Micron Technology, Inc. (MU) Q2 2018 Earnings Report, Transcript and Summary

Good afternoon. My name is Jonathan, and I will be your conference facilitator today. At this time, I would like to welcome everyone to Micron Technology Second Quarter 2018 Financial Results Conference Call. [Operator Instructions]. It is now my pleasure to turn the floor over to your host, Shanye Hudson. You may begin your conference.

Thank you, Jonathan, and welcome to Micron Technology's Second Fiscal Quarter 2018 Financial Conference Call. On the call with me today are Sanjay Mehrotra, President and CEO; and Dave Zinsner, Chief Financial Officer. Today's call will be approximately 60 minutes in length. This call, including audio and slides, is being webcast from our Investor Relations website at investors.micron.com. In addition, our website contains the earnings press release which was filed a short while ago. Today's discussion on financial results will be presented on a non-GAAP financial basis unless otherwise specified. A reconciliation of GAAP to non-GAAP financial measures may be found on our website, along with the convertible debt and capped call dilution table. As a reminder, the prepared remarks from this call and webcast replay will be available on our website later on today. 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'll be attending. You can also follow us on Twitter, @MicronTech. As a reminder, the matters we will be discussing today include forward-looking statements. 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 we filed with the SEC, specifically our most recent Form 10-K and Form 10-Q, for a discussion of 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're 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 you, Sanjay.

Thank you, Shanye. Good afternoon. During the second quarter, Micron, once again, set company performance records across multiple metrics, including revenue, gross profit, EPS and cash generation. We are consistently delivering results that underscore our relentless focus on execution and solid progress on our strategic priorities. Specifically, we are evolving our product portfolio to a richer mix of high-value solutions, enhancing our financial performance and cultivating deeper relationships with more key customers across multiple mega markets. Our growing portfolio of managed NAND solutions and low-power DDR4 products boosted our mobile business to record revenue and profitability during the quarter. We also grew our SSD shares in our second quarter with total SSD sales up 80% year-over-year and sales of cloud and enterprise drives more than tripling for that same period. Continued strong penetration of our highly competitive DDR4 products into cloud applications and our industry-leading high-performance graphics memory portfolio into gaming, graphics and cryptomining applications contributed to a robust 15% sequential growth for our Compute and Networking Business Unit. Strong demand for our DRAM and NAND products delivered record second quarter revenues for us in the automotive market. We continue to execute well on our goal of introducing new products on our advanced technologies, delivering performance, quality, supply and cost advantages to our customers. In NAND, we are transitioning from being the components supplier to becoming a solutions provider. We launched and began qualifications of the industry's first cloud and enterprise SATA SSD drive incorporating 64-layer 3D TLC NAND. We also introduced the 3DFS solutions targeted at flagship smartphones. These solutions are also based on our 64-layer 3D TLC NAND, which has 15% higher performance and double the density of the prior technology. We have qualified a family of these products with a major chipset vendor and we expect to complete customer qualification in the coming months. In DRAM, our focus remains on enhancing our cost competitiveness and accelerating our product execution. We have qualified our 1X nanometer DRAM at three of the world's largest hyperscale customers, with other qualifications underway. We also garnered positive feedback on our 1X nanometer LPDRAM solutions and set industry benchmarks for power efficiency, which is particularly critical to our mobile customers. Our comprehensive and expanding portfolio of DRAM, NAND and NOR solutions has enabled us to achieve record design wins for our automotive business in the first half of fiscal 2018. We believe we are well positioned to continue to support our share leadership in this rapidly growing market. These achievements illustrate our focus and the ability to deliver value to both our customers and shareholders. I'll now discuss some of the trends we are seeing across our end markets, which will continue to expand the significant opportunities for our business in the years ahead. At Mobile World Congress recently, phone manufacturers featured high-end smartphones with larger 4K displays, multiple high resolution cameras and 4K HDR video recording. Capabilities like these have driven increased memory and storage requirements in recent years. But perhaps more most impressive were the multiple implementations of artificial intelligence and virtual reality. OEMs are bundling new artificial intelligence, augmented reality. OEMs are building new artificial intelligence, augmented reality and lifelike virtual reality capabilities into high-end smartphones, including facial and voice recognition, realtime translation, fast image search and scene detection. To support these data intensive capabilities, flagship and high-end smartphones are migrating towards 6 gigabytes of LPDRAM, a trend that bodes well for Micron given our leadership in LPDRAM power efficiency, which is essential for optimizing battery life. Average storage SSDs are also increasing across all smartphone classes with new flagship models using 64 gigabytes of flash memory at a minimum. Micron's portfolio of managed NAND solution is well-suited to address this growing demand, and we are leading the industry in TLC utilization with a portfolio that leverages the strong attributes of our 3D NAND technology. Of course, the growing adoption of AI is not limited to mobile. At the Consumer Electronics Show, several companies showed AI smart cockpits in new automotive models. These systems integrate the instrument dashboard, infotainment and telematic systems with a centralized compute and storage architecture to create a data center on wheels. Voice and gesture recognition, combined with driver alert monitoring capabilities, are making automobiles more intelligent and much more compute intensive, requiring higher capacity and more powerful memory and storage solutions. Micron is already working with automotive customers who will benefit from our highest speed automotive-grade, LPDDR4 solutions in the near term and new memory technologies in the future like our high-bandwidth GDDR6 graphics memory. The new features in mobile, automotive and other connected devices require rapid data analysis in storage and enterprise and cloud servers, including machine learning, training and inferencing to complement the compute taking place at the edge. This is driving significant investments in the data center and growing demand for both memory and high-performance storage. Micron's broad technology portfolio and strong innovation engine position us well for these growth trends. We continue to partner with our customers to ensure our technology and engineering roadmaps deliver the critical features for tomorrow's solutions. Now I will provide an update to near-term industry supply/demand dynamics. The dealer market today is very different from the PC-dominated market of the past. This market now supports a healthy demand environment with several secular demand drivers that I have discussed earlier. More specifically, memory is making possible applications such as AI and VR, and enabling new cloud-based business models which deliver a fundamental value far in excess of a price per bit. Against this healthy demand backdrop, we project DRAM industry bit output to grow in the 20% range for calendar 2018, maintaining favorable industry fundamentals. For the NAND market, we believe the ongoing transition to 64-layer 3D NAND creates the opportunity for a more balanced industry dynamic in calendar 2018 versus the constrained conditions we saw in 2017. We expect industry bit output growth to be somewhat higher than 45% in calendar 2018, providing incremental supply to address the increasing demand created with the further displacement of HDDs in client, enterprise and cloud applications. From a Micron perspective, we continue to make significant strides to strengthen our competitive position through technology and cost improvements. In DRAM, we are focused on accelerating our technology transition cadence and ramped our 1X nanometer technology to mature yield faster than any of our previous technology nodes. We remain on track to achieve 1X nanometer bit output crossover relative to our 20-nanometer node by the end of calendar 2018. We now expect Micron's calendar 2018 DRAM bit output growth to be in line with the industry's 20% range. In NAND, our 64-layer technology continues to ramp very well with yields somewhat ahead of plan. We continue to execute plans to achieve bit output crossover on our 64-layer 3D NAND technology relative to 32-layer in the second half of fiscal 2018. We believe we will be somewhat above industry bit output growth in calendar 2018 for NAND. We expect to deliver qualification samples to OEM customers of both our 1Y DRAM technology and our third generation 3D NAND technology by the end of fiscal 2018. And we continue to expect to ramp initial volume for each of these new nodes in the second half of calendar 2018. For some time now, industry participants have pointed out that the cost and complexity of DRAM and NAND scaling is increasing with each subsequent technology node. Additional space and equipment is required to manufacture the increasingly complex architectures of these leading technologies to maintain data capacity and meet market demand. Accordingly, we are executing plans to add clean room space in our NAND and DRAM SAS network. With the support of the Singapore Economic Development Board, we have finalized plans to build additional shelf space in Singapore, adjacent to our existing NAND Center of Excellence. The primary purpose for this new clean room space will be to transition our existing wafer capacity to future 3D NAND nodes. This location will enable us to drive efficiencies of scale. We expect to build out this facility in phases, in line with our manufacturing requirements and market demands. The first phase of this clean room is expected to be completed by the summer of 2019, with initial wafer output from the facility expected in the fourth quarter of calendar 2019. We are also building out incremental clean room space in our fab in Hiroshima, Japan, which will be available for production at the beginning of calendar year 2019. This clean room space will be used to continue our 1Y nanometer DRAM transition. For fiscal year 2018, we expect our capital expenditures to be in the upper end of our previously guided range of $7.5 billion, plus or minus 5%. Long term, we target capital expenditures as a percentage of revenue to be in the low 30% range. Before we move to the next section of our call, I would like to address a supplier maintenance issue, disrupting nitrogen supply to 1 of our 5 DRAM SAS which occurred on Tuesday of this week. We expect this event will impact our DRAM production output by 2% to 3% for the quarter. Our teams are working around-the-clock to recover from the situation and we expect to return to full production within the next week. Lastly, I would like to welcome Dave Zinsner as our CFO. Dave brings years of experience within the semiconductor industry, and we are happy to have him on board. Dave will now provide details on our second quarter results and third quarter outlook.

Thank you, Sanjay. I'm excited to be joining Micron at a time when the company is accelerating its focus on execution, including the delivery of more high-value solutions and the ongoing improvement of cost competitiveness. During my first few weeks at the company, I've been diving into the details of the business and operations, and I'm more convinced than ever that there's a fantastic opportunity to build an even stronger company while continuing to enhance shareholder value. For the second fiscal quarter, revenues were $7.35 billion, up 8% from the prior quarter and 58% from the prior year. The overall strength reflects a positive business environment and broad-based demand for our memory and storage solutions, particularly for cloud, enterprise and mobile markets. Non-GAAP gross margins for the quarter were 58.4%, up 300 basis points from the prior quarter and up from 38.5% in the prior year. Our ability to drive a richer mix of high-value products, strong execution on our cost goals and favorable market conditions contributed to the gross margin expansion. Non-GAAP operating margin was 49%, up from 46% in the prior quarter and 25% in the prior-year period. Non-GAAP operating expenses were $666 million, up approximately 9% from both the prior quarter and prior-year periods. The sequential increase is primarily attributed to expenses associated with shifting our portfolio to high-value solutions and accelerating our technology and product development. These expenses tend to fluctuate quarter-to-quarter. We're also beginning to incur the impact of solely funding the development of our fourth generation 3D NAND technology. We continue to manage operating expenses tightly and are generally only increasing operating expenses for developing and qualifying new products and technologies. Turning to performance by business unit. The Compute and Networking Business Unit grew revenue to $3.7 billion in the second quarter, up 15% from the prior quarter and 93% year-over-year. Cloud server revenues were up nearly 30% quarter-over-quarter as hyperscale customers continue to invest in data center infrastructure and broaden their service offerings. We also benefited from strong demand for graphics memory with cryptocurrency mining augmenting sales for gaming applications. Operating income increased to $2.3 billion or 63% of revenue, and reflects higher sales of our 1X nanometer DRAM solutions, along with tight supply conditions. The Mobile Business Unit achieved its highest ever revenue and operating income in the second quarter of $1.6 billion and $680 million, respectively. These results compare to $1.1 billion of revenue and $170 million of operating income for the same period last year. Our performance underscores our laser focus to meet customer's needs. The Embedded Business Unit reported revenue of $829 million in the second quarter, in line with last quarter, and up 41% year-over-year. The automotive business had a record quarter, driven by strong sales of ADAS and in-vehicle experience applications. We also saw an increase on our industrial business, driven by the growing industrial IoT markets, expanding factory automation, transportation and surveillance applications. Operating margins were 44% in the fiscal second quarter, expanding by 260 basis points compared with the first quarter. And finally, turning to the Storage Business Unit, revenue was $1.3 billion, up 20% year-over-year, supported by record revenue in SSDs. On a sequential basis, SD revenue declined by 9% with the strong growth in SSDs offset by a reduction in components revenue. The sequential revenue comparison was impacted by a mix shift within our NAND component sales, which I'll elaborate on momentarily. We're continuing to penetrate the SSD market and expand sales across each end-market, consumer, compliant -- client, enterprise and cloud. The growth is most pronounced in the enterprise and cloud SSD portion of the market. Our sales of these end markets were up nearly 30% quarter-over-quarter and more than 230% year-over-year. As we previously noted, product developments for 3D crosspoint solutions is now underway. During the second quarter, and over the next few quarters, we have incurred, and will likely to continue to incur, costs associated with production capacity underutilization in advance of volume ramp of these new 3D crosspoint products. These charges negatively impacted our SBU operating margins by approximately 500 basis points this quarter. Including these charges, second quarter operating margins were 20% compared with 29% in the fiscal first quarter and 7% in the prior-year period. Moving to performance by product line. DRAM represented 71% of total company revenue in the fiscal second quarter. DRAM revenue in the quarter was up 14% from the prior quarter and 76% year-over-year. Sequentially, shipment quantities increased in the mid-single-digit percentage range while ASPs increased in the low double-digit percentage range. DRAM non-GAAP gross margin was 66% in the second quarter, up 4 percentage points from the prior quarter and up 22 percentage points from the year-ago quarter. Revenue from trade NAND represented 25% of overall company revenue in fiscal second quarter. Trade NAND revenue on the quarter was down 3% sequentially and up 28% year-over-year. On a sequential basis, shipment quantities increased in the low double-digit percentage range, while ASPs declined in the mid-teens percentage range. The sequential ASP decline in NAND increased in part due to a meaningful last time purchase of higher price MLC NAND in the fiscal first quarter. This is the mix shift in our SBU NAND components that I have referenced earlier. Trade NAND non-GAAP gross margins were at 47% in the second quarter, down 2 percentage points from the prior quarter but up 16 percentage points from the year-ago quarter. Gross margins for both SSDs and managed NAND solutions increased quarter-over-quarter, offsetting the declines in component margin. This change in mix illustrates the importance of shifting our sales towards high-value solutions. I'd like to take a moment to update you on the impact of U.S. tax reform on Micron. The onetime impact related to the taxation of accumulated offshore earnings and cash was largely neutral for the company. The impacts of this repatriation transition tax were largely offset by our accumulated tax losses and other tax credits. For the remainder of the year, we expect our non-GAAP tax rate to remain in the low to mid-single-digit percentage since we are not yet subject to certain provisions of the new tax code. For fiscal 2019 and beyond, we expect our non-GAAP tax rate to settle in the low teens percentage range. Going forward, we'll benefit from having greater flexibility to access our worldwide cash deposits. Our non-GAAP earnings per share were $2.82, up 15% from the prior quarter and up over 200% from the prior year. As a result of our record performance, we generated $4.3 billion in cash from operations, which represented 59% of revenue. This compares to $1.8 billion in the year-ago period. Capital spending, net of third party contributions, was $2.1 billion, resulting in a very strong free cash flow adjusted for the third-party capital contributions of $2.2 billion or 30% of revenue. This compares to free cash flow of approximately $600 million in the year-ago period. As Sanjay mentioned earlier, we expect capital spending, net of third party contributions, to be at the upper end of our fiscal 2018 guided range of $7.5 billion, plus or minus 5%. As a result of the strong free cash flow, we ended the quarter with approximately $8.7 billion in cash, marketable investments and restricted cash. The face value of our debt increased approximately $200 million to $9.5 billion. A $300 million reduction in debt due to scheduled debt repayments was offset by a $500 million increase in debt at our IMFT joint venture. Since the first of our 3D crosspoint products are expected to launch in calendar 2019, we chose to defer funding for IMFT. Our partner is contractually able to make the funding on our behalf and designated a debt on IMFT's balance sheet. And that debt is then counted as part of our debt for the purpose of GAAP reporting. We still expect to be in a net cash positive position in the fourth quarter, and possibly sooner, depending on the extent and timing of any future convertible note redemptions. This net cash positive position remains a significant milestone in the ongoing strengthening of our financial foundation. We continue to evaluate additional opportunities to accelerate our deleveraging actions that will provide a high rate of return. This strong financial profile is the result of consistent execution and focus across the entire company. Now turning to the fiscal third quarter guidance. As Sanjay mentioned, we had a maintenance issue at one of our Taiwan DRAM fabs this week, which is impacting production. We expect this event to decrease our total revenue by approximately 2% in the third quarter, which we've accounted for in our guidance. Having said that, we continue to experience a strong demand environment and we, therefore, expect fiscal third quarter revenue to be in the range of $7.2 billion to $7.6 billion, and non-GAAP gross margins to be in the range of 57% to 60%. We expect to see an increase in operating expenses, again, associated with product and technology qualifications, and the funding of our fourth generation 3D NAND technology, both of which primarily impact R&D. Considering these costs, non-GAAP operating expenses are expected to be $725 million, plus or minus $25 million. We expect non-GAAP operating income to be in the range of $3.6 billion to $3.8 billion. Based on a share count of approximately 1.25 billion shares, these results should drive non-GAAP EPS of $2.83, plus or minus $0.07. I'll now turn the call over to Sanjay for some concluding remarks.

Thank you, Dave. Micron will be celebrating our 40th anniversary this fall. Innovation has always been a key cornerstone to our success, ensuring that our technologies and products quickly adapt to serve the world's growing appetite for faster data. As we look ahead, we remain focused on nurturing and fostering an accelerated pace of innovation, and I know our team is fired up and ready for the challenge. The opportunity to create a dramatic impact on the world around us is undeniable and I'm excited to be part of this team shaping that future. I'm looking forward to speaking with all of you at our analyst and investor event in May. You can expect us to provide more detail on how we see secular market trends, creating new opportunities for memory and high-performance storage, and why we believe Micron is well positioned to win. We will now open for questions.

[Operator Instructions]. Our first question comes from the line of Rajvindra Gill from Needham & Company.

I was wondering, Sanjay, if you could talk a little bit about the changes in the DRAM industry that you've seen over the past year or so. I think in the past, you had mentioned that memory is becoming a strategic differentiator for high-performance computing. I was wondering if you can maybe elaborate on what specific end markets or behavior patterns that have been changing with some of your main customers in terms of how they consume memory.

Certainly. I think we are seeing the fastest growth through our DRAM memory at large-scale, in cloud computing and hyperscale data centers. And this is where high-performance memory is absolutely becoming essential along with fast storage is becoming essential for the trend such as AI, which are really driving new business models. Whether you go from education and training tailored toward the individual levels of coaching or training to the individuals or to millions of transactions processed realtime in the financial sector, to detect fraud or going to diagnosing and treating life-threatening diseases. Bottom line is we are barely starting with AI in cloud computing and data centers. And to realize the full impact of these solutions and to truly provide this new business model and services and applications to consumers and businesses alike, more and more data needs to be processed. It needs to be realtime analytics, and that requires more fast memory and more fast storage, that means flash as well as DRAM. So we are seeing tremendous growth. And if you look at trends, when we project that 2017, about 145 gigabytes per server going to about 350 gigabytes per server by 2021. Similarly, if you look at flash storage, 1.5 terabyte average in 2017 growing to something like 6 terabyte average with each server by 2021 timeframe. So these are massive, secular demand trends in the cloud computing and hyperscale for memory as well as for flash storage. Similarly, going to mobile, I talked about in my script that at Mobile World Congress, several new phone models were introduced that leverage 4K SDR capabilities that leverage AR and VR and even in our processors that are being introduced for mobile application that actually have the AI unit built into it. So just imagine how much data-intensive applications are now being run in order to provide users smooth experience. That then requires high performance and lot of memory. And you're starting to see now 6 gigabyte phones, 6 gigabyte of DRAM in the phone. So mobile is another large driver of DRAM memory. And of course, it is also a large driver with average capacities continuing to increase for flash as well. And then autonomous driving is barely starting. I mean, industry pundits are talking about robo taxis, intercepting the whole autonomous driving trend and maybe being introduced even in 2019-2020 kind of timeframe. And autonomous driving means, as I said in my remarks, data center on wheels. It's requiring more fast memory to again make all realtime decisions providing for the safe and comfortable and efficient driving experience. So these are really massive trends and these are secular in nature and I believe will continue to drive strong demand for DRAM in the years ahead. And of course, there are other ones, continuing average capacity increases in PCs for DRAM, with more gaming features and VR features. And of course, industrial 4.0 initiatives. I mean, these are all multiple mega markets for DRAM. And we are very well positioned with our product portfolio focusing on cost competitiveness, with our technology advancements as well as what I indicated, low-power DRAM solutions which are becoming increasingly important across a multitude of these applications.

That's very helpful, Sanjay. As my follow-up, your SSD revenue was up 80% year-over-year. Can you talk a little bit about the attach rate for client SSD specifically? Last year, they were put on a temporary pause because of tightness of supply. I was wondering if you could talk a little about that now that we're about a quarter into this year?

I mean, last year's flash was severely constrained and that did somewhat slow down the cash rate of SSD in client computing as well as slow down the march towards higher capacities of SSDs in notebook computers. And attach rates for SSDs in client computing, around 35% to 40% and maybe 40% in 2018, maybe going towards 50%. Over the next few years, this is expected to continue to go toward by 2020-2021 timeframe to 85%-plus attach rate for SSD. So again, this is a large growth driver for SSDs in client computing applications. And we are focused on, of course, expanding our portfolio of SSDs. We talked about significant progress of our SSDs across-the-board in client, enterprise as well as in the consumer market. And we look at opportunities to gain further share in all of these SSD markets in the future as we continue to execute on our product roadmap.

Our next question comes from the line of Chris Danely from Citi.

This is Wayne Loeb [ph] on the line for Chris Danely. My question is when you talk about your plans for acquisitions, what would be the criteria that would make you buy something? And how does M&A fit into your plan in the context of Micron wanting to be a NAND solution provider?

We're not going to speculate on M&A matters here but we are very pleased with the portfolio of technologies and the initiatives we have with respect to continuing to advance our product solution. But of course, we do not rule out, in the future, leveraging M&A toward any growth initiatives. And of course, we'll always look for core capabilities to expand the market opportunities for Micron. And of course, we'll be focused on value in terms of any acquisition that we may entertain in the future, again, not speculating on anything at this point. And of course, always looking for ROI kind of opportunities.

As a follow-up question, can you talk about what Micron's projected cost reductions are for NAND and DRAM this year?

So we don't provide specifics on cost reductions but what I can tell you is that we are making very good progress on our technology. As we indicated, I mean, in our 1X DRAM technology, we have achieved the fastest ramp to mature yield in the history of the company. And similarly, our 64-layer technology, it has ramped to mature yields, rather well and we are continuing, of course, to do very well on our 20-nanometer DRAM technology that we use as well. So we are very pleased with our continuing progress on cost at the technology level, and continuing to focus on advancing our next-generation technology nodes and products. And of course, also very much focused nonmemory costs in our products such as SSD nonmemory costs. So making good progress and all of that is baked into our gross margin guidance that we have provided.

Our next question comes from the line of Mark Delaney from Goldman Sachs.

First question, I hope you can detail a little bit more about that nitrogen issue you mentioned. Did you have to scrap wafers or just idle production? And can you help us reconcile the comment about a little bit less DRAM output for next quarter with the now full year guidance about growing in line with the industry compared to last quarter. I think, Micron was going to grow slightly before -- definitely below, excuse me.

So this nitrogen maintenance issue has not caused trapping of vapors. It has idled or slowed down production. As we said, it's impacting 2% to 3% of our this quarter's DRAM production output. And with respect to our expectation of our output growth for calendar year 2018, that remains in line with the industry estimate of 20%, and this effect is already included in that as well.

Is it fair to assume a better 1X nanometer yield? Is that how Micron is now growing in line with the industry for the full year despite this nitrogen issue?

Yes, that is correct that our production output is expected to grow in line with the industry and that is, of course, as a result of our excellent yield on 1X nanometer node as well as the 20-nanometer node.

Okay. And then one other question for me, if I could. Sanjay, you commented about having a CapEx-to-sales target in the low 30% range. I don't want to parse words too closely, I think it was about 30% as of the last Analyst Day. But the strategy for Micron, as I understood it, had been that the company is trying to keep its net DRAM wafer starts flat and there's a lot of costs associated with getting to these new nodes because of all the extra factory space that you need and need for new clean rooms. Just given your comments about CapEx coming in toward the higher end of the range this year and the comments about that ratio, is there any change about the strategy of Micron, how it's thinking about CapEx? And really just enabling -- getting to those next nodes which are getting more expensive? Or is there a change we need to be thinking about in terms of how Micron is thinking about managing its net wafer starts in DRAM?

I think if you look at last few years and you look at Micron's revenue and you look at Micron's CapEx, you will see that Micron's CapEx, over the course of last few years, is in the low 30% range of the revenue over those last few years as well. So what we have said here today is fairly consistent with what actually has been the case at Micron over the course of the last few years. And in fact, if you look at the industry itself and you look at the revenue of the industry players and you look at the CapEx, over the course of last few years, you will see actually that, that average for the industry as well is in that same range also. So in terms of our own strategy for CapEx spend is absolutely focused on accelerating our technology transitions. So our CapEx is geared toward realizing DRAM and NAND technology transition toward more cost-effective advance technology nodes for our products. And it is not about capacity, wafer capacity production increase for us.

Our next question comes from the line of Karl Ackerman from Cowen and Company.

Dave, welcome to the team. I have two questions, please. My first question is on DRAM demand. We all know that DRAM is more inelastic than NAND but I was curious, what are some signs that you look for to assess if you are beginning to see demand destruction in DRAM demand from higher ASPs, particularly in mobile and PC environments that are more sensitive to price than hyperscale environments? I have a follow-up, please.

So can you clarify the question to me? I didn't totally get the question. I'm sorry.

Yes, I'm just curious how should we assess the potential demand destruction in DRAM demand from higher ASPs in mobile and PC environments over the next few quarters, if there were to be an issue?

So I think what we have to realize is that DRAM absolutely is essential to the experience in the business models that it enables. Whether it is the experience in mobile phones, I talked about all these experiences, AR, VR, 3D gaming, multitude of applications and users absolutely expecting seamless experience that requires -- such data-intensive applications require more DRAM. So it is essential. I mean, it's not like you can offer a model with a less DRAM in it, a high-end model with a less DRAM in it and expect that users will still have the same good experience. So DRAM capacity has really become a key enabler and essential element of mobile. And same, as I talked about earlier, for hyperscale data centers. When they look at what models that they can enable for their end customers, those are all being built on very data-intensive applications. I mean, imagine retailers, and a consumer goes into a retail store and the retailer already knows about what are the needs of their consumer. All of that requires realtime -- for retail, realtime AI applications, which means lot of data that has been processed fast, which means, again, it needs more DRAM memory. So it is actually, when you look at hyperscale data centers, it's not about the cost of DRAM anymore. I think the value that it enables to these cloud applications and hyperscalers is far in excess of any aspect of DRAM price per bit. So DRAM really has become an essential part. This is very different from any time in the past.

And the best indicator of this is that DRAM pricing is strong and DRAM demand is strong right now.

That's helpful. As my follow-up, I was hoping you could elaborate on your comments for OpEx as we think about the trajectory of spending for the next few quarters. Specifically, do you plan on reinvesting the savings you expect to achieve from Micron and Elpida coming together for the first time on 1X development? And how should we think about the timing of any planned prequalification expenses for maybe 1X DRAM or QLC 3D NAND deployment when we make assumptions for OpEx for the balance of 2018?

So let me go back to kind of the commentary and make sure it's clear. So in the second quarter, most of the increase we experienced was around qualifications of various technologies that kind of all came together, all in kind of the second quarter. And it kind of continues on into the third quarter. Those expenses kind of vary over time. And so this just happens to be kind of a couple of quarters in which that activity is pretty heavy. And so we're kind of experiencing kind of a lift in expenses, and I would expect that portion of it to kind of settle down. And when the next set of qualifications are required, it will come back up again. The other piece of the expenses really relate to our fourth-generation 3D NAND where, as we announced earlier, we're taking that on ourselves. About half of that hit us in the second quarter. We'll have the full quarter's effect in the third quarter, and that was about $20 million on a full quarter. So about $10 million left in the second quarter and $20 million lift in the third quarter.

Our next question comes from the line of Tristan Gerra from Baird.

Given the continued strong demand that you see in data center, how should we look at the initial supply/demand outlook in NAND flash for the second half of calendar '18? Should we expect the pricing to stabilize? Any commentary based on trends that you see currently continuing for the rest of the year?

So we are not going to comment on pricing trends in the industry but what I can tell you is that NAND industry does have certain aspect of its end market such as USB flash drive or imaging cards or retail. That tends to be more somewhat seasonal in the first calendar quarter. And as we go forward, that part changes. The most important thing to look at is that as more supply becomes available, it drives deeper penetration of SSDs in client devices as well as in -- gives an even stronger value proposition in enterprise and data center applications. So this is what we expect during the course of the year. And of course, average capacities of NAND in mobile phones, smartphones, continue to increase as well. And we are expanding our portfolio of multichip packages with DRAM and NAND, which is where Micron is uniquely well positioned to expand our opportunities and increase our shares with NAND flash and DRAM-based solutions in multichip packages as well as discrete NAND solutions such as the UFS that I talked about that are in the stages of qualification with our customers. So we look ahead at the year with strong demand drivers for NAND in the industry and growing opportunities for our NAND business for the remainder of the year, calendar year here. And very focused on execution of all our new product introductions and qualifications with our customers because those will ultimately drive our success toward high-value solutions as part of mix of NAND revenue.

Okay. That's useful. And then as a quick follow-up, is it fair to assume that the high double-digit growth rate in better demand for NAND in data center is something that is possible again for this calendar year?

Yes, for this calendar year, for data center, absolutely, NAND bit consumption in data center is expected to be 50%, in the range of 50% or higher. Basically, a data center is where demand will grow faster than the average of the industry. Keep in mind, same thing for client SSDs as well.

Our next question comes from the line of Mehdi Hosseini from SIG.

Sanjay, I have a follow-up. You and others in the memory industry have been discussing opportunities in moving up the stack. At the same time, some of your enterprise customers are also trying to navigate their way and move up the stack. And I'm just wondering, what's wrong with keeping the business as is? Your NAND gross margin is in the 45% to 50%; DRAM gross margin is in the 65% to 70%, and assuming that the industry is rational and we can't avoid excess capacity, why not just focus on making the most cost-effective DRAM and bit and capitalize on the margin profile? And I have a follow-up.

So let me be clear that we are very excited about the market opportunities for DRAM and NAND. All the things that we have been talking about so far over the course of the last 45 minutes here. And of course, our strategy is to continue to strengthen our competitive -- cost competitiveness as well as increase the mix of high-value solutions in our revenue. And by high-value solutions in our revenue, we mean products such as SSDs as well as managed NAND solution because we have both DRAM and NAND and that gives us a unique opportunity to provide management solutions for today's smartphones that are needing more and more of such solutions. So we are absolutely focused on leveraging our core capabilities to drive cost reductions, catch up on the DRAM cost with the rest of the competition and in the NAND, strengthen our portfolio of these high-value solutions. And I have no doubt that there is -- nobody is taking their eye off the ball and we have relentless focus on strengthening the execution engine of the company and tremendous opportunity ahead in that regards for us. It's already implied through the strong results we have demonstrated so far but there is even greater opportunity ahead of us.

In terms of costs, you recently introduced a QLC 64-layer 3D NAND SATA SSD. Is there any way you can either quantify or qualitatively discuss the cost per gigabyte that this particular product offers you? And how we should think about its ability, due to lowest cost, to penetrate and displace existing technologies?

So what we introduced recently is a 64-layer bit TLC SATA SSD. And as we have said before, QLC is certainly an exciting opportunity for Micron in the years ahead, and QLC is in the development stages. And it is not a 2018 phenomena. I mean, that is something that's more like 2019 opportunity, starting in 2019 timeframe.

But should we assume that this offers you, perhaps, I'm just going to give you a number, could it still offer a customer less than $0.20 per gigabyte of cost?

We don't get into cost discussions. And our focus, of course, is to develop QLC solutions that will be, in the future, going toward applications that are very read intensive and somewhat balanced in terms of more write applications. And of course, our goal would be to drive these -- build value in these solutions, especially going toward high-capacity aspects of the storage market, build value in these solutions so that we can be selling them in a profitable fashion and bringing strong value to our customers as well. I'm not going to get into pricing or speculate of the pricing for QLC.

Our next question comes from the line of Hans Mosesmann from Rosenblatt Securities.

Sanjay, if you can just clarify, I think somebody asked the question before but I'll just make it more concise. Are you seeing any despeccing in DRAM or NAND market? And I have a follow-up.

We're not seeing any despeccing. If anything, again, given the nature of the application, the average capacity requirements continue to go up in all end markets arena.

Okay. And then a follow-up, more of a longer-term or midterm question. After 1Y in the DRAM world, how many more node transitions or half transitions do you expect you and the industry to have before you hit a wall, if you will?

We have talked about our 1Z technology node in DRAM and our engineers are working on that. And engineers, of course, always continue to look at opportunities for further scaling. And concurrently, we are working on other advanced technologies of the future as well.

Okay. But there's no letter after 1Z at this point?

There is no letter in the alphabet after Z.

You can go to 1Zb or you can add a plus, plus or plus, plus, plus.

Thank you. We'll take you up on your suggestion.

And our final question comes from the line of Vijay Rakesh from Mizuho.

Just on the NAND side, I was wondering what percent of your NAND was SSDs? I know you mentioned it grew 80% year-on-year and seeing good traction enterprise.

We don't give that breakdown.

Got it. And I know you talked about 3D crosspoint. There's a bit drag on the margins. When do you start to see the drag go away? And I was just wondering, as you look at that ramp by the end of -- by year-end, what proportion do you think that would be of your NAND?

So 3D crosspoint products are expected to come out in -- sometime in calendar year 2019. We will have -- sometimes we'll have underloading charges. It's possible that our partner might take some of the -- of those wafers so that would obviously help on the underutilization. Of course, as we start to release those products, about late 2018, we'll start to build some of those wafers and that will help out on the underutilizations as well.

And I just want to comment on your earlier question regarding SSD. Of course, we don't provide the specifics but clearly, SSD is growing fast and it's increasingly large portion of our revenue. And very pleased with the progress that we have made in increasing the mix of SSD in our portfolio.

This does conclude the question-and-answer session. I'd like to hand the program back to management for any further remarks.

Thanks, Jonathan. As always, we appreciate your interest and support for Micron. I'd remind you that a copy of the prepared remarks as well as a webcast replay can be found on the Investor Relations section of our website later this afternoon. Thank you.

Thank you. This concludes today's Micron Technology's second quarter 2018 financial release conference call. You may now disconnect.