邁威爾科技 (MRVL) 2003 Q2 法說會逐字稿

Ladies and gentlemen, thank you for holding.

Welcome to today's Marvell Technology Group second quarter fiscal year 2003 earnings conference call.

During the presentation, all participants will be in a listen-only mode.

After the presentation, you will be invited to participate in the question-and-answer session.

This conference is being recorded on Thursday, August 22, 2002.

For opening remarks and introductions, I will now turn the call over to Co-Chairman and Chief Executive Officer of Marvell Technology Group, Dr. Sehat Sutardja.

Please, go ahead, Doctor.

Thank you, Justin.

Welcome, everyone, to our second quarter fiscal year 2003 conference call.

Weili Dia, Executive Vice President of the Communications Business Group and George Hervey, Vice President of Finance and Chief Financial Officer, are joining me in this call.

It is unfortunate that we are not even close to be able to predict when the economy rebound will happen.

By this time, we all must realize that the huge excesses of the dot-com heydays will cause the recovery to be more painful than that of the past downturn.

In fact, one can ague that many of the once thought of big market opportunities for semiconductors may never come back any time soon.

At least one measure by our capability to plan things for the future.

While this prolonged downturn may be a disappointment for most companies, it actually gives us breathing room to do proper long-term product planning.

That is, keeping to our true and tested strategy we have used from the beginning to build silicon solutions that most of us need regardless of the state of the current economy conditions and regardless to what is in today.

This is, by the way, the same strategy that most successful and large semi-conductor companies have taken in the past.

The main difference is Marvell focuses on building the next generation core technology that will obsolete the existing core technologies that other companies have built in the past decade.

Our core technology may find applications initially in high-end, lower-volume devices, but by using the same strategy used by successful large companies, namely integrating our core technology into other things that people need anyway, we expect our technology to be widely used in future mass market products.

This is the reason why even when the end markets that we are targeting may be negatively impacted by the economy downturn, as we all have seen in the past year, we are still able to grow quarter-after-quarter by growing our market share and by leveraging our technology into other markets.

In fact, we are probably one of a handful of semiconductor companies that is can claim continuous profitability for the last 15 quarters.

Before I move on, George will give you our Safe Harbor statement and a review of our Q2 financials.

Thank you.

Sehat.

Good afternoon, ladies and gentlemen.

I'd like to remind all participants that the following dialogue will contain predictions, estimates, and other forward-looking statements covering subjects such as data storage and communications market trends, competition, customers, suppliers, products and demand, revenue growth, gross margin expectations, operating expenses, other income, accounts receivable, and inventory.

Such statements will be proceeded by other words like: expects, anticipates, believes, should, will, may, or words with similar import.

The following factors, among others, could cause actual results to differ materially from those described in the forward-looking statements: The unability to further identify, develop and achieve success for new products, services and technologies; increased competition and the effect on pricing, spending, third part relationships and revenues, as well as the inability to establish and maintain relationships with commerce, advertising, marketing and technology providers.

We direct your attention to our annual report on Form 10-K, recent quarterly reports on Form 10-Q, recent current reports on Forms 8-K, and other Securities and Exchange and Commission filings, all of which discuss other important risk factors that may affect our business, results of operation, and financial condition.

Please be reminded that we take no obligation to revise or update publically any forward-looking statements for any reason.

Moving on now to our Q2 report: Today we reported that net revenue for the second quarter of fiscal 2003 was a record $119.7 million, an increase of 74% over the $68.7 million reported for the comparable quarter in fiscal '02, and a sequential increase of 21% from the first quarter of fiscal '03.

Pro forma net income which excludes the effective acquisition-related expenses, amortization of stock based compensation and a special charge related to facilities consolidation was $14.2 million, or 11 cents per share diluted for the second quarter of fiscal '03, compared with pro forma net income of $3.4 million, or 3 cents per diluted share for the second quarter of fiscal '02.

Shares used in computing pro forma earnings per share diluted for the second quarter of '03 were 129.6 million compared to 127.4 million shares for the second quarter of fiscal '02.

Additionally, Marvell's Chief Executive Officer and Chief Financial Officer will certify the Company's financial statements as required by the Sarbanes Oxley Act of 2002 in connection with the filing of the Company's Form 10-Q on or before its due date of September 17, 2002.

Now I'd like to turn the call back to Sehat for comments on our business.

Thank you, George.

I would like to first review our storage business this time around.

This is the area that a lot of our investors are not too familiar with, and as a result, many are baffled by our continued market share gains in an area that is thought to be highly matured.

This is a classical example of our true and tested strategy of developing new core technology to address the high-end market, which in this case is the enterprise storage market.

As our technology matures, we are integrated with other things to create SOC, or System-on-Chip, for the higher volume markets such as the mobile and desktop markets by building more powerful yet more cost effective solutions, we then take market share away from our competitors.

We believe that as we continue to develop more and more advanced core technology and develop IPs to create more complex System-on-Chip for this market, our market share can only grow.

Our superior technology combined with our commitment to long-term partnerships with our customers give our customers the confidence that they can always rely on us to provide them with the very complex devices that they need.

The most recent notable announcement of new devices in our core technology is the 7500 class of rechannel physical devices.

These are the most complex mixed-signal DSP device ever built for any semiconductor market.

This device includes mixing circuits such as an analog to a digital converter that operates at close to two gigahertz sampling frequency, utilizing standard [INAUDIBLE] 0.13 microns digital CMOS process technology.

The DSP circuit itself is equally impressive, as it performs on-the-fly competitions of the digital signal at hundreds of billions of operations per second while using less than two watts of power at the maximum operating speed.

When eventually combined with our SOC technology, the result is a device that will be able to fully extract the performance of the next generation Serial ATA drives that are targeted for the next year desktop PCs market.

Talking about Serial ATA, this is another example of our core technology that we have developed for over the last several years.

We originally introduced this technology late last year initially for the rate market.

Today, practically all target storage devices shown in the industry first, [INAUDIBLE] compatibility test held last week use our Serial ATA silicon.

All of our customers are currently sampling and preparing to ramp up the Serial ATA storage devices for use in mainstream desktop PCs by the middle of next year.

The movement to Serial ATA is accelerating rapidly as PC manufacturers are starting to realize that by next year, PCs with Serial ATA will outperform ones without Serial ATA by about 20-30%.

In the short 4 years since we started shipping our first regional device, we have achieved the largest market share in the enterprise segment, and more recently, in the mobile segment as well.

Now we are on our way to gaining significant market share in the desktop segment.

The desktop segment is an important area for our continued future growth in the storage market.

It is the largest storage market segment, at least until the consumer, the new area takes off in the next two to three years.

We are glad that with the early reproduction of our newest design win is in the desktop market on the way, we can now disclose our biggest design win ever.

We have just recently signed Western Digital to work with us in using initially our stand-alone physical layer devices for their mainstream desktop platform and eventually using our System-on-Chips for lower cost and higher performance solutions.

Meaningful shipments to Western Digital will occur next quarter with some initial shipments this quarter.

We expect full transition to our solutions by the end of next year.

Our goal is that in two years we will be the largest supplier of storage electronics.

To do this, we are investing heavily in building the ATA team and building new IP cores to make sure that whatever our customers need in the future, we will be able to provide them.

In fact, our goal is to be a complete one-stop shop and service for our storage shipment product solutions.

Now let's move to our communication business.

We currently address two main markets.

The first is the gigabyte ethernet and the second is the wireless LAN.

I'll give you a short update on the gigabyte then spend more of my time left on the wireless LAN.

Again, gigabyte ethernet is another example of our strategy of building complex core technology for the masses.

While initially gigabyte ethernet was only needed for the enterprise server market, from the very beginning, we were targeting gigabyte ethernet technology to quickly replace the existing fast ethernet market.

We are proud that in a short two years since we introduced our gigabyte ethernet technology, we quickly became the market leader.

We believe we have now shipped more ports of gigabyte ethernet solutions than all other suppliers combined.

Our next goal is to make gigabyte ethernet the standard communication device for next year's PCs.

Already today, Dell has deployed our gigabyte technology for chips that we developed for Intel in their [optiplex] desktop PCs.

As more and more of our customers realize the numerous benefits of our gigabyte technology by including, among other things, the drastic cost reduction of network installations, we are confident many of our present futures of our gigabyte technology will become the standard PC requirements.

In terms of market opportunity, the gigabyte market will translate into about 300-400 million ports or chip set solutions by the time the transition from fast ethernet to gigabyte is completed.

We do expect that by further driving down the cost of gigabyte to the desktop PC, as well as the entry level enterprise switches, the infrastructure, this transition to gigabyte will happen sooner than most people expect today.

In fact, we will do whatever it takes to make sure this will happen.

This is the reason that we have recently acquired [Sisconnect], a small but very well known company based in Germany, specializing in high-end server solutions.

With this acquisition, we now have the industry's most complete software drivers, device drivers for all major operating system platforms.

Finally, I would like to provide an update on our wireless LAN business.

This past Monday, we formerly introduced the industry's first and only two-chip solution for the 80211B wireless LAN market.

What we mean by two chips, we really mean two.

Not two plus one with the one missing from the announcement.

Our two-chip solution is a follow-on to the industry's first, which is ours, and still the only 802.11BRF solution with integrated power amplifier.

More impressively, all of this is do done in standard digital CMOS process technology.

As you can see by now, this is yet another example of our strategy of building very complex course or mixed signal and DSP circuits that while initially are targeted for niche high-end market will eventually -- eventually, this technology will go into very high volume markets.

In this case, the 802.11B wireless LAN solutions are currently mainly used in higher priced DSL or cable modems, access points and high-end laptop computers.

By driving the costs down in the future, new market opportunities will emerge.

Already the industry is talking about deploying 802.11B into wireless cell phones, IP-based cordless phones, MP3 players, home stereo systems, car-to-home LAN interface, printers, digital cameras, and too many other applications to mention here.

In some sense, this can be a bigger market than the core technology compared to the first two markets we entered earlier.

Potentially, in the not-to-distant future, a billion of such devices incorporating the 802.11B may be produced each year.

We believe for some massive acceptance of the wireless LAN technology into the consumer markets to materialize, two things must happen: First, the cost must be lowered, which is a given requirement in any large market.

By using standard CMOS technology and simply by taking advantage of future process, such as using the emerging 19 nanometer process node, we plan to eventually cost reduce the chip-set solution to around $5 per system.

Second, but most importantly, the wireless LAN reliability must be improved drastically for wide consumer acceptance.

This includes making device that work at longer distances and devices that work in rooms with severe distortions or problems.

This is critical because most consumers will not want to install multiple access points in their homes just to increase the coverage of their home range.

Nobody really wants to have two or three access points, because they need to cover all their rooms including living rooms, bed rooms, back yards, kitchens, with their wireless LAN network.

At all, they will need wired Internet as soon as they use more than one wireless access point, thus negating the reason to use wireless to begin with.

By being the first to actually solve all of these problems, we are confident that we will repeat our success in this new area.

Speaking on reliability, our first generation device already operates with as much as 4-5 times the amount of multi-path delay compared to the most successful solution available today.

Of course, one can ague that this is not a fair comparison, as existing solutions utilize technologies they will develop four or five years ago.

However, this is precisely how we plan to capture market share away from the existing suppliers; by introducing a much more advanced technology, the users will, at the end, make the call.

Will they spend their money on something that sort of works, or spend their money on something that works much better?

I hope you can see by now the reasons why Marvell can continue to grow rapidly in the next several years.

And as we continue to invest in other new developments that we are not able to talk about yet at this time, we will hope to extend our leadership position into the next decade.

Now I would like to turn the call back to George for additional comments regarding our financials and guidance for the next quarter, as well as updated guidance for the year.

Thanks, Sehat.

First I'd like to make additional comments in our Q2 results.

Our Q2 revenue of $119.7 million, again, was a quarterly revenue for the Company, and a 21% sequential increase from Q1.

This compares very favorably to our guidance of a 10-12% sequential increase in revenue from Q1 to Q2.

As we mentioned on our Q1 call, we expected in Q2 to begin the ramp of gigabyte ethernet to the PC client market and guided for a high-teen sequential revenue growth in our communications business.

We're very pleased to report that the ramp is progressing ahead of expectations, and the sequential revenue growth from Q1 in our communications business was 38%, with gigabyte ethernet revenue increasing 90% from Q1.

Additionally, we continue to see signs of improving business conditions in the traditional segments of our communications business.

Moving to storage, there have been a number of data points that indicate overall unit demand in the storage market in Q2 was flat to down from Q1.

While we agree with those data points, we continue to be successful in growing our storage business.

Our guidance or revenue growth from Q1 to Q2 in storage was mid-single digits, and our actual sequential revenue growth was 11%.

Driven by increases in production ramps of SOC solutions and desktop pre-amps.

For the quarter, storage represented 57% of total revenue, with communication products representing the balance.

In my conversations with the analysts and investors during the last quarter, there have been a lot of questions raised about our long-term gross margin percentage.

I'll have more commentary on this subject in a few moments when I update our guidance.

As previously mentioned, our gross margin percentage is primarily driven by product mix, and with a high volume of PC client shipments representing a greater percentage of revenue, gross margin percentage on product shipments was lower than our original quarterly estimate.

But this mix also resulted in the higher revenue, higher gross margin dollars, and EPS upside.

For Q2, the 21% sequential revenue increase resulted in gross margin percentage on product shipments of 54% versus our guidance of 55%, which had been based on a 10-12% sequential revenue increase.

Also early in Q2 we experienced some manufacturing inefficiencies associated with the ramping of production to meet our revenue demand.

Those inefficiencies were corrected, and we exited the quarter with production returning to normal, but the impact of the inefficiencies from early in the quarter resulted in our overall gross margins for Q2 of 53.2%.

Operating expenses as a percentage of revenue in Q2 were 41.3%, a decline of 450 basis points from Q1, which is approximately a 320 basis point greater decline than our guidance.

As evidenced by the increase in our operating income percentage over the last several quarters, we have begun to see then positive operating leverage that we expected from our rapid revenue growth.

Since Q2 of last year, we have increased our operating income from 2% to 12%, which brings us significantly closer to our long-term model of 18%.

During the quarter we completed the acquisition of [Sisconnect GMBH].

The net effect of this acquisition on our Q2 results was non-dilutive, and we expect that the [Sisconnect GMBH] acquisition will remain non-dilutive to our results for the balance of fiscal '03.

On the balance sheet. our cash and short-term investments declined by approximately $7 million to $250 million.

Cash flow from operations was break-even, as we invested a significant working capital to fund our rapid growth.

Additionally, during the quarter we made several strategic investments in private companies.

DSOs for Q2 were 53 days, which is consistent with our guidance of high-40s to low-50 days.

DSOs will most likely remain at our high end or our range during these periods of rapid revenue growth.

As we mentioned last quarter, our days of inventory had declined to 45 days, which was below our target of 65-70 days.

During Q2 we were very successful in ramping production to meet the gigabyte ethernet revenue demand, build sufficient inventory levels to support increasing demand for gigabyte ethernet going forward, and load the front-end portion of the desktop storage ramp.

This resulted in our Q2 days of inventory increasing to 79 days.

As we balance these production ramps during the second half, we would expect our days of inventory to range between 65-70 days.

Now turning to the future.

As we enter the second half of fiscal '03 and analyze the significant revenue opportunities that we are engaged in, we would like to take this opportunity to update our guidance regarding the financial outlook for the Company for all of '03.

As we have mentioned previously, we have significant revenue growth opportunities which are not directly tied to the growth of the end market.

Two key examples are the continued adoption of gigabyte ethernet as the replacement upgrade from fast ethernet, and the second trend is the further expansion of SOC devices in the storage market.

In addition, during the second half of '03 we will begin realizing revenue from our new desktop storage opportunity, Serial ATA, and earlier revenue from wireless LAN.

We currently expect our revenue for fiscal '03 to range between $490-500 million, which at the mid-point is approximately a 71% growth from fiscal '02.

We would expect our storage business to represent approximately 55% of revenue, with our communications business representing the balance of 45%.

Gross margin percentage for the second half of fiscal '03 is expected to increase from Q2 and range between 53.5 and 54% , which is comfortably above our long-term model of 52%.

As we continue our technology investment, operating expenses will increase in absolute dollars but will decrease as as percentage of revenue as we make progress towards achieving our long-term model of 18% operating income.

We are positive very regarding our outlook for the remainder this year and expect our pro forma EPS to range between 46 to 49 cents per share, which at the mid-point is approximately a 200% growth from fiscal '02.

Now moving more specifically to Q3: Based on the rapid growth or gigabyte ethernet, we experienced during Q2 we feel these will trends continue into Q3 providing us the opportunity for accelerating growth.

We targeted the company revenue for Q3 to increase 10% from Q2.

We'd expect our communications revenue on a percentage basis to sequentially grow in the high teens and storage in the mid-single digits.

Our backlog entering Q3 was at a record level for both our businesses, and the backlog percentage to our revenue guidance improved significantly, especially in our communications business.

Due to the continued volatility in our markets, we believe it's prudent to anticipate that our actual Q3 may range within plus or minus a few percent of our target.

As I mentioned, our gross margin percentage for Q3 is expected to increase from Q2.

Entering Q3, the manufacturing inefficiencies we incurred in Q2 have been corrected.

Our manufacturing organization continues to focus on reducing the costs of our devices, and the product mixed change will not be as significant as Q2.

Taking these into account, we expect Q3 gross margin to be 54% plus or minus 25 basis points.

As we continue our new product investments and structure the company for continued long-term growth, operating expenses in absolute dollars will increase in Q3.

But as we have previously mentioned, we are committed to increasing our operating income percent.

Consequently, Q3 operating expenses as a percentage of revenue will decline at least 130 basis points from the Q2 level.

During these times of reduced interest rates, the return on our cash assets has decreased.

We would anticipate Q3 interest and other income to be approximately $400,000 lower than we experienced in Q2.

In summary, considering the various factors impacting our business, we anticipate Q3 EPS to be 13 cents per share, an increase of 2 cents per share from Q2.

Now I'd like to turn the call back to Sehat.

Thank you, George.

That completes our commentary.

Justin, would you please poll for questions.

Absolutely.

If anyone would like to ask a question today, you may do so by pressing the star key followed by the number 1 on the telephone keypad.

Again, to ask a question please press star 1 at this time.

And we'll take our first question from Michael at CSPB.

Congratulations.

A great quarter.

Talking maybe for a minute on margins and thinking about this as an established company; this could be way on the gross margin a little bit, but looking on the operating margin side, should it be higher?

That's correct, yes.

And you did incremental margins around 21%; is that something you think is a decent level going forward?

No, operating income was 12%.

Incremental?

Incremental growth.

As we said, we clearly are targeted at 18% as our long-term model and come from 2-12 within a year.

When you run out, you know the guidance we just gave you, that 12% is going to improve significantly between now and the end of the year.

We probably are creeping up on that quickly.

You talked about a number of new applications where you are ramping up in the second half of the year, what percentage do you think that could be for the second half of this calendar year?

Well, wireless LAN will be less than 1% of revenue.

We don't really break out any specific pieces, you know, within the product lines with the exception of gigabyte, but I think over time we would expect Serial ATA, as Sehat mentioned, to be mainstream in PCs next year with significant volumes.

Great.

On the ethernet side, you have an opportunity to see some pull through and motherboard ramp; when do you expect to see that?

If you talk to PC manufacturers, you'll find out that everybody is working to put gigabytes to all the future desktops.

The question is when those products will be introduced, and that depends on the OEMs and OEM basis.

What we do feel very comfortable about is that the time has moved to the directions of gigabytes.

Nobody is talking about lower costs any more, so the modems will be very high.

The question is what there is going to be -- what the ramp curve is going to look like.

We don't know how to predict, but this is the reason why we have built these products.

We have to build more than what we think the markets will need.

I think we still remain conservative in what we would expect overall for gigabyte, which, by the way, is a dramatic increase as the ramp has started here.

I don't think we've, again, seen the full anywhere near the impact of what it could be as we move into next year.

Also, you know, again, most of the growth right now is still coming on the client side, and you know hopefully by the end of the year as more and more client units are in the market this will accelerate the infrastructure side.

And I know that Cisco made positive comments on their call regarding CAT4 and 3550, which are products that our gigabyte, you know, is included in.

So we think the trend there is very positive.

Great.

Thanks.

Again, if you would like to ask a question today, you may do so by pressing star 1 on the touchtone phone.

We also ask that you please limit yourself to one question initially so as to have time to get to everyone's question.

We'll move next to Nathaniel [Cohen] of Goldman Sachs.

Great.

Thanks.

Nice quarter.

I wonder, Sehat, if you could talk a little bit more about the Western Digital opportunity that you just announced today.

You said you're starting to ship this quarter, and it's the rechannel only.

At what point do you see the SOC transition happening, and are you going to be selling other products besides the rechannel in there?

Thanks.

First of all, Western Digital is one of the three largest desktop manufacturers in the storage market.

So this is an important design win for us.

We've been working, as we mentioned in the past, to put our effort, concentration of our efforts to gain market shares in desktop.

The initial design wins happens to be on the stand-alone rechannel first, and reason for that is because this is the fastest way to ride the softwares, and then we do expect this will be only just for one generations and from then on, all subsequent generations will be using the System-on-Chips.

There might be some intermediate half-way generation that will still use the stand-alone devices, but that yet has to be seen, because all the efforts are being put today, both by us as well as by our customers, to write codes to -- for the SOC.

So conservatively I will say second half of next year or when the ramp-up SOC will take place.

Okay.

Do you -- I assume then with the SOC -- do you already have design in place and the activity is going currently and it's just a matter of transitioning to next generation platforms?

Or how would you classify what's happening on the SOC front?

Classify as heavy [INAUDIBLE].

They have the chip we did for them, and it's a matter of when they will bring that across into a platform.

Great.

And on the preamp side, you are shipping currently in [INAUDIBLE], I believe.

Can you just give an update on how your success is going on the preamp side?

You mentioned -- you are asking about other components.

Yes, we are shipping preamps to other suppliers.

We have to talk about specifics, but we also are in the process of final developments or soon to be introduced modem controllers, what are called [combo motor] controllers functions into a single chip device.

So by sometime middle of next year, we'll have a complete set of Silicon, everything you need into the drive, we'll be able to provide.

And will you be shipping that by middle of next year ,or will you be shipping that sooner?

Depends on some customers will probably will ship earlier than others.

The [motor] controllers will probably start shipments initially second half of next year with more volumes ramp-up by the end 2004 time frame.

Okay.

Great.

And then just one real quick question on -- for the gross margins, George, your model you talked about it at 52%, sort of a longer-term model, can you -- and the guidance, I guess with the glitch this quarter is for gross margins to be back to the 54% range -- can you just comment on how you see gross margins, given the guidance that you've given, how you see gross margins trending over the next several quarters?

Thanks.

Sure.

Well again, I think we've defined in some pretty good detail what we believe for the back half of the year, which is they should range between, you know, 53.5 and 54.

You know, again, when we developed the long-term model, and I think as probably most people know, we actually changed the long-term model down to 52% to basically take into account how our business is changing.

I think Sehat went through a very detailed discussion of where the opportunities are that are going to lead us into these very, very, large markets.

We took that into account when we analyzed where we should set our model moving forward.

So, you know, the model contemplates obviously product mix, which is the single biggest factor, but it also takes into account pricing trends.

And we believe we have some pretty good visibility into what pricing looks like on our, obviously, our existing business and a very opportunity that we talked about today, obviously we know what the pricing environment is there and the things we are going to be looking forward in the future.

And then of course, the cost side of the equation is also a key component.

As I mentioned, our operations folks are working very hard to continue to give us the lowest cost devices possible.

So, you know, when you put all those things in, those are the factors that we consider, certainly, when we kind of put that model out there.

Now if I look out into, you know, well beyond the end of this year and look out into next year, you know, I honestly don't see factors that actually would drive a point quarter down to, you know, actually 52%.

So most likely, you know, the earliest that we believe we would see an actual gross margin, potentially -- and we have to say, there's a lot of factors that may never get there -- but, you know, given what we know today, it's clearly beyond the end of next year.

And we'll move to Sandy Harrison with Bank of America.

Good afternoon, gentlemen.

Again, nice quarter.

Sehat, you talked a little bit about the transition at WBC from the stand-alone rechannel to the SOC.

Could you help us understand a little bit how that's going to work from an ASP perspective?

What are you looking at right now for ASPs in the rechannel versus what you could be getting from the SOC?

Sure.

The partition to SOC --[INAUDIBLE] in the first design win, it's a stand-alone device.

So that's for the mainstream platforms.

So when it transitions to the SOCs, most likely it will transition to practically all the platforms, with the exception maybe from the very, very high-end niche market segments.

That will translate approximately, I will say, 1.8, 1.9 times of the ASP of a single device or stand-alone device.

So in the future, it combines with 3Ms and motor controllers, this will translate into a huge business for our storage business group.

So what you're saying, to make things easier, a dollar today for a stand-alone would go to 1.80 for a system-on-a-chip or 1.90, something like that.

Somewhere in that range, yes.

And then you get the volume kicker.

Right.

By the way, that's what's been happening with Samsung.

I mean Samsung, which was our existing main desktop guy, we started converting them to SOC late last year, and in our commentary, I think, you know, a lot of the revenue growth that we're experiencing this year right now in these challenging environments is coming from the fact that we're taking that -- seeing that 1.8 multiplier take effect at Samsung.

Gotcha.

And then just a quick follow-up: I mean, obviously with the success you guys are having in the markets, is this there a concern, or what are the views from the competitive landscape here that you guys, being so successful and the markets growing, that you see a greater competitive threat coming in.

And has the competitive landscape changed given the success you guys have had in the last two quarters?

I think the competitive landscapes have not changed much in terms of the technology.

If anything, we are seeing our competitors is getting further behind in technology introductions.

However, the prolonged downturn consequently probably maybe a slight increase in pricing pressure because of, among other things, excess capacity in the manufacturing lines.

This, I believe, is a short-term issue, because as we move to the future, a lot of those people that have their own factories, their own manufacturing lines will not be able to simply just fill up the capacities as they have also have to go to off-site foundries to build their products.

So I will say that we don't see that much change in the competitive landscape.

Of course, the weaker ones we are seeing in the [INAUDIBLE] space, the weaker -- the people in the fast ethernet are losing the market shares very fast, because the market is moving to gigabytes.

And storage, landscapes very much left to us and the [INAUDIBLE].

It's just a battle between the two us.

All right.

Thanks, guys.

And the next question we'll take with will come from Quinn Bolton at CIBC World Markets.

I have a few questions.

One, just wondering about the wireless LAN opportunity.

I know you were initially targeted in the consumer space.

But I wondered if you could comment -- Intel it looks like next year will be rolling out the new [Bannias] platform and their Prescot Pentium 4 platform second half of next year, both which will be looking at integrated 80211AB combo solutions.

Wondering if you can address that opportunity and what your plans are for it, and then I've got a couple of follow-on questions?

Sure.

If you notice in my -- for today's presentation, one of the key, very important things to notice is that the wireless LAN -- for wireless LAN to be widely accepted in consumer market --, when I say consumer, I mean the people that will buy this product for their homes, including for their PCs and their laptops -- the solution has to be much more robust than what you have today.

As far as I know, as far as what we have seen in the market through extensive benchmarks verifications with other products that are available in the market, we are significantly better, specifically in the range and amount of distortions, multi-path distortions.

We are four to five times better than the industry leader in this market.

We have the highest integration solutions.

Our chip is the only one in the market -- we have one chip has everything in it including power amplifiers in one chip, and we have the other chip that has the microprocessors.

We integrated the -- we have the only one in the industry to have an online class of controller into our [MAG].

We have baseline processor, we are integrating all the analog circuits into the second chip.

So we do have the, basically, the lowest chip count in the business.

And because of that, the solution is applicable for people that don't care about the price but care more about quality such as the laptop manufacturers, because they cannot afford to get phone calls from the customers complaining why the chip doesn't work to all the way to consumer devices where cost is more important.

So the [INAUDIBLE] in these chips.

Whatever you do, you get both, at least these two chips.

So we are comfortable that we should be able to take significant market shares when the dust settles.

How about specifically in terms of an A-B dual mode or combo solution?

Because I think that's what it looks like Intel will be trying to push into the PC market beginning next year.

Uh-huh.

SO, we talk about next year, we have all of those as well.

Okay.

We're not preannouncing products, though.

You have to be careful.

So we should just stay tuned.

We are announcing products that are ready for production today.

We are not in the habit of introducing products that takes another nine months or six months to be ready.

The markets that our customers are asking for are is more than sufficient for all the applications.

By the way, DSL is only running at [INAUDIBLE] megabytes.

So 11 megabytes per second is about 20-30 times more than you have, more than you have in homes.

And if you look at -- the immediate migrations will be the G. The G will use the same [RF] as the B. The only difference will be the A-base band, which is mainly just a [DSP] change and maybe some slight analog change.

So the transition to G will happen first, and then the A will happen later, specifically because the cost will be higher due to the in fact that you need two different antennas and two different RFRNs.

Then two, sort of, quick follow-ups.

Recently, Outside Logic and Infineon announced a joint venture to start designing chips for the hard disk-drive market..

Wondering if you could comment on that.

And then also just wanted to see if you wanted to give us a quick sense of what the mobile and enterprise market conditions are.

Are those markets more flattish with the ramp in desktop driving the growth, or are you still seeing continued growth in mobile and enterprise?

Thanks.

Sure.

Four years ago when we entered the business, we had about two dozen competitors, and these are companies that have a lot more experience than Outside Logic and Infineon.

We're talking about people that shipped hundreds of millions of units of products into the market before we entered the business.

We entered the business by building products that are significantly more superior compared to what's available then, and by using better technology and lower cost product, lower powered products, we are able to future market share away.

Now the game has changed to integrations, system-on-a-chip integrations, so now we think we have the most advanced stand-alone rechannel solutions, as we have announced this week.

We also now have in-house technology for system-on-a-chip.

So Infineon is a company that has been trying to enter this market for the last four years.

They are hiding the business under their memory divisions.

They realize that without the system-on-a-chip technology, they won't be able to get into the markets.

I will say it's a long shot for two large companies to work together and to be successful long-term.

If I'm the customers, I will not trust two companies to be able to work together for such complex devices.

It's way, way too complex.

And then on the other part, as we indicated in the commentary relative to Q2, you know, overall storage market conditions were actually probably down unit-wise quarter-to-quarter.

The most current view we have at the moment is Q3, from a market standpoint, appears to be flat, you know, with Q2.

We don't see -- our customers have not indicated, you know, that we should expect overall up units.

Although, they are indicating that by the time we get to the fourth quarter, we should expect to see up units in storage.

Before the Christmas season.

And we'll move to the next question from [Ambrish Rustada], Gerhard Clower.

Hi, thanks.

A couple of quick questions for you, Hervey, in terms of what was the depreciation in capex. and then also any update on the better than 10% customer.

And the second question for you, Sehat, is on the Serial ATA, you mention, and I didn't catch what acquisition or short was.

You mention that most vendors were displaying your Serial ATA silicon.

Any light on the competitive landscape in that area that would be great.

I don't happen to have those specific numbers here for the call.

We can follow up with you later on the capex and depreciation.

On the actual top-ten customers, again, we break them down by the two businesses.

We'll start first, again, with the storage guys, and they're the traditional names that we talked about previously: Toshiba, Hitachi, Samsung and Seagate are all 10% customers, and now we do have a 10% customer on the count side, which is obviously Intel.

Let me cover the Serial ATA.

The Serial ATA, as I mentioned, we introduced late last year, originally targeted to address the rate markets, people that cares about performance.

What we found rapidly that the market is moving forward to the desktop, as people are realizing that they need higher speed communication device between their storage device to their PCs.

So we saw huge activities in the last couple months on the -- on our customers to build systems because of the expected compliance fest that was held last week.

In that [block test], we found out that about 80%, I think, 80 or so-plus percent of the devices available out there using our silicon.

In fact, there's only one that doesn't use our silicon.

And the comments from the people that run the block test is that for the ones it works real well are the ones using our silicon.

So in terms of the competitive nature of the Serial ATA, everybody in the storage business is working on Serial ATA, they all promise to deliver the silicon, it's just the nature of the business.

People must have the silicon, but today, the one that is shipping this type of silicon is only us.

There's another company that I don't want to mention that is shipping a small -- very, very low volumes post device, who's device bridge, but we haven't seen that in actual competitions yet ourselves.

So even on the host device, our customers are using our device, as well.

And we'll take our next question from [Arnub Shandun], from Lehman Brothers.

Impressive performance here.

Couple of questions, I guess.

First is a tag-along question.

Is there a memory strategy in the SOC in the future?

And then a couple follow-ups, if I could.

Our chip always have embedded memory.

Question is how much.

[INAUDIBLE] more embedded [INAUDIBLE], is kind of what I was --

Oh, okay.

So you're saying, Is there any plan to increase the cost beyond what our customer can afford.

The answer is, no.

Good.

That's simple enough.

The next question is that I think we have a bit of a positive surprise in the gross margin in terms of the guidance, and just was wondering if you could give us clarification on how you are increasing or your reducing your costs.

Is it something to do with the back end, is it a fond relationship, your design, process, technology; just, could you shed a little light on that.

Sure.

One of the key things I think that is becoming very obvious is that Marvell is becoming a much bigger company.

And of course, as you become a bigger company, you have more buying power in the market.

And clearly, you know, we're becoming a more important customer to our foundry partners, to our assembly and test partners.

You know, if we can provide them more business opportunities, you know, we will always be looking to see if we can get more favorable costs associated with that.

So you should -- but that's just a normal part of running our business.

That is not something that we all of a sudden we woke up and decided we needed to start doing.

This is something that our operations people are absolutely focused on.

So clearly, yes, those things are taking place.

The yield is very important, and our yields are very good, and as products mature, they will get even better.

We are not doing the geometry move at this point.

We don't see a need to have to do that.

It's not clear that even if we went to .13 micron we'd get a better cost structure than what we currently have right now.

That's something you saw our .13 micron gigabyte products that we announced a couple weeks ago, you know, those will be production next year when the cost is beneficial to us.

So all in all, to us it's business as usual.

But business as usual is to come up with a low cost.

Thanks.

Last question, if I may: In terms of the storage business, can you talk a little bit more specifically about what you are expecting from the Serial ATA business next year -- sounds like you're pretty excited about that -- and how this sort of -- if there any opportunity with the two that are left behind, I guess, as not customers, [INAUDIBLE] and Seagate, is there a Serial ATA strategy or what you are doing there, specifically?

Thanks.

Our goal is actually to use the Serial ATA together with our rechannels and other IPs that we have and aggression engines and so on to capture the electronics into the drive.

If that means integrating the Serial ATA and rechannels and other stuff into one chip, so be it.

So the intermediate steps obviously is enabling the existing suppliers even though they are not customers to use our Serial ATA technology.

We do want to make sure that all the PCs out there will interoperate with the target devices, and by ensuring that all those devices use our technology, in the future the transition from the ATA to Serial ATA will be smoother.

So when our customer needs to have those integrated, what we hope is that they will call us or we will work with them to make that happen, to integrate the Serial ATA into the SOC, and we'll give the cost reductions at that time.

So because this is our long-term goal, I cannot give you the predictions what the volume of Serial ATA on a stand-alone basis, because eventually they are going to be integrated anyway with our SOC.

Thanks, Sehat.

And the next question comes from Mark Lipisis with Merrill Lynch.

I have a clarification and a question.

On the Western Digital, Sehat, when you talk about full transition by the end of 2003, does that mean we should model you guys taking 100% of the market share and rechannels and their drives by the end of '03?

Well, as mentioned, most of the drives, with exception maybe is if there are any specialty drives that we run at low volume or requires special controllers.

But I would say 100% of the mainstreams.

So that would be like 80-90% of their total unit shipments, something like that?

Should be at least.

Fair enough.

Ant then, when you talk about 90% sequential growth in gigabyte ethernet, which is clearly impressive, you typically think, you know, gosh, these guys may be building inventory.

What do you do to make sure Intel is not building inventory or your components and you potentially see a hiccup in orders from them?

The first indication -- you really can do -- two things that we can do.

One, we have a close relationship with the largest user right now, which is Dell, because we also supply our networking chips to Dell as well.

We've gotten some good indications that there is not any abnormal inventory being created here, so we feel comfortable that there is sell through happening on the PC/OEM front.

And then of course as evidenced by our guidance, the demand from Intel hasn't -- is not flattening out.

It's continuing to increase.

So you would think if they had overbuilt or something like that. we wouldn't be seeing an increase that in fact we are seeing.

Okay.

Fair enough.

And then on -- I mean, has Intel made any indications to you that they plan to design you out of any of their DD programs at all?

We've received no communication from Intel.

Okay.

And finally, George, on the balance sheet, other non-current assets were up by $13 million?

What's included in there, and can you [ Audio inaudible or unavailable ] they went up?

Bear with me a second.

The other -- that should be, and I reserve the right to double check afterwards, I believe that has to do with the software licenses for our tool that is we signed.

I'll have to double check that.

Okay.

Thank you very much.

Okay.

And the next question comes from David Lou from Bush Morgan Securities.

Good afternoon.

One classification, I have a question.

The classification is what kind of manufacturing new problems did you encounter earlier in your second quarter and the question relates to you mentioned about 802.11B going to .09 micron CMOS, which will yield to you a cost of $5 on the present two-chip solution.

I wonder what time frame we are talking about, and have you taken into consideration, I guess, the TSNC had problems with micron migration?

How confident are you on that .09 micron for TSMC and also on the same issue, if Intel were to go at .09 micron to the base and the Mac of their future generation of mobile processor and leaving only the RF outside the microprocessor chip, would that present a major problem for you?

Okay.

I'll take the first spot and let Sehat deal with the rest of it.

Semiconductor processing is a very dynamic manufacturing process.

You know, we literally on some of our especially on the gigabyte ethernet, we were going from zero units to millions of units.

Sometimes those start-up have a few bumps along the road.

So we hit a couple of those little bumps.

You know as the good manufacturing controls we have, we were quickly able to address those, correct them and move on.

I won't be specific what they were but they were clearly things that we were able to quickly fix and move on and, by the way, be able to meet the objectives that we had which were substantial for Q2 both from the amount of revenue we shipped to Intel but also feeling of our manufacturing flow to allow us to ship even greater numbers of products this quarter.

Because of the levels of the engagement, the visible engagement, we do need to put more -- spend more time testing to make sure the products we are shipping of the highest quality levels.

You overkill it doing the testing and so on.

Over time, those inefficiencies go away.

As you have more confidence level, you can cut down the amount of test time and so on.

Now regarding the wireless, today our Silicon authority is using 1.15 micron technology.

Not far away from 9 nanometer process technology.

So the migration to 113 and eventually to .09 will happen.

What we are doing is we are doing that for targeting the higher volume markets.

For the markets we are talking today for the PCs, for the laptops, there's no need to go to 9 nanometer because they don't need a $5 solution at this time for that kind of volumes.

What we talk about $5 solution that we are talking about applications for cell phones that goes to half a billion units in markets where every dollar cost counts.

And markets not just develop overnight.

That will take probably two years to three years timeframe, and by that time the process will be mature.

The .13 problems you talk about, we hope to see those problems to be resolved by sometimes next year.

If not, we have to consider because majority of products we are buildings are optimized for .15 microns anyways, today.

What's the other questions?

[If Intel] on their next round is going to put their base in Mac and integrate of their technology at .09 micron and leave only PE outside the microprocessor, how does it affect your opportunities to combo into that space?

Let me answer.

I don't know what Intel plans.

I would never put this into my [INAUDIBLE], because it's going to give me more headache.

The Mac for the wireless LAN is changing basically every quarter.

Every few quarters it changes.

The new standards are coming along.

The G, the A, too many things changing, the security is changing, and to integrate very, very powerful DSP together with sensitive analog front ends is very tricky in the market world.

And the other one is you do not want to qualify your motherboards through FCC one by one.

What you will see is wireless LAN will be adopted for PCs and laptops as a module, as an add-on module, so you'll only have to qualify it, the [INAUDIBLE] in different countries once.

When you're done, you populate it to different mother boards instead of the FCC qualification one by one.

If you look basically look at it from the basic point of view, not the technology point of view what makes sense and that is how we drive our road maps.

And not from what's sexy to talk about.

Thank you.

And the next question comes from Jim Liang at Pacific Growth Equites.

Thank you.

George, can you give us some rough numbers for the Western Digital opportunity for Q4 and potentially for next year?

We never break out, Jim, the revenue opportunities by customer.

I think the guidance that we gave, you know, for the year truly now and for the rest of this year clearly incorporates the beginning, as Sehat said, of the volume shipments beginning in the fourth quarter.

So there's clearly some amount of contribution, but it also is going to take a year for it to get all the way to, you know, as he said, 80-90% potential.

You know, there have been a number of numbers banted around that the WD opportunity if you had the majority of their businesses somewhere on a rechannel basis approaching $100 million.

And of course SOC would be, you know, significantly higher to that.

So I think that's the ultimate opportunity.

You know, we will -- we're looking forward to getting all that business, but it's going to take a few quarters to get all the way there.

Great.

Can you give us an update on further market share gain opportunities in the desktop beyond Western Digital?

Any progress update on that front?

Obviously.

You're aware of the rest of the suppliers that we haven't to talk about.

But at this point, I think -- it's sufficient just to talk about Western Digital first.

It's too early to talk about the others.

We will continue to work with them so they can partition to our products, but we don't know yet at this point when they'll be ready for that.

But one thing we know is that we have the technology, in fact superior technology, to make them to be able to build better products, but not everybody will build better products -- I mean, switch to us just because the need better products.

Sometimes schedules or relationships they have builds or others dependencies that they have made or agreements they have done where their current suppliers may prevent that to happen soon.

But we are hopeful and confident that in the long run we will be the leader in this business.

Thank you, Sehat.

One more question for George.

Can you talk about the giggy sequential growth expectations for Q3?

I don't think you guys mentioned that.

Maybe I missed it.

Well, we said that our communications business we expect it to grow, again in the high teens.

So I think that, you know, the majority of that growth we expect still to come from gigabytes.

Although, the switching part of our business is also showing very nice strength right now.

But again, the majority of our growth will come from gigabyte.

Okay, great.

Thank you very much.

And we'll move to Karl Mody with Wachovia Securities.

My questions have been answered.

Thanks, guys.

And we'll move on to Jeremy Bunting with Thomas Weisel Partners.

Mitch [INAUDIBLE[ for Jeremy.

George, I'm sorry.

I missed the 10% customers.

There's the traditional guy: Toshiba, Hitachi, Samsung and Seagate.

Right.

My next question is for either Sehat or Weili.

We've been hearing that Cisco may start asking vendor, its [five-side] vendors, to maybe integrate some of its switching IP and deliver custom-[ASIX] type chips.

Do you see this happening, and do you see that being a significant portion of your business in the future?

I think as far as in the gigabyte space, as you know, are the current productions, the product they introduced by Cisco roughly over a dozen platforms 100% Marvell [INAUDIBLE].

Obviously, we have accomplished a huge amount and enabled Cisco in the gigabyte space.

As far as the details, developments, strategies within Cisco, I think we just have to wait until such product is introduced.

Then it's the right time for us to address.

But we are heavily engaged with the [INAUDIBLE].

So whatever they need -- this is what we can say: Whatever they need, we will provide them.

We have the capability doing SOC, but I think that as far as Cisco strategy, the best is to wait until they introduce to the marketplace.

Right.

Okay.

Thank you very much.

And that's all the time we have for questions today.

Did anyone have concluding remarks?

All right.

If there's no further questions, I would like to thank all of you.

This completes our Q2 fiscal 2003 conference call.

I would like, again, to thank all of you for joining us and looking forward to updating you on the next quarter.

And that will conclude today's conference call.

We do thank everyone for their participation.