Good day, ladies and gentlemen, and thank you for standing by. My name is Huey and I'll be your conference facilitator today. At this time, I'd like to welcome everyone to Micron Technology's second quarter 2013 financial release conference call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there'll be a question-and-answer period. If you would like to ask a question during this time, you may press star then one on your touch-tone phone. If you would like to withdraw your question, you may press the pound key. 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. Welcome to Micron Technology's second quarter 2013 financial release conference call. On the call today is Mark Durcan, CEO and Director, Mark Adams, President, and Ron Foster, Chief Financial Officer and Vice President of Finance. This conference call, including audio and slides, is also available on our website at micron.com. If you have not had an opportunity to review the second quarter 2013 financial press release, again, it is also available on our website at micron.com. Our call will be approximately 60 minutes in length. There will be an audio replay of the call, accessed by dialing 404-537-3406 with a confirmation code of 21046896. This replay will run through Thursday, March 28th, 2013, at 5:30 P.M. Mountain Time. A webcast replay will be available on the company's website until March 2014.
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. Mark Durcan. Mark?
Thanks, Kipp. I'd like to start today with an overview of the key developments during the quarter and follow up with a few strategic and industry updates. I'll turn it over to Ron for a financial summary, and we'll close with Mark Adams discussing key developments in our business units, operations, and market conditions. Our second fiscal quarter was highlighted by improving market fundamentals in memory, along with solid execution from an operations and financial perspective. We had significant growth in key segments, including SSDs and server DRAM. Personal systems, a segment which has been a drag on our financial performance for some time, has rebounded significantly due mostly to structural supply shifts in the market. As we suggested in the last earnings call, we took a significant step forward with the successful restructuring of our supply agreement with Inotera, effectively doubling the wafer output we receive from them.
We're pleased with the increased scale and profit opportunity this restructuring process provided. As you will have noted in the release, revenue in the quarter was up about 13%, with strong growth in DRAM and NAND. Gross margins were up significantly from 12%-18%, again with improvements in both DRAM and NAND. Ron will get into more of the detail for the quarter from a financial perspective, but we're building a very large and diversified memory business and remain focused on making capital and segmentation decisions to optimize margin and free cash flow over time. The Elpida acquisition process remains on track. Creditors voted in favor of the reorganization plan, and the Tokyo District Court approved the plan at the end of February. In addition, all of the country and regulatory approvals have been completed.
The closing of the transaction remains subject to the satisfaction and waiver of certain conditions, including finalization of the Tokyo District Court's approval order under Japanese bankruptcy rules and recognition of Elpida's reorganization plan by the U.S. Bankruptcy Court in Delaware. We believe Elpida's financial performance is consistent with current market industry trends. A successful close will increase our overall trade memory capacity by about 45% from today's level without impacting industry supply. This capacity initially will be all DRAM. However, our 300-millimeter capacity is fungible and can support other advanced memory manufacturing technology. We're taking steps to enable flexibility to transition a portion of our capacity to NAND, should market demand and returns on investment warrant. I mentioned last quarter that we would continue to look for ways to streamline all of our manufacturing operations, including the eight-inch frame capacity.
With that in mind, we made the decision to sell our 200-millimeter imager sensor fab in Avezzano, Italy, in Q2. We believe this is a positive step to keep us focused on the memory market and optimize our overall margin structure. We don't have any other plans to announce today, but we'll continue to look for opportunities to streamline our operations. In terms of the memory market, supply dynamics in particular appear to be healthy looking forward. We believe the DRAM industry wafer capacity will be down in both 2013 and 2014, and process technology upgrades are being stretched out, thereby reducing supply growth compared to historical trends. In the NAND market, we see a favorable supply and demand balance. We have the benefit of significant demand growth in several applications, such as SSDs and smartphones, coupled with generally stable supply conditions due to slowing process technology changes.
I'll stop here and turn it over to Ron and Mark before turning for Q&A.
Thanks, Mark. The end of our second quarter of fiscal 2013 was on February 28th. On our website, you'll find a schedule containing certain key results for the quarter, as well as guidance for the third quarter. That information is also presented on the following slides. For the second quarter, we reported a net loss of $286 million or $0.28 per diluted share on net sales of $2.1 billion. These results compare to the previous quarter's net loss of $275 million or $0.27 per diluted share on net sales of $1.8 billion. We made a reporting change in the second quarter to reclassify gains and losses on foreign currency from other operating income and expense to other non-operating income and expense. This non-operating classification provides better comparability with the results of our semiconductor peers. The historical periods have been reclassified to match the current period presentation.
Accordingly, operating income for the second quarter, although a loss of $23 million, improved when compared to the operating loss in the first quarter of $97 million. Included in the $23 million operating loss for the quarter is $62 million charge in other operating expense for the projected loss on the pending sale of our Avezzano, Italy, facility, which we announced on February 25th. Included in this transaction, which is expected to close in the third quarter, is the assignment of our wafer supply agreement with Aptina. This sale will be the second spin-off of 200-millimeter production capacity, following the sale of our Japan fabrication facility in 2011. Going forward, we may restructure or dispose of other assets as we continue to optimize our manufacturing operations and focus on our expanding core business opportunities, as Mark mentioned.
In the non-operating category, we recognize losses on currency hedges relating to the Elpida Rexchip purchase amounting to $120 million in the second quarter. This is a result of the yen significantly weakening in the quarter. The hedge program we entered into at the signing of the sponsor agreement provided protection from a strengthening yen while allowing for partial participation in any purchase price reduction if the yen weakened. With the weakening of the yen, approximately one-half of the savings from the lower purchase price for Elpida is reflected as a loss on the hedge and has been included in our income statement for the quarter. Any reduction in purchase price will not be recognized on our balance sheet until the transaction is closed using the exchange value at that time.
We do not provide any guidance on hedging gains or losses, given the extreme volatility of the yen at this time. Non-operating expense in the second quarter also includes a $31 million charge associated with the repurchase of $464 million face amount of 2014 convertible bonds. I will touch on this in more detail in a few minutes. Also, in the results for the second quarter are favorable non-recurring adjustments to income taxes of $19 million relating to two non-U.S. jurisdictions. Turning now to operating results and outlook. Trade NAND bit sales in the second quarter increased compared to the prior quarter, primarily as a result of production increases from higher output of 20-nanometer devices. Trade NAND average selling prices were relatively flat quarter-to-quarter. We continue to see growth in SSD sales, which increased just over 40% in the second quarter compared to Q1.
Mark will provide more detail on these trends in a few minutes. Guidance for Q3 trade NAND is as follows. Estimated trade NAND selling prices will be down mid-single digits quarter-to-date, including the projected effect of mix for the quarter. Bit costs are expected to be down mid-single digits, while bit production is expected to be up to mid-to-high single digits. Key trends for Q3 affecting this guidance are, we expect a higher density product mix in both the NAND Solutions Group and the Embedded Solutions Group, lowering both ASP and cost per gigabyte while increasing volumes, resulting in similar margins overall. In ESG, the density increase is coming mainly from automotive and consumer applications. In NSG, we expect to sell an increasing number of SSD drives with greater capacity per drive, which slightly lowers the average ASP and cost per gigabyte.
We continue the migration to 20-nanometer process technology, resulting in lower costs and increasing bit output. We expect to see bit crossover this quarter. DRAM revenue in the second quarter increased 24% compared to the previous quarter, primarily as a result of a 38% increase in bit sales, partially offset by a 10% decrease in average selling price. The higher DRAM production was achieved in the second quarter due to strong operational performance and higher volumes from Inotera. The DRAM average selling price decrease in the second quarter is primarily caused by a higher concentration of sales into the PC market as a result of increased production taken from Inotera. These sales have lower per-bit selling prices as well as lower per-bit costs compared to other DRAM products in our portfolio.
Wafers acquired from Inotera in the second quarter increased 26% as the new supply agreement took effect January 1st. DRAM bit cost decreased 18% quarter-to-quarter, driven primarily by the higher concentration of lower-cost products from Inotera. Overall gross margin across all our DRAM products increased quarter-to-quarter by about six points. Note that our cost of wafers from Inotera will vary under the new supply agreement as market prices for DRAM products increase or decrease. The new Inotera structure is expected to be a net positive to our income and cash flow in Q3, as it was in Q2. Guidance for Q3 DRAM is as follows: quarter-to-date selling prices for DRAM products, including the projected mix effects for the quarter, are up mid-single digits compared to the Q2 average.
Projected bit costs are down mid-to-high single digits, and projected production volume is up a couple of %. Key themes influencing this guidance are 4-gigabit DDR3 volume continues ramping with the effect of improving margins while muting the 30-nanometer bit growth and comprises over one-half of DRAM bit sales in the second quarter. We're seeing a somewhat higher percentage of our total bit production from Inotera with improving product mix and costs that move directionally with prices in the market minus model. Output in the third quarter is muted as we plan to reconfigure a portion of our current DRAM capacity to enable NAND flexibility, as Mark previously mentioned. NOR sales decreased 14% in the second quarter, in line with our seasonal expectation.
NOR revenue and costs in the third quarter are expected to be relatively flat compared to Q2. SG&A expense in the second quarter was below our guided range as a result of lower costs associated with pending legal matters. We expect SG&A expense in the third quarter to be between $135 million and $145 million. R&D expense was $214 million in the second quarter and is expected to be between $225 million and $235 million in the third quarter. With the termination of the DRAM joint development program with Nanya, virtually no DRAM development costs were shared with Nanya during the second quarter. The level of R&D expense in any given quarter can vary based on the timing of product qualifications and the volume of development wafers processed. The company generated $234 million in cash flow from operating activities in the second quarter, roughly flat compared to the prior quarter.
Cash flow from operating activities in the second quarter includes a negative shift in working capital accounts within the normal scope of fluctuations that we see quarter-over-quarter. Total inventories decreased slightly in the second quarter compared to Q1. The second quarter ended with cash and investments, including non-current investments of just under $2.8 billion. Expenditures for property, plant, and equipment in the second quarter were $353 million, and we still expect total expenditures for the fiscal year to be between $1.6 billion and $1.9 billion. Note, however, that quarterly expenditures can vary depending on the timing of equipment receipts. We successfully completed an offering of $600 million in convertible notes the first week of February. We also entered into cash calls to effectively minimize the dilution of these convertible notes.
The bulk of the proceeds were used to repurchase $464 million face amount of our 2014 convertible notes that are coming due, effectively refinancing them out to 2018 and 2020. We paid holders a price very close to the market price for the notes. However, the accounting for the early redemption resulted in a non-cash charge of $31 million I mentioned earlier. Depending on market conditions, we may repurchase additional 2014 notes prior to their maturity. Concurrent with a convertible note offering, we published some recent financial information on Elpida, including pro forma financial summaries. These pro formas present historical financial information for Micron and Elpida as if the acquisition of Elpida Rexchip had already occurred, including the effect of purchase accounting.
Elpida's US GAAP financial statements for the fiscal year ended March 31st, 2012, reflect an impairment charge of $2.8 billion to write down property, plant, and equipment to its estimated fair value as of that date. As part of the pro forma adjustments, PP&E is assumed to be further written down in purchase accounting by an additional approximate $2 billion to a net value just over $1 billion, resulting in the fair value of the net assets acquired equaling the purchase price. As a result, the Elpida Rexchip acquisition represents an approximate 45% increase in Micron's wafer capacity, with only a 15% increase in depreciation expense. Now I'll turn it over to Mark Adams for his comments. Mark?
Thanks, Ron. I'm going to provide some more detail on our second quarter operating performance, as well as share some observations about the overall market environment in the memory industry. Our NAND Solutions Group recorded a 15% quarter-on-quarter revenue increase. Coming out of the holiday season, we were watching for potential downward pressure in market pricing. However, NAND pricing was favorable in our Q2 as end markets for SSDs, smartphones, and tablets remained strong demand drivers for increased NAND bits. Micron's trade NAND was basically flat, even while somewhat limited by growth in the high-density solid-state drives that have a decreasing effect on the ASP per gigabyte calculation. Margins improved in the quarter, benefiting by increased shipments of our industry-leading 20-nanometer node, improving our cost per gigabyte quarter-on-quarter. We shipped 1.1 billion gigabytes in our trade NAND business in the quarter.
We continue to grow our presence in the storage category with products aimed at segments such as personal storage, cloud and data center storage, and IO accelerators for the data center. We shipped roughly 1.3 million solid-state drives in our second quarter, up over 40% when compared to our first quarter. In Q2, SSDs represented 20% of our trade NAND business. If you include NAND component sales to SSD providers, around 40% of our trade NAND bits go into solid-state drives. Our development of enterprise-class SSDs is progressing well. Revenue roughly doubled last quarter, albeit off of a low base, and continues to be an area of investment going forward. We began customer shipments of our P320h SLC-based PCIe SSD and continue to strengthen our SATA and SAS product roadmaps, positioning Micron as a broad provider of enterprise storage-class products.
On the technology and operations front, we are making steady technical advancements with both our planar and 3D NAND technologies. We began shipping our 20-nanometer TLC NAND flash and continue to increase our 20-nanometer MLC production as well. Our 20-nanometer ramp is on track, and we expect production crossover in the current quarter. Keep in mind, some of our more value-added segments, like SSDs, often require additional time before utilizing the most leading-edge process. This quarter, we will commence volume sales of our 20-nanometer based SSDs, including our recently announced M500, which ships in terabyte-class densities with very competitive price points. MLC represented about 80%-85% of our wafer production in Q2, with SLC and TLC essentially splitting the remainder.
Sales for our DRAM Solutions Group in the second quarter reflect the improving ASPs Ron referenced, combined with strong growth in bit sales, buoyed by increased volumes, resulting from our new arrangement with Inotera. We are very pleased with the additional scale, stable gross margin percentage, and operating margin leverage in a strong market which this new structure provides. Revenue for DSG was up 26%, and our gross margins for DRAM overall improved by close to 600 basis points quarter-over-quarter. We were pleased with the direction of several of our DSG segments. On the server side, we increased share significantly at key OEMs and cloud service providers. Q2 was a record quarter in this segment, with shipments of over 200 million gigabit equivalents. DRAM bit shipments and networking were up 20% compared to our first quarter and now represents roughly 17% of our overall DRAM revenue.
We had another strong quarter and made progress to expand beyond the traditional large OEM business to smaller customers and distributors in the sale of our networking portfolio. DDR3 volume eclipsed DDR2 volume for the first time, and we are now sampling DDR4. RLDRAM continues to gain momentum, and we had a strong design activity with enablers of and OEMs for HMC, Hybrid Memory Cube. From a market standpoint, U.S. carrier CapEx for LTE deployment is strong, and data center networking remains a bright spot for investment by key OEMs. On the operations front, we hit bit crossover for 30 nanometer in Q2. We anticipate that we will be able to ship 25-nanometer technology in volume during our second half 2013 calendar year.
In addition, our team continues to make progress on the development of our 20-nanometer product as we look to start sampling that product technology by the end of the calendar year. DRAM pricing has strengthened from the bottom and is up mid-single digits from the Q2 average quarter-to-date. Obviously, as with any recovering pricing environment, the channel or spot market starts out in front of our OEM contract pricing. While there remains a gap between spot and contract pricing, it is narrowing, and we see that as a continuing dynamic in Q3. Channel inventory is very tight, and we are currently on allocation to several channel and OEM customers. The DRAM supply and demand balance appears to be favorable, and we are well-positioned to capitalize that going forward.
Sales by the Wireless Solutions Group were down 19% quarter-over-quarter, with a slowdown in the sale of feature phone-driven products, including wireless NOR and SLC NAND. As mentioned in our last call, we are growing our presence in the mobile DRAM space in particular, with low-power DDR2 design at two of the top five smartphone makers. We are also now sampling low-power DDR3 with major system-on-a-chip suppliers, which will enable end customers' design and activity. In anticipation of the Elpida close, we continue to work on rationalizing our product lines from a margin perspective as we seek to blend the best of the Elpida portfolio with our Micron products and technology. Our focus will continue to shift from feature phones to smartphones and tablets going forward, with a product portfolio focused on low-power DDR2 and DDR3, as well as a growing portfolio around eMCP and eMMC.
From a market perspective, improving pricing trends in PC DRAM and NAND is starting to positively impact the mobile market. We expect smartphone units to be up close to 40% in 2013, which is better than our initial forecast coming into the year, driven by improved growth for entry-level smartphones. In addition, higher-end smartphones continue to be significantly growing memory content, in particular for DRAM, which is expected to increase roughly 30%-40% year-over-year. Our Embedded Solutions Group had a record revenue quarter in Q2 of roughly $282 million, with strong continued operating margins of 23% in what has traditionally been a slower seasonal quarter. This was our fourth consecutive quarter of record revenue. We benefited from strong regional demand in Japan and the Americas. DRAM was a particular highlight driven by automotive and industrial, medical, military segments.
We are ramping 45 nanometer NOR, which now represents over 60% of our high-density shipments, and we are growing our presence in eMMC for the embedded segment, with that revenue up 50% quarter-over-quarter. Excluding the extraordinary charges that Ron alluded to, we were encouraged with the improvement in our operating performance in the second quarter. In addition, we continue to make progress on our overall cost structure. SG&A and R&D were in line with our guidance. We continue to drive our inventory lower. In fact, inventory was down $110 million quarter-over-quarter, which represents a decrease of over $350 million year-over-year over the past 12 months. The sale of our Avezzano imaging fab is a good example of the steps we are taking to improve our long-term operating efficiency.
Our management team continues to work on integration plans as we prepare for the close of our Elpida transaction. As we've interacted with the Elpida team members in planning the new Micron, we have grown more impressed with the people and resulting product and technology capabilities. Our customers have been extremely supportive of the acquisition and Micron's overall strategy. They clearly understand how memory is evolving to new applications and form factors as part of a solutions orientation away from a standard commodity module. We are optimistic that end markets such as mobile, server, networking, enterprise, and embedded will continue to drive strong demand for our products, and thus feel that both DRAM and NAND will stay in good supply and demand balance in Q3. With that, I will hand it back over to Kipp.
Thanks, Mark. We would now like to take questions from callers. Just a reminder, if you are using a speakerphone, please pick up the handset when asking a question so we can hear you clearly. With that, we'd like to open up the phone lines.
Thank you, sir. Ladies and gentlemen on the phone lines, as a reminder, to queue up for a question, please press star then one on your touch-tone phone. If your question has been answered or wish to remove yourself from the queue, you may press the pound key. Our first question comes from the line of Monika Garg with Pacific Crest. Please go ahead. Your line is open.
Hi. Thanks for taking my question. First question is, given that Inotera has spent pretty low CapEx in the past two years, and they're also guiding to low CapEx in 2013, could you maybe comment on where do you think Inotera is in its node transition roadmap, and especially comparison to Micron's node transitions?
This is Mark. The Inotera fab is really just in the last stages of completing the 30-nanometer ramp. That CapEx is really pretty much completely funded at this point. The forward-looking CapEx for them will occur late in 2013 and into the first half of 2014 to drive a 20-nanometer transition. I wouldn't expect that to be overly steep, but could occur throughout the totality of 2014 with most of the CapEx coming in the front half of that year.
All right. Thanks. The last question here. Elpida, could you kind of give us any idea on what Elpida spent on CapEx last year? Do you think that CapEx needs to ramp up since Elpida was not able to spend enough money due to its bankruptcy?
We're not going to comment on what Elpida has been doing or on their internal operations. What we have told you before, relative to forward-looking CapEx for Elpida, is that on average, it should look similar to our own technology migrations and capital intensity, which we don't believe is overly burdensome.
Sure. Thank you so much.
Thank you, ma'am. Our next question comes from the line of James Schneider with Goldman Sachs. Please go ahead. Your line is open.
Good afternoon. Thanks for taking my question. First of all, you mentioned or alluded to the fact that you will be shifting some DRAM capacity to NAND over the coming quarters. Can you maybe give us any more color on what the magnitude of that shift might be and in what facilities?
We're not going to comment on particular facilities, really what I intended to convey to you was that we were making steps to provide ourselves the flexibility to make those moves, not necessarily a commitment to any particular timeline or magnitude of those changes. I think what you'll see is activity going on at
Micron fabs in order to lay the groundwork so that we have the flexibility to do that. In the short term, what that means is more engineering activity and a small amount of incremental tool installation and minor disruption to ongoing operations, no significant capacity decisions are made at this point.
Fair enough. Thanks. Maybe as a follow-up, can you maybe update for us your assumptions about what the NAND industry will do in terms of bit growth for the year overall, and where Micron expects to come in within that, given what you just talked about in the last question? Thanks.
Jim, this is Mark Adams. I think the numbers we base our planning around is in the 30%-40% range of NAND bit growth. I would expect Micron to be to the upper end of that range, basically right around there.
That's an incremental tick down for the industry relative to what you talked about before. Is that correct?
Just slightly, Jim. Yep.
Great. Thanks so much.
Thank you.
Thank you, sir. Our next question comes from Glen Yeung with Citi. Please go ahead. Your line is open.
Thanks. Can I just clarify a point? When I hear your statements about ASPs for the quarter in DRAM, is it right to say that what's really happening is as you take on more Inotera wafers, it actually impacts the ASP increase because of mix, but it actually is still contributing solidly to profitability?
Yeah, I think that's right. There's really two things driving the cumulative impact on ASPs for us. One is the relative lag contract to spot, and the other is the change in our mix that has occurred with the addition of the Inotera increment. Over time, obviously, we'll start moving some more of that capacity into value-added segments, but the initial slug of capacity came in the PC area.
Okay, fair enough. That's an important point. Second thing, Mark, actually, just to clarify, you made a statement about channel inventory being very lean. I just wanted to clarify, when you look into distributors, you don't feel like those guys are trying to build inventory as they sometimes do when pricing is rising. Just sort of related to that, are you also on allocation for NAND, you suggested you were selling DRAM.
The answer to the first question is we don't think that the distributors are in a good inventory position at this point relative to 6-12 months ago. We think it's pretty lean there as well. Quite honestly, there's just not that much capacity available for those channels for them to take that position. As it relates to the NAND piece of the equation, there are certainly pockets of NAND that are in tight constraint as well. A lot of that's driven by the production mismatch sometimes with what the requirements are on higher density SSDs. We're seeing really tough pockets on limits of available supply for some of the SSD categories.
Okay, just one last quick question. Thinking about the pull that you're seeing for DRAM from now, a variety of end markets, not just normal PC, I wonder if you can rank the end markets now in terms of what areas are pulling DRAM at the greatest or fastest growth rate, be it server, networking, mobile, or PC.
What we're seeing in terms of overall raw capacity, it would be mobile certainly would be at the top of that list. PCs are probably second when you factor in tablets, and it depends where you put tablets in the overall ranking. Servers is actually pretty significant. You've got single-digit growth, but pretty significant server DRAM content growth. That's becoming more and more of a growth category in terms of bit consumption. Of course, you've got the automotive and networking as you go down that list.
Is it fair to say that PC demand ex tablets is not so good right now?
In balance, remember now, it's kind of funny, as you just highlighted, when you're a manufacturer of DRAM, you've got to make those calls on production bits on where those products are going. You can't just take a mobile bit and make it available to a PC guy or vice versa in a given production period. The unit growth in PCs might not be super high, but the balance of PC bits and PC demand is pretty healthy right now.
Got it. Okay, thanks so much.
Thank you, sir. Our next question comes from John Pitzer with Credit Suisse. Please go ahead. Your line is open.
Good afternoon, guys. Thanks for letting me ask the question. Congratulations. Mark, I'm just kind of curious on the NAND pricing environment. Typically, this time of year is a seasonally weak period. In addition, you've kind of got the largest NAND buyer in the world going through a well-publicized, kind of weaker than expected demand cycle, inventory work down ahead of perhaps a new product launch in a few months. I guess I'm curious, your expectation for NAND pricing for the balance of the year, and at what point, independent of getting the Elpida asset, do you think about raising the core CapEx of Micron from that $1.6 billion, $1.9 billion to maybe something higher?
This is Mark Adams. I'll take the first part of the question. The NAND business is kind of funny when you think about where it was and where it is today. It used to be a very seasonal business, right? You had memory cards and USB and MP3 players, all of it driven by either back to school or the holiday sales. Today, where you see NAND bits going, that's more diversified and less seasonal. You referenced some changing customer dynamics around NAND, but in general, the demand drivers are still really solid coming out of the holiday period, as I mentioned in my opening comments. We think it looks pretty good balance when you think about 30%-40%. You still think about strong demand across a growing SSD environment, smartphones not only growing in units, but taking on some more content.
We think overall, when you couple that with tablets, it's a real healthy outlook for us in the remainder of 2013 calendar. Let me take the CapEx piece. We're going to be very disciplined on CapEx here moving forward, there's a lot of good reasons for that. First of all, of course, I already mentioned we've got options relative to the capacity we have in terms of how we move that around between the various technologies and market segments to optimize margin. That can all be done relatively cost-effectively. Think in terms of, that's our most desirable way to adjust our capacity to match any imbalances we see.
Beyond that, there's a couple of reasons why I think you will see Micron in particular, and maybe others, be somewhat reticent as they think about putting new capacity online, and that's this whole technology environment that we see today relative to. We have 20-nanometer in volume production. I think most of the folks are working on technologies down as low as maybe 14-nanometer for planar. Beyond that, most of the competitors are working on 3D, and there's just a big mismatch in the type of capacity you put in place for 3D versus what you put in place for those end-of-life planar technologies. I think people are going to be pretty circumspect about how much capital they go spend on NAND capacity over the next couple of years while they're prepping for a 3D transition sometime out in 2015, 2016.
Mark, that's helpful. Would you guys ever think about putting NAND inside of Inotera, or will that stay predominantly a DRAM facility?
We're not going to talk about which fabs we would or wouldn't consider. Actually, all our 300-millimeter fabs, I think, are candidates, and we're not going to be forecasting which ones we may or may not do.
Guys, one last quick one for Ron. Ron, when you look at the OpEx growth that you guys are kind of guiding to this quarter, how much of that is just improved revenue outlook versus maybe some one-offs, either on the legal front or some R&D stuff that you have to get done that's driving that? I'm just trying to get a sense of what's variable growth versus maybe some stuff that you just have to get done.
Yeah, John. In terms of the OpEx this last quarter, we had a couple of favorable effects I called out in the SG&A side regarding legal expenses. They just happened to be lighter, related to some of those legal activities that are hard to predict, and they have lumpiness and accruals. On the R&D side, it's varied, as I commented, related to wafer quals and timings of releases of products to production. Some of that's a shifting between COGS and R&D. It's a question of which bucket we put it in. You need to understand that we're, in many cases, just moving dollars around. The guidance I gave you for the third quarter is roughly in line with where we've been historically, and I would see it as a more normal trend line.
Perfect. Thanks, guys.
Thank you, sir. Our next question comes from Stephen Chin with UBS. Please go ahead. Your line is open.
Great. Thank you for taking my questions and nice job on the quarter. First question is on the conversion of DRAM capacity to NAND. I think a couple of years back when some of your peers were doing that every so often, I think it took them about a month or so, maybe a little more than that, to get to full yields on one SSD conversion. Can you talk about how long it would take for you guys to achieve the same sort of mature yields after converting to a NAND capacity, given the pretty fully utilized clean room space in most of your fabs?
Well, I can't vouch for your time frames on what the competitors have done previously relative to mature yields. However, I would say that anything that we would consider would be on a much lower frequency than switching back and forth to take advantage of short market cycles. It'd be much more along the lines of permanent transitions as opposed to high-frequency changes to capture small changes in temporary market conditions.
Thanks, Mark. My other question is just in terms of memory content, can you guys talk about what the current NAND content is in SSDs in terms of the sweet spot, and where you expect to be at the end of the year? Similarly, can you also run through the same for mobile DRAM and PC DRAM? Thank you.
Sure. I'm assuming you're referring to the volume segment, the client business. On average, it's kind of in the 127, 128 gigabyte on the client side. When you look at the enterprise part of the business, obviously a little bit higher towards somewhere in the mid-250s on the enterprise side, and we think that will continue to go up. As I mentioned earlier, it's interesting because it has a little bit of a negative impact on the ASP per gigabyte calculation as you go to higher density drives. We think that that will continue to go, as far as the density consumption, that will continue to go up throughout the year, both in the client SSDs and enterprise.
How about for mobile DRAM products going to smartphones and also PC DRAM?
Well, I think what we're seeing in general on that side is continued increase in content, albeit at a much lower amount or quantity, if you will. If you look at gigabytes and gigabits in the DRAM space, on the gigabyte per your handset, in the smartphone business, we're projecting it's up about 37% year-on-year. If you look at NAND in the premium segment, it's up double digits as well. We're seeing some of the tail off in some of the feature phone environments. We think DRAM and NAND will continue to get pretty solid growth in content. Again, it's nowhere near comparisons to what SSD volume is in terms of storage content. Again, increasing content nonetheless.
Thank you, Mark.
Thank you, sir. Our next question comes from Doug Freedman with RBC Capital Markets. Please go ahead. Your line is now open.
Great. Thanks for taking my questions, guys. If you could talk a little bit about helping us understand how much of the business is presently exposed to contract-based pricing versus pricing that we're able to see in the marketplace on spot.
Well, Doug, as you know, we've increased our capacity a lot in terms of the NAND business with our restructure of the Intel relationship. On the DRAM side, this is the first quarter that we're seeing some of the effects of the Inotera volume, and obviously Intel, the Elpida deal hasn't closed. We're still more heavily weighted to OEM contract pricing in the volume markets specifically, and in Japan through especially. Our model today is pretty heavily focused on contract, and that's why, as I mentioned in my earlier comments, that lag effect shows up in our ASP at Micron relative to what you see in the DRAMeXchange. We see that gap closing in Q3 like it normally does in a recovering market.
Is there any chance that we get to a point in the industry that you guys feel comfortable enough with how much business is being done and the visibility that you have on that business, such that you'll offer revenue, gross margin and EPS guidance?
Doug, this is Ron. That's a good question about stability. One of the challenges we have in trying to give hard guidance, if you will, is just the variability in our pricing. Historically, it's been related to wild variability and volatility in ASPs and PC DRAM pricing, for example. Going forward, we have the added challenge of a very nice mix of products that we've been able to develop in the recent years, and it continues to expand. As you've seen us call out in conference calls, we have a lot of mix adjustments we have to do in our messaging. With that combination of ASP volatility, which at least is still with us to date, and variation in mix, it's pretty challenging to try to call it.
We do try to point you, as you know, in the direction of where we think things are headed as best we can.
Great. I've got a few more, but I'll jump back in the queue and leave it to my peers. Thanks, guys.
Thanks, Doug.
Thank you, Mr. Freedman. Our next question comes from Daniel Berenbaum with MKM Partners. Please go ahead. Your line is now open.
Yeah. Hi. Thanks for taking the call. Real quick on the sale of the Avezzano fab. When does that close, and does that mean that the sort of imaging category will finally just go to zero for you guys?
It'll be very close to zero post-close, which should happen sometime in the May timeframe.
From August on. That, I assume, is going to be accretive to your overall results. Can you talk how accretive would that be to gross margin and how much OpEx comes out?
Daniel, first of all, it's a May close, after May, we wouldn't have any volume. It would shift over if that's when the close actually occurs. In terms of impact on our financials, the imaging business is a piece of our other non-segment category in our financial statements if you look at our Form 10-Ks and Form 10-Qs, and it's the predominant share of it, but it's not everything that's in there. If I look at the recent quarters, we've been running relatively negative on margin. Not large amounts. It's zeros to tens of millions of dollars kind of range per quarter, and often around zero because that's the targeted structuring. If you think about it, in recent history, it's probably a slight improvement, marginally a few million dollars per quarter.
Okay, thanks. Then going back to DRAM, Mark, you'd mentioned that you thought DRAM wafer capacity was going down, if I heard that correctly. In the industry, what gives you confidence that industry wafer capacity goes down? Then what industry supply growth do you think we can get from technology shrinks alone?
The technology nodes are becoming increasingly complicated to drive the shift below 25 nanometers. If you look at, depending on exactly what your process technology node is, you look at anywhere from 10%-30% increase in complexity moving to a 20 nanometer node in terms of overall floor space loading and a total number of wafer moves. There's a big range there because people are moving from different places with different tool sets, et cetera. The trend is the same for all of us, which is we need more lithography, we need more dry etch equipment, et cetera, to facilitate scaling. Unless somebody's going to go out and build a new fab, which I mentioned a little while ago, relative to NAND, you've got some impediments with 3D. On the DRAM side, you've got EUV technology over the horizon.
Are you going to go build a bunch of pitch doubling DRAM capacity in advance of a potential technology shift to EUV? I think there's a lot of reasons why people may look at just living within their existing four walls, and that would tend to drive wafer capacity down while still driving an increase in bits.
Okay.
Sorry, was there a second part to your question?
That makes sense. Just sort of related to that is, if people live within their own four walls now, as we go through shrinks, what kind of bit growth do you think the industry is on? What kind of path do you think the industry is on purely from shrinks with no additional equipment?
Yeah, I think it's in this 20%-30% range that most people are prognosticating, that looks about right to me.
Okay, great. Thanks very much.
Thank you, sir. Our next question comes from Steven Fox with Cross Research. Please go ahead. Your line is open.
Good afternoon. Just two questions from me. First of all, on the server DRAM, I think you guys mentioned that you were gaining share during the quarter. I was just curious what you would attribute that to. Secondly, just back on SSDs. Relative to the growth rate that you talked about, the unit growth rate, how does that compare to the market, and what would you consider your expectations for unit growth on enterprise and client going forward? I know you talked about average capacities, but if you could sort of flesh that out a little bit more, that'd be helpful. Thanks.
On the first question as it relates to server, our growth has really been limited to our own capacity, which is now increasing. If you looked, we've actually had record quarters the last four quarters, I believe, in server. This is the first time we were over 200 million gigabit equivalents, and a lot of it has to do with the amount of server-grade quality we've been getting out of our factories, which again, has been going up dramatically. Micron's been a leader in terms of the server DRAM business for some time, and our customers are asking to continue to scale there. We continue to see growth opportunities, and it's just as we ramp more server product, we're getting it out to the market. Can you please ask the second part of your question on the SSD side?
Yeah, I'm just trying to understand the unit growth that you gave on SSDs a little bit better in terms of perspective on how you think the market's growing. Then what is the sort of outlook for SSDs going forward from a unit standpoint? You talked about average capacity, but I'm just trying to see.
Yeah, sure. I think the overall SSDs, we see year-on-year unit growth 2013 over 2012, in the 60%-70% range. Of course, ours is growing much faster than that. I think part of that is, anytime you have a new technology like SSDs, there are a lot of competitors in the market, some of which just weren't able to compete long term. That opens up market share for people who are in it for the long term, such as Micron. We've always said that we believe people with access to the NAND manufacturing, the NAND are going to have an advantage, we think as driven by our growth in the SSD market, that's playing out pretty well. We're optimistic we're going to continue to outgrow the market in the short term.
Great. Thank you very much.
Thank you. Our next question comes from the line of Ryan Goodman with CLSA. Please go ahead. Your line is open.
Hey, thanks for taking my question. Congratulations on the quarter. Question on the capacity. If you were to shift or convert some of the DRAM capacity to NAND, I know in the past you've talked about this being able to happen at a discounted price, but how should we be thinking about incremental cash requirements to get that new NAND volume to a competitive geometry?
I think the easiest way to think of it is, there's a spend for 1,000 wafers per week or 1,000 wafers per month that we would typically make to move a DRAM fab forward a technology node or a NAND forward, fab forward a technology node. It's about the same. If we want to take a leading-edge DRAM fab and move it to a leading-edge NAND node, it's about the same spend as it would be to take that same fab and move it forward to the next node in DRAM. It's that same sort of capital intensity. It's just choosing a slightly different node to transition to.
Okay. I guess as a follow-up to that then just a little bit. If you were to make shift from DRAM to NAND, it's a positive for margins at face value, but you're shifting into a sector where you probably have one or two shrinks to catch up if you're going just from where you are in DRAM to NAND. Costs are already getting a bit higher in NAND. There's some major transitions around the corner that are still unknown with vertical 3D, whatever it turns into. I guess, how are you thinking about the long-term return potential on these conversion costs, just between the conversion and the shrink versus, especially now with the DRAM, it's looking pretty profitable. I guess how are you thinking about balancing those two opportunities?
Well, we're all about margin, we're going to take a look at where we think the margin's going to be for us on a relative basis, and it's got to be sustainable. That's what I meant when I talked about low frequency and short frequency changes here. We're not going to be jerking our manufacturing capacity around to take advantage of short perturbations in the market, but we are going to try and align our capacity with the long-term demand and where we think we can drive higher margin business. If we identify an unfulfilled segment of the SSD market, we may want to go out and put capacity in place to service that. Likewise, if we see something particularly attractive in package-on-package DRAM or something of that nature, we want to make sure we're there, too.
We're going to identify those high-margin opportunities and move our capacity to intercept them.
Okay, great. Thanks.
Not too fuzzy for you, that's really all I can tell you.
No, that's great. Thank you.
Thank you, sir. Our next question comes from Alex Gauna with JMP Securities. Please go ahead. Your line is open.
Good afternoon, everybody. I was wondering, could you clarify, did your exposure to commodity PC DRAM actually go up on the quarter? Can you give us a sense of maybe what that exposure is percentage-wise for DRAM? Also how you feel about that exposure. I know you mentioned that there's a favorable supply-demand dynamic going on, maybe if you could punch through that and talk a little bit about what you see for PC and market trends. Thanks.
[inaudible] can help you a little bit with that. It did increase, our exposure to personal systems memory, which is what you're referencing, went from around 13% of total revenues in Q1 to around 18% or high teens in Q2. Maybe I'll let Mark talk more around the pricing dynamics and the demand dynamics, those are the data points for it.
Yeah, I'd like to just reinforce the point I touched on a little earlier, which is, a mobile DRAM part is not necessarily perishable to a PC DRAM part and vice versa. As we look at the overall market, both mobile DRAM and PC DRAM are pretty healthy right now. Having said that, as you think about the market conditions, even though the PC unit growth is projected to be low single digits to flat year-over-year, it's in pretty good balance right now. The PC market is not oversupplied as we sit here today in March. The mobile market's pretty hot. We're watching that balance because, as manufacturers, we have to call it right to keep the balance and get good favor, and that's what's, I think, helping us both in the PC space.
It's not end-user demand for PCs are up, it's just the economics of supply and demand are providing a pretty favorable business for us in PCs.
Okay. If you could help me understand with Haswell coming up, is there another SKU change there for you in terms of maybe DDR3 to DDR4, or how should we look at the timing of Haswell and what that means for you in terms of picking the right SKU?
I think we'll just continue to work with our customers to work on adoptions. We get a little concerned when we try to think too much about what's out in the press and the marketing of these types of technologies. It's really about customer adoptions and making our manufacturing output aligned to their overall demand there.
We've got qualified DDR4 products out there to Intel and other systems enablers.
Okay, thank you.
With that, I think we have time for one more question.
Sure thing. Our final question for today's event will come from the line of Kevin Cassidy with Stifel Nicolaus. Please go ahead. Your line is now open.
Hello, this is Dean Grumlose calling in for Kevin. Thank you for taking my call. What are your expectations for gross margin between the various segments, especially DRAM? Do you expect these will equalize in the longer term or remain stratified in some way?
Wow, good question. I guess to Ron's earlier comments, we generally have a difficult time ultimately predicting ASPs, which obviously are 50% of the impact to gross margin. Maybe I'll just end the call or give you one more chance with questions without trying to lead you to gross margins, particularly. We do have a few views, but maybe I'll just let that for market speculation for now.
Okay. Regarding Inotera, do you have the flexibility to run different types of especially DRAM within that agreement, or is there any reason this stays tied to the PC area?
No, we can use that capacity as we see fit.
Okay, one final thought if I could squeeze it in, is does the new Inotera agreement change the equity method treatment in the income statement, or does that remain the same?
Dean, this is Ron. It remains the same. We're about a 40% owner, and we take 40% of it on our equity line.
Okay. Thank you very much.
Thank you, Dean. Really appreciate it. We'd like to thank everyone for participating on the call today. If you will, please bear with me. I need to 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 the 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 second quarter 2013 financial release conference call. You may now disconnect.