使用警語:中文譯文來源為 Google 翻譯,僅供參考,實際內容請以英文原文為主
Operator
Please stand by.
Good day and welcome to Intel Corporation's second quarter Earnings Conference Call.
A reminder that today's call is being recorded.
At this time, I'd like to turn the call over to the Assistant Treasurer and Director of Investor Relations, Doug Lusk.
Please go ahead sir.
Doug Lusk - Assistant Treasurer and Director of IR
Thank you and welcome to the Intel second quarter earnings conference call.
Attending from Intel are CFO, Andy Bryant and President and COO, Paul Otellini.
I would remind everyone that the earnings release and this call are available on our IR website at www.INTC.com.
For those of you who may not have seen the quarterly earnings report, revenue in the second quarter was $6.8 billion and earnings per share were 14 cents.
The second quarter earnings report discusses Intel's business outlook and contains forward-looking statements.
These particular forward-looking statements and all others statements that may be made on this call but not historical facts are subject to a number of risks and uncertainties and actual results may differ materially.
Please refer to our press release for more information on the risk factors that could cause actual results to differ.
The specific forward-looking statements cover expectations for product mix and demand, revenue, gross margin, revenue, expenses, tax rate, interest and other income, capital spending, depreciation and amortization of acquisition-related intangibles and costs.
These statements do not reflect the potential impact of any mergers, acquisitions, divestitures or other business combinations that may be completed after July 15, 2003.
Lastly, if during this call, we use any non-GAAP financial measure as defined by the FCC in reg G, you will find on our website, ww.INTC.com, the required reconciliation to the most directly comparable GAAP financial measure.
I now want to introduce Andy Bryant who will discuss the second quarter earning results.
Andy?
Andy Bryant - CFO
Thanks, Doug.
Our financial results were consistent with our outlook in April and June at the high end of expectations.
The positive news was that we saw double-digit year-on-year growth in microprocessors and chipsets.
Unfortunately, this was partially offset by continuing weakness in flash and networking.
Typically Intel's revenues declined in the second quarter, but this year revenue of $6.8 billion was approximately flat 1% higher than first quarter.
Compared to the second quarter of 2002, revenues were up 8%, the best year-to-year growth we have seen since the third quarter of 2000.
Unit share month of Intel architecture microprocessors and chipsets were approximately flat with the first quarter, while those of motherboards were higher.
Flash memory (inaudible) shipments were lower.
There was a slight increase in unit shipments of Ethernet connectivity products.
All of the year-on-year growth came from the Intel architecture business, comprised largely of sales of microprocessors, chipsets and motherboards.
Its revenue of $5.8 billion was approximately flat with the first quarter, and of 12% from the second quarter of last year.
Revenue for the wireless communications business declined 2% sequentially, to $465 million, a year-on-year decline of 13%.
And in the networking business, revenue was $508 million, approximately flat with the first quarter, and down 5% year on year.
Gross margin percentage of 50.9% is sequentially down from 52%.
The decrease is due to a combination of factors, primarily driven by higher start costs in second quarter for a 300 millimeter logic factory, and the benefit from the sale of previously reserved inventory in first quarter that did not recur.
Margin at 50.9 is four points higher than a year ago, due to margin benefits of revenue being higher in the current year, combined with the fact that last year's margins were depressed by reserves taken for winding down the Intel on-line services business.
Operating income of $1.3 billion was down 8% from first quarter, and up 100% in the second quarter of 2002.
All of the year over year growth and profit came from the Intel architecture business, where operating income was $1.8 billion, down 4% sequentially, and up 35% year on year.
In the communications businesses, operating losses were $143 million for the networking group, and $123 million for the wireless group.
Weak unit demand in a wireless group lowered revenues, and increased factory under-utilization charges in the second quarter.
Spending on R&D and MG&A of $2.1 billion was at the high end of our outlook of 3% from first quarter and 1% from the second quarter of 2002.
The size of the work force at the end of the quarter was 78,700 employees, compared to 79,200 in March, and down 5% from 83,200 a year ago.
The total for interest income, other incomes, and gains and losses in equity investments was a loss of $5 million, better than the loss of $20 million expected in June.
In the category of income statement, interest and other income was $53 million.
The net loss on equity investments was $58 million, primarily as a result of impairment charges of $64 million.
Interest income and other income was higher than forecast.
The tax rate was 29.5%, lower than our forecasted 30.5%, due to a tax benefit related to a divestiture on a prior acquisition that closed during the quarter.
The impact of this tax benefit was about $13 million.
Fully diluted earnings per share of 14 cents, which include potential dilution attributable to employee stock options was flat with the first quarter.
Basic earnings per share, which does not include potential dilution, was also 14 cents.
This is double earnings per share of 7 cents in the second quarter of 2002, when net income reflected $106 million and reserves related to Intel on-line services and $112 million write-off of acquired intangibles.
For the quarter just ended, average shares were calculated on diluted earnings per share were $6.6 billion.
During the quarter, we repurchased 51.8 million shares at a cost of $1 billion.
On the balance sheet, inventories overall were approximately flat, with a 6% increase in finished goods.
Cash, short-term investments and fixed income trading assets ended the quarter at a total of $13.4 billion, an increase of $1.2 billion.
After approximately 900 million in capital spending and a 1 billion on stock buyback.
In summary then, for the second quarter, revenue of $6.8 billion, up 1% sequentially and up 8% year on year, gross margin of 50.9%, operating income of $1.3 billion or 19% of revenue, fully diluted in basic earnings per share of 14 cents.
I return now to the outlook for the third quarter.
Please keep in mind that the forecast data do not include the effect of any new acquisitions or divestitures that may be completed after July 14th, with the exception of an anticipated divestiture affecting the third quarter tax rate, which I will discuss in a minute.
I will use the new point of forecast ranges when making comparisons to prior periods.
We expect rev fly in the third quarter to be between $6.9 and $7.5 billion.
Revenue of $7.2 billion would be up 6% from the second quarter consistent with seasonal patterns.
In the previous five years, revenue growth in the third quarter has ranged from 3 to 14% and averaged 7%. $7.2 billion would represent year-on-year growth of 11%.
Our forecast for gross margin percentage in the third quarter is 54% plus or minus a couple points.
The potential 3-point increase comes from three primary sources.
The first is higher revenue, the second is lower start-up costs, as resources focused on ramping 90 nanometer technology, transition to developmental works for next generation process, and the third is continued progress in reducing unit costs.
Intel is reviving a program to provide Intel employees each with one home PC for personal use.
The plan should be completed by the end of this year with most of the remaining costs occurring in the third quarter.
Last in the year 2000, the home PC program was put on hold at mid course in 2001, as we cut nonessential spending to protect profitability.
Spending for R&D and MG&A should be approximately $2.2 billion in the third quarter, higher than 2.1 billion in the second quarter, and higher also than the average for recent quarters.
In addition to the purchase of PCs to complete the home PC program, the projection for higher spending, (inaudible) growth and R&D spending due to transition of technology development spending from cost of sales to R&D.
Keep in mind that as we develop making transitions from ramping to non-animated process to developing the next generation 65 nanometer process, the spending will move from cost of sales to R&D.
Our gains and losses from equity investments and interest and other income is a net loss of $25 million.
This includes a net loss on equity investments of approximately $60 million, primarily as a result of impairment charges.
We expect a tax benefit in the third quarter from a divestiture that should close shortly.
Our forecast assumes a tax rate of approximately 24% in the third quarter, and approximately 30.5% in the fourth quarter.
Depreciation is expected to be approximately $1.2 billion for the third quarter, and amortization of acquisition-related intangibles and costs should be $70 million.
As we look ahead to the rest of the year, we are adjusting our outlook to reflect improving profitability in the Intel architecture business.
We expect gross margin for the year to be approximately 54% plus or minus a few points, higher than our previous expectation of 51%.
This improvement comes from a combination of higher than anticipated revenue and margin in the first half, projected seasonal growth revenue for the second half and a previously mentioned transition of costs to R&D.
Depreciation for the year should be $4.7 billion, slightly lower than our previous expectation of 4.8 billion.
We are also revising our outlook for R&D spending from $4-4.2 billion.
As I discussed a minute ago, this reflects a transition of resources from the ramp of animators to development of 65.
Revised spending does not assume to change in the size of the work force, which we plan to keep approximately flat for the rest of 2003.
Capital spending for the year is on track with $3.7 billion, plus or minus $200 million.
During the quarter, (inaudible) a $2 billion dollar development facility for new process of ( inaudible ) is developing Intel 65 animated process technology, which will be used in manufacturing (inaudible) processes in the 2005 time frame.
In summary, this was another quarter of solid financial performance with improving profitability and revenue growth.
Inventories are disciplined, retirements are lower and we continue to drive costs down.
The financial benefits of aligning resources to core competencies and higher return activities are becoming more apparent.
We are encouraged that for three consecutive quarters now, the Intel's architecture business (inaudible) results with the high end of seasonality with that I return over to Paul for additional comment on the business.
Paul Otellini - President and COO
Thanks, Andy.
As Andy mentioned, Q2 revenue was a bit higher than Q1, a good result in what is normally a seasonally down quarter.
Our Asia-Pacific region led the way.
Setting an all-time revenue record, which included a very modest impact from SARS.
Our Intel architecture business was at the high end of seasonal patterns with units approximately flat from the first quarter, and we continued to gain market segment share in a number of areas.
Our communications businesses continue to experience softness, with overall revenue for these groups flat.
On a geographic basis, the Asia-Pacific region delivered another record quarter, with revenue up 5% sequentially and up 17% year over year.
This region continues to benefit from growth and local consumption, especially in emerging markets, as well as from its growing role as a manufacturing and design center for the PC industry.
This region represented greater than 40% of Intel's total revenue for first-time ever.
China saw single digit sequential growth, but was up double digits on a year over year basis.
Taiwan was particularly strong, driven by chipset and motherboard demand.
Japan had its best quarter in more than two years.
With revenue up 22% sequentially, and 42% year over year.
Growth was driven by notebook PC exports and higher demand for camera phones.
Europe, which typically has its weakest quarter in (inaudible) was down slightly less than seasonally normal and up 10% year over year.
Emerging markets in Russia and Eastern Europe turned in the best results for the region.
Our America's region was approximately flat quarter over quarter, as Latin America generated some sequential growth despite strikes in Brazil and Peru.
In North America, server growth and some increases in corporate buying helped to offset ongoing weakness in communications.
Intel's channel basis remained strong and finished strong, driven by growth in our emerging markets.
The Intel architecture business performed at the high end of seasonality with microprocessor units and ASP's approximately flat.
Chipset and motherboard revenues were higher.
Revenue for the Intel architecture business was up a solid 12% versus the second quarter a year ago.
We believe our channel presents and strong product lineup allowed us to gain a point or two of market segment share in CPUs and several points of share in chipsets.
On the desktop, we began volume shipments of new Pentium 4 processors and chipsets supporting Intel's hyper threading technology, 800 megahertz front side bus and DDR 400 memory.
The new Intel 865 and 875 chipsets are on track to be the fastest ramping chipset family in our history and brings substantial performance increases to both business and consumer platforms.
We shipped over 2 million HT enable processors during the quarter and expect HT to be in over 50% of our performance desktop shipments by the end of the fourth quarter.
We also accelerated our Value Lineup during the quarter with Celeron processors at 2.5 and 2.6 gigahertz and we believe the sits comfortably above the sweet spot of competitive offerings.
Mobile platforms continue to drive the PC segment.
With double digit unit growth, versus a year ago.
In the first half of this year, versus last year, our mobile unit shipments have grown at almost twice the rate of the desktop.
In Q2, we surpassed our goal of shipping a million Next generation Pentium M processors.
The majority of which were sold as Centrino brand solutions with Intel's 802.11 wireless product.
We extended the product family with more ultra low voltage versions and boosted the top of the line to 1.7 gigahertz.
We plan to ship over 2 million Pentium M processor units in Q3 as we market Centrino to the consumer segment for the first time in conjunction with the back to school season.
We continue to expect to have over 50% of our performance notebook processors on the new technology by the end of the year.
Intel's enterprise products group set a new record revenue during the quarter with ongoing strength in high performance Xeon processors for dual processor servers.
We also strengthened our multiprocessor lineup with large cash versions of the Xeon processor running at speeds up to 2.8 gigahertz.
The highlight for the quarter was the introduction of the new Madison versions of the Itanium-2 processor.
The new processors deliver speeds up to 1.5 gigahertz, and cash memories up to 6 megabyte boosting the performance of existing system designs by as much as 30 to 50%.
The number of server and workstation models based upon the Itanium-2 processor is expected to double this year.
An important milestone was reached in the quarter with the introduction of production release of Microsoft's Window Server 2003 Operating System and Sequel 2000 Enterprise Edition database, both of which are optimized for Itanium.
The introduction of large FMP platforms based on our new Itanium II processor has brought new levels of competition to the highest ends of the enterprise computing.
The record for the highest performing TPCC nonclustered system went unbeaten for 18 months.
Following the announcement of two Itanium II records this quarter, performance leadership changed hands an additional four times as the Itanium II and Power 4 based systems competed for the lead.
As a result of this new competition, performance was driven up by 67% this quarter, and price performance improved by nearly a factor of two.
The Itanium II processor is having a similar effect on the world's largest supercomputers.
Last year, they listed the world's top computers contained three supercomputers based on Itanium processors.
In the past six months, the number increased to 21 supercomputers, based on Itanium II processors including out first entry into the top ten.
To summarize, our IA business is clearly benefiting from strong new products and technology such as hyperthreating and Centrino, allowing us to drive demand creation and gain market segment share.
We expect our competitive position to improve further in coming quarters as we begin to ramp 90 nm process technology, which is healthy, and yielding well.
We are currently sampling both Dothan and Prescott with customers and expect revenue shipments of both in fourth quarter.
Moving to our communications businesses, the Intel communications group had flat sequential revenues with slightly higher unit shipments of Ethernet connectivity products.
The group shipped approximately a million wireless LAN products for the Centrino platform in Q2, and is on track for revenue shipments of dual band 80211A and B products in Q3 and BG product by the end of the year.
We also announced that we are developing silicon-based on the 802.16A or Y-Max standard that will provide a wireless broadband alternative to existing last-mile technologies such as cable and DSL.
Revenue for our wireless communications and computing group was approximately flat with the first quarter.
In flash units, we were slightly lower as we experienced some SARS-related impact in Asia.
Demand for high-density flash continues to grow driven by new phone features such as color screens and cameras.
Our blended flash ASP was slightly higher as a result, but the pricing environment remains difficult, especially at the lower densities.
During the quarter, we began shipping our new .13 micron (inaudible) memory optimized for cell phone applications and expect this to be our best flash growth opportunity in the second half.
We continue to work on growing our market segment share at our customers.
In the applications processor business, we continue to (inaudible) momentum and we now have dozens of cell phone design wins based on our X scale technology.
Including initial wins for Manitoba, which is sampling and on track for revenue shipments this year.
We also announced that we will be working with Sun Microsystems to optimize Java on our X scale technology.
In summary, Q2 was slightly better than we expected and substantially better than a year ago.
Our core business is performing well and delivering growth in our key investment areas.
In our communications business, the launch of Centrino mobile technology has moved us into a wi-fi industry leadership position.
Our move towards 90 nm, 300 mm silicon technology remains on track for delivering further gains in both performance and cost.
In short, while the markets for our communications businesses are still lagging, our Intel architecture business is performing better than our expectations.
With that, let me turn meeting back over to Doug.
Doug Lusk - Assistant Treasurer and Director of IR
Okay, thanks, Paul.
We will now open the call for Q&A.
In the interest of moving along and giving other questioners a chance, please limit yourselves to one question.
Thank you.
Operator?
Operator
Thank you.
If you'd like to ask a question today, please signal us by pressing "*", then "1" on your telephone keypad.
Thank you and that's "*1" on your telephone to ask a question.
First up, will be Mark Edelstone at Morgan Stanley.
Please go ahead.
Mark Edelstone - Analyst
Hi, good afternoon, and nice quarter, guys.
Andy, really just had a question on the margins.
If you look at your guys' guidance, you're suggesting roughly an incremental gross margin of about 100%, and I guess I wanted to just get a little more insight into that if I could.
You mentioned on your comments that there was a charge in the quarter for wireless, as you just hadn't had I guess full of absorption of the fabs.
Is that something like 50 basis points to Q2, and then also we look at your operating margin for IAG in the quarter, it was actually down about a 180 basis points sequentially.
Does all that come back so that you get the IAG operating margins back to the Q1 levels in Q3?
Andy Bryant - CFO
You have to help -- don't hang up, because I'm going to need you to remind me of the questions.
The last one's the easiest.
You talked about the decrease in year-over-year profitability, quarter-to-quarter profitability in IAG.
That's because of start-up costs for 300 millimeter in the second quarter.
That's also what goes away in the third quarter in our forecast.
So increase in start-up costs caused IAG profitability, operating margin to be down.
As you look to the third quarter, what we show there is margins.
In the third quarter, we see margins growing for three reasons.
Part of it is start-up costs will be lower again, as those costs are moving into the R&D space.
Part of it is we see better unit costs and part of it we see better revenue.
Did I miss something in that piece?
Mark Edelstone - Analyst
The only other question, then, Andy, you mentioned there was at least some margin hit in the second quarter from wireless.
Is that something like 50 basis points for the company as a whole?
Andy Bryant - CFO
It's for flash (ph).
I really can't give you a specific number, but in order for me to call it out, it has to be a point where it would make, you know, several tenths (ph) of a point of margin.
Mark Edelstone - Analyst
Thanks a lot, guys.
Operator
Moving on to our next question.
This is Adam Parker at Sanford Bernstein.
Adam Parker - Analyst
Can you help me with a little bit kind of jibing --you say IAG's performing better than expectations in the three straight quarters and you are reinstating something like this home PC program, which seems like an optimistic view.
Yet, at every conference, Barrett and Andy, you guys sort of talk about that macro conditions remain weak and there's no set of huge commercial upgrade underway, or the macro backdrop is weak.
Can you kind of position those two things that appear a bit inconsistent?
Andy Bryant - CFO
Well, I hope that -- well, what we're trying to communicate is we're not seeing signs of an economic recovery.
So I'm not seeing a big upgrade cycle, I'm not seeing IT budgets being raised suddenly.
However, what we are seeing is, if you look at a 10-year history for Intel, inside 10-year patterns, Intel -- second halves grow more than first halves shrink.
If you look at the last two years, what you have seen is second halves and first halves offsetting each other with no growth.
We think we're returning to more of a normal seasonal pattern, second half growth, hopefully more than the first half shrinking, but not any breakout.
So if we saw an upgrade cycle, if we saw IT budgets going up, you will see us become more positive.
As far as the home PC program, we made a commitment to employees three years ago.
We put it on hold for half of the employees, half the employees had already received the PC.
What we're trying to do is fulfill that commitment in a time when we see margins going up.
We see ISP's have been stable for first half, market segment share is improving a little bit.
We think we can take this window to finish -- to fulfill that to employee, one time, and be finished with it and be done.
Adam Parker - Analyst
So OK.
Soryr, one other question sort of separately here.
Andy, what are you assuming about -- you talk of your 10-year analysis and return on capital so forth for evaluating businesses.
Can you help at all what you are assuming about the flash memory business, in terms of units or pricing of any inputs you can share in terms of why you still believe that's going to be a good business overall?
Andy Bryant - CFO
Difficult for me to -- no, the answer is no, I can't give you any specifics.
Recognize the way the accounting rules require you to do it is by reporting group.
So inside wireless communications, with both the PC, A-architecture, it's got the XScale press in there, as well as the flash business.
If you take it together you do a 10-year growth rate, you do have to have pricing assumptions.
You just have market segment share assumptions.
Assume we'll be successful in whatever business will be there for flash.
Plus, we think will be successful in PCA as well.
Adam Parker - Analyst
But you, sir, said you do it individually not by reported business segment.
So you still believe that flash is going to be a good business over the long term for Intel?
Andy Bryant - CFO
Actually I do believe flash is going to be a good business over the long time for Intel, but for the reporting requirements you actually put PCA and flash together.
By the way, I think it's appropriate in this case because as I think -- as you put flash into the PCA space, what you will see is a blending of the two businesses.
It will be hard to separate though.
Adam Parker - Analyst
Gotcha, right thanks.
Operator
Next up is Ben Lynch at Deutsche Bank.
Ben Lynch - Analyst
Hi.
Just looking at the way that depreciation charge and cap ex have evolved this year, you started the year with a 4.9 billion depreciation forecast and then you went to 4.8, and then 4.7, but you haven't changed your cap ex forecast.
Can you explain the dynamic behind this?
And then I have a follow-up question, please.
Andy Bryant - CFO
What I can tell you is it's difficult to give you much better insight.
What you're trying to do in doing your depreciation forecast is you try to predict not only when the equipment arrives, when it's installed, how long it takes the computer systems to be geared up and to start to work.
So, what you're seeing is a little bit of timing difference more than anything else.
Also, recognize we're right around break point.
So, if you look at the -- you know, 1.2, we had in the second quarter, it was actually like 1162, the hard round up and you try to figure out where those round points are.
So, there's no fundamental change in the business, just the estimate getting a little bit better.
Ben Lynch - Analyst
OK, great.
Then the follow-up was, if in the IA group, both processor units and ASPs are basically flat, overall that division grew 1.3%, so there was clearly big growth in motherboard and chipset revenues.
You said they were up, but looks like it was quite vague.
Was this more motherboard units or chipset ASPs and would you expect this trend to continue in the third quarter?
Andy Bryant - CFO
It was a bit of both on chipsets and motherboards.
As I said, we gained a few.
We believe we gained a few points of market segment share in chipsets and gained a little share in motherboards.
And the pricing on our new chipsets reflecting the new feature sets was better than the pricing of some of our older chipsets.
Ben Lynch - Analyst
OK.
So we should probably expect this to continue a bit more in the third and fourth quarters?
That's the question.
Doug Lusk - Assistant Treasurer and Director of IR
I thought it was a statement.
Hard to forecast at that level of detail.
In general, we see -- well, you see in the Q3 forecast.
Ben Lynch - Analyst
OK, thank you.
Operator
Now from Merrill Lynch, this is Joe Osha.
Joe Osha - Analyst
Hi, guys.
Yeah, a couple questions.
For starters, just some simple back-of-the-envelope math would appear to indicate that you're looking for quite a substantial additional improvement in margin in the fourth quarter.
I wonder if I could get a comment on that.
Secondly, as we look further out ,looking at again the timing of this depreciation, you had mentioned the timing -- does that mean maybe we're rolling some depreciation out to next year?
I'd be curious as to your comment there.
Andy Bryant - CFO
In terms of the fourth quarter, clearly, if we hit the midpoint for both the third quarter and the year, as you see, improved margins again in the fourth quarter, which is what you would expect if you're getting full factories, which are helping keep (inaudible) down and if you see seasonal revenue growth.
So it's predicated on seasonal revenue growth, which keeps factories full, which gives me hope of the cost benefit.
Joe Osha - Analyst
Can I get you for the fourth quarter quantify what you mean by seasonal?
Andy Bryant - CFO
Unfortunately, no.
But it's been pretty predictable.
If you go back and look at five or seven or eight years of history, you get a pretty good idea of what we think these days.
Joe Osha - Analyst
Okay.
Andy Bryant - CFO
We do actually give a little more insight than just looking at history, but it's a pretty safe haven to use these days.
As far as the timing of depreciation, well, certainly if the capital is not changing, the depreciation is a little less next year, it could be more next year.
But also keep in mind that we have had some -- some capital that's been retired this year, and as you take -- you sell a business, some depreciable assets go with it, so you see some changes with that as well.
Joe Osha - Analyst
Well, that's what I'm trying to figure out looking out onto determine next year in particular, what do these moves mean in terms of your depreciation burden?
Andy Bryant - CFO
Not a whole lot.
You see, there's a little bit less because of some dispositions and there's a little bit pushed out for timing.
And that's it.
No great insights in depreciation line.
Joe Osha - Analyst
OK, sorry, I will go away in a minute here.
Did I hear you say that you think depreciation will be up next year?
Andy Bryant - CFO
No, I didn't measure forecast or I said that when things slip, what that would tend to do is make depreciation come later.
However, take a piece of equipment that's going to be amortized over four years.
If it starts depreciating in November instead of September, next year's depreciation is going to change.
Four years out, you pick up an extra few months of depreciation.
So it really doesn't have much impact on next year other than this stuff that we disposed of.
Joe Osha - Analyst
Understood.
Thank you very much.
Operator
We have a question now from Tom Thornhill at UBS.
Tom Thornhill - Analyst
Good afternoon.
Andy and Paul, if we could go in a little bit into the drivers on the margin, the gross margin impact.
You mentioned start-up costs coming down.
I would like a comment on 90 nanometer yields working well or a little ahead of plan.
It would seem like that would also fit with unit costs coming down in terms of manufacturing leverage on that side of it, and then on that side of the element driving the margins.
Then on the revenue increase, besides unit and seasonal patterns, can you comment on anything related to mix, new Pentium M, new hyperthreading processors?
If those mixes are positioning more toward mobile, that that has an impact on revenue growth and margin trend.
Andy Bryant - CFO
You do the gross margin and I'll do the products?
Tom Thornhill - Analyst
Sure.
Andy Bryant - CFO
So on the start-up costs, Tom what you see named -- what is it?
Three things hopefully driving approximately three points of margin improvement in the fourth quarter.
Start-up costs, lower unit costs and limited (ph) growth.
Those are all about equal in size, plus or minus a little bit.
So that gives you a magnitude of what is being pulled off.
As for the 90 nanometer, obviously I don't have projection parts (ph) off of it.
It's doing well, in fact better than I thought it would at the June of the year.
If you'd asked Dunn and Lind (ph), he'd probably say it's exactly what he expected, that we always put a little conservatism in case it takes a bit longer.
Certainly as it comes online, it gives us 300 millimeter which will help.
The big benefit to unit costs, though, to the back half of this year is filling -- running the factories that we have at a higher levels of utilization.
Paul Otellini - President and COO
In terms of the 90 nanometer yields, I can't give you specific numbers, Tom.
But - the SRAM yields are right on target.
We debugged the processor of SRAM and the logic yields, which we are now starting to see reasonable volumes on s we ran samples and move quantities on, are looking quite good.
So I would characterize the yields and the process ramp or move towards ramp as right on target.
Tom Thornhill - Analyst
Can you calibrate it relative let's say to the 130 nanometer ramp or are you moving faster than that, for example?
Paul Otellini - President and COO
Well, remember, Craig always shows those files at analyst meetings to show one generation of technology hitting mature yielding faster than the last one.
I would hope to show you that similar curve as soon as we have enough volume to predict it.
Tom Thornhill - Analyst
OK.
Paul Otellini - President and COO
On the mix, that's a very tough one because we have two sort of secular trends offsetting each other.
The one trend is emerging markets where the desktop predominates.
And, you know, there we focus on selling as rich a mix of desktop products as possible, but that allows a much lower entry point for people in computing as they get into their first and sometimes even second machines.
So you have that trend in emerging markets, and they have been quite strong for us a number of quarters in a row now.
Mobile is much more appealing in the mature markets of the United States and Western Europe and to some extent Japan.
And that's where we see the mobile growth really outpacing the desktop, particularly in corporate purchases, and now increasingly in consumer purchases as they make their second and third buying decisions over their life in computing.
Both of those trends and the trend of mating the new high-performance chipsets with HT-enabled processors at different price points is giving us a very nice lift overall.
And we're very happy with the mix and the ramp of both sets of products.
Tom Thornhill - Analyst
Do you see Pentium M on schedule or ahead of schedule?
Paul Otellini - President and COO
Well, Pentium M -- in terms of our ramp?
Tom Thornhill - Analyst
Right.
Paul Otellini - President and COO
It's right where we wanted it to be.
Tom Thornhill - Analyst
Thank you.
Operator
Moving on to question from Hans Mosesmann from SoundView Technology.
Hans Mosesmann - Analyst
Thank you.
Andy, what was the percent of notebooks relative to desktops in the quarter?
Thanks.
Andy Bryant - CFO
I don't that we want to give out that level of -- we have never given out that level of information.
And what I gave you today were a couple of facts to indicate the relative growth in the overall markets and that's probably as much granularity as we can give you right now.
Hans Mosesmann - Analyst
Didn't you give before 20% as a number?
Andy Bryant - CFO
I didn't.
Doug Lusk - Assistant Treasurer and Director of IR
There's been published by Gartner Group and others that say mobile as a percent probably was in that range.
Andy Bryant - CFO
You have to be careful.
Back to what I mentioned to Tom, in general, Gartner et al tend to underestimate the size of emerging markets which are predominately desktop.
Hans Mosesmann - Analyst
I see.
In one follow-up, in terms of ASPs does that include Xbox processors?
Andy Bryant - CFO
Yes.
Hans Mosesmann - Analyst
Okay, Thanks.
Operator
A question now from Smith Barney's Clark Westmont.
Clark Westmont - Analyst
Hi, guys.
Nice job.
I am wondering on the ASP mix, you have covered a lot the issues there.
I just want to come back to it.
I would think with the desktop mix or the mobile mix coming up and your server strength and what not, would we hope to see an ASP improvement in the second half of the year, or is that left to our devices?
Andy Bryant - CFO
We would leave that one to your devices.
Certainly, whatever our thoughts are in that space are included in our forecast, but now, if you look at the first half where, you know, Paul has reported it approximately flat two times now, given the long-term trend of down, that's pretty good.
I am happy where we are.
Clark Westmont - Analyst
I guess I set my self up for that answer, but let me try a different angle.
On the server side you had mentioned you had seen a little bit of improvement in North America in the server business or maybe I didn't catch the language exactly, but could you maybe drill in on that at all?
Andy Bryant - CFO
Well, I said a couple of things.
In total, our server revenue was a record, and secondly I said that in North America we saw the server sales, and to some extent corporate upgrades, were offsetting some of the seasonal slowness you would see in the second quarter in general, and also some of the softness we have seen in the communications businesses.
So the fact that servers are starting to ship in the U.S. and in on a worldwide basis is very encouraging and that was driven by a new products in a stronger mix towards DP products.
Clark Westmont - Analyst
Okay.
I will just take one last follow-up on that and that is, there was a company that announced some difficulties supplying server chipsets last week, so that didn't constrain your shipments at all in server processors, or could he have done better?
Any detail on that?
Andy Bryant - CFO
I don't, I was never called to expedite anything, so my sense is it was not constrained.
I didn't hear that from our customers.
In fact, the opposite is that one of the things that's buried in the server revenue growth is that our server chipsets are doing quite well.
Clark Westmont - Analyst
You have got a next generation coming out in the second quarter of next year, could you give us any update on that.
Andy Bryant - CFO
Next generation what?
Clark Westmont - Analyst
Server chipsets, sorry.
Andy Bryant - CFO
Not right now.
Sorry.
Clark Westmont - Analyst
OK.
Thanks.
Operator
Dan Niles from Lehman Brothers is next.
Dan Niles - Analyst
Congratulations on a great quarter, guys.
Given a lot of positives and, Andy, knowing how you like to put some things in detail sometimes, you talked about the last three quarters of revenues being at the high end of normal seasonality.
And then you stated up-front, I think in your prepared comments, that revenues normally for Q3 are up to 3% to 14%, up 7% on average, but midpoint of the revenue guidance you gave us was up about 5.6%.
Am I getting too precise on these numbers or for the fourth quarter, I guess, or the fourth series of quarters, why do you not expect this to be at the high end of seasonality or am I reading too much into this?
Andy Bryant - CFO
Obviously I'd love to be at the high end of seasonality.
We recognize that what we're trying to pick something close to what we are think is historical average seasonality.
We have talked to customers.
We are also coming out of a pretty strong first half and we want to be careful to compound growth on top of growth and not get ourselves in trouble.
Paul Otellini - President and COO
Yeah, I have actually - the range (inaudible) I'm pretty comfortable with the midpoint at 7 to up you know almost 6%.
It's good enough.
It's good place to start.
Dan Niles - Analyst
Okay.
So I guess assuming some sort of conservativeness is a good thing at this point.
I guess, I am just going back to the gross margin thing for a second.
I mean, sort of using, you know, you had the five-year average of growth for Q4 revenues and then working our way backwards to a gross margin in Q4 would have to be -- it implies your gross margin in Q4 mathematically is somewhere around 57% to 58%, which means you're going to get a heck of a lot of flow-through, again, Q3 to Q4 in terms of -- every dollar of revenue you generate flows through to your gross margin line.
Is there -- is a lot of that just -- I mean, I normally look at incremental margins at running around 70% to 80% when things are going pretty well on a normalized basis.
Is a lot of the extra that gets you closer to 100% almost over the next couple of quarters, Q3 and Q4, is a lot of that just costs shifting out of costs of goods sold into R&D, especially in the fourth quarter, or is there stuff on top of that?
Paul Otellini - President and COO
I wouldn't say a lot of it.
There's still, you know -- we don't have startup costs to zero.
So there's still some room for improvement in the fourth quarter as well.
Keep in mind, Dan, --I feel like I'm bragging on us.
It's not my nature.
But we've done a lot of good things.
We really have driven factory costs.
If we can hold factory spending and fill the factories, you're going see higher than normal incremental margins.
The same thing Paul has been talking about.
We made R&D investments over the last two years doing things like hyperthreading, Centrino, server products.
So great things are happening which allow us to take some share, allow us to have a couple quarters now where ASPs held flat.
A lot of good things are happening now.
Paul and I were talking about this earlier.
He said, this is actually what we designed starting to work if you get a seasonal revenue up-tick or seasonal demand up-tick.
So I'm pretty comfortable if the demand comes, the rest of it will fit because of all we have provisioned in the last two years in terms of cost and technology.
Dan Niles - Analyst
OK.
Then maybe Paul for you, you made some -- you've made sort of two sets of comments.
One, you talked about servers obviously being record sales, corporate upgrades in North America offsetting, you know, some the normal weakness that you would have normally expected to see.
I think last quarter you commented on Europe, seeing trends there with servers if I remember correctly.
Paul Otellini - President and COO
No, those are client upgrades.
Dan Niles - Analyst
Sorry, client upgrades, excuse me.
How does that sort of square back to sort of the statements I think Andy you've made about, you know, not really seeing any signs of corporate upgrades?
Because it would seem you are sort of seeing some of that, on top of, obviously, taking share against AMD.
You're seeing it in servers, you're seeing it in clients, you're seeing it in a couple of different regions, North America and Europe.
Can you just kind of get into that a little bit more then?
Paul Otellini - President and COO
Well, Dan, last quarter I think I used a word -- there were anecdotal signs of corporate refresh in Europe, but nothing that I -- was either large enough or consistent enough to call a trend at that point.
Ill say we have seen more of the same.
Maybe a few more anecdotes.
The fact that servers are, you know, growing along with the clients I think is a very good sign as people deploy their e-Business solutions and are now focusing on lowering costs.
In many cases, moving their work force towards mobile, as we talked about earlier.
All those are good indicators, but it's not a Tsunami yet.
Dan Niles - Analyst
Then I'll go away after this one, Andy; you had two tough quarters in flash.
Do your customers gonna forgive you in Q3 and do revenues sort of ramp up again with that combined with sort of the SARS impact fading in the background?
Andy Bryant - CFO
Certainly that's where I'd like to see some growths in flash in the third quarter.
I actually thought I was going to see some in the second quarter and we saw it pretty flat (inaudible).
You know, Dan, to be - you hate to admit these things.
But to be absolutely honest, I don't think there's a lot of clarity in the flash business right now.
We know you get the cell phone market in China is reported to be soft.
You get less good data there from any place else.
We are the first ones out with the flash business so I haven't seen what my competitors are doing.
It will be important for us to watch and see what happens in the competitor's basis (ph) to get a handle on whether there's inventory overhang there out there.
In truth, I actually expect to see some recovery in the third quarter, but it's -- it's based on assessments through the murky crystal ball.
Dan Niles - Analyst
Great.
Thanks a lot, Andy and Paul.
Operator
Moving on now to Nimal Vallipuram at DRKW.
Nimal Vallipuram - Analyst
Yes, this is Nimal here.
Just a question on the gross margin side.
I think there was a question regarding this.
I'm trying to understand how the gross margin is going to improve in the second half.
On a sales side, you're indicating that you're expecting a seasonal improvement as you have done in the last five to six years.
If you look at the gross margin side, the sales not growing more than seasonally, but your margins are going to improve as it was indicated before with incremental gross margin of 70% to 80%.
You indicated that that's because of the cost containment actions you have taken in the last two or three years and the fabs.
I guess my question is that, is that sustainable going beyond 2003, and if the corporate market and the overall demand for PCs comes back more strongly, how far can these actions take your gross margin up to an extent where is it possible with your current manufacturing model that you can go back to the historical gross margins you enjoyed probably in the mid 1990s and once back in year 2000?
Andy Bryant - CFO
No, I can't make a next-year margin forecast yet, but you did say something that was important to note.
We are in a fixed cost environment which means if you have more demand than you expected, which we saw in the second quarter, and based off that we see the second half of the year in third and fourth quarter, it allows you then to get pretty high incremental margin on that extra revenue.
You cannot sustain that forever.
You eventually will break your factory network if you do that.
So if this continues on, we will have to put some more costs in place, both, you know, fixed costs as well as variable costs, people.
So no, I would not take this and extend it all through next year.
What I would say to the back half, I think I see that I can get more volume on the same assets if the growth continued next year I'd have to put more costs in place.
Nimal Vallipuram - Analyst
Just a follow-up on that, Andy.
Does that imply - I don't want to read too much into what you're saying.
Does that imply that you have moved back -- you have moved back some of the cost inward in your manufacturing from this quarter to the future?
Andy Bryant - CFO
No, we have been reducing costs in the factories.
So as we talked about, they tell you, yes, every quarter --Bryant says, this is how much I need you to save me, and we have been doing it quarter after quarter after quarter.
What's what you do a downturn.
When you are to invest in your R&D, you're trying to invest and you indicate abilities, you take all of your -- we'll say non-technology areas, so some of your factory order, certainly some of your (inaudible) areas.
You don't let them grow, and if you can you shrink them.
As you get a leaner network, when demand pops, you can actually get a good return.
But you can't sustain it.
You will - again, over time you'll have to put more cards in place.
Nimal Vallipuram - Analyst
Just a follow-up for Paul, if I may.
Just to go back to the corporate environment and this so-called mythical pickup in the corporate on the replacement cycle.
Is it possible for some reason -- maybe the replacement cycle is happening at a rate which is probably not discernible by the investors that we probably might not see the word tsunami you used in a short period, to call that a pickup.
But it's probably going to be at a very low level?
Paul Otellini - President and COO
You're asking me to make a prediction on the rate of a corporate refresh and I really don't think I'm in a position to do that, Nimal.
I mean, one view of the world is that we are seeing a refresh happening now and is happening more slowly than we've seen in the past.
Another view of the world is that IT budgets are still a bit constrained and as they get less constrained we're going to see a more traditional refresh.
You know, we'll have to watch which one pans out over the next year or so.
Nimal Vallipuram - Analyst
So I guess my question is which one of the views does Intel subscribe to I suppose?
Paul Otellini - President and COO
We're prepared for either.
Nimal Vallipuram - Analyst
Thanks, Paul, thanks Andy.
Paul Otellini - President and COO
Operator, we'll take two more questions.
Operator
All right.
The first of those will be David Wu, Wedbush Morgan.
David Wu - Analyst
Andy, can you help me with one thing and that is -- if I recall you showed a slide at the analyst meeting that you're going to move a lot of your microprocessors on the 90 nanometers and 300 millimeter fabs.
Will that increase your -- call it the start-up costs in the first half of next year as you move that process quickly or does it not matter that much in terms of that thing?
And Paul, on the Japanese business, Japan is pretty sick as an economy, but your business has been not been so strong for quite a while.
What happened in the Japanese market that accounts for this jump in the Japan's business in Q2?
Andy Bryant - CFO
Depending on the start-up costs.
Once the factory qualified and qualified product running in it, it stopped becoming start-up costs.
It becomes inventoriable costs.
As the fabs transition later this year into a high volume, ramping it will then not be a start-up cost.
So, no, I don't expect to see a big start-up cost.
When we start ramping the fabs in Ireland some time next year, then you'll see an increase in start-up costs.
Paul Otellini - President and COO
Then on Japan, David, the principal reasons we saw -- we saw sequential growth were sales of mobile processors and chipsets for mobile processors into Japan and Japanese companies.
We don't yet know how much of that for Q2 was for re-export or indigenous consumption.
We haven't done that detailed analysis at this point in the quarter yet.
My sense is it's a bit of both.
And then secondly, our flash business was relatively good in Japan, driven by the higher density demand on camera-based phones.
David Wu - Analyst
I see.
Thank you.
Operator
We'll take our final question from Sumit Dhanda at Banc of America Securities.
Sumit Dhanda - Analyst
Hi, guys.
I had a quick question on the gross margin.
Sorry to beat a dead horse here, but how should we interface this transfer resources that you had talked about, number one given that it was relatively sudden, because you hadn't talked about it in the mid quarter update.
Secondly is this in any way or form an indication of perhaps cutting down a little bit on the 90 nanometer transition, and focusing more on a 65 nanometer transition?
How should we think about that?
Andy Bryant - CFO
You shouldn't over-think this.
It has nothing to do with backing off the 90 nanometer, it's what Paul said, which 90 nanometers, we're very comfortable with.
You are only trying to make an estimate in your financials when this transition will take place.
The people who developed 65 nanometers some day will transition in the cost of sales as they (inaudible) 65 nanometers.
All of this is trying to call the point they switch over, they're switching back to R&D a little more quickly than we expected.
It's really nothing more complex than that.
Sumit Dhanda - Analyst
OK.
Thanks.
Paul Otellini - President and COO
OK.
We'd like to thank everyone for listening to today's call.
A recorded play back of the call will be available at approximately 5:00PM.
Pacific time tonight.
Those interested should call 719-457-0820, in reference pass code 523195.
Thank you.
Operator
That will conclude today's conference call.
Thank you for joining us.