超微半導體 (AMD) 2005 Q1 法說會逐字稿

Ladies and gentlemen, thank you for standing by and welcome to a AMD's Q-1 '05 earnings conference call. (OPERATOR INSTRUCTIONS).

As a reminder, today's conference call is being recorded.

I would now like to turn the conference over to your host for today -- or opening speaker -- Director of Investor Relations at AMD, Mike Haase.

Please go ahead

Thank you, everyone.

Welcome to AMD's first quarter earnings conference call.

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

Bob Rivet, our Chief Financial Officer; 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 776654.

The telephone replay will be available for the next 10 days, starting at 7 p.m.

Pacific time tonight.

For your planning purposes, I want to call to your attention two upcoming AMD events: On April 21 we will be hosting our second AMD Offline Anniversary Event in New York City.

And that's next Thursday, April 21st.

Also, on June 10, AMD will host a technology focused Analysts' Day, also in New York City.

For that event, we'll be sending out invitations next month.

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 current expectations.

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 risk factors and our business [inaudible - background noise].

You will 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.

With that, I will turn it over to Hector Ruiz

Thank you, Mike.

The first quarter of 2005 was a continuation of the same two themes that defined our fourth quarter of 2004.

We saw strong momentum across our microprocessor business, leading to a quarter of record sales and record operating income for AMD on that side of the business.

And we saw continued challenges in the flash memory market, including a seasonally down quarter, a continued state of over supply and strong pricing pressure.

We have taken aggressive action to improve both our immediate and long-term prospects for our memory business.

First, through both process and design enhancements, we have brought in the performance footprint of our 256 megabit MirrorBit product and expanded shipments to include three of the top 10 wireless phone manufacturers.

We experienced a rise in [inaudible] unit shipments in the quarter, a sign of steady industry demand for our new flash memory products.

In addition, we achieved substantive revenue for our 110 nanometer MirrorBit products.

Second, Spansion filed an SEC form S-1 today.

This is a continuation of a strategy that we began putting in place over two years ago when we recognized the need to enable completely different business models in each of our two businesses, as well as the need to expand our options to fund this capital intensive operation while enhancing AMD shareholder value.

After filing the S-1, we have undertaken a number of actions to significantly reduce our expenses, including the streamlining operations, continuing to align manufacturing utilization to our level of demand, and working to reduce costs associated with AMD and Fujitsu service agreements.

As a result to this filing, SEC rules place limitations on our [inaudible - highly accented language) concerning the flash business in this session.

Our processor business is not just growing.

It's accelerating, thanks to the strong adoption of the AMD 64 processor platform across channels, segments, and geographies.

The first quarter of 2005, processor revenues were up 31% year-on-year, and sales of both our AMD Opteron and AMD Athlon 64 processors, more than doubled from the period a year ago.

As a reminder, the growth rate is actually higher than our overall processor growth rate in 2004, and we have grown faster than the processor market in both 2004 and believe we have in the first quarter of 2005 as well.

We are particularly proud of AMD 64 adoption among our strategic OEM customers.

In the first quarter, Sun Microsystems launched their second generation V20z and V40z servers, featuring AMD Opteron processors models 252 and 852.

HP expanded their Enterprise class offering with the addition of the ProLiant DL 385, the BL 25 p, and the BL 35 p server blades and the xw9300 Workstation.

Based on the strength of our AMD Opteron brand, we continued our unprecedented growth in the server and enterprise segments, adding new global enterprise class customers, including Akamai, Cable and Wireless, EDS, Goodyear, Honda, Lucas Films, MBNA, Met Life of Mexico, and Syntel, among many others in the first quarter.

We now claim that over 55% of the companies in the Forbes Global 100 have satisfied AMD 64 customers.

That is up from 40% last quarter.

AMD 64 base processors now represent 63% of our processor business, the result of expanding demand, continued strong yields, and world-class manufacture and execution.

By the end of 2005, we expect nearly 100% of all CPG processors to be based on our industry-leading AMD 64 technology.

Our transition to 90 nanometers is ahead of schedule with better-than-expected yields, resulting in continued capacity expansion to meet the growing demand for AMB products and solutions.

Our 536 remains on schedule and on budget.

And as a matter of fact, [inaudible] silicon started this quarter with production plan in the first half of 2006.

We continued our innovation leadership in the processor industry with the first public demonstrations of AMD 64 dual-core systems on existing Sun, Cray, and HP servers and workstations.

As some of you may know, we are ahead of schedule on our dual-core AMD Opteron processor.

In fact, we have been shipping samples to customers and partners since January, and unlike our competition, AMD 64 was built from the ground up to be multi-core.

And unlike our competition, AMD 64 dual-core fits inside the same thermal envelope as our single-core processors.

Therefore, in response to customer demand, and because of our rapid qualification cycle, we can confirm that we will indeed be launching our highly-anticipated AMD Athlon dual-core processor, the industry's first x 36 dual-core server processor, in New York next Thursday.

We'll also preview our AMD Athlon 64 dual-core processor for desktops and notebooks.

With AMD 64 dual-core technology, consumers and enterprise customers are going to be able to use their PCs, servers, and workstations in ways that they have never before.

And to complement our growing server leadership, we're expanding our footprint in the enterprise with the launch of the AMD Turion 64 mobile technology, the first 64-bit processor family designed specifically for thin and light notebooks.

Finally, we're thrilled that there [inaudible] manufacturing by Microsoft of their Windows XP Pro 64-bit edition was announced in the last couple of weeks.

The first customer reactions and the reactions of those developers that have downloaded this edition believe it to be outstanding and phenomenally robust.

And I would like to take the time now to thank, personally, our team on Microsoft, our partners and friends who have worked so diligently and hard to be able to make this a reality.

We continue with the great support of our partners on a steady path toward re-inventing the competitive dynamics of the microprocessor industry.

Our Spansion team has taken swift and effective action to position themselves for immediate success, and we remain and plan to create the conditions for sustainable success in Spansion with the filing of the S-1.

At this point, I would like to ask Bob to review the results of the quarter as well as the outlook.

Thanks, Hector.

As Hector pointed out, the first quarter was a quarter of contrast across our two major businesses.

Unfortunately, the negatives more than offset the positives, and we were disappointed in our results.

For AMD, sales were $1.2 billion, essentially flat with the first quarter of last year, and in a seasonally slow quarter, down slightly, 3% compared to the fourth quarter of 2004.

We recorded a disappointing operating loss for the quarter of $46 million, due to the performance of our Memory business.

Our loss per share was $0.04.

Gross margin declined to 34% for the quarter, below the 41% we reported in the fourth quarter and a 3.6 percentage point decline from the first quarter of last year.

This decline was driven by a combination of very competitive flash memory pricing, intentional factory slow downs in the flash memory business, and to a lesser extent, the Euro foreign exchange effects.

Research and development spending was $253 million in the quarter, flat from the prior quarter.

As previously guided, first quarter marking general and administrative costs decreased 14% as compared to the fourth quarter of 2004, closer to third quarter, 2004 levels.

Cash flow from operation was $251 million for the quarter, and EBITDA was $331 million.

Now, I'll switch to the business overview for the quarter.

I'll start with computation product group.

In a typically seasonable down quarter, CPG sales were $750 million, a new record, and a 31% increase over the same period a year ago.

It was a 3% improvement from the fourth quarter.

Unit sales in the quarter increased over the same period a year ago by 28%, and we saw strong demand in high-growth regions, particularly greater China.

Overall, ASPs increased compared with the fourth quarter, and a new sales record were established both our server and mobile processor segments.

CBG's operating income of $92 million established a new high watermark in the quarter, up from the record levels of fourth quarter.

This represents the seventh consecutive quarter of profitability for CPG.

CPG gross margins slight -- decreased slightly from the fourth quarter, but with a healthy 54%.

Now, switching to our Memory business.

As you know now, Spansion filed a form S-1 today and as Hector outlined a few moments ago, we are executing a strategy on which we embarked on two years ago to make Spansion a financially independent organization.

This will allow us to create a consistently successful flash memory business.

Flash memory is a a capital intensive business and our strategy is designed to give Spansion direct access to the capital markets going forward.

In a typically seasonable down first quarter, flash memory sales were $447 million, down 29% from the first quarter of 2004, and down 11% from the fourth quarter of last year.

Despite industrywide oversupply, we shipped approximately 10% more units compared to the fourth quarter.

We recorded a $110 million operating loss in our memory group in the quarter, a $71 million greater loss than we recorded in the fourth.

Through active expense management, first quarter spending was down from the fourth quarter, but this was more than offset by ASP decline and planned factory slow-downs.

Our memory group's gross margin declined significantly in the quarter, approaching zero.

Turning to the balance sheet: Cash balance was ended the first quarter at $1.1 billion.

First quarter capital expenditures were $518 million, up from $470 million in the fourth quarter.

We managed inventories down for the fourth quarter by $29 million, due to planned factory slow-downs in our flash memory business.

Now let's talk about the outlook. [Inaudible] are based on current expectations.

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

In a typical seasonable down second quarter, AMD expects processor sales to be flat or down slightly.

Due to Spansion's filing of the SEC form S-1 today, AMD is not providing guidance for the flash memory business.

Second quarter operating expenses for AMD, which include research and development, marketing and general administrative expenses, are expected to increase by approximately 5% in the second quarter, mostly due to Fab 36 startup costs.

In summary, we're pleased with our computation -- for CPG performance and look forward to executing on the strategy for flash memory business over the coming quarters.

With that, I'll turn it back to Hector.

Thank you, Bob.

Let me close by make a few comments on AMD's growing role as an innovation leader in our industry.

Those of you who have followed us for awhile know that AMD has always been a fiercely competitive company.

Many have said that we're the poster child for competition, and we're proud of that.

We also hope that you have noticed that with less than 10% of the resources of our competitor, we have methodically seized the mantel of innovation leadership in our industry.

Our flawless conversion to 90 nanometer is our most recent example of innovation in manufacturing.

Thanks to our automated precision manufacturing, our APM software, we can very efficiently ramp to mature yields leading to reduced time-to-market [inaudible] technologist.

Meanwhile, we have extended our innovation leadership into the design area, employing a customer-centric approach to design that has resulted in innovations such as PowerNow, the advanced power management technology that results in cooler microprocessor systems, years ahead of the competition;

Hypertransport, our industry standard high through-put chip-to-chip interface; our Direct Connect Architecture, AMD innovative approach to system architecture that eliminates the bottlenecks that would otherwise hold back performance improvements in 64-bit and especially in multicore processor implementations; and of course, AMD 64, the first seamless migration path to provide the 64-bit processing power, and the only one designed with multi-core in mind.

The industry has taken notice.

Last month, AMD was the processor company on hand representing the hardware industry for Microsoft's partner, Summit, celebrating the launch of their 64-bit OS with the developer community.

Next week, AMD will keynote the National Association of Broadcasters' conference and demonstrate how the power of our 64-bit technology allows the nation's top graphics and recording artists to be more creative and more productive than ever before.

We will also celebrate the second anniversary of AMD Opteron, where we will officially confirm our time-to-market leadership in the deployment of x86 based dual-core micro processors.

We are leading the industry by bringing new ideas to market, we're turning technology into concrete benefits for our customers and their end-users, and we're gaining the recognition of industry leaders for the value of our innovations that they'll bring to the market and to end-users.

But for our Company to be able to advance our innovation leadership position, and for end-users to gain the benefits of these innovations, we must compete in an environment of fair and open competition, where our and their customers enjoy the freedom to choose the best technology in the market, without the intervention of illegal business practices.

And that is why we applaud the work of the Japan Fair Trade Commission in their recent decision to condemn anticompetitive behavior in our industry in Japan.

From this point forward, we are expecting a much more balanced competitive playing field in Japan, and we're hopeful that regulators and business leaders around the world will follow suit to protect their consumers and create an environment of freedom of choice and fair and open competition in this highly competitive crucible of innovation.

Thank you, and I want to turn this back to Mike Haase for the Q&A.

Thank a lot, Hector.

For the Q&A session, any flash questions regarding the first quarter, we will happily address.

However, due to S-1 filing, we will not answer any forward-looking Flash Memory questions.

With that, I'll turn it back over to the Operator and we'll begin the Q&A session.

And thank you. (OPERATOR INSTRUCTIONS) Our first question comes from the line of Joe Osha with Merrill Lynch.

Please go ahead

Hi, folks.

Congratulations in filing the deal.

A couple of questions.

First, Hector, can you give us a little more color on the transition to 90 nanometer and maybe, if you can, comment on 60 -- what 65 nanometer looks like in terms of what the plan for starts and outs is.

Okay, Joe.

Let me start by saying that, you know, we had a plan to achieve complete conversion by the end of the second quarter in terms of ways for starts.

We're very much on target to do that.

We expect to accomplish it this quarter.

The one area where we have experienced a -- better than we have planned is in the area of defectivity.

For those of you in the industry, that turns into yields, and we had hoped to reach defectivity levels at the point in time that we're competitive with 130, then that actually, we ended up doing that considerably ahead of that.

And so we're very pleased with those -- the yields as a result of that.

We had previewed that, as I said in my remarks, to the automated precision manufacturing technology that we've developed for many years.

So, we're pretty -- very much confident that we're going to get there here pretty quick.

We're talking about just a few more weeks before we're totally converted to 100% away for starts in 90 nanometers.

In terms of 65 nanometers, the work we're doing is jointly developed with IBM.

We have been running now engineer and silicon use and study grants to provide the technology.

We're very excited about the progress we're making there, and we expect that early in the life cycle of Fab 36, which is expected to be in the first half of next year, we will be building some 65 nanometer product.

I can't tell you anymore than that now because, frankly, it's a lot of customer and product [inaudible] that we're making and it will be later on in the year where we will be able to have more clarity on that

Okay, thanks.

Just two quickies and then I'll go away.

First, can you say anything qualitative about Opteron, either revenue or volume?

And then secondly, question for Bob.

I see there are still some pretty substantial losses attaching to this personal connectivity business.

Do your comments about reducing costs in any way apply to those?

And what should we expect in terms of the operating losses there?

And that's it for me.

Joe, this is Henri.

With regards to some color on Opteron, I'll just tell you that we grew both revenue and units double digits.

Is that sequentially?

Yes.

All right, thank you

On your final question, Joe, on the PCS business, at the $30 million, you can see our cost structure is clearly well north of that.

The key issue is volume.

We continue to be -- have favorable press and opportunities for presenting ourselves in the PC area. but that's the key to that business, turning it around.

We're comfortable with the level of investment we have at this point

Okay, thank you very much.

Thank you for your question.

Our next question comes from the line of Adam Parker with Sanford Bernstein.

Please go ahead.

Hi.

A couple of questions here.

I'm not sure I understand the moving parts for your operating margins in the CPG.

I mean, it looks like you're, what, 12.2% here, down from 13.2%. 13.2 a couple of quarters ago, despite the additional revenue.

So I'm trying to understand the moving parts.

Is it mix or pricing, given you said you had record units, server, notebook, [inaudible] ramp, better yields, better -- how do I think about the operating margins not going up here over the last couple quarters?

We are building Fab 36

So it's just depreciation and other costs -- startup costs?

It's not depreciation --

Well, it's 4 million

It's just pure startup costs associated with starting up a brand new 300 nanometer facility of hiring people and putting in chemicals and gases, et cetera

All right, so, given you said revenue's flat to slightly down in Q-2, and you said 5% up, ex growth, should I assume that your CPG operating margins are lower again in the second quarter?

I don't want to give guidance in either direction there

But your revenue's flat-to-down and your costs are up.

Is that correct?

That's exactly what we said.

Yes

Okay, and lastly, what do you guys plan on doing with your tools in Fab 36?

A couple of years from now, once they're no longer sort of suitable for making microprocessors?

I think you probably meant Fab 30

30.

Sorry.

Yes.

Yes.

You know, we actually feel very optimistic about having such a wonderful facility.

First of all, it's a world-class manufacturing facility with great people and as we expand the footprint of x86, which is rapidly gaining acceptance outside of the more traditional PC business -- and I mean that for example in industrial embedded as well as consumer electronics -- our projections are that we will be able to benefit from having an outstanding facility in place there.

And to put it just in a brief sense of summary, our plan is to fill that sucker

So, you're going to use those tools -- you know, once they're no longer good for processors, what are you going use the tools for?

For those things.

Because these are pretty state-of-the-art tools.

They're capable of going down to 90 nanometers as they are today, because that's what we're building, and in some cases, where we're talking about products that may not require the, you know, the bleeding, leading-edge performance, they're cable of actually doing better than that.

So we expect that [inaudible] of that factory to be quite considerable.

And with minor investments, much less significant than you would expect from a new factory, we believe we have a 10-year life in that factory, easily.

And so, what -- the depreciation of the tools will be how many years?

We depreciate wafer fab, 200 millimeter wafer fab in five years

Five years Okay

It's quite depreciated by the time we get out in time.

Okay.

And then, sorry, one last thing.

On the CapEx and depreciation for CPG over the next few quarters, can you give any kind of guidance there?

It's unchanged from where we were before.

We're still talking total corporation of $1.5 billion for the year, with the bulk of it being for CPG to build up Fab 36

So if you did 500, around, you expect the same amount here and then the big dropoff in the back half?

In CapEx?

That's right.

Okay, thank you.

-- loaded as we put it in the first set of tools.

Okay.

Thanks.

Thank you for your question.

Our next question from Ben Lynch with Deutsche Bank.

Please go ahead

Yes, hi, guys.

You alluded to this, Hector, a bit, in your discussion, but there has been increasing talk about all these dual-core introductions over the last few weeks.

Could you give AMD's perspective on how important you think dual-core is in 2005 and sort of separately for the server and the desktop segments, and also what you see as the -- I know you touched on this a little bit -- the respective pros and cons of the AMD and the Intel approaches.

Let me make a couple of comments on the sequence of the choice we made in how we introduced dual-core and then I will ask Henri to make other comments relative to what the product and the market match to needs and applications.

First of all, we made the decision to put all our emphasis in the beginning in the server part of the market because that is the one that is ready to -- to jump on the exploitation of this technology.

And our customers, as you will see later when they become more public next Thursday with the announcements of their products, it's -- there's an awful lot of excitement about the performance, the capability, the power footprint, the value of what dual-core's going to bring to the server and the enterprise.

We believe that over time, this technology will find incredibly powerful uses also in the client, but that, we believe that time sequence makes sense because there's today today no software that can really optimize this product in the client space.

So that's the sequence that we chose and though you will see as we get a more public and crisp in our announcement next week, that because of the dependence of that client being an offshoot of the core of the server, that the time in between those two actually can be quite small.

Henri, could you comment on the other part?

Yes, Hector.

Well, I think you already positioned a large part of the answer.

Clearly, in the service space is the immediate usage and benefits for the end-users of the old core technology, providing it's implemented correctly.

And in the desktop space and the client space in general, I think there is -- we don't see a single-core performance being dethroned by dual core until such time that the software system will be ready to take advantage.

At which time we're very comfortable that our technology will be a leader, like it is today, in single-core absolute performance.

And then, for the last part of your question, Ben, I think -- to use an analogy, can you take a car and then put two four-cylinder engines into the hood.

That doesn't make it a V-8.

And so I think that there's a very large difference between AMD 64, and the fact that it's been architectured from day one to be a multi-core architecture, and Intel's approach of taking two processors and gluing them on a chip with a bus that's already insufficiently sized for a single processor.

Great.

I have a followup then on -- sorry, not far, but just related question on Flash.

Q-1, of course, not Q-2.

Operating profit deteriorated by something like 70 million and rev deteriorated by 60 million.

That says a lot there.

If I assume that -- you said units were up 10%.

I don't know if this is a device units or megabit units, typically megabit units go up faster than device units, but could we think that pricing pressure was as bad as 21%, Q-on-Q?

That's just looking the revenue and the unit numbers you gave would suggest?

It's a tad high, but we're definitely in those big double-digit numbers

Okay, and there was no ability via MirrorBit or just shrinking a little bit more to really offset any of that?

Henri could probably add a little color

Yes, I think you have to remember that this is the first quarter of the year in the the Wireless base.

The large customer, the strategic customers are renegotiating their yearly contracts.

So actually, the price pressure is two-fold.

There was a tremendous amount of pressure in the market just as the business was going on.

But also you also you had a reset of some of the large contracts with the market makers in the Wireless space.

So some of it is pure erosion and price pressure, and some of it is just readjustment of the pricing levels in the market as new technologies come to market

Great.

Thank you very much.

Thank you.

Our next question comes from the line of Mark Edelstone with Morgan Stanley.

Please go ahead.

Good afternoon, guys.

Hector, a question.

You gave a little bit more detail on the 90-nanometer ramp in terms of where the starts are going.

But for the first quarter, can you give us what the average outs were?

So, what percentage of outs on average in the quarter were 90-nanometer versus 130?

You know, Mark, I don't think I -- off the top of my head I would know the answer right now.

All I can tell you is that, you know, the wafer starts, which -- it's kind of the more traditional thing we look at -- will be 100% here in another few weeks by the end of the quarter.

And that's -- you know, and maybe extrapolate back.

I would say the ramp has accelerated in the -- between -- in the first and second quarter, is probably the steepest part of the ramp.

So, we're -- I don't even want to guess the number.

But those first and second quarter are the steepest accelerations

Okay.

I guess what I'm trying --

Mark, I'll try to give you a little bit different color on that.

You know with -- I'll call it a relatively good cycle time, but cycle times in the 12-week kind of time period, you know, basically you will have all the die output in the third quarter being close to the 90 nanometer.

So, we're pretty far down the pike as we -- as we closed out the first quarter.

As Hector said, most of the starts in the second will convert all into 90-nanometer and the bulk of the output in the third quarter is all 90.

So, per die, you've actually [inaudible].

Okay.

I guess what I'm struggling with a little bit, Bob, is from a product perspective, your mix was richer in the quarter, yet the gross margins in CPG are down maybe 200 or 300 basis points sequentially.

I'm assuming that that's all because you basically were trading off larger -- AM 64 die, basically, and that the full leverage of the transition, though, 130 -- from 130 to 90 nanometer, hasn't fully flown through the PNL yet

That's correct.

I mean, we're still in the early stages of seeing the positive impact of the reduced cost structure to 90 nanometer.

So there's more to come over the next two quarters from the conversion of 90 nanometer

Okay.

So, I guess to look at it another way, if you're mixed in change in Q-2, I'm expecting then, unless we had a dramatic shift in pricing in the industry, that your gross margins and CPG would expand,

That's a reasonable assumption

Okay.

Just two quick questions on Flash, if I could.

Again, relating to the quarter.

Were there any either write-downs or reserves taken in the flash business to deal with the adverse pricing conditions?

Not on pricing.

We were -- on inventory, of course, we continue to take the appropriate levels of reserves required of inventory, and we did step up inventory reserves in the quarter in the memory business

Okay, I'm guessing that that partially explains some of the gross margin deterioration there?

Yes

And then just lastly, our contacts would suggest that at least from a -- a little momentum perspective, the flash business started to improve as the quarter was progressing.

Is that a fair statement?

Can't answer that one, Mark.

Sorry.

Okay, thanks a lot, guys.

And nice job on the Spansion spin,

Thank you.

Thank you.

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

Please go ahead

Thanks.

I was wondering if you could just clarify, Bob, with respect to the expenses that you suggest will be up 5%, whether that relates to overall AMD expenses

No, it's primarily -- I kind of think of it this way.

I mean, as Hector said, you know, we started the first wafer start in the engineering mode and Fab 36, so it's mostly associated with Fab 36, of actually exercising the tools, so it's mostly R&D.

Some of it will be marketing.

As usual, that moves with the seasonality of the business, but the bulk of -- it's all CPG and it's mostly Fab 36.

And with respect to the flash in the first quarter, do you have a percentage that the MirrorBit represented in the first quarter?

Tim, MirrorBit was in the same range that it's been the last two previous quarters.

Double digit

Okay And then, if you could possibly comment on the Turion and how we should think about that in terms of a framework for expectations.

Sure.

Well, you know, we announced the Turion 64 platform during Q-1.

We made very good progress in terms of the design and activity of [inaudible] and you should look at Q-2 as the quarter where you are going to see those design wins turn into platforms launched into the market by our customers with, you know, acceleration in terms of the volumes in Q-3 and Q-4.

So, Q-1 involves the technology, Q-2, you're going to see the products, and then Q-3 and Q-4, we're going to start to see the ramp of sales

Just one other thing for Bob.

Just, on the shape of the year and the progression of the margins, how should we think about the operating margin progressing in CPG as the costs go lower for Fab 36?

Well, Fab 36 actually continues to increase throughout the year as we continue to ramp the capacity, hire more employees, et cetera.

So we've got a cost increase that is building for Fab 36.

On the other side, though, on the positive side, we'll see the full benefits of the 90-nanometer conversion and the continued penetration into the richer markets on the ASP fronts.

So you've kind of got balancing between both sides of the equation.

Thank you

With additional capacity, you know, unit capacity that comes online, so we're not at the top of the house.

If you could just clarify to the extent you're able comment on it, the goal entering the IPO is that you will IPO a percentage of Spansion and then distribute the rest to shareholders?

Is that the structure, why you can't provide any details, that you're looking at?

No, that's not what the document says.

Could you clarify that then

Basically, the document says we will do the initial IPO for the appropriate structure to capitalize that business.

AMD will still be the significant owner of that business.

And the majority shareholder?

Yes.

So, it will be consolidated within your results going forward

We -- as we stated in the S-1, our goal is not to consolidate that business.

We think through that process of actually doing the IPO, that will probably move us to a level that consolidation would not be allowed

To be a minority interest line in your income statement

That's correct

Thank you very much.

But, please read the document.

I think it's fairly clear in the document.

So, for modeling purposes, that's what we should frame as?

Yes

Thank you

And thank you for your question.

Our next question is from the line of Michael Masdea with CSFB.

Please go ahead.

Yes, thanks a lot.

Maybe Henri can give us a little bit of color on Japan, the markets, [inaudible] a little bit behind us, hopefully.

Is that a market where your market share's a little bit lower?

And what's the dynamics of that market in terms of mix that you typically see over there?

Well, I think what is really relevant today in the Japanese market is the introduction of our Turion brand.

This is a very competitive offering that we've got.

Clearly, some of the recent events are allowing us to take a different stance, in terms of how we want to attack the opportunity that's in front of us in the Japanese market.

We have a very good product that is available, and I'm obviously expecting some market share gains in the Japanese market and with our Japanese OEM partners.

Okay.

And then on the dual-core side, we've seen some pricing from your competitors start to leak out there.

Help us understand kind of what the cost structure looks like, going from single-core and dual-core, and I believe it's a little different between you and your competitor.

And then, what are you expecting for pricing?

Is pricing going to move much relative to the single-core of similar specs?

Well, you know, I cannot talk about my competitor's pricing strategy, but our strategy is very clear.

We believe that dual-core brings significant value to the enterprise in the server space.

They're probably not proposing a product there because they can't.

There is a significant premium for performance in the server environment, and we intend to take advantage of that opportunity that is in front of us.

As far as the client's space, as I stated before, we need to wait until the software ecosystem is ready to take full advantage of dual-core implementation.

And at that time, based on our customers' feedback, we'll be ready to take advantage of that opportunity.

Great.

This last question's for Hector.

Hector, you said earlier you've got 10% of the resources of your competitor.

You've obviously made a number of shifts in the computing landscape with some technology you introduced recently.

You feel like you can continue to survive with that kind of 10% rate, or do you look to expand that over time, now that you're more focused on the processor business?

Well, we've always stated our goal is to become a relevant and significant player in the industry.

And so long as we're operating in an environment of free and open and fair competition, we have complete confidence that we'll we continue to out-innovate the competition and gain the trust and business that our customers are going to wish to give us.

Great, thanks.

Thank you for your question.

Our next question is from the line of Krishna Shankar with JMP Securities

Lost him.

Between Sempron and Athlon 64, in terms of the PC mix, and for the new dual-core Opteron launch, will the new product support things such as PCI Express, DDR-2 memory, and will it incorporate your PowerNow savings technology?

I'm sorry, Krishna.

You were cut out at the beginning.

I didn't hear the beginning of your question, but I guess that you were asking if there was a shift between Q-4 and Q-1 in our mix between Sempron and Athlon 64?

Is that the question?

Yes, how did your value as a performance mix do, Q-4 to Q-1?

As Sempron --

Well, we're seeing an increase.

There was an increase in AMD 64 base product overall.

And within those architecture, obviously, the Athlon 64 grew faster than Sempron.

As I indicated, if you recall, in Q-4, we had a very good surprise, a very strong success on Sempron, but in Q-1, we came back to a more balanced ratio of Sempron to Athlon 64, in favor of Athlon 64.

And then, the second part of your question, relative to dual-core, I guess you wanted to understand our strategy from the DDR1 to DDR2 transition and PCI express support?

That's correct

Okay.

Well, PCI express, two sets are available to date that support the AMD64 platform, across all variation of it, in the mobile space, in the desktop space, and in the server space.

Relative to DDR1 versus DDR2, as you know, the direct connect architecture places a very different problematic, in terms of system performance, on the table.

We don't need DDR2 to exceed the performance of our competitor on DDR2.

We have a planned transition to DDR2 in 2006 that we're working in concert with our [inaudible - heavily accented language] partners and customers to make in a timely fashion, when finally DDR2 will provide a performance enhancement to our architecture.

We are very confident that we will continue to have superior benchmark capability, based on DDR1, well into 2006.

Okay.

And can you also comment on your chipsets for Turion 64?

Who are your partners in terms of the mobile chipset and Wireless LAN solutions.

Sure.

Well, you know, we took an approach of a mobile platform for Turion 64.

The chipset partners that we have are the usual chipset partners that you find in the [inaudible] namely, ATI, nVidea, VIA, SiS, and others.

And as far as the wireless connectivity, you've got of course Broadcom, Atheros, and others.

So, it's the usual ecosystem that you find around AMD platforms

And can you articulate a little bit your own sort of open platform strategy, versus your competitors' more closed platform approach?

Can you talk a little bit about your platform strategy going forward?

Well, let's talk about it in the -- for at least the next 12 months.

I believe that there is an awful lot of technology transitions occurring in the platforms, particularly in Wireless space where the standards keep evolving in improvement and performance and a lot players are making a concerted investment in the core competency they have.

Let me use the example of [inaudible] just as one.

Here's a company that's totally focused on being a premier company in that space.

I believe it's our duty, our responsibility, to offer our customers the choice of best-in-class in each of these categories, and I believe that, at least in the foreseeable future, you're going to see players that -- like the ones Henri mentioned, where customers will have the opportunity to actually select the platform.

They will have best-in-class in every one of the components, not only just on the CPU, but on the peripheral circuitry that they will want and need

Great.

And finally, can you give us some sense for how rapidly the Turion 64 could ramp up in the second half of the year and the ASPs there versus your premium Athlon 64 ASPs on the desktop?

You know, Krishna, it's a little early.

As I indicated, our customer platforms are going to be out in the market in Q-2.

What we have seen to date is from an acceptance perspective in the OEM community, it's our fastest design win, visible velocity in the mobile space, because it's a very competitive offering.

I will translate into end-user and corporate users buying the product of our customers.

It's too early to tell.

But we expect to see, obviously, an increase in share in the mobile space, thanks to the Turion 64 platform

Thank you.

Thank you for your question.

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

Please go ahead.

Thank you very much.

Hector, I think I heard you say that you were going to be at 100% 64-bit architecture at some point.

I didn't catch the timing of that.

Could you just reiterate that?

That's right.

We believe that by the end of the year we'll be at nearly 100% of all of our CPG microprocessors, be AMD 64-based

Okay.

And then maybe, Bob, to expand on that, if all else were to be equal, can you talk about what kind of impact that would have on gross margins and then sort of as a subtext to that, is there a way to think about the mix between Athlon 64 and Sempron?

Is there a target mix that you guys are looking for?

It's not -- I'll make a couple of comments and then Henri can add color to that, too, as well.

We build our strategy on some fairly powerful and unique platforms inside AMD.

For example, the process technology is identical for all of our products and the core technology is fairly similar so that you can think of all the AMD 64-base processors as being relative to each other, so if we dial them in as we build the products and our desire, as we've shown for the last couple of quarters, since the gross rate of servers and mobile have been faster, as would expect, because that's what the market is -- the direction the market is going.

We expect that those two segments, server and mobile, to continue to outpace the growth of any other segment, and for us to be able to dial that in as appropriate, based on our customers.

I, believe that all of that leads, then, to an improved utilization of our factory, an improved utilization of our technology, and -- as expected -- and our gaining position in the strengthening in the market, leads to improved margins.

Our expectation is that margins will continue to improve

Great.

Actually, I need one more for Henri.

As we think about dual- core, we've done some work that suggests that your competitor's dual-core parts are something like 80% bigger, but recognizing that they have a different architecture or a different approach than you have.

Can you maybe give us a feel for relative to that bogy of 80%, whether or not your dual-core is above or below that watermark?

You mean -- is that a die size question?

Die size, yes.

Yes.

Clearly, we know our die size is smaller than our competitor.

But again, I want to stress the the fact they don't have dual-core parts.

They have two processors glued on the same chip.

Exactly.

I guess I mean relative -- not relative to their part, but relative to your single-core

Oh, no.

I don't have that off my head.

But again, I think that as you will see the competition forum in the desktop space, relative to dual-core or two processor on the chip versus single-core, the high-end of the market, where the performance matters, single-core processor, particularly as our Athlon 64 FX brand continues to gain wide acceptance in that community, will outpace any dual-core offering for quite a while.

So, you know, again, in the performance market, we see our single-core processor retain the performance lead.

As far as the penetration of dual-core in the value markets, we believe that it will eventually happen, but it's not immediately and that we'll be prepared by then to have a competitive offering when our customers require it

Okay.

By the way, I'm an FX 55 user and it's plenty fast.

So, it's great.

Thanks.

Well, thank you for your vote of confidence.

Thank you for your question.

Our next question comes from Chris Danely with J.P. Morgan.

Please go ahead.

Hey, thanks, guys.

Just a quick clarification and a few questions.

So, can we assume that the IPO is going to happen this quarter?

Or has that not been determined?

[Laughter.] Good try

Hey, you know.

Zero for 1.

How about a couple of non-flash questions.

On the CPU business, gross margins in the mid-50s, obviously, a pretty good level there.

You know, traditionally, been up and down.

What do you see as sort of the normalized gross margin on your CPU business and what do you see as the range?

I mean, we're talking 50 to 60%?

Or 45 to 55?

Can you just give us your thoughts on that?

Yes.

This is Bob.

I'll give you my thoughts.

I mean, 50, 55, is -- 54, what I reported today -- is fairly normal.

Clearly with volume, depending on exactly what mix we shift, what die we shift, et cetera, we'll move that around fairly quickly on a small revenue base.

But our goal is to be north of 60.

So, you know, where we continue to try to move the organization to 60, and that's through continued cost reductions of conversion -- conversion to factories of 90 nanometers, and continued penetration in the markets we've never played in before

Sure.

And then can you guys comment on the 300--millimeter ramp?

What we can tell you right now is that the factory build occurred on schedule and on budget.

And we started engineering silicon this quarter, so we're running currently engineering in the factory and it's already buttoned up and it's a gorgeous factory.

Really executing fairly well.

And our plan is to start running actually prototype silicon towards the end of the year and be in production the first half of next year

Great.

And then, last question, what is the inventory mix between flash and CPUs?

We don't give that level of granularity, but clearly in the flash business, part of my comments of some of the gross margin issue is we slowed the factories down because the demand environment is so slow.

So bottom line is we got a little bit more inventory on the flash side than we do on the microprocessor side

Great, thanks.

Thank you.

Our next question comes from the line of David Wong with A.G. Edwards.

Please go ahead.

Thank you very much.

With IPO -- and presumably there is going to be quite a bit of change to the balance sheet -- Bob, do you have any goals in terms of your maximum level of debt or minimum level of net cash over the next year to two?

Not ready to actually talk about that, since we have a long way to go through the SEC process, et cetera, et cetera.

But that's something I will talk about in the future as we go through that.

But we still are working.

Overall, we think a healthy company should be having debt-to-capital ratio of about 20%.

So, long-term, we have not taken our sights off that.

So that's what we will continue to target.

But obviously, that is not going to happen in a very quick period of time, but that's where we're trying to get to.

Just one quick followup on that.

If I understand correctly from the S-1 filing, AMD is not selling any of its portion in this IPO.

There will be no cash infusion as a result of this IPO, but there will be balance sheet improvement because debt goes away.

Is that correct?

Yes.

The IP -- the cash proceeds from from the IPO are for the Spansion organization

Yes.

Great.

Thank you.

Thank you.

Our next question is from the line of Michael McConnell with Pacific Crest Securities.

Please go ahead

Thank you.

Bob, I don't know if you already addressed this.

If we were to back out the startup costs in Q-1, what would operating margins have been in CPG?

Don't want to go there

Okay.

Looking at kind of the above-seasonal growth you have seen in Q-1 and Q-2 -- and you're expecting in Q-2, is this more a function of you taking advantages, do you believe, with some of the product shortages from your competitor, or is it more just a healthier market and just some ASP improvement as well?

We believe -- and again, Henri can add to that -- but, we -- you know, if you look at our growth rate in 2004 over 2003, [inaudible] we grew 29% in the microprocessor business.

We believe that is significantly higher than the market for the processors.

In the first quarter, we drew 31% year-on-year.

We also think that is also higher.

That, I think, is an indication that the products that we are creating and building are products that are allowing us to win.

And winning is gaining share and doing it while making money.

I think the processor business has demonstrated that over the last several quarters.

And just how would you characterize the overall PC market heading into Q-2, just from a higher level, desktop versus mobile?

Well, right now what we're seeing is a healthy market.

You know, one of the characteristics of the first quarter was that it was very linear for us.

As you know, second quarter is always a little more back-end loaded.

It's a tough quarter to close because of some of the seasonality factors in Europe.

But you know, again, to build on what Hector said, we have two growth engines that were not traditionally in our business model.

First, our server business is definitely growing and we're winning big-time, particularly in the Fortune 500 space.

And secondly, our high growth markets.

A large portion of our growth in the first quarter was derived from exceptional growth in the high-growth markets, where we believe we have a very solid offering, and -- markets like India or greater China or Russia are proving to be extremely good markets where seasonality is less effective than in the mature markets of Europe or North America

And has China -- the strength in China continued thus far into Q-2?

Absolutely

Thank you.

And thank you for your question.

Next question from the line of Chris Caso with FBR.

Please go ahead

Yes, hi.

I was wondering if you could give some color, and to the extent that you can at this point, I realize it's early.

But into the cost structure of AMD, once the IPO is completed and, more specifically, in terms of where the operating expenses might lie for a CPG-only business

I think that's, to be honest, I think that's too premature to answer that kind of question.

There is so much information in the S-1 that we'll have to execute to, so I don't think I can answer that question at this moment

Okay, that's fair.

I'll try another one.

With regard to, you know, performance of the CPG group in Q-1, if you could give some color on the desktop business as a whole.

Was that in line with what you'd expect normal seasonality?

A little better, a little worse, what you can tell us in terms of units and revenue in the business?

Well, I'm not going to give you too [inaudible] except to say that we had an ASP increase prior to Q-1, and that was led by the desktop business, as we continued to take advantage of the excellent positioning of Athlon 64 and Sempron.

We also had a very strong quarter in the channel.

It was typically a desktop-driven piece of the business.

So, the desktop business continues to do very well for us, but clearly, we see a lot of growth opportunity in server and mobile

Okay, thanks.

Thank you for your question.

Our next question from John Barton with Wachovia Securities.

Please go ahead

Thank you very much.

On the topic of ASPs, you commented that obviously, they were up sequentially in the first quarter.

If you were to narrow that calculation down to just AMD 64, quarter-on-quarter, how did ASPs track?

I'm going to give you more granularity than what I said, which is, you know, our ASP increase was led by desktop.

And we had stable ASPs on the two other segments.

Okay.

And, Bob, you had commented that there were operating cost savings in the March quarter, in the first quarter for Spansion.

Could you give us some idea of the magnitude of the savings, quarter-on-quarter?

They're in the tens of millions of dollars.

But obviously, wasn't anywhere near close enough since the loss expanded by $70 million, quarter-on-quarter

Right.

Then final question, if I could.

Obviously, tax benefit in the recently closed quarter, how should we be thinking about tax as we look at the next several quarters?

We're still at the same point.

We're, effectively, 10% kind of range.

It will move around a little bit, depending on actually where the jurisdiction of the tax consequence takes place, but 10% for modeling is not a bad number

Thank you.

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

Thank you.

Our next question comes from the line of Steve Lister with UBS.

Please go ahead.

Yes.

This is Steve Eliscu for Tom Thornhill.

Follow-up to a prior question regarding filling Fab 30.

Wanted to better understand what you feel is the market opportunity for x86 processors in the embedded and the consumer markets.

We believe that there is a fairly large opportunity and a strong desire for people in the industrial segment of the market.

These are people like Telcos, people who like make switches for telecommunications environment, to really exploit the benefits that Opteron has as an embedded product, and we have seen a very large interest in that.

We believe that market is large, and when you look at the total embedded potential market for x-86, frankly, it's up in around $2 billion.

So it's a fairly sizeable market.

We think we deserve a big chunk of that today.

We're a very small player in it.

Therefore, that is one factor.

But when you go to the other extreme of the curve, you are now into the much lower cost and lower power and smaller, you know, footprint applications, where you look at a variety of consumer electronics and that brings us into the digitalized -- digital convergence.

I'm going to ask Henri to make some comments on digital convergence, because we believe that, for us, it gives us a great opportunity to expand the footprint of x-86, on quite a different strategy than the competition has

Yes.

What we have seen in the consumer electronic space is obviously a set of market makers that are not traditional PC market makers that have seen a business model develop in the PC industry that is not a favorable -- you know, it's basically a vendor-driven industry rather than a customer-driven industry.

Consumer electronics, historically, has been a customer-driven industry.

They want to retain that business model.

And so we're seeing a lot of interest for x86-based solution manufactured by AMD, because we are seen as a friendly partnership-oriented company in that space.

Great.

As a followup, regarding 64 bits, just clarification on Sempron.

So, you plan to actually market that as 64-bit processor.

Is that correct?

Yes, we have always been clear that there is a differentiation between Athlon 64 and Sempron.

It's not based on whether it's 32 bit or 64 bit.

They address different markets.

Sempron is for the value-conscience buyer, Athlon 64 is for the performance-driven buyer.

But we've also made very clear that our ambition is to lead this industry, the pervasive 64-bit computing, and as such, it wouldn't make sense for us to keep alive 32-bit processors which, frankly, are rapidly becoming something of the past and, therefore, you should expect that at some point in time the entire lineup of AMD 64-based processors from AMD will be 64-bit enabled

And that's before the end of this year?

Most likely

Okay.

And just one final question.

If you could comment if you believe you'll be shipping material volumes of dual-core desktop and notebook processors this year.

The answer is yes.

Okay.

Thank you.

Thank you.

Our final question comes from the line of Simona Jankowski with Goldman Sachs.

Please go ahead

Hi.

Thank you very much.

I just wanted to ask you on the inventory.

What do you think your inventory will be in Q22 directionally?

Do you think [inaudible] going to be slowing down some of your fabs?

And also, what is your target for later on in the year as far as basic Inventory?

First, I can't comment on Flash.

That would be forward-looking guidance, so I can't comment there.

Micro processors, you know, we clearly build to the market demand, and as we have stated before, we're executing quite well in that business and so we're bringing on additional capacity to service the market so that we will continue to be extraordinarily higher growth rate than our competition.

So, inventory to me is not an issue in the microprocessor business.

What is your comfortable range for basic inventory?

We're very comfortable.

Actually, I don't give that level of granularity, but we're very comfortable in the microprocessor business

Thank you very much.

Okay, that concludes the call.

I want to thank everyone for participating and hopefully a number of you will show up at our dual-core launch next Thursday as well.

Thank you very much.

Thank you, and Ladies and gentlemen, today's conference call is being made available for replay starting today at 7 p.m.

Pacific time and running through Saturday, April 23, 2005.

You may access our service by dialing 1-800-475-6701 and entering today's access code of 776654.

Internationally, you may reach us by dialing 320-365-3844, again with the access code of 776654. [OPERATOR INSTRUCTIONS].

That does conclude our conference call for today.

We thank you for your participation as well as for using AT&T's executive teleconference service.

You may now disconnect.