美光科技 (MU) 2009 Q3 法說會逐字稿

Good afternoon.

I will be your conference operator today.

At this time, I would like to welcome everyone to the Micron Technology third quarter 2009 financial release conference call.

(Operator Instructions).

Thank you.

It is now my pleasure to turn the floor over to your host, Kipp Bedard.

Sir, you may begin your conference.

Thank you very much and welcome to Micron Technology's third quarter 2009 financial release conference call.

On the call today is Steve Appleton, Chairman and CEO; Mark Durcan, President and Chief Operating Officer; Ron Foster, Chief Financial Officer and Vice President of Finance, and Mark Adams, Vice President of Worldwide Sales.

This conference call, including audio and slides, is also available on Micron's website at Micron.com.

If you have not had an opportunity to review the third quarter 2009 financial press release, it is available on our website at Micron.com.

Our call will be approximately 60 minutes in length.

There will be an audio replay of this call.

You can reach that by dialing 706-645-9291, confirmation code of 14636400.

This replay will run through Thursday, July 2, 2009, at 5:30 p.m.

mountain time.

A webcast replay will be available on the Company's website until June 25, 2010.

We encourage you to monitor our website at Micron.com throughout the quarter for the most current information on the company including information on the various financial conferences that we will be attending.

Please note the following Safe Harbor Statement.

During the course of this meeting, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company and the industry.

We wish to caution you that such statements are predictions and that actual events or results may differ materially.

We refer you to the documents the Company files on a consolidated basis from time to time with the Securities and Exchange Commission, specifically the Company's most recent Form 10-K and Form 10-Q.

These documents contain and identify important factors that could cause the actual results for the Company on a consolidated basis to differ materially from those contained in our projections or forward-looking statements.

These certain factors can be found in the Investor Relations section of Micron's website.

Although we believe that the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee future results, levels of activity, performance, or achievements.

We are under no duty to update any of the forward-looking statements after the date of the presentation to conform these statements to actual results.

I'll now turn the call over to Mr.

Ron Foster.

Ron?

Thanks, Kipp.

Let me provide a brief summary of the financial results of the third quarter which ended June 4, 2009.

The results for the third quarter reflect an operating loss of $246 million and a net loss of $290 million, or $0.36 per diluted share on net sales of $1.1 billion.

Costs of goods sold was just under $1 billion resulting in a gross margin of $107 million, or 10%.

No lower cost of market write-down was recorded in the third quarter and the gross margin reflects an estimated benefit of $240 million from sales of products in the third quarter that were subject to previous write-downs.

The cost of goods sold in the third quarter also includes approximately $30 million of charges from Inotera and IM Flash Singapore for the cost of idle capacity.

Total sales in the third quarter increased 11% compared to the second quarter as a result of higher sales of both memory and imaging products.

Memory sales also include royalties and technology fees of $32 million.

DRAM sales increased 14% compared to the second quarter to $553 million, primarily due to an 18% increase in bit shipments.

NAND sales were flat compared to the second quarter at $426 million as a 20% increase in bit sales was offset by a decline in the NAND average selling price compared to the second quarter.

Imaging revenue was $127 million in the third quarter, a 53% increase compared to the previous quarter.

This signals that we may have seen a bottoming out of demand in the mobile markets combined with some seasonal effects.

Looking at NAND in more detail now, the average selling price in the third quarter decreased approximately 17% compared to the second quarter.

However, trade ASPs on sales to customers other than Intel were up about 13% while the ASP for products sold to Intel from IM flash, which are on a cost basis, declined markedly with the successful ramp of our 34-nanometer technology products.

Notably, trade ASPs on component sales increased significantly more than the 13% for all non-Intel trade customers.

Mark will elaborate on this in a few minutes.

This new 34-nanometer technology node correspondingly drove down our overall NAND bit cost by 35%, well ahead of our prior projections.

These cost declines exclude the effects of lower cost or market write-downs and their flow-through benefits, as well as the costs of the idle facilities of IM flash Singapore.

The conversion to the 34 nanometer NAND process is now complete for all wafer starts which is providing us the lowest bit cost in the industry and has placed the entire NAND business in positive gross margin territory in the third quarter.

Q4 NAND cost reductions are expected to be in the high single digits and bit production is forecast to increase in the mid teens range.

On the DRAM side, DRAM average selling price in the third quarter was relatively flat compared to the second quarter.

Pricing for core products as well as specialty DRAM products each increased in the third quarter.

However, the increase in bit shipments of DRAM products came primarily in core products which continue to shift in mix away from specialty DRAM products that generally command a higher per bit selling price.

This shift in mix offset ASP increases in the quarter.

DRAM production costs, excluding the effects of the lower cost or market write-downs and idle facility charges, decreased 11% compared to the second quarter.

The lower production costs were primarily due to ongoing improvements in manufacturing efficiencies, including cost reductions and technology migrations.

You'll recall that Inotera has significantly reduced its production following Qimonda's default under its wafer purchase commitment.

This triggered charges for idle capacity to Micron under the wafer supply agreement.

To date, Inotera has not produced any products for Micron.

However, we expect to begin taking some volumes in Q4.

Micron will evaluate the volume of purchases from Inotera on a monthly basis.

We anticipate only ordering products when the purchase and subsequent resale are expected to be cash flow positive compared to the idle facility charge.

Fiscal Q4 DRAM cost reductions are expected to be in the low teens and bit production is forecast to increase in the mid to high teens range.

These projections exclude charges associated with idle capacity at Inotera but include any product that Micron receives from Inotera.

Turning now to operating expenses.

We continue efforts to scale down operating costs as reflected in the aggressive reduction in SG&A costs which were $80 million in the third quarter.

Staffing levels in SG&A areas have decreased 11% from their high early in the fiscal year and they are continuing to decrease.

R&D costs in the third quarter of $162 million were lower than previous guidance as some new products, particularly NAND products, were transferred out of development and into production in advance of the original plan.

We anticipate SG&A expense in the fourth quarter to be in the $75 million to $80 million range and R&D expenses to be in the $140 million to $150 million range.

Reductions are primarily driven by the carve-out of Aptina which is expected to occur a third of the way through our fourth quarter.

Other operating expense in the third quarter includes an estimated $53 million loss on disposal of Aptina and $28 million of losses from changes in currency exchange rates as the US dollar weakened substantially during the third quarter.

Relating to third quarter results from our joint ventures, the loss from equity method investments reflects our share of Inotera's net loss, the non-controlling interest in net loss line primarily reflects the allocation of TECH's loss to the minority shareholders.

Turning now to the balance sheet.

The third quarter ended with $1.3 billion of cash which includes approximately $475 million in net proceeds from our convertible debt and equity offering completed in the quarter.

The increase in the reported balance of inventories at the end of the third quarter is principally due to the effect of the inventory write-down at the end of the second quarter.

Inventory balances excluding the estimated effect of write-downs decreased 8% in the third quarter compared to Q2.

Cash flow from operating activities in third quarter continued to remain positive at $151 million, well ahead of competition.

The debt to capital ratio, both before and after the issuance during the quarter of the convertible notes and common shares, remained around the 30% range, consistent with our long-term capital structure target and also significantly better than most of our competitors.

Capital expenditures during the third quarter were $93 million, bringing the year to date total to $566 million.

We estimate capital expenditures in Q4 of roughly $100 million.

For fiscal 2010, capital expenditures are expected to be in a range similar to 2009.

This range can vary based upon market conditions as we go forward.

With respect to our joint ventures, in the third quarter we made a capital contribution of approximately $100 million to continue the 50-nanometer conversion at TECH Semiconductor which was financed through loan proceeds.

In addition, the third quarter had distributions to JV Partners of approximately $125 million consisting primarily of payments from IM Flash to Intel.

As previously announced, we signed an agreement to sell a majority interest in our imaging solutions business, Aptina Imaging Corporation to Riverwood Capital and TPG Capital.

As part of the agreement, Micron will retain a 35% minority stake after the partners contribute significant debt-free capital to the independent privately held company.

The transaction is expected to close around the first part of July.

Partially due to operating losses in recent periods and changes in the balance sheet occurring in the normal course of business, we now estimate the loss associated with the sale of the majority interest in the imaging business to be approximately $53 million.

Since we now treat Aptina as held for sale, we are required to recognize this estimated loss in fiscal Q3 in advance of the transaction's closing in Q4.

After the close of the sale, the financial position and results of operations of Aptina will no longer be consolidated in Micron's financial statements and we will have equity method accounting for the ongoing investment.

Micron will continue to manufacture products and provide services for Aptina through supply and service agreements at its worldwide facilities.

And, with that, I'll close here and turn the commentary over to Mark Adams.

Thanks, Ron.

We experienced a strong increase in bit shipments in both our DRAM and NAND businesses during our third quarter.

The bit increase was driven by stronger than anticipated demand across many of our key market segments.

Given this performance, we continue to grow share of the DRAM and NAND markets.

The pricing environment was more stable in Q3 when compared to our prior two quarters, and thus we are more optimistic about the overall memory sector heading into the back half of calendar 2009.

Our overall DRAM shipments grew 18% in bits shipped quarter over quarter.

As market forecast suggests bit demand and present computing to be up in the low to mid teens, we continue to grow share from competition in this segment.

(inaudible) shipped in our third quarter increased 31%.

Additionally, we have seen some strong upside forecasts in the server space which we anticipate will bode well for us in fiscal Q4.

In Q3, we saw a shift in demand towards DVR 3 memory for both server and PC applications.

Bit shipment of DDR 3 product increased approximately 70% quarter over quarter and represented about 25% of our overall core DRAM sales.

Micron's technology and cost leadership in DDR 3 puts us in a unique position to capitalize on the growth in demand coming from both PC and server-related customers.

Due primarily to some new customer wins, our networking shipments were up quarter over quarter as well.

As large corporations and governments continue to invest in infrastructure, we see the enterprise networking segment as a growth opportunity for our specialty business through 2009 and through 2010.

The mobile segment saw a decline in bit shipments quarter over quarter, down 12%.

While the overall handset market continues to be challenged, Micron saw strong growth in shipments of our MCPs, multi-chip packages, where we experienced a 35% increase from previous quarter bit shipments.

We continue to feel Micron has a unique position to grow our bit shipments into smart phone segment which is projected to grow 15% year-over-year.

Micron has a broad portfolio of products aimed specifically at the segment, including MCPs, PSRAM products, EMMC, NAND based technology and mobile memory cards.

As Ron mentioned, due to a mix shift in our business, our overall DRAM pricing was relatively flat quarter over quarter.

That notwithstanding, PC-related DDR 2 pricing saw an increase in our fiscal Q3.

In addition, we saw a strong increase in DDR 3 pricing toward the end of our quarter driven by a surge in DDR 3 demand.

It is important to note that as of today, Micron is only one of a few suppliers of DDR 3 memory.

We are leading in DDR 3 technology development with 1 gigabit and 2 gigabit products based on our 50-nanometer process.

In the third quarter, our NAND shipments were up 21% from the previous quarter due to stronger than expected demand from the mobile, USB, MP3, photo and SD markets.

This quarter, we shipped in excess of 90% of our (inaudible) MLC products based on our industry-leading 34-nanometer technology.

The 34-nanometer conversion is now completed and we believe this technology gives Micron a sizable cost advantage versus our competition.

Our retail business, which includes our industry-leading eCommerce site, crucial.com, made strong contributions to Micron's overall operating performance.

We had a number of significant wins in the retail channel with Lexar and Crucial branded products.

We continue to gain global market share growth leveraging Micron's NAND technology and a strong brand portfolio in the consumer markets we serve.

Shrink component pricing for NAND was up significantly in the quarter.

As the industry continues to migrate to higher density chips, we feel bit consumption will continue to grow and that the NAND market will be in fairly balanced supply and demand position throughout our fourth quarter.

We remain optimistic that the market segments such as SSDs, notebooks and NAND for mobile phone will continue to drive strong bit consumption.

Overall, we are very pleased with our NAND performance in Q3.

As we suggested in our last call, the pricing environment in both DRAM and NAND did in fact improve in Q3.

We remain hopeful that almost two years of capital expenditures cutbacks and reduction of supply in the industry will improve the profit potential for Micron.

We feel Q3 was a good step in that direction.

With that, I'll hand it back over to Kipp.

Thanks, Mark.

With that, we would like to now take questions from callers.

Just a reminder, if you are using a speaker phone, please pick up the handset when asking a question so we can hear you clearly.

(Operator Instructions).

Your first question comes from Gary Hsueh with Oppenheimer.

Hi.

Thanks a lot for taking my question.

Just a first question -- it seems like the biggest variable in my model for you guys in fiscal Q3 was NAND pricing due to the impact from Intel.

What's the expectation in the August quarter for Q4 in terms of NAND ASP per bit?

Well, right now in the market today, we expect to be kind of flat to modestly up in the quarter until we see what some of our customers (inaudible) look like going into the holiday season, which normally we get a good sense of in August.

I guess my question is, this is just guess work here, but if you're rolling off Intel heavy Q3 and maybe lightening a little bit up in terms of the mix in, in terms of Intel in Q4, what -- I mean does that effect have any impact on our assumptions for ASP per bit in Q4?

Well, I'm not sure of the assumption that we're rolling off Intel in Q4 is correct.

It's basically a flat split.

Okay.

If that--

Okay.

That doesn't change quarter over quarter.

Okay.

So the Intel content in your NAND flash business is basically roughly flat quarter over quarter as a percentage of the mix?

On a volume mix, that's right.

Okay, great.

Then if I, just kind of look at the royalty business, I think with Nan Ya projected to start kicking in some time next year in terms of their production.

Just wondering how should we be modeling basically our thinking about that royalty and licensing revenue stream in 2010?

Do we expect R&D kind of sharing part of that revenue stream to start declining or does that start picking up because of the units coming out of -- sorry, from Nan Ya in 2010?

What's the directionality and what's the magnitude of that industry in 2010?

This is Ron.

We've got some technology agreements that are already in place and that's showing in the revenue numbers that I mentioned that were $32 million this quarter, roughly the same last quarter and that continues forward.

Then also, as you point out, when volume picks up with, with product volumes out of Inotera there will be licensing royalties associated with that capacity as well at typical market rates you might say.

We haven't specifically called it out.

Okay.

Okay, great.

Just my final question, just in terms of your CapEx spend in 2010 being flat with 2009, what are some of the [big growth] assumptions that you have for production for DRAM and NAND?

So this is Mark.

For DRAM bit production, I think we would expect to see bits up in the mid to high teens in the fourth quarter and NAND bit growth in the mid teens.

How about 20 -- fiscal 2010?

Oh, sorry.

Longer term I think the NAND bit growth will depend a little bit on our 2X node migration which we intend to start late in the third quarter, early in the fourth quarter of -- calendar quarter of 2009.

But overall, I think you could look for something in the 10% range quarter over quarter in line with a normal shrink path for NAND technology on a go-forward basis.

For DRAM, a lot of it's going to depend on the timing of the Inotera ramp and as you know we're bringing that capacity online now.

If the market goes 40% next year, we'll be well in excess of that.

Great, thank you.

Your next question comes from Tim Luke with Barclays Capital.

Thank you very much.

Can you hear me?

Yes, we can hear you, Tim.

Excellent.

Just to clarify, if your (technical difficulty) order remains stable from here, what should we be assuming with respect to DRAM for the fourth fiscal quarter?

Quarter to date, we've got DRAM ASP pretty flat, Tim, and quarter to date we have the trade portion of our NAND business down about 5%.

And can you talk a little bit about how you perceive the licensing opportunity associated with the coming fiscal year, particularly as Nan Ya makes increased production?

And could you also just remind us of the framework for what we should be thinking about in terms of minority interest loss level associated with Aptina or what that might have been in the fiscal third quarter, and those are the two questions.

Tim, this is Ron.

In terms of licensing, sort of as I mentioned earlier, we've got ongoing technology agreements that will be roughly at the level we've seen and there will be, as I said, typical royalty levels paid on incremental volumes where our technology, our stack technology is being applied in Inotera.

Haven't called out those specifics, but there will be some expansion of that in the future as the stack technology gets adopted in Inotera which is, as Mark mentioned, is scheduled out in the coming quarters.

In terms of Aptina, loss levels of course depends on their financial performance on their P&L going forward and so we can't exactly call out what that would look like or our share of it, but we would be taking a 35% share of whatever that financial result would be on a quarter by quarter basis with a slight lag effect on the timing.

Let me jump in with what we're talking about on Inotera.

I think it's fair to say that we view Inotera as historically one of the most efficient and forward-looking probably the most efficient DRAM manufacturers that roll out a 50-nanometer technology.

When I talked about bit growth just a minute ago, there's bit growth obviously associated with the volume coming back online, but as they roll out our 50-nanometer technology moving into calendar 2010, we think that they are going to be a very, very focused and very efficient producer leading the industry from a cost perspective.

If I may, you talked about your CapEx being flat for the coming fiscal year.

Could you just remind us what that -- what level that is off and could you just outline as you move forward what you guys perceive to be the key variables now in terms of improving the gross margin dynamics from here.

Thank you.

This is Ron.

In terms of the CapEx that's the level similar to 2010, which was the $650 million to $700 million kind of range is what I would peg that at and that's based upon our current assumptions relative to our expansion plans, but I'll turn it over to Mark who can elaborate a little bit on our sort of thinking of our competitive position vis-a-vis our CapEx investments.

Thank you.

First of all, let me say on the CapEx that we plan going forward, it's a relatively small number, we believe, primarily because we've been pretty thoughtful about how we deployed our capital as we rolled out what is pretty efficient new 300 mm capacity.

I'll give you a couple of examples.

As we roll out 300 mm capacity, Dell came out with Damascene copper processing.

So we don't have that hurdle in front of us as we move to advance shrink geometry.

We also deployed NAND technology at 300 ml to a large extent with the double patterning tools we need to drive level of 34-nanometer now.

We're well down the path in terms of having all the tooling in place and that's why you see a relatively low maintenance capital spend for Micron on a go-forward basis.

Now, obviously if there's significant changes in the marketplace, we'll take a look at whether or not there's additional things we want to do from a CapEx perspective, but we think we're in great shape to continue advancing technology and driving down costs with a relatively small CapEx spend.

On gross margin, relative to the gross margin question you asked, we feel very good about our road map on a go-forward basis.

You really haven't seen much of the impact yet from our 50-nanometer DRAM transition.

In fact, that's all really in front of us on a go-forward basis on the DRAM side.

There's still some follow-through as we look to fiscal Q4 impact from the 34-nanometer NAND ramp is now complete.

As I mentioned earlier, we're well positioned to roll on into our 2X nanometer mode.

We think we're in good position to keep driving costs down.

We've had -- we've demonstrated some good success here over the last three or four quarters and we don't plan to change anything relative to our trajectory.

I was just going to add, Tim, that the NAND gross margins I mentioned is in positive territory, even excluding the effects of inventory write-down carry-through, we're still in positive territory in NAND margin on an operating basis for the first time.

So a significant accomplishment, and as Mark mentioned, I think there will be more to come as we continue out our 34-nanometer ramp and right on the heels of that will be the 2X nanometer technology.

Okay.

Your next question comes from Shawn Webster with JPMorgan.

Yes, good afternoon.

I -- maybe I missed it.

You were giving us lots of good information.

What was your bit production sequentially in NAND and DRAM?

We saw DRAM up mid teens, Shawn, and we saw NAND production bit growth up over 30% Q to Q for Q4.

Okay, and then you said your cost reduction was down 35% sequentially in NAND, is that correct?

Correct.

And does that track precisely with what you send to Intel in terms of pricing per bit?

So your Intel pricing, for example, was down 35 as well?

It's approximately the same formula.

It's on a cost basis, so, yes, that's essentially correct although there is some minor variation based upon timing.

Right, so I think that one of the things I'm struggling with is when we observe the spot pricing and even the contract pricing market, they would suggest that you would have had better DRAM pricing and better external NAND pricing.

Can you help, maybe help us understand why there's a difference there?

Yes, sure.

This is Mark.

The issue that's really driving that is that some of our specialty markets have lagged in terms of demand and that's had a little bit more price pressure in the current quarter.

And when the core commodity markets, both in DRAM and NAND rebound, those prices lag and the demand is starting to pick up.

We have seen some favorable uptick, as I said earlier, on the server market for example while it wasn't strong throughout the early part of the quarter, the last month of the quarter was very strong.

We're seeing good demand going into the back half.

So a lot of it has to do with mix as we look at that.

Our 32 gigabyte bit part on the NAND side was out in the market earlier, for customer trials and what have you and we just started to get that established in the marketplace and we're starting to see a better take on demand in pricing there.

Okay, and then as far as what your PC OEM customers and maybe even your consumer electronics customers are saying for calendar Q3, what kind of bit shipments are they guiding you for in terms of DRAM consumption on the PC side?

What we're seeing on the DRAM side is somewhere in the mid teens and the NAND side slightly lower than that today.

In terms of demand for calendar Q3?

Yes.

Okay.

How about channel inventories?

Can you give us an update there on both the DRAM and NAND side?

Sure.

I've read a lot about people suggesting that the first half of the year was a lot inventory replenished in the channel.

We're not seeing that.

We're actually seeing fairly robust demand, albeit at pricing and DRAM, which we would like to be higher, but the -- we're not seeing an inventory buildup in the channel.

I think people -- primarily I said this on the last call and I still believe it, I think people in this current economy are so focused on the cash flow operations of their business that they are not allowing themselves to take these inventory positions that are above and beyond what they can manage.

I see.

And then just maybe one more quick one.

What would your wafer production be sequentially?

It was up a couple percent.

Okay, thank you very much.

You bet.

Your next question comes from John Pitzer with Credit Suisse.

Thanks for taking my questions.

Just a couple quick ones here.

If DRAM pricing were to stay flat from these levels, what would be the expected utilization charge you guys might have in fiscal fourth quarter relative to Inotera?

I didn't quite hear all that question.

Could you repeat that for me, please?

Yes, if DRAM pricing were to remain flat from current levels, what would be the expected underutilization charge in the quarter for the Inotera JV?

The Inotera charge is an idle capacity charge.

Doesn't have anything to do with pricing, so if we are not utilizing the capacity then there's a charge that comes through related to that, but it's not related to pricing in the market per se.

Now, if you're asking what happens if we do or don't take wafer volume, as I mentioned, that will be predicated on whether we have incremental profit contribution that we get out of taking the wafer supply in the near term out of Inotera vis-a-vis the idle capacity charges.

Well, I guess that's my question.

We had a price level in DRAM, one would expect those idle charges to decrease, stay flat, or go up quarter on quarter?

So the way I see it is right now we view we're in a place where we can take supply of their trenched capacity and make incremental contribution out of it at current pricing.

That's why we're taking volume, but we will make decisions on a monthly basis whether to start wafers based on our view of the economics going forward.

And then getting back to the CapEx being flat, going into fiscal 2010, can you give me a sense of how much of that's going to DRAM versus NAND and is any of that going to Inotera to help transition from trench to stack, or are they beholding to support themselves on their own CapEx needs through that transition?

Yes, it's slightly more DRAM than NAND and I don't have the numbers in front of me, but probably roughly 65 DRAM, 35 NAND, something on that order.

And relative to Inotera, there's nothing in that CapEx number to support Inotera.

There are feelings that we'll be able to have a successful financing to take care of the capital needs independently.

And to be clear, we have a minority equity interest in Inotera, so it would not show as consolidated CapEx for us anyway.

It would be, it will be some kind of capital contribution that would be required for that were to happen, which is not at all contemplated at this time.

Inotera's going to take care of their own technology transition through their own financing.

And then, guys, you commented in the prepared comments that enterprise demand was looking better going into the calendar third quarter.

And I'm wondering if you could elaborate on that a little bit.

Is that just all based upon the demand you're seeing around DVR 3 for servers?

Are you seeing that pickup in corporate PCs as well, or what's giving you confidence that enterprise is upticking here?

The networking sector for us is more around networking server and enterprise networking equipment that we're seeing some uptick on.

We've had new customer wins.

Primarily a lot of this is over in the Asian market.

For us, my comments were really around our business and our growth in that segment versus the overall industry's trajection.

And then, guys, last question for me just relative to your views of the soap opera that is the DRAM industry around consolidation.

We've gone through another quarter now with rumor, speculations and some hard facts around Elpida and TMC and I'm just kind of curious, as things sit now how you view the consolidation in the industry relative to more players being shaken out or if you would give me a sense from peak production, you know, in Q3 of last year to where you think things finally shake out, how much capacity do you think will permanently exit the industry?

It's difficult to know at any given moment where it stands.

I think that my perspective is that it stalled here for a little bit, that effectively Qimonda is factually being liquidated and there's no return of that.

And that the DRAM pricing is such that on the margin, I think some of the Taiwanese are making a bet that it's going to get better but at the current levels it's not helpful for them.

Maybe equally important is to look at the debt structure that they are trying to deal with and it probably gives you a little more insight into what are the real alternatives that they have moving forward.

I think that the probabilities of them going out and doing anything in terms of new capacity is essentially zero.

The probabilities of them actually even really being able to upgrade much at the current capacity is pretty low, and they are all struggling with (inaudible) on how to deal with the debt.

And, by the way, that's also something that we have to encounter in the event that we have any kind of consolidation discussions with any of those players.

So when you look at the debt repayment structures, they are pretty heavy in the next 12 months.

That's where Micron now has an advantage because, as Mark Durcan already described, we've spent a lot of money, we're pretty good positioned, don't have to save much more money to advance our cost structure, but these guys do and they don't really have it.

So their cost structure's getting less and less competitive every quarter that goes by and lots of speculation as to what will happen.

There's really -- at the end of the day there's three major facilities that exist in Taiwan.

It's the Inotera facility, which obviously you've heard about Micron and (inaudible) it's an awesome facility and it's state of the art and it will be low cost but (inaudible) it is a benchmark facility.

Then you have [pro moss] which probably, again, just anecdotally from what I've heard in the industry, had their output number down in the 80% range, so you have a facility that could produce, call it 100K a month and it's -- it was producing something closer to 20 and maybe it's inched its way back up to 30 or thereabouts, so it's still way, way underutilized, but that capacity is actually a pretty rapid rate now because it's getting less and less cost effective.

And then the other big producer, of course, is power chip.

And you probably continue to hear about the saga between power chip and Elpida and [rex chip] and so forth.

But power chip had also I think cut their production down by the 80% range, but they are probably closer back up to 40%, 50% capacity now, trying to figure out where this market's going.

All -- and rex chip, by the way, never did cut its capacity that much and neither did Elpida and neither did Micron, so the output numbers are relatively stable.

In aggregate, kind of wrap it up here and quit rambling, in aggregate, the total capacity that came off line we think was in the neighborhood of 30% to 35% of the loan capacity for DRAM came off line.

You can do the math yourself, but essentially there's probably maybe 5% points of that is thinking about making its way back on the line.

You can even argue as high as 10 percentage points with Inotera, power chip and pro moss have all done in combination.

So it's not a lot.

And even from our perspective on what's going on with the Inotera capacity, the pricing has to get a lot better than it is now to provide much incentive for a whole lot of this capacity to come back on line until it gets converted or until the pricing gets a lot higher we've obviously got our own plans on Inotera but there are just no dollars available for anybody else to do much and we think that there's not a lot of capacity to come back on line here in the short-term and actually we don't think the current pricing there's only a couple of us where the structure looks pretty good.

So I don't think it's very advantageous for them at this level.

Perfect.

Steve, thank you.

Very helpful.

Yes.

Your next question comes from Jim Covello with Goldman Sachs.

Great, guys.

Good afternoon.

Thanks for taking the questions.

Just to clarify, one follow-up from the answer there, you said 5% or 10% of that capacity could come back online.

Is that 5% or 10% of total industry capacity or 5% or 10% of what was taken offline in the first place?

No, it's 5% to 10% of the total industry capacity.

Okay, great.

That's very helpful.

Thank you.

And then only other question for me, because you've covered a lot, is what are you seeing in terms of trends of content per box?

On the PC side, it's relatively flat through Q3.

The net book impact on the overall PC segment has a modest negative hit on it, but it's modest and so it's relatively flat going -- up until today when we see better things down the road as windows 7 takes and gets out in the marketplace.

And could you just comment on that for a second relative to windows 7, I know there's some concern from people that there's going to be despecking around windows 7 because you don't need as much content to run it, but do you obviously have different thoughts on that?

Yes, we do think there's a -- there's an upside from an application perspective when these applications come online and get over to the windows 7 platform and so when you look at some of the third party analysts who are looking at it as well, they are thinking the same thing, that content per box will continue to go up probably through the holidays, assuming all things go well in October, things go well through the holidays and through 2010.

We're pretty positive on that.

Thank you very much.

Good luck.

Your next question comes from Uche Orji with UBS.

Thank you very much.

Let me just go back to an earlier answer you gave about channel inventory and you said it is a good build-up of inventory.

What is your estimate of how many weeks of inventory held in the channel and held by most of your customers?

If we understood the question, it was about memory in the channel, is that correct?

That is correct.

How, how can we quantify, if you will -- tell me how much of the inventory in terms of weeks and if you can just compare that with what it has been historically, just for me to be able to quantify the answer that there wasn't much channel inventory.

We don't normally release that information because it's a lot of customer, proprietary information that we have that we're using on our own internal data.

But as I said earlier, we don't see any abnormal from kind of current state and what we expect in the channel.

We haven't seen inventory build up, we're kind of reluctant to give data that's customer related to answer that type of question.

Very well.

In terms of NAND, what are you seeing in terms of -- would you just give us any update as to what your development plans are on that (inaudible) for NAND?

Yes, this is Mark.

So, obviously as with all participants in the NAND space, we have active programs in 3 bit per cell and other end bit per cell and have had for a long time.

We will have 34-nanometer call it end bit per cell for competitive reasons right now in the marketplace late this calendar year.

Later this calendar year?

Yes.

Would that be volume or would that just be [samplings].

Well, it's not going to be a huge piece of the volume because we don't see that it really addresses a large piece of the market.

So, I don't think you should think of it as having, in this timeframe at least, a significant impact on Micron.

Okay.

To follow up on that, when you look at where we're selling our NAND applications, there's not a lot that will take the 3 bit per cell in the near term, so our application base to our customers both internal with Lexar and Crucial as well as our OEM customers, we don't think there's a utilization they pay a value premium to the current technology we have in the markets.

All right.

And just in terms of Inotera, I appreciate you've answered most of the question already, but when -- what is the migration pattern then for Inotera?

When should we expect 50-nanometer stack reduction at Inotera as it starts to ramp based on the plans you have and just financially, do you think -- what kind of funding requirements do you think Inotera will need just to be sure that they can meet their own part of the other funding as it migrates with stock?

This is Mark.

Let me take the first bit.

Then maybe Ron can talk to the financing.

On the technology deployment, we have equipment on order now to support the introduction of 50-nanometer technology.

We'll see first silicon out late in the calendar year, but you should expect volume of a significant amount in late calendar Q1, early calendar Q2.

Okay.

In terms of the financial funding, as mentioned earlier, they are a separate public company focused on doing their own financing, so we worked with them to develop the view on what the capital requirements are as they integrate that with their total business view and have communicated numbers, but be aware that they are really numbers owned by Inotera and they have viewed it and we concur that they can scale the buildup of their capacity over time and scale the CapEx investments relative to the market need, their customer requirements, including Micron, etc.

So the way to think about it is that they have communicated that there can be a phase in the $700 million, $800 million kind of range and then up into a full conversion cost, depending upon the timing they want to do that, to be around $1.6 billion, $1.7 billion.

Those are numbers coming from Inotera.

Okay.

Again, just lastly, on DDR 3, what should we model as a percentage of your DDR 3 exiting calendar 2009?

And then, also, what kind of premium do you get on DDR 3 over DDR 2 and when do you think we will achieve a cost of -- between DDR 3 and 2 in terms of bit production?

Three questions in there.

We didn't quite hear the first part of what you were interested in modeling.

In terms of how much DDR 3 as a percentage of the bit production we should model for exiting calendar 2009.

Maybe I should take this, a shot at that.

I think as we exit calendar 2009, I think you would expect that at the core, core DRAM business DDR 3 would represent roughly 50% of the output.

Okay, and then in terms of the DDR 3 over 2 premium, what is it now?

Is it in a good price premium at the moment, DDR 3 over DDR 2 and what is that?

That premium, which I stated in the earlier comments, is up towards the end of our quarter.

It's up in the 15% to 20% range.

All right, great.

That's it for me.

Thank you very much.

Your next question comes from Hans Mosesmann with Raymond James.

Thanks.

Going back to the 30% of the global capacity that was taken offline earlier this year and some of it that may be coming back, how much of that capacity is DDR 3 capable or viable?

Of the capacity coming back online, almost zero probably.

I'm sorry, that was zero?

Yes, essentially zero.

And what--

I just had some comments on that.

One thing that -- I think one thing that has occurred that is we're thinking about is the, in particular most of the time on these producers didn't do a DDR 3 design on their current or now what I would consider to be year old process technology.

I think they all thought that they were going to be spending money and going to the mixed node and since that hasn't happened, they don't really have DDR 3 designs.

First of all, it wouldn't be very cost effective or might not even fit into the form factor for the application given the current technology.

So I think that's one of the reasons that's the case as I stated almost zero coming back on DDR 2.

And then as a follow-up, how much would it cost them to come up with a design, just for kicks?

Oh, I don't think the cost of a design is all that significant.

I think it's the ability to get to the next process node to produce it on is where you need the money.

Okay, thank you.

Thank you.

Your next question comes from Daniel Berenbaum with Auriga USA.

Oh, hi.

Thanks for taking my call.

This is actually JD for Dan.

I just have a quick question on your cash usage.

Other than capital expenditures, how do you plan to use your cash on hand going forward?

Well, you say other than CapEx obviously the big cash usage we have is CapEx over time, but if you have noted from our quarter over quarter, some of the commentary from Ron and Mark, operationally, we're in pretty good position on cash.

So, I think we're, first and foremost, going to protect the balance sheet and protect the solid financial shape the company is in, but then I think what it does allow us to do, as we've noted earlier when we were raising additional money is just be prepared for other opportunities to come along in the event that we want to take advantage of them.

And that doesn't necessarily mean, by the way, that we're going to go out and spend a bunch of cash on an acquisition, but in the event that we participate in M&A activity and we become larger, then obviously we would desire to have a larger cash position with which to operate a larger company.

So I think that's how we're looking at it today.

Okay.

That's great.

Thank you.

Your next question comes from Bob Gujavarity with Deutsche Bank.

Hi.

Thanks for taking my question.

Just curious if you've kind of adjusted your loading between DRAM and NAND for the back half, any big changes there?

Any preference for one or the other?

We haven't done anything of significance.

There's always a little fine tuning going on (inaudible).

Okay, fair enough.

And within the NAND market, are you seeing anything particular in terms of the different markets you may sell into, particularly maybe spot versus OEM versus kind of retail?

Is there any one area that's particularly strong or are they all pretty coincidental?

I think the OEM market has started to pick up a little bit as people start to position for the holiday not knowing what the market will be like.

You've seen some price fluctuation over the last 90 to 120 days in NAND and I'm not sure that the market quite understands what pricing will be as people start to build in supply for their Christmas build.

So I think the OEM market seems a little bit stronger.

The spot market right now is sitting watching that and waiting to see.

I think we commented earlier that the market's down about 5% quarter to date and that's primarily driven by the spot market.

The retail sector is kind of so-so right now.

It's not great and it's also going in to a quiet period end of June, early July before back to school starts.

So that's kind of the landscape right now in NAND heading into the back half of the year.

Okay.

Thanks, guys.

Your next question comes from Atif Malik with Morgan Stanley.

Hi.

Thanks for taking my question.

Nice job on the margins, especially for NAND.

So I'm kind of surprised that your CapEx for next year is going to be more skewed towards DRAM than what I would have expected Inotera to take most of that burden.

And so the question is, how do you sustain your leadership in NAND costs if (inaudible) and what is the next milestone for Micron beyond 34-nanometer?

Hi, Keith.

This is Mark.

So, on a pro wafer basis if you think about the technology transition that has to occur, we need essentially to convert all of TECH's 50-nanometer, so that's why there's a slightly larger burden there as well as MTV being split 50/50.

Now, when you look at NAND, why is that build so low?

Well, the build was so low by design when we rolled out -- in fact, I think it's okay to say now when we rolled out 50-nanometer NAND we contemplated our strategy through 34 and 2X as being a double pattern strategy and put those tools in place, so rather than having a lot of retooling at this stage, we really have relatively small amount of retooling to do as we transition to the 2X nanometer node late this year.

And, and that's the short story.

Relative to 3 bit per sell or whatever, there's really not much CapEx associated with that at all -- design effort as opposed to technology transition.

Okay.

Then a follow-up on the image sensor, after you sell the majority, Steve, how should we think about the normalized revenue rate from this business and normalized OpEx rate?

Well, first of all we're going to continue, as Ron noted, we're going to continue to make the wafer.

I think the way to think about it is the additional margin that occurred from the wafer making its way into the marketplace will now be picked up by Aptina and will account for that on an equity method basis, but in terms of the revenue sale of the wafers, that will continue on as it has.

Okay, great.

Thanks.

Your next question comes from Tristan Gerra with Robert W.

Baird.

Hi, guys.

This is Scott [Herland] calling in for Tristan.

Thanks for taking my questions.

Real quick housekeeping question.

What was the breakdown between core and specialty DRAM?

As a percent of revenue?

Yes.

The core DRAM was in the area of 30% more or less and especially DRAM was around 17%, 18%.

Okay, thanks.

And you were talking a little bit about how your NAND patterning, you don't need new equipment, a lot of new equipment for that.

Is that why the DRAM CapEx is high is that you do need new equipment for the 50 nanometer node and how long do you think that that equipment can last?

Can that last a couple of transitions or is that going to be needed to replace in one to two transitions?

So I didn't, I didn't mean to leave you with the impression there's no CapEx for NAND.

There is a split that goes to NAND.

Yes, for DRAM, there is lithography CapEx required.

There is also some lithography CapEx required on the NAND side.

And at this point, the way we have staged our CapEx, we're buying essentially end of the road map tools.

So as we buy lithography tools, they really are 1.35 NA emergent scanners and there's nothing beyond that purchase at this point.

Okay.

And you talked a little bit about kind of guidance relative to what you're going to be without Aptina.

Do you have what the OpEx number for Aptina was this quarter?

Well, if you look at the data that Ron had referenced earlier, in fact I'll turn it over to Ron and let him call them out.

Yes, if you look at our typical imaging segment which, as Steve already mentioned, will be parsed into what's remaining as we go forward, most of the OpEx will go away in the new construct.

We'll still have some amount of R&D and SG&A, but relatively small percentage of the total that's there, and rough numbers, we run around $30-plus million in R&D and $9 million or so in SG&A in the total imaging segment.

So we'll drop to a relatively small number on both of those scores and, as Steve mentioned, the product will be shipped on a foundry basis so, in general, we would expect to have gross margins that are a little bit lighter in the OpEx is significantly lighter in the new model.

Okay, and so we should expect an additional $10 million drop-off in the November quarter because you're modeling for one month of OpEx in the August quarter?

Is that how we should think about it?

No.

We are contemplating closing around early July and so we have two months of the effect in our fiscal fourth quarter and only have an additional month of the effect in the subsequent fiscal first quarter.

So we got two-thirds of the effect already in the guidance I gave you for Q4.

Oh, okay.

Sorry.

I misunderstood that.

And then are you currently producing both DDR 2 and DDR 3 above cash costs right now?

And given the strong shift to DDR 3 in the market, what are your expectations for DDR 2 pricing for the rest of this year?

The answer to the first part of the question cash flow wise is yes.

And as you probably know, historically we don't try and predict the ASPs on the outward quarters.

We'll leave that to you guys.

Okay.

Do -- what do you expect I guess for the industry then?

You're talking about 50% of your core DRAM's going to be -- DDR 3 exiting this year.

What do you think the industry will do on that front and what percent of current capacity out there do you think is capable of producing DDR 3?

Well, just to clarify one thing, we will be 50% of our core by our fiscal year end.

We think the market looks like, if you're looking at third party data, will probably be in the 25% to 30% range at the end of the calendar year.

We have a significant advantage technologically against all of our competitors in the DDR 3 market.

So we have led that.

As a rough guess, you've probably got a couple of competitors and their capacity that can actually make a solid 50-nanometer DDR 3 product out there.

To Steve's comments earlier, there's a substantial amount of current DRAM capacity that has not been brought forward and will cost substantial amount of money for them to get into that DDR 3 market.

Okay.

And then on the 50-nanometer and the 2X nanometer for the DRAM and NAND, if I'm looking at two equivalent wafers, assuming 100% yields or whatever, what's the bit growth that is just from going from a 34 to the 2X or the 68 to the 50 nanometer?

I think I'm going to keep that one in our pocket for now.

The 50-nanometer data is actually publicly available.

You can get a die size on a 50-nanometer.

I think I'll just save publishing that in case anyone else is listening in on the call.

And we're not talking about the die size on our (inaudible).

Okay.

Thank you guys very much.

And with that, we would like to thank everyone for participating on the call today.

If you would please bear with me, I need repeat the Safe Harbor protection language.

During the course of this call, we may have made forward-looking statements regarding the company and the industry.

These particular forward-looking statements and all other statements that may have been made on this call that are not historical facts are subject to a number of risks and uncertainties and actual results may differ materially.

For information on the important factors that may cause actual results to differ materially, please refer to our filings with the SEC, including the Company's most recent 10-Q and 10-K.

Thank you.

Thank you.

This concludes today's Micron Technology third quarter 2009 financial release conference call.

You may now disconnect.