customers want and need.

And sometimes there mix and our view of the mix is slightly different, but we usually reconcile and find a way to have a win, win situation in the market.

That is why we have been able to maintain such high-growth rates with good ASPs across the three product segments.

We're not doing any arbitrary decisions.

We're working with our partners to make sure that end-users around the world get the AMD technology they desire.

I guess the follow up of that, in the most recent quarter is there any business that you guys have turned away?

Not that I am aware of.

Operator, we going to take two more questions please.

John Lau with Jefferies & Co.

I was wondering if you can comment more on the profile of strength in the emerging markets, is it more high-end or low end, or is it similar to the profile of the U.S. markets?

The only granularity I will give you is actually we're seeing an improvement of ASP in the high-growth markets that is faster than in the mature markets.

Of course, coming from what was traditionally a much lower ASP situation.

So we are seeing a mix of products in the high-growth market that is getting increasingly favorable.

As a follow-up, there has been -- I was wondering if you can comment also on the reception for your Opteron X2 product, especially on the server sides with the dual-core?

And when do you believe the operating system will be available to take advantage, full advantage, of those dual-cores inside?

First, there is no such thing as an Opteron X2.

There is an Opteron dual-core for a server and workstation, and there is Athlon 64 X2 for desktop.

So is your question for the server market?

Server.

For the server the operating systems are available.

And most operating systems today are taking full advantage of multicore technology.

At the application level there is an increasing level of applications that can take advantage of it.

But of course as you know there is long development cycles to optimize compilers.

The thing, you know, the scalability of our architecture is such that customers get immediate benefit.

I notice that our competition just announced a product that they claim can scale up to 50%.

It is interesting to know that our product scales from 50 to 90%.

So there's benefit immediately available through Opteron for users as long as they use the right operating system and the right applications.

In terms of what we want to look for on operating systems of desktops, is that what we look for in Vista (ph) next year?

Is that what will take for us to do -- take advantage of that on the desktops?

I think on the desktop you'll have to rely more heavily on application tuning than simple operating system tuning.

Of course the operating systems can take some partitioning of the work, but really you're going to have to get to the application level, because you're not running as many threads as you are on a server.

Our final question will be from Michael McConnell with Pacific Crest Securities.

Looking at the margins in Flash, it looks like -- if I looked at it year-over-year revenues right now roughly the same, and operating margins slightly below -- I figure about 3% operating margins last year, negative 10% in Q3.

Can you talk about with MirrorBit up 35% sequentially, why we're not seeing the Flash margins improve off the same revenues year-over-year?

There are a number of ways to look at that.

First of all, if you look at the third quarter performance versus the second quarter performance you see a rather significant improvement in the operating results of Flash, given a rather stable pricing environment.

Which means that if you look at the incremental sales in the (indiscernible) which was roughly about 75%, it means that the business unit, the people in that business has done a damn good job of managing through these very difficult times.

And that to me is an indication of the potential these units have as the top line growth will improve over time, as you would expect it.

So I believe that the comparison of second to third quarter is a good example of what the potential of these units is, and it gives you an indication of the strength of the product.

And as this MirrorBit technology continues to expand, which is mostly going to be based on the next node of technology at 90 nanometers, it is reasonable to expect the continuation of that improvement.

What percentage of Flash was based on MirrorBit this quarter?

Can you break that out?

It's about -- yes, I will give you that number.

It is roughly -- we have talked about this and it has continued to move up, and it is approaching in the 25% zone.

Thank you very much.

We look forward to seeing everyone on November 15 at our Analysts Day.

Thank you.

Thank you.

Ladies and gentlemen, that does conclude our conference for today.

Thank you for your participation and for using AT&T Executive Teleconference.

You may now disconnect.