Good afternoon. My name is Karen, and I'll be your conference facilitator today. At this time, I would like to welcome everyone to Micron Technology's third quarter 2016 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, please press star, then the number one on your telephone keypad. If you would like to withdraw your question, press the pound key. Thank you. It is now my pleasure to turn the floor over to your host, Ivan Donaldson. Sir, you may begin your conference.
Thank you, Karen. Welcome to Micron Technology's third quarter 2016 financial release conference call. On the call with me today are Mark Durcan, CEO and Director, and Ernie Maddock, Chief Financial Officer. This conference call, including audio and slides, is also being webcast from our investor relations website at investors.micron.com. Our website contains the earnings press release filed a short while ago and supplemental information including quarterly operational and financial metrics and guidance, GAAP to non-GAAP reconciliations, slides used during today's conference call, and a convertible debt and cap call dilution table. Today's call will be approximately 60 minutes in length. A webcast replay will be available on our website for one year. 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.
You can also follow us on Twitter at MicronTech. As a reminder, the matters we will be discussing today include forward-looking statements based on the environment as we currently see it. These forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from statements made today. We refer you to the documents the company files with the SEC, specifically our most recent Form 10-K and Form 10-Q, for a complete discussion of these important risk factors and other risks that may affect our future results. 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 today's date to conform these statements to actual results. I'll now turn the call over to Mark.
Thanks, Ivan. For fiscal Q3 2016, Micron posted total revenue of $2.9 billion with gross margin of 17%, a non-GAAP net loss of $79 million, and a non-GAAP loss of $0.08 per share, all within our guided range. Operating cash flow was $389 million. Top-line results were primarily impacted by continued weakness in the PC segment and the mobile qualifications we discussed last quarter. With recent data points indicating some improvement in channel pricing and expectation of finalizing our mobile qualifications and continued progress on our technology and operational milestones, we remain confident about our opportunities. Today, I'd like to provide a brief overview of our progress in each of the businesses. Ernie will cover other business unit performance details. In our compute and networking business unit, we returned to revenue growth despite pricing pressure in the client segment.
This was driven by the ongoing ramp of our 20 nanometer products, which exceeded 25 nanometer shipments on a bit basis for the first time. We enjoyed continued 20 nanometer qualifications across all CNB market segments led by our eight gigabit DDR4 product in enterprise, cloud, and client. In May, Nvidia launched the world's fastest consumer graphics card designed with Micron's GDDR5X. We're excited about the prospects for this high-performance memory product. In enterprise and cloud, we saw initial customer announcements based on our NVDIMM product offering with high-performance persistent memory. Turning to our mobile business unit, our results continue to be impacted by the timing of product qualifications as we transition customers to 20 nanometer versions of LPDDR4. We have successfully concluded some of the delayed qualifications that we discussed last quarter, and we anticipate finalizing the remainder during fiscal Q4.
We are ramping the output of these products throughout the quarter and into fiscal Q1 2017. We anticipate continued demand growth in mobile market in the fourth quarter in terms of both NAND and LPDRAM. In our embedded business unit, the automotive segment delivered record revenue driven by volume increases in DDR3 and increasing density mix in the eMMC. Density activity remains strong with recent qualifications of automotive consumer and connected home applications. In the consumer MCP business, we saw some recovery that we expect to continue into fiscal Q4, aligning with seasonal demand. Finally, our storage business unit is in the midst of refreshing its SSD portfolio with a higher capacity 3D NAND memory technology. We also introduced Micron Accelerated Solutions, which are enabled by our enterprise SSDs and advanced DRAM, integrating compute and storage to improve efficiency and performance in a variety of storage applications.
We announced our new 1100 SATA client SSDs, leveraging triple-level cell 3D NAND for class-leading performance and power efficiency. To summarize, our leading-edge technology deployment continues to progress throughout manufacturing for both DRAM and NAND, and we are on track with our bit growth and cost reduction targets. We believe the combination of new products with more efficient manufacturing on advanced nodes will drive improvement in our competitive position in the rest of 2016 and beyond. Turning to the memory industry more generally, we believe that the DRAM industry supply growth will be in the low to mid 20% range in 2016, which is consistent with our prior commentary. If wafer output declines in the latter half of the year, as some parties forecast, we would expect to exit the year on a slower run rate, and 2017 bit supply growth could be in the mid to high teens.
This compares to our long-term bit demand forecast in the low to mid 20% range. The significant improvements we're seeing in channel pricing are not currently impacting other segments, and as a result, we continue to take a conservative view of the market environment. For NAND, we estimate 2016 industry bit supply growth in the mid 30 to low 40% range, with a similar range in 2017 as early 3D conversions create some temporary supply constraints. Over the last several quarters, we've experienced strong demand coupled with aggressive pricing as suppliers have been driving to increase penetration rates and densities. Similar to DRAM, the current channel pricing environment appears to be improving, but has not yet significantly impacted across other segments. Our long-term bit demand forecast is in the low 40% range as lower costs and higher performance 3D NAND solutions enter the market.
From an operations perspective, Micron remains focused on a few key priorities. For DRAM, we successfully achieved our targeted 20 nanometer crossover during fiscal Q3, and also enabled our 1X node in manufacturing. We expect to ramp 1X nanometer DRAM in volume starting in 2017. We are still forecasting Micron's fiscal year 2016 and 2017 DRAM bit growth in the 20%-30% range, which is likely above the market. We achieved 22% bit growth in fiscal Q3 and expect even stronger bit growth in Q4. In light of current market conditions, we have no plans to add DRAM wafer capacity. As noted above, we are migrating to advanced technology nodes in order to achieve cost reductions and adjust higher density designs for mobile, cloud, and enterprise segments. For NAND, we continue to make great progress on our gen 1 3D NAND and are reaching mature yields ahead of expectations.
We still expect to achieve 3D bit crossover by this fall, which will allow us to take advantage of the cost benefits that this technology provides. Our 2nd generation 3D product is also on track with initial production this quarter. Equally exciting is that we expect TLC to be the majority of our 3D bit output within the next few quarters. In aggregate, we are forecasting Micron's fiscal year 2016 and 2017 NAND bit growth in the 30%-40% range. We expect to be somewhat below the market in 2016 and somewhat above the market in 2017. TLC-enabled 3D NAND technology is a big step forward and a significant driver relative to the progress we expect to make in our NAND business. This technology progress must be complemented by enabling product solutions for key storage and mobile segments.
We've outlined a storage product roadmap, which includes our recently announced client SSDs, followed by cloud drives later this year, and enterprise solutions early next year. We're also evaluating a number of mobile product opportunities for 3D NAND in 2017. Relative to 3D XPoint, we're working with market enablers across a number of market segments and continue to believe this innovative technology will be a strong contributor to Micron's future success with revenue in 2017 and beyond. Micron has a committed focus on the deployment of advanced technology to drive manufacturing efficiency and enable innovative new products for our customers. While we haven't finalized our fiscal year 2017 business plan, we are approaching that plan with prudence and conservatism and carefully reviewing our capital investments and projected operating cash flows to ensure the appropriate balance. Now I'd like to turn it over to Ernie.
Thanks, Mark. I'll start off by sharing technology and business unit details and circle back to the overall company results for the quarter, followed by the guidance for the fiscal fourth quarter. DRAM represented 60% of our total revenue with the following segmentation. Mobile was in the mid-20% range. The PC segment represented about 25%. The server business was in the low-20% range, and specialty DRAM, which includes networking, graphics, auto, and other embedded technologies, was in the high-20% range. In our non-volatile memory business, trade revenue represented 31% of total revenue with the following segmentation. Consumer, which includes memory cards, USB, and components, represented about 55%. Mobile and SSDs each represented approximately 13%. As a reminder, eMCPs are accounted for in the mobile segment. The automotive, industrial, multi-market, and other embedded applications were in the high-teens percent range.
Moving on, I'll share a brief operational summary of each of our business units. Micron's Compute and Networking Business Unit posted fiscal Q3 revenue of $1.09 billion, up 4% from the previous quarter, primarily driven by our 20 nanometer shipment growth across all segments and partially offset by lower average selling prices. Our non-GAAP operating loss was $63 million, or 6% of revenue. In the Enterprise Segment, demand for our 20 nanometer 32 gigabit DDR4 RDIMM was driven by the launch of Intel's latest server platform. We also saw solid growth in the Cloud Segment with continued transition to DDR4. In Graphics, we saw strong growth driven by our GDDR5 and GDDR5X products and are approaching a seasonally strong period for graphics applications, including virtual reality, and expect good performance from this segment for the next quarter. In Networking, our business was somewhat flat.
We made significant progress enabling our 20 nanometer products and also announced our eUSB 3.0 solution. Finally, within the Client Segment, we continued enablement and volume ramp of our 20 nanometer 4 gigabit DDR3 and 8 gigabit DDR4 solution to all major OEMs. In Micron's Mobile Business Unit, we posted fiscal Q3 revenue of $561 million, up 12% from the prior quarter, as we continue to ramp 20 nanometer LPDDR3 and LPDDR4. While the low-density eMCP market is still under pressure, a move to higher density designs in FQ4 and early FY 2017 should help stabilize and improve this segment. Our non-GAAP operating loss was $17 million, or 3% of revenue. We have successfully concluded some of the delayed qualifications that we discussed last quarter, and we anticipate finalizing the remainder during fiscal Q4.
This work will allow us to more fully ramp our 20 nanometer mobile products in fiscal Q4 and into next year. Both LPDRAM and NAND content continue to increase in mobile devices, which when combined with even modest unit growth, will result in very solid bit consumption. In our Embedded Business, we posted fiscal Q3 revenue of $487 million, up 6% from the previous quarter, with a non-GAAP operating income of $107 million, or 22% of revenue. The results were primarily driven by continued strength in the Automotive and Consumer Segments, offset by softness in the Industrial Multi-Market Segment. In our Automotive Segment, we achieved record revenue, increasing 6% quarter-over-quarter and 10% year-on-year.
We continue to see increasing demand in both DRAM and eMMC applications that include infotainment, instrument cluster, and advanced driver assist systems, and continue to see strong demand from our EMEA customers, which is more recently complemented by growth in Korea and demand recovery in North America. Our portfolio of leading-edge solutions is enabling major 2018 platform automotive design wins. Our Industrial Multi-Market Business declined 9% quarter-over-quarter, primarily due to global market softness in the Manufacturing Infrastructure Segment. We continue to see healthy demand for our NOR and NAND-based MCPs used in machine-to-machine wireless communication modules. Our Consumer and Connected Home revenue was up 3% quarter-over-quarter, with some softness in the set-top box business, offset by strong demand for NAND and LPDRAM MCPs to support action camera and home automation applications.
As we enter into a seasonally strong period, we expect demand to continue to grow. Customers are also beginning to design in and ramp our 20 nanometer DDR4 products into set-top box applications. Micron's storage business unit posted fiscal Q3 revenue of $719 million, down 20% from the previous quarter, with a non-GAAP operating loss of $62 million, or 9% of revenue. As we transition to lower-cost 3D NAND products, we continue to optimize our product mix. In client and consumer SSD, consecutive quarter bits sold were down 20% as we reduced production of planar NAND-based SSDs while ramping volume production of 3D NAND-based SATA and PCIe client and consumer SSDs. These new products will enable the company to enhance its competitive position. In the enterprise and data center SSD segments, consecutive quarter bits sold were down 10%.
As we've previously noted, our 3D NAND solutions will improve our product portfolio in this segment, enabling us to participate more significantly in this important growth business for the company. Looking at the company overall, as Mark noted earlier, revenue for the third quarter was $2.9 billion, which was near the midpoint of our guided range and roughly flat compared to the prior quarter. Fairly significant increases in volume shipments for DRAM were offset by decreases in selling prices, while trade NAND shipments declined as we are in the middle of a significant conversion from planar to 3D NAND. Gross margin for the quarter was 17%, within our guided range. The non-GAAP net loss for the third quarter was $79 million, or $0.08 per share, slightly better than the midpoint of our guided range.
As a reminder, Micron includes both amortization of acquisition intangibles and stock compensation expense in our non-GAAP results. Taken together, these two items represent $0.05 per share for the recently completed quarter. Let's look at results by product line. DRAM revenue increased 9% compared to the second quarter as a result of a 22% increase in bit shipments, partially offset by lower selling prices. As the result of our 20 nanometer ramp and ongoing mobile qualification timeline, DRAM finished goods inventory increased during the quarter. DRAM gross margins for the third quarter decreased approximately two percentage points to 18% as decreases in ASPs outpaced significant cost reductions. Our non-volatile trade revenue decreased 15% compared to the second quarter, reflecting a 10% decrease in bit shipments combined with a 6% decrease in ASPs.
Gross margin decreased a couple of percentage points to 17% as ASP reductions outpaced cost per bit reductions. Non-GAAP operating expenses for the quarter came in at $523 million, below our guided range due to the reversal of accrued cost for variable compensation plans, which were suspended in the third quarter. The company generated operating cash flow of $389 million, and we ended the quarter with cash and marketable investments of approximately $5.7 billion. Expenditures for PP&E during the quarter were $1.7 billion, and we continue to expect our fiscal 2016 capital expenditures to be in the range of $5 billion-$5.5 billion net of partner contributions. During the quarter, we received approximately $2 billion from the issuance of secured notes and an additional $114 million in equipment financing. We resolved a long-outstanding tax matter, which resulted in a $52 million benefit to the tax line.
This benefit was offset by the write-off of a related $30 million receivable that was reflected as non-operating expense. In the third quarter, we also acquired Photronics' interest in our captive mask operations for $93 million, resulting in 100% ownership of the mask operations. Moving on to our guidance for the fourth quarter. On a non-GAAP basis, we expect the following: consolidated revenue in the range of $2.9 billion-$3.2 billion, gross margin in the range of 15.5%-18%, operating expenses between $580 and $630 million, an operating loss ranging between $135 and $55 million, and an EPS loss ranging between $0.24 and $0.16 per share based on 1,036,000,000 diluted shares. Operationally, we are on track to achieve the bit growth and cost per bit reduction that we've previously shared as we continue to ramp our 20 nanometer DRAM and 3D NAND production.
In recognition of the current business environment and the need to accelerate focus on the company's key priorities, we plan to implement a cost-saving program, which we expect will save the company approximately $80 million per quarter in fiscal 2017. The savings will result from a combination of a more focused set of projects and programs, the permanent closure of a material number of open headcount requisition, and a workforce reduction in certain areas of the business, as well as other non-headcount-related spending reductions. About half of these savings will appear in the gross margin line of the company, while the remainder will be reflected in operating expenses. These savings are baseline against our previously planned 2017 fiscal spend levels.
We expect to take reserves approximating $70 million for the cost of this program, the majority of which will occur in fiscal Q4, with the remainder in the early part of fiscal 2017. As we complete our fiscal 2017 planning process, we are mindful of the need to effectively balance period spending, CapEx, and free cash flow, and we continue to explore other opportunities to improve the company's financial performance. Finally, at this time, we don't have any new information to share relative to Inotera. As we stated in our press release, the transaction will not be closing in mid-July, and we expect to provide an update during the latter part of the calendar year. With that, I'll turn it back to Mark.
Thank you, Ernie. To summarize, we continue to navigate challenging market conditions but remain confident in the long-term health of the industry and the company's strategy to improve our relative competitive position. The decision to implement cost-reduction initiatives is always difficult and is never made without thoughtful consideration about the short-term and the long-term impacts. However, to ensure that we can continue to place emphasis on our most important company priorities, we believe the steps Ernie outlined are prudent and will help deliver the best long-term results for the company. I'd like to take a moment to thank our customers, partners, shareholders, and team members for their continued support. Operator, we're now ready for Q&A.
Thank you. Ladies and gentlemen, if you have a question at this time, please press star followed by the number 1 key on your telephone keypad. If your question has been answered or if you'd like to remove your line from the queue, you may press the pound key. Our first question comes from the line of Vijay Rakesh from Mizuho.
Yeah. Hi, guys. Just a couple of questions here. On the DRAM side, obviously, good bit growth there. What's the mix of 20 nanometer that you're shipping now, and how do you see that as you progress through the August quarter? Do you see further cost reductions there? Thanks. I have one follow-up.
20 nanometer is a majority of the bits that we are shipping now. When we talked in our prepared remarks about strong bit growth above Q3 levels, you saw what the relative growth was in Q3, and you assume there'll be a little bit north of that as we look at fiscal Q4.
Got it. On the 3D NAND side, obviously, looks like bit growth was a little light. Obviously, looks like with the transition. What's your mix of 3D NAND that you're shipping now, and when do you see the transition and the bit growth start to open up for you on the NAND side?
Yeah. I'd say low double digits today. As we commented, as we get into the back half of the calendar year, we expect a crossover. That's coming up pretty soon. It's a pretty significant ramp from here over the next couple of quarters.
Got it. Obviously, you guys are giving a point estimate on the gross margin now, which is great. As you look at your mix with 20 nanometer costs coming down and 3D NAND TLC starting to ship, any thoughts on where you see those margins start to bottom out? Thanks.
We typically don't comment on margins because of the function of a pricing environment that's moving around on us every day. Really, what we have said is that the cost
Curves that we articulated in our Analyst Day are still legitimate. As you can imagine, when we do 3D crossover in NAND in the fall, that's when you'd expect to see the accelerated cost reductions begin to occur there. On the DRAM side, with the kind of bit growth that we delivered in fiscal Q3, as well as what we're forecasting for fiscal Q4, you can get some idea of what the cost reductions would be there.
Great. Thanks a lot.
Thank you. Our next question comes from the line of John Pitzer from Credit Suisse.
Yeah, good afternoon, guys. Thanks for letting me ask the question. Ernie, maybe just to follow on to that margin question. I appreciate that margins are a function of pricing, which is hard to predict, but you've got a lot of elements moving in the right direction. Revenue is growing sequentially into the fiscal fourth quarter. You've got 20 nanometer mix. You've got more mobile. NAND is starting to turn your way. Yet you've got gross margins that are likely going to be down sequentially. Can you help me understand just kind of some of the factors around mix that are in play? Kind of the narrative that we were hoping for is that this quarter, the May quarter, and the August quarter, would start to show some tangible evidence of you kind of closing the relative cost gap to some of your peers.
Just given the gross margin guidance, it looks like that's not happening, and I'm trying to figure out why.
John, maybe I can take that one instead of Ernie. On the pricing environment, as you noted, we're seeing there's obviously encouraging things going on in the channel. Before we start baking that into any sort of guidance we would give, we want to actually see that as it moves through other end market segments and out in the contract pricing, et cetera. Really this is a question about the tailwinds we have, and I think you did a nice job articulating for us. I think we've done a good job on the 20 nanometer ramp, and we'll continue to see some improvement for that as we move through the next quarter.
We articulated pretty strong bit growth moving forward in both DRAM and in particular, maybe a couple of quarters out in NAND as we really see that transition kick in and the 3D TLC part of that transition kick in. We've got a lot of good tailwinds where we also articulated we're making a lot of good progress on qualifying and re-segmenting those more advanced technology nodes in the right market segments. We do believe that you'll see progress relative to our competition in terms of how our margins move. We never know exactly what they're going to do, and it's tough for us to predict what their results are going to look like given the pricing environment.
That's helpful, Mark. I appreciate it. Ernie, maybe as my follow on, just on the OpEx guidance for the fiscal fourth quarter, a little bit higher than we thought. Help us understand, are there some period expenses that are coming in? As you talk about the restructuring and the cost savings, given that you kind of have that cost gap and you have a lot of initiatives on your plate for investing, how do we get comfortable that you're kind of taking costs out of the right areas and yet still investing in areas you need to invest?
Sure. To the first part of your question, there are some very significant pre-qual expenses that we expect to incur in Q4 that are influencing the guidance that we've provided for the quarter's OpEx. Relative to the cost reduction, the only comfort perhaps I would offer you, which I hope is a very strong comfort, is that we actually think very carefully about what we're doing here. The decisions that we made with respect to the cost savings program were deliberated very thoughtfully, debated, and at the end, we believe were the right balance to both help improve the financial performance of the company, while at the same time, not jeopardizing any of the company's potential in terms of revenue opportunities.
John, maybe I can add one point there just so everyone on the call is clear. I know you are. The pre-production R&D costs Ernie's referencing are really associated with a lot of new products in qual with these technology nodes that are now ramped. Some of those are sort of costs that flow through the R&D now, as opposed to directly on the cost pool and inventory.
Perfect. Thanks, guys, appreciate it.
Thank you. Our next question comes from the line of Romit Shah from Nomura Securities.
Yes, thanks. Mark, you mentioned that channel pricing's improved. At what point would you start to see that show up in contract and in some of your other businesses? Because it looks like, just based on the guidance for August, that you're not expecting the overall pricing environment to be better relative to May.
It's a very dynamic environment. Again, we want to make sure we're giving you what we believe is realistic guidance and guidance that is achievable, and we're going to have to see what plays out in the markets over time. When spot pricing starts moving, that's a great indicator of where the markets might go. What has to happen, obviously, is there has to be enough liquidity in the market that it can carry through into larger volume orders associated with the contract market. Inventory has to burn off. We're seeing good indications of those types of things, but we need to just make sure the pudding's fully baked before we serve it.
Okay. That's helpful. Just on the pre-qual expenses that are boosting OpEx this quarter, is that more one time in nature, or do you expect expenses to sort of remain elevated through the course of the year as you ramp these new products?
Well, certainly, they will occur as there is a concentration of new products. You shouldn't expect that they're going to occur each quarter. I would tell you that in Q4, they are particularly high relative to what we've seen over the recent history as a reflection of what Mark said. It is likely that sometime during the course of 2017, you're also going to see these occur as we release new products. Again, we see a particularly high concentration in FQ4 2016.
Okay. Thank you.
Thank you. Our next question comes from the line of Timothy Arcuri from Cowen and Company.
Thanks a lot. I guess I had two questions. First of all, demand bits are really moving around a lot, obviously given the planar to the 3D move. Maybe you can hold our hand a little bit, A, on the 3D transition, because it does sort of look like there's been some issues in the 3D transition, looking at the bits. Maybe also talk about what the bit outlook is for August in NAND to just sort of give us some comfort that you're sort of getting through this transition.
Tim, I'll take the second part of your question first, then we'll circle back around to the first. The stake in the ground is we've said a few things which were actually, we reiterated today that would kind of be in the opposite direction of thinking there were problems with our 3D NAND transition, right? We're seeing mature yields occurring at rates that are slightly faster than we anticipated. We had originally, probably six months ago, talked about bit crossover for 3D at the end of the year. Today, we reiterated again that that was going to occur in the fall, so a pull forward of a couple of months. We're not going to give you a bit guidance for the August quarter.
We do think that these data points will hopefully give you comfort that in fact, the transition really is occurring slightly more rapidly than we have expected it to occur.
The only thing I would add to that, Tim, is keep in mind, we've also talked about the fact that we want to deliver more of our NAND bits in actual solutions to the customer. As we do that, there's a natural tendency to stretch out the supply chain in aggregate and some lengthening of the total manufacturing cycle time as we move through that transition. Part of this, I think, is also reflective of the fact that we're making some fairly significant progress relative to getting these bits into the types of end products that we want to deliver to the customer.
Yeah. I think one final point that may be also contributing to the overall impression is we do see planar bits coming down fairly significantly as we move away from the planar SSD markets that we have sold into in the past. There are many dynamics sort of at play here relative to what comes out as an aggregate bit number.
Got it. Then just as the follow-up to that, Ernie, you talked at the Analyst Day, I think you said that $800 million of the CapEx this year would be for fab shell spending, so that sort of net of the partner contribution, you'd be sort of in the $4 billion-$4.5 billion range. I'm sorry, net of the partner contribution and also net of the fab shell spend, you'd be in the sort of $4 billion-$4.5 billion range. So I'm just kind of wondering, as you look to next year, and I'm not asking for guidance, but I'm just wondering how you think about the maintenance CapEx level of the company, given some of these initiatives you have in DRAM and also in 3D NAND. How do you sort of think about how to balance that thing?
Well, you really asked kind of two different questions, right? How do we balance? Obviously the balance has to come in as we think about the aggregate business for the company and the cash flow that we plan to generate, and a number of other factors. As we look at maintenance CapEx, we sort of said that we were normalized in that low $4 billion range ex Inotera. We wouldn't move away from that in a very substantial way as we're thinking about things today, bearing in mind that we're still in the midst of a ramp of 3D in the Singapore fab. It isn't exactly a normalized environment for us. Even taking out the shell, we're still outfitting that shell with productive capacity.
Got it. Okay. Thank you so much.
Thank you. Our next question comes from the line of David Wang from Wells Fargo.
Thank you very much. For the second generation 3D NAND, when you hit crossover in the fall for 3D NAND, will this all be on second generation technology, or will it be first generation? What's the difference in layer count between first and second generation, please?
David, it will be primarily a significant majority of that, in fact, almost all of it will still be gen one. Gen two, we start to see significant output in the second calendar quarter into the summer of 2017. We haven't said what the layer count does, but we've indicated that it's roughly a 30% cost reduction moving from gen one to gen two, and that it's roughly double the bits.
Great. Thanks.
Thank you. Our next question comes from the line of Christopher Danely from Barclays.
Thanks very much for taking the question. I guess, first of all, building on David's question, as you're seeing 3D ramp across the industry, how do you view your competitive positioning? At your analyst day, you laid out some beliefs that you were going to have cost leadership. Is that holding as you're seeing the competition ramp?
Yeah, this is Mark. Yes, absolutely. We're very happy with the way the ramp's going. We're very happy with the technology we're delivering and its ability to scale going forward. We think it's the best solution in the industry for almost every end market application.
Okay. I guess to the harder question, what needs to happen for you guys to get back to sustained profitability? I applaud the restructuring efforts, but I look at the savings you've announced, and that falls short of the losses you put up the last 2 quarters and what you guided to the next quarter. Are you confident these changes are going to be enough to get you back to profitability next year?
Well, we think we're doing all the right things, and we think we're making a lot of good progress on our relative competitive position on technology. The products are all coming along nicely now, and the qualifications are progressing well. There's always a wild card around ASPs, but we think we are making very significant progress, and we think we'll see that every quarter as we move forward, you'll be able to measure us versus our competition and hopefully confirm that we're doing all the right things. Exactly when that results in profitability, we can't tell you, but we think it's coming.
Thanks much.
Thank you. Our next question comes from the line of Kevin Cassidy from Stifel.
Yeah, thanks for taking my question. On the mobile side, as you start getting more qualification on the LPDDR3 and LPDDR4, I guess, what % of your revenue will be those products versus eMCP, and what is it today?
Let me see if I can find those numbers. I don't have them right in front of me. For the mobile DRAM portfolio, generally, eMCP probably represents.
Well, I'll jump in real quick, Mark. The way we break it out, Kevin, is mobile DRAM discrete is reported in our DRAM business. That's about mid-20% of our total DRAM. Our eMCPs are reported in our NAND reporting segment, and we said mobile was in the mid-teens % of the NAND business, low to mid-teens.
Okay, maybe as you ship more of the LPDDR3 and 4s, are you expecting that your average selling price would increase?
It's really challenging to give you a forecast of what's going to happen with the pricing environment. In terms of aggregate revenues, we would expect that as we move through these qualifications, that you're going to see the mobile business maintain or slightly improve in terms of a percentage of its overall revenue to the business. As we noted in our comments, as we see eMCP market moving to higher density, we think that will be a sort of a starting point for an improvement in that aggregate market, which has been weak, actually, in the last couple of quarters. That's a quarter or so off as we roll into calendar 2017.
Okay, great. Thank you.
Thank you. Our next question comes from the line of C.J. Muse from Evercore ISI.
Good afternoon. Thank you for taking my question. I guess first question on the mobility qualifications side, how much, I guess is the uncertainty around timing there, whether it's the month of August or September, October, impacting your gross margin guide? Then as part of that, how should we think about the incremental gross margins as you start selling down your inventory going forward once we do get the qualified?
C.J., this is Ernie. We're pretty clear on the qualification timeline. There is more uncertainty around the timing of shipments because at the end of the day, that is going to be a function of each of those customers' sort of current inventory dynamics with respect to the inventories they have on hand, their success in the marketplace, and how they want to position their inventory. In addition to that, we are going to have, as we already talked about, significant quarter-on-quarter growth in production, some of which is directed toward that mobile market because we believe that it's fairly strong. Mobile in general has a better gross margin profile than the aggregate business.
As that product flows through, although there will be some early pressure as a result of that inventory having been built at a time when costs weren't fully baked as the result of the volume ramp, eventually it will flow through into the margin profile of the company.
Okay, that's helpful. I guess as a follow-up, when you think about your 3D NAND cost structure, curious how you would compare that today, versus the price leader, in planar NAND, and when you think you'll cross over. Is that gen 1 when you're fully ramped at high volume, or is that more gen 2 in the summertime next year?
I think it is probably most closely timed to an increase in our TLC mix, which is probably a couple of quarters out, and then we'll get a second surge sort of in calendar Q2 and Q3 of next year as the gen 2 kicks in a significant way.
Great. Okay.
Again, we think we'll be the leaders. We fully implement gen 2 and TLC.
Great. Thanks, Mark.
Thank you. Our next question comes from the line of Joseph Moore from Morgan Stanley.
Yeah. Hi, thank you. I just want to clarify, when you talk about Q4 DRAM bit growth being better than Q3, that's sequential bit growth north of 20%. I guess that then implies when you look at revenue not growing very much, that ASPs are down similarly. I know you don't want to give a pricing forecast, I just want to make sure I understand. I'm finding it hard to get to your revenue number with my pricing assumptions.
The first part of your assumption process, Joe, is correct. Again, we're not going to comment on ASPs.
Okay. Separately, can you talk about how the Inotera, the new pricing agreement, it factors into this? Is that a positive or a negative at this point, relative to the old pricing agreement? Maybe if you could compare that versus, a wholly owned situation, if that deal, in fact, moves through. Just if there's any context you can give us around that dynamic.
Sure. In the current pricing environment, what we've contemplated as we look forward to Q4, it's pretty much a wash to be honest with you. Either way, the results to Micron would be roughly the same.
Got it. Okay. Thank you.
Thank you. Our next question comes from the line of Harlan Sur from JPMorgan.
Hi, guys. Thanks for taking my question. I'm still a little bit unclear as to the gross margin decline or the slight gross margin decline in the August quarter, given that we've seen some stabilization in DRAM and NAND prices starting kind of in the month of June. Your 20 nanometer mix and your 3D NAND mix is moving higher. We'll assume that your blended costs are coming down nicely. You mentioned other parts of the market that may not be participating on the pricing stabilization that we're seeing in the PC market. If I look at the end market demand parameters, data center, networking fundamentals seem to be improving into the second half. Embedded fundamentals seem to be seasonally up. Can you guys just articulate what segments are still showing aggressive pricing declines?
Yeah. Let me try to characterize that a little bit for you. It is important to recognize that these data points are really happening in real time. As we look at how we actually entered this quarter, we were actually starting the quarter below the average of the prior quarter, and we're now seeing some upward trends that are causing everyone to have the enthusiasm that we're talking about. It is important to understand that what we've been reading about in the last two weeks has, A, not impacted the entirety of the quarter, and B, hasn't reflected itself yet as we look at forward pricing. As Mark mentioned a couple times, we really are wanting to provide prudent guidance in the context of an environment that's moving around very significantly.
The other point I would make is, it is absolutely true that we are seeing strong bit growth and cost reductions on the DRAM side. Remember that we said the majority of the NAND cost reductions occur as we achieve that bit crossover, which is later in the fall of this year. We are getting quarter-on-quarter cost reductions in NAND, but I think you might be perhaps pulling forth that bit crossover comment into the full quarter that we are coming up on, and that is just not going to be the case.
Okay. Appreciate the insights there. Maybe more of a product question from me next. If we include 2-in-1s, I think SSD and embedded NAND attach rates and notebook PCs are at or slightly above 50% now and continuing to climb. I think you guys had a target of having your 3D NAND-based client SSD in the marketing calendar Q3. Maybe if you can just give us an update there. I think you guys also had a target to have your 3D NAND-based SATA solution, your high-capacity drives for hyperscale guys, starting in calendar Q4. If you could give us an update there as well. Thanks.
Well, they are both in the market, and we are going to just have to see what the demand looks like. So far, we like the reaction we are seeing.
Great. Thank you.
Thank you. Our next question comes from the line of Rajvindra Gill from Needham & Company.
Yeah. Thanks for taking my question. On the whole 3D NAND conversion, I just wanted to get a better understanding of, in your estimation, what would be more cost effective, the 3D on 32 layer versus the 20 nanometer that's currently on 2D NAND?
Oh, definitely the 3D, 32 layer is
Cost advantage relative to 20 nanometer, even 16 nanometer planar NAND. We believe that as we move through time, the market is going to appreciate incremental value in 3D NAND bits as well.
Can you give us an idea in terms of if your 3D NAND is actually fitted for some of the applications, is it will be available for applications such as mobile or embedded or cars?
Yeah, we have a full product portfolio coming for 3D NAND that'll address, and I think I mentioned this in my commentary. We have 3D NAND products for consumer client, data center, and enterprise, as well as evaluating mobile applications on a go-forward basis.
Okay, just last question on the mobile DRAM qualification issues. What specifically can you talk about? Has there been quality issues with respect to your mobile DRAM, which is preventing you from getting qualified at certain customers and as a result losing market share to Samsung?
No, I think we've gone through a very significant transition here, in some cases from 30 nanometer to 20 nanometer, in some cases from 25 nanometer to 20 nanometer. From a timing perspective, in some cases, we were later with a particular either density or IO or interface that the customers were wanting. It's just a matter of working through that process and qualifying those products in the various fabs that we're ramping. I don't think there's anything untoward or unusual in that process. It's just a matter of a fairly significant ramp across two high-volume fabs and taking the time to qualify a full suite of products across all those technologies.
Rajvindra, if I could just squeeze one more. In terms of competitively as it relates to your other major competitor moving to a smaller process node, and you guys perhaps perpetually playing catch up in terms of the process node transition. What are some of the lessons you think you guys have understood from the 20-nanometer transition and trying to rectify that when you go to the other process node transition? Just in general, any comments there would be helpful. Thank you very much.
Yeah. Certainly, we think 3D NAND has gone very well. For us, that has been really a model where we take technology we've developed in Fab 4 and move it into a consistent equipment set in a historically Micron environment where the equipment all matches, and things go very well. What we've experienced in 20-nanometer is we're matching a disparate set of equipment given the Elpida acquisition. We've got one tool set in Hiroshima, one tool set in the MMT fab in Taichung, and a separate tool set in the Inotera fab, and that's just more complicated. Every time we make this transition, we get more and more of those tools aligned, and we feel like actually we're in pretty good shape now.
As we move to the 1X node beyond that, it's going to be a lot more seamless and a lot easier to execute on a go-forward basis. We think we're the leader in 3D, and we think we're getting closer in DRAM, and 1X is coming along now running at sort of pre-production levels in both Hiroshima and Taichung.
Thank you very much.
Appreciate it.
Thank you. Our next question comes from the line of Mehdi Hosseini from SIG.
Yes, thanks for taking my question. Mark, I'm just looking at your NAND gross margin. It's been below corporate average since mid-2013, and you have all of these new product qualifications coming up. I'm just wondering, is there a strategic alternative here to take a more dramatic action? It's been more than two years that NAND has been underperforming compared to the corporate average. In case these product qualifications don't go well, is there an alternative strategy here? I have a follow-up.
Yeah, we're always open to looking at lots of different ways to optimize the business for the shareholders. Having said that, we like our NAND position. We like the technology. The early feedback from the customers is they like what we're doing. They like not only the 3D technology, but the product portfolio. We think we're making progress. We are kind of encouraged with the progress we've been making. Yes, we've admitted for 18 months now that we completely missed the boat on planar TLC. We think we've taken significant steps to remedy that on a go-forward basis. We intend to do so.
We're going to play it out, and as we do that, we'll look at lots of different strategic options for the DRAM business, just like we do for the NAND business, just like we do for the DRAM business and 3D XPoint and all the other interesting new technologies we're developing.
Any way you could share with us qualitatively or big picture what some of those options are?
Well, I think we'll just keep those to ourselves, Mehdi. I think the main point I would want to make to you is we think it's important to have a diversified set of products or technologies to support memory system solutions for our customers. We think we're in a strong position there. We like the growth profile associated with the non-volatile memory business generally, and that includes not only NAND, not only the 3D XPoint business that we're currently developing, but also other advanced storage class memories we're working in. We'll look at lots of different ways of optimizing the value, in all those different technologies. We may find different solutions in different segments, but we're always looking for those opportunities to create value-added partnerships or new relationships.
Got it. Quickly for Ernie, how should we think about working capital, inventory, and accounts receivable? It seems like at least on the inventory side, it may not come down significantly until February quarter. Is it a fair assumption or not?
I think it's reasonable to think about inventories being within a pretty narrow band for a couple of quarters, I think that's a reasonable thing to think about. Obviously, we pay a lot of attention to working capital and intend to do so on a going forward basis. My statement about inventories was on a dollar basis. Obviously bits are going to move around as costs move around.
Okay. Thank you.
Operator, I think we have time for one more question.
Certainly. Our final question for today comes from the line of Tristan Gerra from Baird.
Hi, good afternoon. You've mentioned on the call today that the 3D NAND cost structure is linked to the TLC mix, and you've mentioned a two-quarter out timeframe for that mix ramp. Is that in line with the fiscal Q4 target, in terms of cost improvement in your NAND business that you provided last quarter, or is there a little bit of a change in terms of where you expect an inflection point in your cost structure in NAND flash?
No, there's no change. I think we're pretty consistent. Fiscal Q1 is aligned with our crossover in the fall, what we talked about today, which is a little earlier than we had originally talked about. That drives a surge in the cost improvement, but not necessary to drive the improvements that we've outlined relative to our guidance.
Okay, just as a quick follow-up, any steps that you can talk about that you're taking to accelerate the node migration, notably at Elpida?
The migration of DRAM from 20 nanometers to 1X?
That's right. Mm-hmm.
We're not completely done with 20 nanometer there yet, but as I mentioned a minute ago, we have 1X running there. It's kind of where we've been doing our pilot line activity, and we've transitioned, or we've transferred then that technology into the Taichung fab, and they are progressing with that. We wouldn't expect to see volume start in either Elpida or Taichung until later this year.
Great. Thank you.
Thank you. That does conclude our conference for today, and we thank you for your participation. You may now disconnect.