邁威爾科技 (MRVL) 2007 Q4 法說會逐字稿

Welcome to the fourth quarter 2007 financial results conference call. [OPERATOR INSTRUCTIONS] As a reminder, this conference is being recorded for replay purposes. I would now like to turn the presentation over to Dr. Sutardja, Chairman and CEO of Marvell Technology.

Thank you. Welcome, everyone, to our fourth quarter and fiscal year 2007 conference call. Weili Dai, Chief Operating Officer; and George Hervey, Vice President of Finance and Chief Financial Officer are joining me on this call.

Today we reported Q4 revenue increased 27% from the prior year to $622 million. Sequentially our Q4 revenues increased 20% from the prior quarter given the contribution of our newly acquired application and communication process business. For fiscal 2007, our revenues were $2.24 billion, which is a 27% increase from fiscal 2006 revenues of $1.67 billion. We are pleased with our results for the quarter. In Q4 our business has stabilized as we expected, and we are now poised for a resumption of solid growth. As we start into the new fiscal year, I am very excited about our near-term opportunities as well as our position to grow much faster than the market over the long term. We continue to focus on leveraging our core mixed signal, DSP, RF, and processor technology to strengthen our leadership position in our served markets as well as to enter new large market opportunities. With this effort and our execution we are well on track on becoming a truly industry leading and diversified semiconductor company. I will elaborate more about our opportunities and our business progress, but first I will have George produce more insight into our Q4 and fiscal 2007 financial results.

Thank you, Sehat. Good afternoon, ladies and gentlemen. I would like to remind all participants that the following dialogue will contain predictions, estimates, and other forward-looking statements covering subjects such as enterprise, consumer, and emerging market trends, competition, customers, suppliers, products and demand, and expectations regarding revenue growth, gross margin expectations, operating expenses, other income, accounts receivable, and inventory. Such statements will be preceded by words like expects, anticipates, believes, should, will, may, or words with similar import. These statements include those relating to our guidance for the first quarter of fiscal 2008, the pace of our business for fiscal 2007, and fiscal 2008, and the impact of the continued adoption of our solutions on our revenue growth.

The following factors among others could cause actual results to differ materially from those described in the forward-looking statements. They include the inability to further identify, develop, and achieve success for new products, services, and technologies, the inability to successfully integrate the application and communication processors acquired from Intel, increased competition and its effect on pricing, spending, third-party 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-Q, recent quarterly reports on -- I'm sorry, 10-K, recent quarterly reports on Form 10-Q, recent current reports on Forms 8-K, and other Securities and Exchange Commission filings all of which discuss other important risk factors that may affect our business, results of operations, and financial condition. Please be reminded that we undertake no obligation to revise or update publicly any forward-looking statements for any reason.

Now, moving to the Q4 and fiscal 2007 financials, as we indicated in our press release, we will only be reporting revenue for Q4 fiscal '07. Today we reported that net revenue for the fourth quarter of fiscal 2007 was $622 million, an increase of 27% over the $489 million reported for the comparable quarter in fiscal 2006, and a sequential increase of 20% from the third quarter of fiscal 2007. The sequential revenue increase of 20% from Q3 fiscal '07 was consistent with our guidance for Q4 '07 of $620 million to $630 million. Q4 fiscal '07 is the first quarter that includes revenue from the acquisition of the X-Scale products from Intel. The acquisition was completed on November 8, and the revenue contribution to Q4 was consistent with our expectations of $90 million to $100 million.

Revenue from our storage product stabilized during the quarter as the component inventory held by some of our large storage customers returned to normal levels. Our Prestera networking products and printer ASIC products contributed another solid quarter as we continue to gain market share. In the PC market, we did experience a decline in demand for our PHY controller products. In the Embedded market, our wireless LAN 802.11B and G products experienced a slight increase in revenue as ramps in new gaming platforms were mostly offset by the end of life programs on older platforms. We experienced an improving demand for 802.11N products as the excess of products in the channel declined.

We had two 10% customers in Q4, Western Digital and Toshiba. While we cannot provide detailed numbers for our gross margin and operating expenses, the following are trends that we experienced during the quarter. As we indicated in our Q3 conference call total Marvell gross margin percentage would decline as a result of the acquisition of the XScale business but we did expect the gross margin percentage of our existing business to improve slightly from Q3. For Q4 the product mix of shipments was consistent with expectations and manufacturing performance was above expectations resulting in a slight upside to our gross margin percentage expectations.

Operating expenses for Q4 include the expenses from the XScale business and ongoing costs related to the internal review of the Company's historical stock option practices by the special committee. As we indicated on our Q3 call we are reviewing discretionary spending to ensure that those levels more closely reflect the growth of our business. We realized some savings from the review during Q4, with actual operating expenses slightly below our expectations.

Our balance sheet continues to remain very strong. As expected, our Q4 cash declined to $596 million from $876 million in Q3. During the quarter, we paid Intel $600 million for the purchase of the XScale business and raised $400 million through a term debt financing. Additionally, during the quarter we made advanced payments on certain product supply. As we commented on our Q3 call our inventory trends were as follows. Due to the lower than originally planned shipments for Q3 the quantity of product in our inventory increased. In late Q3 we took steps to bring our days of inventory down to support our ongoing business. We made progress in Q4 and would expect to reach an appropriate days of inventory in Q1 fiscal '08. Now I'd like to turn the call back to Sehat for comments on our business outlook.

Thanks, George. First, I would like to start off by thanking our employees for all of their hard work and dedication. This quarter we successfully integrated the application and communication processor business acquired from Intel into Marvell. Even with such a large integration effort underway, we continued to have strong execution in servicing our customers, completed a successful launch of a family of our next-generation high-performance application processors, and made great progress on our fab migration strategy.

Now, by combining our XScale and Ferocian processor engineering teams we have assembled the critical mass to internally develop even more advanced application processors ranging from very low-power processors for handhelds, to ultra high performance processors that address the increasing computing requirements for a variety of high-volume consumer applications beyond just cell phones. This quarter, we successfully launched three handheld processors with the fastest processing speeds in the industry at 806 megahertz. Additionally, we are already demonstrating 1.25 gigahertz processors which we are actively working on commercializing with customers. We are excited about our strong positioning in the market, given our tremendous performance advantages that our technology brings to our customers. Our solutions truly enable our customers to differentiate their funds in the critical areas that consumers recognize and value.

As high end applications such as video, Internet browsing, e-mail applications, and MP3 music players are loaded into cell phones, the role of the application processor continues to become more significant. More MPS allows the phone to handle more sophisticated operating system, advanced applications, and provide faster response times and richer multimedia. Also, our industry leading low-power technology results in longer battery life. As cell phones continue to evolve, the integration and performance of such advanced applications will be a key differentiator in what sells the phones. Consumers care how quickly their application run such as page loads and how long their battery lasts. Our application processors allow our customers to offer the highest performance along with the longest battery life. And, of course, we are not sitting still as we see tremendous opportunities to expand our leadership by leveraging the distinct technical strengths of our broad portfolio of processor technologies.

Along with our leading application processors, we also have very advanced voice, baseband solutions. Our communication processors are battle proven and accepted around the world. This month we announced full approval of our advanced 3G communication processors in the Vodafone networks. And look forward to continued market acceptance of both of our application and communication processor solutions in this huge market. Beyond the cell phone market we are also seeing tremendous opportunities as we leverage the XScale application processors into many other high-volume consumer applications.

With the integration of the XScale business we are able to now offer our customers platforms that combine our Wi-Fi, Bluetooth, and application processors. Examples include Seagate's recently announced a DAVE product utilizing our new Monahans processor, and wireless LAN which allow customers to attach hard drive storage and wireless connectivity to a variety of consumer devices in a very low-power and a portable credit card size form factor. Our embedded wireless LAN solutions, including our single chip wireless LAN with Bluetooth are receiving strong design wins as the consumer market continues to adopt wireless technology. Given our recent success we look for fiscal 2008 to be another strong year for our embedded wireless solution in gaming and cell phones. Additionally, we are seeing strong adoption in MP3 player as well as for now automobile applications. In the 802.11N market we are seeing the inventory issue that hampered growth last year recover and look for a strong year for 802.11N for both retail as well as for new enterprise-class applications.

Now moving to gigabit, we continue to enjoy solid growth from our gigabit infrastructure solutions as the Marvell Total Solution continues to allow our customers to differentiate their products as well as to provide tremendous time to market advantages. The migration to gigabit continues its strong adoption as a replacement for faster ethernet equipment which plays right into our strength. Additionally, consumers are pushing for new video services and intelligence in the network which is also driving our switching and advanced packet processors in the enterprise and metro applications.

Before I turn it back to George, I would now like to discuss two new product initiatives. First, at CES last month we introduced the first of our video products for high definition digital entertainment. We have been making a substantial investment in advanced video technology for a number of years given the growing significance of video for both fixed and mobile consumer electronics ranging from next-generation high-definition televisions and high-definition optical video players to video for portable handheld devices and cell phones. We are very excited about the technology that we have developed and I would like to stress that what we have developed is state of the art technology and is clearly superior to any other video solution in the market today. Our first product we just announced at CES is a 1080 progressive high definition, video format converter processor and is the first of a number of products we will introduce based on our advanced Quiet Video technology. In a short time after announcing and demonstrating our very advanced video technology it has received great praise by the media as well as increasing customer traction. Our video quality is so distinctive that we now have customers upgrading their existing designs with our video processor to further enhance and improve their existing video quality.

We are entering the market at a time when the demand for big screen high definition TVs are becoming increasingly popular. Big screens are nice, but the larger screen makes it easier to see the artifacts included in the existing video content. Our Quiet Video technology solves this situation by making the video look sharp and pleasant. Our solution makes a huge difference, it is very noticeable, and as a result we are receiving some designs in the high-definition video market. Lastly, today we announced another exciting new market opportunity for Marvell. By leveraging our storage expertise and robust portfolio of enterprise hard disk drive interface technology we are now going to aggressively expand our solution to also service the storage enterprise class -- host class controller and RAID on motherboard market. As we look at the competitive landscape we are very excited about this opportunity. We are entering this market with very high performance storage technology, coupled with the industry's smallest footprint and lowest power.

In addition to our very advanced silicon solution, we have also internally developed a robust RAID and management software capabilities. We believe our entry into this market is well timed to fully take advantage of the growth of the serial attached SCSI segment, or SAS for short, and it is a natural extension for -- of our technology. I would now like to turn the call back to George for additional comments regarding our financial guidance for the next quarter.

Thank you, Sehat. Beginning with fiscal '08, we will be providing guidance only for the current quarter. As we enter fiscal '08, we are encouraged by our design win momentum and reasonable business conditions in our major markets. As I mentioned in my Q4 commentary, revenue from our storage products stabilized during the quarter as the component inventory level held by some of the large storage customers returned to normal levels. This improvement in inventory levels is positive as we enter the seasonally weaker first half of the year and expect seasonality to have minimal impact on our storage products. Q1 will be the first full quarter of revenue from our XScale products and our design win momentum should again offset the impact of seasonality in the handset market. Market conditions remain stable for our communication products such as Prestara and printer ASIC products and overall demand for embedded wireless LAN products is expected to increase as new platforms are introduced to the market. Based on the above factors we believe that our Q1 revenues should increase to approximately $640 million to $650 million. We expect Q1 gross margin percentage to be up slightly from Q4 and growth in operating expenses to be slightly below the rate of revenue growth. Now I'd like to turn the call back to Sehat.

Thank you, George. That completes our commentary. Would you please poll for questions?

Yes, sir. [OPERATOR INSTRUCTIONS] And your first question comes from the line of Cody Acree of Stifel Nicolaus.

So it sounds as though the inventories have corrected. Are you finding any pockets of lingering inventories out there? Is there anything that still is weighing on this first quarter outlook that you think still has got to be cleaned up yet?

Not of anything significant, Cody. I think the two big areas that we had challenges in the back half of the year were obviously in the storage market and in the 802.11N market and both of those looked much better than they did 90 days ago. So we are encouraged in having essentially a good position here in Q1, and that storage market is reflective of the recovery there.

Great. Then I know there's been some recently fairly high-profile speculation about your position, your design win position with the XScale processor. Can you talk about how design win momentum has been recently and what you believe your market share trends will be throughout the remainder of this year?

Yes, the XScale processor design momentum is very, very positive, so we're very happy. We're also very happy with the fact that we had just already have started receiving samples from our outside foundry for some of the XScale products, so they will help us with our capacity requirements for the future, as well as obviously for the cost reduction that we need to serve this market. So our customers are seeing that progress as well and they also are very, very excited about that progress, and this is one of the reasons that we're getting more design wins, not mentioning that, of course, the processor is the best in the class in the industry.

And Sehat you mentioned, you brought it up here, that the foundry position -- any update on when you expect that to happen?

We already start receiving the first silicon -- not just that, but we have only been having this business for, what, several --.

90 days.

90 days, right. So it's -- I consider that as a very, very good progress. So as we continue over the rest of the year, we will continue to move additional devices into the foundry as well.

Then very lastly, any update on the options investigation that you can give us?

Well, not much more than we can -- is publicly out there. We do believe that we are moving towards the conclusion. The special committee is moving towards the conclusion of their efforts. So we'd expect in the not too distant future to have something more to report on that.

All right. Thank you, guys.

And your next question comes from the line of Michael Masdea of Credit Suisse.

Can you give us an idea on visibility? It sounds like it's improving, but maybe give an idea of how orders have been trending both on the term side and overall. Has there been any change in lead times now that you have got your inventory down a little bit?

Not really, Michael. I think that the -- in general, we've always been, certainly in our storage business, have a very strong backlog position entering a quarter, and I would say that Q1 is no different than our historical trends, and maybe a little bit better on the comp side, but again, I think when we talk about turns, we know the programs that we're involved with. It's just a matter of the fill-in coming in. So it's not that much of an issue, or at least historically has not really been that much of an issue. But clearly we know where the end markets are going, and we know where we are in relationship to that And we feel better about that.

Great. Then it sounds like you have pretty broad growth but is there any one area that stands out, and especially is there any difference between the core business and the rest of the business in your forward guidance?

Well, one of the area, we don't talk a lot about it but I will say when we acquired the printer ASIC business from Avago we were very excited about that and indicated we thought that over a three-year period that business ought to double, with an objective of getting back to having the majority market share. Its performance so far has been very good and we're definitely on track to this point. Haven't even had it a year yet but at least as far as where we are today since we acquired it in May, we're very pleased and that's been a stand-out for us.

I would like to also add (indiscernible) we are leveraging as our core processor, high performance processor technology into that business as well.

Got it. Last one, Sehat, the two businesses that you introduced this quarter can you give us an idea of what timing expectations are for you and what kind of markets we're potentially talking about for Marvell in terms of size?

Are you talking about the video, Michael?

Yes, the video and then the storage side also.

The video processor.

The video products -- this is just one of a series of products that we are introducing to the market. We're entering the market through the very, very high-end segment, so over time, what we will like to do is also to move this technology to the mid end market and eventually to the entry level as well. But we felt that it's important for us, coming to this market, somewhat late -- but nonetheless it's a huge market opportunity -- we want to put ourself at the highest end as the leader in technology. As we got those design wins on the high end we will try -- we will work on addressing the market. So it's a huge market opportunity, and we do expect to develop and introduce more products into this market.

On the initial products -- people that visited our booth at CES, we would -- they are fairly quick time to market on the initial scale up products, so hopefully not a super large amount but some meaningful revenue in the back half of the year.

The processor, to give an idea, the video product actually is a video post processor.

Right.

So it communicates with other people's, let's say HD decoder. And over time we will introduce our own HD decoder as well.

That's helpful. Thank you very much.

Your next question comes from the line of Shawn Webster of JPMorgan. Please proceed, sir.

Yes, thank you very much. Can you give us a sense of what the percentage mix was of your storage and communication segments in the quarter?

We don't break our business out that way, Shawn. We basically are more of a split among consumer end markets and enterprise end markets although I think it's probably fair to say that still our storage products are the largest part of the Company.

Can you tell what your storage business did sequentially?

We don't break out the individual products growth.

Okay. Can you give us -- tell what you say your total employee headcount is now?

Yes, we have targeted with bringing 1250 people over from the -- with the XScale business, that we'd be right around 5,000 people, and that's right around where we came in.

Around 5,000? Okay. Thank you very much.

You're welcome.

And your next question comes from the line of Allan Mishan of CIBC World Markets.

Quick clarification. George, did you say that 802.11N increased for the quarter?

Yes.

How much of your Wi-Fi business is now derived from N, just roughly?

It's still pretty small, Allan. The majority is still the B and G in the embedded space. It was up for the first in three quarters.

Did you say that both PHY and controller products were weak in the quarter, or was it only the controller products?

Well, it's the combo product, but it is the combo product for the PC space.

I see, it's not the switch PHY.

No, no just the -- the switch PHYs were very strong in the quarter, but I think most people understand in that the PC market is totally converted to 8-gigabit at this point and there is price erosion and so forth. So that market is not strong anymore. That is going away.

Sure, it makes sense. Do you still expect the Intel acquisition to be accretive by Q4?

Yes, we had a very good performance in Q4 this past quarter as I indicated in our slightly better than expected gross margin performance and my commentary about slightly better improving margins in Q1 is a reflection of the progress we're making there.

Last question. Any update on the optical storage business? Still expect material revenue, maybe in the second half?

Yes, okay, the progress -- okay, it's moving back to on track, specifically related to the firmware, the software related to the optical side there. But in terms of revenue, it is sometimes a little later on -- but the design wins are progressing well though. It is just it takes a little bit longer than expected. The complexity of the software is tremendous. But we have consumers -- we have a different set of customers who are working on their own software as well, as well as we have [tech bodies] -- assistant providers are working with us on the additional software supports. So we are moving forward.

Okay, great, thanks a lot, guys.

And your next question comes from the line of Louis Gerhardy of Morgan Stanley.

A question on the ultra portable consumer opportunity in HDD like the Seagate win that you announced, is this opportunity going to be large enough where you have already got on your roadmap a plan to share resources with some of the HDD Silicon along with the XScale and the wireless LAN functions that are in that?

Yes, so when we mentioned about the -- we talk about the immediate opportunity that we got for the Monahans and the Wi-Fi combo products to address different market and one of the markets was just the -- an example of the market was the HDD small form factor device that combines a low-power processor and Wi-Fi. So that is just an example. This is not to say that, okay, the cell phones are going to go -- it's going to got to use HDD. We had that kind of hopes several years ago on micro drive in a cell phone, but with a price of flash going to the level that we are at, at this point, I don't think we could expect any micro sizing of cell phones in any significant volumes, with exception for the very, very high end high capacity cell phones.

Okay. If I could just switch gears and ask you about the cost savings of the business you acquired from Intel, of the total cost savings you expect to get over the next year, how much of that would be from changing wafer fabs versus back end savings like probe assembly and test? I'm just looking for a comment like it's 50/50 or something like that.

You are probably going to probably be disappointed. I think that again, you identified, Louis, the things -- there's three things that are very important to taking the gross margins on this business which we initially said were in the mid 20s to something that starts with a 4. And they are going to be accomplished with three things. First of all, bringing the redesigned chips into the product mix. We have announced those chips now, and we anticipate that they are going to be in our mix beginning in Q2. But also we are moving as aggressively as we can to converting to a true foundry model with Intel. Our assumptions at this point even though we're very excited about the progress we're making with our own foundries, right now our assumption is that our production comes from Intel during '07. So then when we move to a foundry model that's going to be very beneficial to us. Then of course, that allows us to do the back end processing in that we believe we can do it at a more cost effective basis, so it's really -- the cost of moving to the foundry model is a significant contributor but also having the redesigned chips in the mix is a very significant contributor.

Thank you.

And your next question comes from the line of Seogju Lee of Goldman Sachs.

George, Sehat, over the past couple of years you've done very well on the design wins for Wi-Fi into cell phones, yet the revenues and the uptakes have been somewhat lagging, obviously the markets themselves aren't there yet and we see a lot of activity. In 2007 do you think that the cell phones for Wi-Fi will start to become a material contributor, and what's the outlook there?

Well, it's an interesting situation. There's a lot of parallels to the gigabit ethernet market where there's a lot of design activity, the technology is very applicable for the platform, but the SKUs themselves just don't have a lot of volume associated with them yet. I think that's still the -- we have got designs with Nokia and Motorola, and Samsung, and actually another major guy we haven't talked about yet. We have yet to have them really ramp them in any meaningful quantity so we're kind of waiting at this point.

One thing I think you probably -- we're also learning ourselves as we enter the cell phone business, we have customers that are even still shipping those XScale products. They were designed like three years ago, so that indicates that it takes a little bit longer time than -- in the cell phone business to ramp new products compared to, like I say the PC business. But I think the customers are a little bit more conservative -- symptoms of getting all the software, all the systems debugged before they are turning on the switch to higher volume business. But at the same time, because we were very, very early in the game, we feel comfortable that we have a better position in -- we have a better position in the customer base. So as the customer -- as more and more products are being introduced with wi-fi capability, those are most likely to base on our technology and as the usage model proven in the market you will see that more and more people will have to race to introduce or to ramp up their products into the market. So we felt that actually we were in a good position in this space.

Just a follow-up question. Does a product like the iPhone help spur the market or accelerate the rollout of these products with Wi-Fi to the market and has it impacted the design win activity as well? Thanks and good luck.

I think we all hope, everyone do well on the products because it's just going to help the industry.

Okay. Then just one last clarification item. On the XScale revenues for the first quarter, George, do you suspect it will grow on a full quarter basis or is it just the full quarter versus the partial quarter? Thank you and good luck.

We are not going to get specific, I think the comment I made was that the design win momentum is good, allowing us to come through a quarter which normally could have seasonal -- a seasonal downdraft to it. So it means that yes, we have a full quarter, which means there's more revenue than there was in Q4 but I think it's a seasonally challenged quarter.

Thanks. Good luck.

[OPERATOR INSTRUCTIONS] Your next question comes from the line of Shaw Wu of American Technology.

Thanks. Two questions. Regarding the SAS opportunity, you talked about doing different types of silicon, including disk drive controllers and host controllers. Do you plan to do the HBAs themselves? That's question number one. Then question number two is do you have a long-term gross margin model? Now I believe it's below 50%, well below, I am just wondering what is your longer term gross margin model? Thanks.

Yes, the host bus controller, it is a, as some people call it an HBA, so our solutions that we introduce into the market -- smallest in terms of the footprint and also the lowest power and it is a [bundle]. It takes about 1/3 the space of our competitors. So this solution is not just suitable for the HBA market, but it is extremely suitable for the RAID on motherboard because of the footprint that -- the space that it takes, without sacrificing any performance. In fact, it has higher performance than the competition. So this is the market that we're targeting.

I think he's asking is it a chip versus the whole end solution? Oh, yes, we have customers, they are building HBA using our chips.

You don't plan to build the HBAs yourselves, right?

Not at this point but we have customers that are building HBAs using our chips.

Thanks for the clarification there.

Then on the second one, Shaw, we indicated when we did the Intel transaction that we believe that the long-term model which probably won't get to us until sometime next year is a 50% gross margin target for the Company, 26% operating expenses, 24% operating income so we still have -- our first objective is to get the XScale business to be accretive for us, and then we can move above that, although I will say we're very pleased with the gross margin progress that we've already seen.

Just a follow-up. I think it was Allan Mishan's question earlier, do you still expect XScale to be accretive by fiscal Q4?

Yes.

And your next question comes from the line of David Wu of Global Crown Capital. Please proceed, sir.

Good afternoon. Thanks for taking my call. The -- I had some questions for George on one thing, which is the -- you mentioned that the gross margin was slightly better last quarter than you planned, and this quarter than you planned. Is that coming out of that Intel XScale business? And the second thing, really on operating expenses, you indicated that it's going to improve this quarter. Since the revenue growth was minimal, I guess I was wondering what are -- after you add the people from Intel that internally at the Company there really is no head count addition to speak of. That's how you get the operating expenses down that quickly. And lastly, I just have one thing for Sehat. Samsung is no longer one of your top three 10% customers. Can you talk about that business whether there is a particular issue with Samsung this quarter, or has Samsung been losing market share in the storage space?

There are a lot of questions there, David. We'll take them one at a time. I think on the first one, we were pleased with the XScale performance from a manufacturing standpoint. So, yes, they contributed, not the sole contributor, to the slightly better manufacturing performance, but they clearly were a contributor to that and they will continue to be a contributor here in Q1 as well, although, we do expect the other products that I indicated that would be increasing will also be positive margin contribution.

On the expenses, we are -- as we had indicated before, in -- that we will start to see a roll-off in the -- we believe the amount of money we're spending on a quarterly basis from the options, special committee process. We did actually see a little bit better than we anticipated in -- level in Q4, as I indicated our operating expenses were a little bit favorable to our expectations. So we're starting a little bit ahead of the game there, and we would expect to see some more. So, there's a balance there going forward to the level of investment that we need to make going forward with the other businesses.

And -- okay.

I think that's the last one.

Samsung.

Samsung is the -- I think just -- with Samsung slightly from the 10%.

The number is bigger.

The number is still a big number.

Right. We are still overall, we won't be totally have everybody back in proper form from the inventory until we get through Q4. So I wouldn't read anything, David, into that, them not being a 10% customer.

Okay. That was really my concern. And Sehat, while I have you on the phone, can I ask one more question? If you're going into the video processing side from the digital side, there's all the way into set top box and also LCD, SOC business, is that something that you want to move into in the future?

Let's take a look first at the HDTV. When we say HDTV, we start -- we entered as I said earlier, we entered through the introduction of the post processing. This is an area where our customers can see the difference, immediately can see the difference -- the difference in quality of our solutions and the existing solutions. After all, the customer, the end users -- the consumers are paying multi thousand dollars for the TV. So even for let's say a Blu-Ray player, at least, what, $400 or $500 for the Blu-Ray player so it is important that the quality coming out from these boxes are the best in class, so that's how we get the design wins. Now, eventually, we also want to introduce the companion chips, the decoder, the 200 and so on, but this is over the next year or so after we completed our development, then we will introduce those companion chips as well to lower the cost of the overall solution from us while at the same time maintaining the quality that we introduced at the high end.

Okay. But that's mostly on the digital side. You don't really want to get into the image processing side?

We're talking about image processing right now. Right now our entry point is through the image processing. Is the processing of the image that was produced by other people's chip, and then instead of sending directly to the TV, we massage the images to make it more pleasant, cleaner, quieter. We call it Quiet Video technology. It really -- there's a really big difference between the pictures that goes directly to the TV versus the digital signal that goes to our post processor before it goes to the TV set. You really have to see it to appreciate the benefits of our image processing technology.

I was just wondering with the difference between yours and AMD's and Trident Microsystems approach and Broadcom's approach to that HDTV business.

Okay. Our video post-processing technology, if you think about it, just the -- this is the -- if you're familiar with the term [Faruga] in the past, this is the next generation of that -- those kind of technology. If you look at the -- when other people are talking about the video they are talking about the front end. They are talking about the DMOT, the MPEG decoders. That's when they are talking about the video. What we're talking the video is the post processing of the image that those people produce, so that leads to an opportunity that down the road we can also provide a more complete solution.

Okay. Thank you.

Maybe I should add a little bit, the front end are the standards based, so like MPEG2, SG.264, VC1s, these are all standard based technology. Where we are talking about video processing is nonstandard based technology. This is the black magic of the art of producing better quality videos.

Well, that's the way the margins are. Yes. Thank you.

All right.

And your next question comes from the line of Chris Caso of Friedman Billings Ramsey.

George, I was just wondering if you could clarify an earlier comment with regard to the storage market. It sounds like that you do still expect customers to be burning some inventory this quarter. Do you think you're shipping below actual consumption in this quarter and then you kind of go back to a steady state next quarter? Is that the right way to interpret your comments?

Well, I think it depends on what quarter. I think that was -- if you made those quarters Q4 and Q1 then I would agree with the comment.

So then inventory basically burned off by the Q1, the April quarter?

No, no, the January quarter.

January quarter, sorry. Okay. With regard to the XScale business, as you guys continue on to where your goals are in that business, I think you previously had a number of $150 million a quarter by the end of the fiscal year, if you can confirm that's still your goal. And if you could talk about how much of that growth you would see in unit growth as compared to content growth and where do you stand right now in terms of shipments of the communication processor as compared to the applications processor?

Probably some questions there that we won't give all that detail but I would tell you that we absolutely feel comfortable with the revenue numbers that we've put out which is basically going from the $100 million plus where it currently is to $150 million by Q4. Clearly that's something we believe from everything we can tell is very achievable targets for us to be pursuing. We have -- the new products are very exciting, including the combo chip that we have done, so that would be moving into the mix more as we get into the second half of the year. So the mix does become richer, when we move in that direction.

Combo chip is the 3G and application processor combo.

Right.

Are you willing to give us an order of magnitude of how much the combo chip represents now? Is it any meaningful percentage of revenue now?

We don't want to get to that level right now, Chris.

Just one final question with regard to the margin improvement you talk about with the XScale business. I guess the largest component of that margin improvement will be when you start moving the bulk of that business to a foundry. How should we look at the ramp of the margin improvement in that business? Is that more sort of a calendar 2008 phenomenon? How should we think about the ramp up of the margin improvement in that business?

Again, none of this is new from what we had said previously, which is that as we -- from the acquisition date literally through the end of calendar '07, the three things I talked about, the new design of the chips and then Intel -- moving Intel to a foundry model, are the things that will take us from the mid 20s to a 40%. To improve our margins beyond 40% we need to move into our own foundry model type of business, and that will happen by the beginning of '08.

Okay.

Calendar '08.

Right. Okay. Thank you.

And your next question comes from the line of Ruben Roy of Pacific Crest Securities. Please proceed.

Thank you. Sehat, I had a quick follow-up on a previous question on the video processor. You mentioned some customer traction that you are seeing already. I was wondering if the attraction there is mostly today on the source side, perhaps on -- within Blu-Ray type products, et cetera, or if you're seeing some traction within actual TV OEMs as well.

Yes, actually, you're right. Originally the traction was started on the Blu-Ray, because that's the most obvious place where people can -- when people are spending $500 plus for their Blu-Ray box they do really want to see the best quality from their player. As I said earlier, the problem with the big screen TVs is that any architecture that is part of the nature of the compressions, in like say, standard DVDs, when you play back on HDTV, it will look just horrible, if you don't do any additional processing. So a lot of our -- a lot of the high end Blu-Ray players, realizing that they're first generation products, are having a lot of complaints from the customers when they play back older content. So a lot of customers quickly are redesigning their box. The next generation box will include advanced post-processing technology to take care of this problem. Now, that's the first case.

Now, we already are seeing some customers on the very high end TVs are also starting to realize that when they're selling $2000, $3000 dollar TVs, they also want to make sure that people that use low-end, let's say optical players, or even the older DVDs, okay, will still have very good qualities on their TVs. So the benefit actually is important on both sides, both on the source side as well as on the TV side. So I don't see any way to get around this because there will always be low-end players, so there will always be high end TVs they want to look good, even though people are using low-end players. Vice versa, always people buy high-end players but maybe their TVs are still like two years old or last year model that doesn't have this capability, so they also want to have better picture quality.

Okay. Thank you, Sehat. And I had had another question on the applications and communications processors. I was wondering if you can give us an update on the structure of your -- one of the advantages of your products is your architectural license agreement with ARM. I was wondering if there was any changes now that the products are completely under Marvell in the sense that, are there any restrictions on the timing of how long you have that license or any other changes that may have come about now that the products are completely under Marvell?

Yes. On the architecture license, I believe, I could be wrong here, but I believe we are the only company in the world to have the architectural license from Arm, that span from V5 to V7. So V5, V6, and V7 instruction sets. So we have no restrictions whatsoever that I know of in terms of develop -- no restrictions at all. So we can build anything we want, we can build single issue devices, we can build dual issue devices, we can build thousand issue devices if it makes sense. So we have -- we have the flexibility to develop anything that we want for this market and we are doing so, obviously, carefully to make sure that we don't overkill it so we don't build something that's not needed in the -- but I can tell you that we are actively working on new generation of processors with the goal of making the processor to run even faster and even lower power at the same time. Our goal is to be able to allow our customers to, as I say earlier, to incorporate very sophisticated operating systems in the phones, and to make the phones appear to be usable and long battery life at the same time.

Thank you, Sehat.

And your last question comes from the line of Ross Seymore of Deutsche Bank.

Couple questions. Just a couple questions. First on the XScale side of the business when you talked about the design wins are more of those design wins on the base band or the aps processor side of the business?

Both.

So pretty much 50/50?

Basically there are people that -- the way you look at it, there are people that don't have the time to switch the base band, but they want to have much better quality for the application. They are using the application processor first, so they will develop the [Inaudible] or the Linux platforms on top of that to support all of the application into the device, build the user interface and so on. For people that wants to have -- people that have the bandwidth, okay -- the combo chip with the integrated base band and application processors to get smaller footprint and smaller form factors instantly. So we have both types of business. Now, in the long run, people, if they use even the application processor will, one by one over the next few years, we'll hope that they will also move to the combo solution as well.

Okay. I think you mentioned on the combo chip that it was a 3G solution. Can you give us more color on what generation of 3G? Is that a UMTS or HSDPA? What level is that that you're offering?

The production is the UMTS, the HSDPA is in the qualification stage.

So you are sampling HSDPA. Is that full speed, half speed, single chip, base band, or single core base band?

I don't have the right details. It's obviously single chip, and it meets the requirements. That's the only thing I can say.

Then the last question on the margin side, George, on the gross margin you were pretty clear about it being a little bit better both in the quarter you just reported and looking forward. Is that solely because of the improvements that you guys are finding on the Intel side or is there stuff with the foundry pricing being a little bit more favorable, or mix, or something along those lines than the core Marvell business as well?

I think, Ross, it's a combination of, again, you have to remember, we had the operating experience with the Intel group so we were making estimates of what we thought would actually take place and now we have some actuals. We're pleased to say that the actuals were a little bit better than what we had planned, so that's the impact on the XScale stuff, but, no, we just continue -- I wouldn't say it's foundry pricing. It's more of just that our own products continue to do very well in manufacturing.

Thank you.

Traditional Marvell products. And thanks.

I would now like to turn the call back over to Dr. Sutardja for closing remarks. Please proceed.

Thank you. This completes our Q4 fiscal year 2007 conference call. I'd like to thank everyone for joining us and look forward to updating you in the next quarter.

Thank you.

Thank you.

Thank you for attending today's conference. This concludes the presentation. You may now disconnect. Good day.