Micron Technology, Inc. (MU) Earnings Call Transcript & Summary

Hello. Good afternoon. Thanks, everyone, for joining. I'm Mark Newman, Senior Analyst at Bernstein covering memory as well as energy storage. And it's a great pleasure today to introduce you to Sanjay Mehrotra. He is the President and CEO of Micron Technology. He is also the Co-Founder and previous CEO of SanDisk up until the time of its acquisition from Western Digital. Sanjay, thanks very much for joining us again at our virtual SDC today. I hope everything is going well for you and your family considering the current pandemic.

I think just to open up, it will be great to give an overview of -- given all this uncertainty around the virus and the recent pandemic, talk about how you think that's impacting the memory market and Micron? And how you see the second half of this year shaping up and also next year? And if I could just say one more thing to all the investors tuning in, just a quick reminder, there is a link on the left-hand side of your screen, a Pigeonhole, where you can submit questions. I've got a screen here that's picking up questions and your votes. You can also vote on those questions, so I can see which ones are the most popular. I've also got my own list of questions I'm going to start with, but I'm looking at Pigeonhole, and we'll try to get your questions, particularly if there's a lot of votes to it. And there's also a Procensus link in there, which if you could complete at the end, which is a survey, which will be very helpful. And you'll see the results of that yourself as soon as you have finished that. And so Sanjay, if I could go back to you in terms of how things are shaping up for Micron and the memory market?

So first of all, Mark, good to be here virtually connecting with you and our listeners today. And I hope you and all the listeners are doing well and staying safe. Before I start, I just want to remind everybody that during the course of the presentation today, I will be making certain forward-looking statements, and please do refer to our SEC filings, which we make from time to time regarding the risk factors associated with our business. Going back to the question that you asked, Mark, I just want to remind everybody that pre-COVID, when we looked at the industry, the supply at the suppliers as well as with the customers was at normal levels, and we looked at pre-COVID calendar year '20 to be a year of continuing increase in demand and healthy demand and supply fundamentals. Of course, COVID impacted the industry. COVID impacted the industry demand. As we have spoken about before that we have seen some acceleration in cloud workloads, data center demand as well as surge in demand related to enterprise PC and e-learning and notebook computers as well. On the other hand, COVID environment has impacted consumer, has impacted the smartphone demand, and certainly, has impacted the automobile market as well. However, when we look at the long-term trends, these trends tend to be strong. The demand drivers related to AI, 5G, will continue to drive growth in our industry. Near term, we may have certain uncertainty that we have talked about before that COVID has introduced in the near-term environment. What I want to also highlight is that for our fiscal Q3, we just, a short while ago, released an 8-K, where we have shared that our business is tracking favorably in both DRAM and NAND, and particularly in markets that are exposed to the work-from-home and e-commerce economy. With a couple of days still left to go in completing our fiscal Q3 quarter, we are expecting our revenue to be in the range of $5.2 billion to $5.4 billion versus our previous expectation of revenue between $4.6 billion to $5.2 billion. The pricing environment has also been favorable compared to the assumptions that were baked into the guidance that we had provided before, which should improve pricing environment as well as improved revenue, should drive gross margins for fiscal Q3 in the range of 33% to 34% versus our guidance range of 29.5% to 32.5%. So these results would yield non-GAAP EPS between $0.75 to $0.80, again, well above the high end of the range we gave in March of $0.40 to $0.70. While we still need to finish the third quarter and complete a more detailed analysis of our expectations for the fourth quarter, our business continues to be stable. As a result, we are cautiously optimistic about sustaining our current business performance in fiscal fourth quarter as well. Of course, given that F Q4 is a 14-week quarter, that would suggest some sequential revenue growth in F Q4 as well. Having said that, there continues to be ongoing uncertainty from COVID-19, so we will defer providing more detailed F Q4 estimates until we release our F Q3 earnings on June 29. I would also mention, Mark, that while our business performed better than our expectations over F Q3, it is meaningfully lower than our pre-COVID expectations. And COVID-related macro will likely continue to be a negative factor in second half of calendar 2020 as well. And much will, of course, depend on the pace and rate of recovery around the globe from the lockdowns and the COVID situation and will also depend on the strong government stimulus measures continued across the world. And let me just provide some further color that as the economy fully recovers from the pandemic, hopefully, by some time in 2021, we expect healthier market conditions. On the demand side, we have certainly seen an acceleration in digitization of the economy due to COVID, which arguably should benefit long-term demand growth. Some companies have indicated cloud demand actually has pulled ahead a couple of years. And looking ahead, there are positive demand drivers with recovery in consumer, 5G and stimulus-related spending on AI and 5G. And on the supply side, certainly, there has been more discipline related to CapEx in 2020. And that should lead to less growth than pre-COVID. And as is also well-known that certain tool deliveries, et cetera, had been delayed, given the commentary that has been made by ecosystem players, and that also impacts the supply side of the industry as well. So I just wanted to provide this color to you as we launch into further questions.

That's great, Sanjay. So it sounds like F Q3, a lot better than what you've previously guided. So that's excellent news. So just unpackaging that a little bit, it sounds like -- am I right assuming that pricing is the main driver of that, that pricing has improved a bit? Or is there some mix or volume element in that as well?

So what I'll tell you is that we -- when we provided our guidance in March, of course, there were considerable concerns on the supply side on a global basis. You make recall that when we had our earnings call in March, we were just coming off of some of the lockdown and its impact in our Malaysia operations. So certainly, in March, we had concerns around supply stability itself. And therefore, we built a significant amount of conservatism in our guidance, again, related to our supply concerns. The demand, as I have shared with you, of course, has been impacted by the COVID environment. But certainly, the pricing compared to the assumptions we had baked into the guidance has been more favorable in the industry. So our -- the guidance that we just provided, the revised guidance that we have just provided for F Q3, really points to our continuing strong execution on the supply front compared to the conservatism we had built in the guidance at that time as well as healthy pricing environment in the industry.

Got it. And just clarifying, it was 33% to 34% gross margin, $0.75 to $0.80 non-GAAP EPS. And what was the new revenue guidance? I didn't quite get that.

$5.2 billion to $5.4 billion versus the guidance we had provided in March of $4.6 billion to $5.2 billion.

Got it. Excellent. Well, that's -- so that's definitely a good start to the call, Sanjay. So I appreciate that update.

I think this reflects our ongoing strong execution at Micron. And really, even in a challenging COVID environment, the team has done a great job in managing our supply chain operations and our sales teams and business units working with the customers to manage the mix of the demand in terms of positioning us well for our Q3 results that I just shared with you, in terms of the revised guidance.

Got it. Excellent. Excellent. And it does seem that clearly the overall pricing dynamic in the current environment is a lot better than expected as you hinted on. Clearly, we're seeing DRAM -- both DRAM and NAND, but particularly DRAM price has been very, very strong in the last few months. I'm sure that's a huge part of it. So actually, my questions, next. I'm going to ask a few questions around demand. So clearly, that's a bit of a concern that you touched on, the virus is impacting, particularly, the consumer applications, and particularly mobile. So for smartphone today, estimates vary from down high single digits to down over 20%. I've seen some industry analysts predicting that for the full year 2020. And I think much of this obviously depends on second half recovery. So what's your thoughts on this? What have you heard or your sales and marketing team heard from customers in the smartphone, mobile side?

So as we have shared before, I think the smartphone market certainly impacted by COVID. And on a full year basis for calendar year '20, pre-COVID expectations were for unit growth in smartphones. From 2016 to 2019, actually, there had been continuing decline in total units sold of smartphones. 2020, pre-COVID, was expected to be an increase, particularly driven by 5G. Of course, with the COVID environment and its impact on macro global economies and the consumer sentiment, the smartphone unit sales are expected, by industry analysts, to be down from 10% to 15%, and that seems reasonable overall. But as you know, China, for example, when it got hit by COVID, January and February sales there were substantially down. March onward, the consumer is back there, and sales have increased. In fact, in April, the unit sales in China for smartphones increased on a year-over-year basis as well as, of course, on a month-over-month basis. And in China, May is strong as well in terms of unit sales of smartphones. Of course, as COVID spread to other parts of the globe, it has impacted the sentiment and the smartphone sales in other parts of the world as well. But as the different countries control COVID at different rate and pace and as the economies open around the globe, we do expect that the pent-up demand for smartphones will be back. This estimation of 10% to 15% reduction in smartphone unit sales on a worldwide basis has been taken into consideration by the smartphone manufacturers. And of course, it has also been taken into consideration by Micron, as we shared in our March call that we did shift some of our supply from mobile phones toward the data center applications. I would also like to point out that 5G unit sales during calendar year 2020 are expected to be around anywhere 175 million to 200 million. I think pre-COVID estimations were somewhere around 200 million or 200 million-plus, maybe a little bit less in 2020 due to now the impact of COVID, yet the estimations are 175 million-plus. 5G is going to be a long multiyear growth cycle for smartphone. It will reignite the smartphone unit sales. And 5G also drives increased content for memory and storage. I just want to point out that even the lowest end of 5G smartphones, the $300 price point 5G smartphones have 6 gigabyte of DRAM content in them. And of course, the higher-end 5G smartphones have 12 and 16 gigabyte of DRAM. So average content in 5G smartphones will be higher for DRAM, and similarly, for NAND as well. What was more like 64 and 128 gigabyte capacity points are shifting toward 128 and 256 gigabyte capacity points in 5G phones. So when we look at longer-term trends, 5G will be a significant growth driver for multiple years in terms of increased unit sales as well as increased content for both DRAM and NAND. Of course, in the near term, while we have this COVID environment, of course, there is element of uncertainty. But the point is that, as we have seen in China, in the rest of the world, too, as the world starts opening up, the pent-up demand for smartphones will start coming back. I can't exactly project when, but we do know that COVID that has impacted smartphone unit sales, post-COVID, that will -- the sales growth will resume for smartphones.

Yes. Also, I'd just like to point out, your Head of the Mobile Business Unit just did a great presentation, which I believe is on your website about this specific topic about 5G content and the lift-up in both DRAM and NAND. So if anyone wants any more details on that, you can find that on the Micron website. So Sanjay, a follow-up here on mobile. What are you hearing about inventory levels at customers, particularly mobile OEMs? That's one area we've heard is a bit of a concern. Is it something we should be concerned about? What have you heard or have seen from your mobile customers?

So I think as I pointed out that when the economy has experienced a sharp drop due to COVID, of course, the sell-through of smartphones dropped as well for our customers. And for example, that happened in the January-February time frame for our customers in China and in the rest of the world, again, depending upon different time frames. But the point is that those customers have adjusted to their demand expectations for the full year basis. We, too, have made those adjustments. And I would like to remind you that in few months, there will be launches of new phones. Along with 5G, new phone launches, as the world recovers from COVID environment, will spur up the demand growth in COVID. So other than perhaps one large customer in China, given U.S.-China trade tensions that may be carrying some extra levels of inventory potentially, in general, I think, the point is that customer ecosystem has expectations of lower demand that I mentioned earlier for the full year 2020 basis and has made adjustments in that regard. We, too, have made adjustments in our own supply, shifting it from mobile to the data center side. Calendar Q2 from global economic GDP output point of view, is expected to be the low point, with economies starting to recover from C Q3 onward. And as the economies recover, we do believe that smartphone sales will resume their growth as well.

But is it true, though, that the adjustment down in purchasing mobile OEMs is not as much as the adjustment down in sales? So meaning that they've been buying more than what they've been using in the first half. Should we be thinking about it like that?

Yes. I think it's important to understand that they don't make sharp changes in their buying patterns because they're looking at longer-term expectations as well, and that's why I'm highlighting that they have adjusted to their overall demand expectations for the year being lower. And customers may carry inventory for a short period of time, but they have made their adjustments, and we have made our adjustments accordingly as well in terms of our supply.

Great. Moving on to data center demand. This is another hot topic. So we've seen a lot of strength in this area. How much of this do you think is catch-up from the underspend last year versus how much of the data center strength that we've seen recently related to the virus COVID-19 and this work-from-home dynamic playing out? Any thoughts on how you think about those 2 drivers of this great strength we're seeing at the moment?

So within data center, particularly within cloud, pre-COVID, the customer inventory levels had returned to normal, and the CapEx spend from data center customers, cloud customers was expected to be strong through calendar year 2020 time frame based on some of the comments made by them. And then ultimately, the trend of overall adoption of AI, et cetera, driving greater cloud adoption and greater demand from the data center point of view. So the COVID environment, as we mentioned, has actually resulted in acceleration of the digital economy, and that has seen a surge in workloads related to work-from-home economy on the data center side, particularly on the cloud side. So it's not just about shift of workloads from on-premise to cloud. It is -- cloud is now also a growth engine for services that are AI driven, and these are long-term trends. And we are at the sweet spot of the cloud growth cycle here, and it's a long-term, next 10-year kind of growth cycle for cloud, and for certainly, memory and storage as well because AI-driven applications require more memory, they require more storage. New CPU platforms have more cores. They can work with higher-density memory. So what you are seeing is that the modules for DRAMs are actually shifting to higher average capacities. There are more channels, more cores, more attached possible for memory, applications that are AI-driven are requiring more memory. So memory and storage is becoming more relevant than ever. DRAM content as well as NAND content continues to grow. And certainly, during the COVID environment, we have seen strength in data center-related demand, and we expect that to continue.

So specifically on that, and this is actually a question that's got the top votes from clients. It was on my list as well, but client -- investors have also a lot of interest in this, which is how long do you expect the data center strength to remain? Obviously, calendar Q1, extremely strong; calendar Q2, from everything we've heard, extremely strong. How long should we expect that? How long are you expecting that to remain?

So as I mentioned, I think, we are at the sweet spot of the growth cycle in data center. And data center is a long-term growth driver for the industry. Of course, there can be variations from one quarter to the next quarter. But important thing is that compared to 2, 3 years ago, then it was a very small part of the overall memory and storage industry. Today, it has become a larger part and continuing long-term growth trends in data center, in cloud, driven by AI machine learning as well as CPU architectures that I referred to earlier that bode well for more DRAM attach. And data, all with intelligent devices at the edge and greater cloud connectivity, 5G is going to further accelerate those capabilities, all of that really enables more real-time data analytics and more intelligence from the data, and that again bodes well for long-term memory and storage trends. So as I mentioned, we have seen strength in the data center, and we expect this to be a long-term growth driver for the industry.

So no comments on the dates, like how many more quarters you expect? Because there's a bit of lumpiness. I understand your comment, which is well taken that there's an underlying structural growth to this demand story in data centers. Absolutely, that's well taken and that continues to play out longer term. But then the question is more specifically about the cycle because the hyperscalers or the cloud service providers, these companies tend to be very lumpy in their purchases. They tend to have 5 quarters of huge strength and 5 quarters of weakness. We saw 2019 was very, very weak. And so far, this year, extremely strong. So that was -- I mean it's -- even just trying to get to that, how many more quarters do you expect? It's been maybe 2, maybe 3 quarters. I think the strength really started Q4 last year. So it's not -- it doesn't look like it's near the end. But I think this is a critical kind of thing that investors are trying to figure out, how many more quarters have we got of this data center strength?

So look, I mean, I will definitely tell you that, as we have shared before, that given the COVID environment, it certainly does bring lower visibility and does bring some element of uncertainty in all industries and -- in terms of overall macroeconomic environment. So we are operating under that kind of environment. But compared to the past that you're referring to, I would like to point out that data center customers, and let me refer to the cloud customers here, cloud customers definitely have matured in terms of their buying patterns and understanding of the industry as well. And from our dialogue, we work closely with this these customers. We are a major supplier to these customers. We work closely to understand what their demand and product requirements are, and Micron is -- they are valued partners to us, the cloud customers, and we are valued partners to them as well. In general, I would like to say that the kind of volatility that was experienced in 2018-2019 kind of time frame, our -- those cloud customers also would not want to have that kind of volatility because it is better for them as well as for us. So I do believe that their overall forecasting method, their buying patterns, their inventory strategy has matured further. And I do believe that 2018-2019 scenario versus now is somewhat different. Having said that, I cannot claim that I have perfect visibility into this. The COVID situation certainly complicates matters further. But what we continue to see is that overall applications and the trend of applications driven by AI and cloud environment is a long-term secular trend here in -- as a demand driver. Is there lumpiness from one quarter to the next for one cloud player versus others? Yes, those kind of dynamics will exist, but in aggregate, the trend is one of healthy demand growth for the industry on a long-term basis.

Yes. Got it. And your point is interesting that as they mature, they should figure out their demand a bit better, maybe in the previous peak, they weren't quite sure. So I think that's...

And we work closely with them too, and it has become a bigger part of it. I just want to make sure that mature doesn't get confused with growth rates maturing. Growth rates are continuing. It is becoming big. So as a big operation, I think, us working closely, we are in a better position to manage this business.

So -- yes, so that leads me to the next question really, which is -- you partially answered it, which is considering this rising dominance of these hyperscale data center customers, they're going to continue to get bigger and bigger portion. I think data center overall is about 1/3 of DRAM consumption at the moment and hyperscale is, I believe, a little bit over half of that and growing. So these customers tend to be very opaque, very lumpy. So further thoughts on how Micron and the memory industry, more broadly, can deal with this.

So again, I think, it's important that we work closely with them, and this is what we are absolutely doing. And this is with respect to the cloud as well as with respect to the enterprise data center customers. We fully understand what their requirements are, and we continue to advance our product portfolio to meet their requirements. And as I mentioned that data center customers, particularly on the cloud side, as that becomes a larger part of the overall industry and becomes even bigger portion of their overall operations. In terms of their forecasting, in terms of their inventory strategies, I believe, they, too, are interested in an environment of less volatility, and therefore, those -- their forecasting, inventory strategy, buying patterns are maturing as well. So we remain extremely focused on working closely with them. Again, on the enterprise side, related to COVID, perhaps there could be some weaknesses. But again, those are the trends that we continue to monitor working closely with them as well.

Right. Last question on demand, which came in through Pigeonhole, it was around game consoles. So can you provide any comment on whether you expect -- whether expected game console releases will be material to results in the second half of this year?

So new game consoles definitely have higher content, certainly the higher DRAM opportunity as well as they are shifting from HDD to SSDs, which bodes well for the overall industry as another growth driver for the NAND industry. So yes, I mean, new game console is -- again, I think, the point is that there is need for more memory, more storage in the new applications. In the new game consoles, our GDDR6 memory is well positioned. It's really high performance, and some of the leading players actually have pointed out to the benefit of GDDR6 kind of memory in terms of the performance capabilities that it provides in those game consoles. So Micron is well positioned with its product portfolio to continue to address the growth opportunities in this particular part of the market for the future. And we, of course, continue to advance our product portfolio, not only for the gaming or game console part of the market, but across the board, both in NAND and flash.

Great. So moving on to supply and CapEx. Again, given all this uncertainty we've talked around -- talked about around demand, how are you thinking about your capacity and CapEx plans for the rest of the year and into next year?

So first, let me remind you that in fiscal year '20, our CapEx is meaningfully reduced from our fiscal year '19 CapEx. Fiscal year '19 was $9.1 billion. And for fiscal year '20, we have said our CapEx is $7 billion to $8 billion. And that includes a significant portion of CapEx going toward building of the clean room shelves to implement the technology transition on the installed wafer capacity base. Our CapEx is not directed toward any increases in wafer capacity. And in fact, in fiscal year '20, our CapEx is down by more than 40% for equipment spend. So our CapEx strategy is to focus on supply growth CAGR to be in line with demand CAGR expectations. And we remain extremely focused on making sure that we are managing our CapEx. We are managing our supply growth carefully to make sure that our supply and demand are in balance, and we are extremely focused on ROI as well. So pre-COVID, we were expecting that our fiscal year '21 CapEx would be higher than fiscal year '20. Again, note that fiscal year '20 was meaningfully lower than fiscal year '19. But as we encountered the COVID situation, we have made adjustments to our plans. And our CapEx expectations at this point are lower -- for fiscal year '21 are lower than our CapEx expectations pre-COVID. So you can see that we are continuing to review this, continuing to make adjustment. We'll continue to monitor this over the course of next few months as well. And of course, as we get closer to talking about fiscal year '21, we'll provide you further specifics related to CapEx. And I just wanted to point out that with respect to managing supply growth, longer-term supply growth considerations, CapEx is important. In the near term, it's really about utilization that we can adjust at our factories as well to make sure that our supply and demand are aligned. For example, Manish Bhatia, our Executive Vice President, in one of the webcasts recently talked about how we adjusted some of the automotive production. It has a small impact on our supply growth, but we made adjustments to our wafer production related to automotive supply, given the weakness in the automotive market in the current COVID environment. So we have knobs that we can turn that go beyond just CapEx in terms of impacting near-term supply growth, which ties to the utilization that we can impact.

So is it -- it's possible then that you could see a reduction in capacity if demand remains very, very weak post this virus?

I'm just pointing out that, that is always an opportunity, and Micron -- an opportunity to make sure that our supply growth and overall supply is aligning with demand. So just since you had asked about CapEx, I just wanted to point out that managing supply, another knob, is, of course, underutilization, which we did use in the last fiscal year time frame. On the DRAM side, we had turned down some of the utilization until December of last year.

And so to clarify on that, the utilization cut you'd made slight tweak, right, 5% or so, I believe, last year. Is that now back to 100%? Or what is the latest on that?

Yes. That has been back to 100%, except what I just noted related to some automotive demand, we have cut back some of the utilization. It is a small impact on the supply growth because automotive bits are -- while revenue on automotive side on a per bit basis is very different, but on a bit basis, it's a small percentage of our total mix.

And on the CapEx, just to clarify, you're saying that the number is likely lower than what you're originally thinking, but you're not giving like any like real number. Your -- but pre-COVID-19, the FY '21 was supposed to be meaningfully higher than '20. Now you're saying maybe not meaningfully higher, but you're not saying if it will be higher or not. So can you just clarify on that?

We're not providing the specific numbers at this point. We typically provide those numbers at the end of the fiscal year when we discuss full year results, and we'll provide you that consideration at that time.

Got it. Got it. Okay. Sure. So moving over to technology now. I think we'll start off with NAND flash because there's actually a question that some of the audience was also asking, which is, first of all, what's the progress on replacement gate node? And interestingly, we thought that floating gate seemed to be running out of any potential future migration as evidenced by Micron moving away from floating gate. And no other company has been putting floating gate. But interestingly, Intel recently had a presentation, I believe, in Korea, where they announced their road map on floating gate well into the 100-plus layers. So just curious what you think about that? And thoughts on why Intel keeps pushing on floating gates?

So with respect to progress on our own replacement gate technology, we are making good progress. In fact, in March, we had mentioned that we will be starting production of our replacement gate technology in fiscal Q3. And I'm pleased to report you that we have started production of replacement gate in our fiscal Q3 and that will translate into revenue opportunities starting in fiscal Q4. And by the end of this calendar year, our replacement gate technology will represent a meaningful part of our production. Of course, we'll continue to ramp through the course of next calendar year as well. As we have pointed out before, the first generation of our replacement gate technology has limited deployment across our product portfolio because of its cost competitiveness capabilities. And we intentionally chose the second generation to be applied broadly. Second generation of replacement gate is what will provide us stronger cost reduction capabilities going forward as well. So just as a reminder that our floating gate NAND inventory, we continue to carry, and that inventory will be used to meet our shipment and demand requirements going forward as well. But overall, replacement gate, I'm pleased with our technology readiness as well as technology production capabilities. I'll point out that we have a good, strong replacement gate road map, and we chose to switch from floating gate to replacement gate because our strategy is to be able to address well-diversified, multiple end markets. And replacement gate is a better-suited technology for multiple end applications with respect to performance, with respect to the reliability considerations, replacement gate is better suited to enable us to address multiple markets. And that's why we chose to switch like bulk of the rest of the industry to go to replacement gate technology, and we are doing well. And we believe we absolutely made the right decision. And I really cannot comment on what Intel's plans, et cetera, are. We have a good road map lined up for our replacement gate technology as well as our focus on products and building high-value solutions with this replacement gate technology.

Got it. Okay. And then on DRAM, what's the latest progress on your 1z nanometer? And thoughts on what's next and timing for latest thoughts on EUV?

So with 1z DRAM, we continue to execute both on the compute side as well as on the mobile side. And again, I'm proud to say that as part of our high-value solutions strategy, 1z nanometer multi-chip packages that have both NAND and DRAM in them, we were the first in the industry to begin to ship those with 1z nanometer DRAM. We are shipping today 1z nanometer for mobile, and we are shipping 1z nanometer for compute as well. And we continue to make good progress on our next node, which is the 1-alpha node, and we have talked about previously that after 1-alpha will come the 1-beta and the 1-gamma nodes in DRAM. And we see a technology road map that will give us cost reduction capabilities, well aligned with overall industry capabilities that do not require EUV through 1-gamma node. Having said that, we, of course, are continuing to evaluate EUV, continuing to invest in R&D of EUV technology as well. And in the future, when we see EUV becoming the cost-effective technology node for DRAM production, we, of course, would consider switching to that. But through 1-gamma, we see our multi-patterning techniques in place of EUV serving us well with respect to our product road map as well as cost reduction capabilities.

Got it. But any like thoughts on why -- actually, we've talked about this before. So sorry to repeat this because this is a common question, but I'll have to ask it again because it's pretty pertinent here. Why Micron is so much later on EUV versus Samsung? So Samsung has a bit of EUV on their 1z nanometer. And it's going to be -- and they've admitted that it's not -- and there's no cost benefit in using it in 1z, but the 1a, which is the one that comes after that, they reckon there is a cost benefit to using EUV, whereas you're commenting that 1-alpha, 1-beta, 1-gamma -- so it's even through 1-gamma. So after that, so 3 nodes later. So just any reason why that would be?

I think you've got to realize that Micron has had a history of tremendous innovation in multi-patterning capabilities. We have always led the industry in this regard. And that's continuing to serve us well. And we have assessed that with our technology innovation and our capabilities, we will be well served continuing with multi-patterning approaches and advancing them for the needs of the future technology node. And we see our cost competitiveness and our cost effectiveness well served by multi-patterning approach rather than EUV until after the 1 gamma node. But as I said, we do continue to evaluate our road map, and we'll deploy EUV as and when it makes sense from cost competitiveness and cost reduction point of view. Right now, we feel very good about the road map that we have.

Got it. There's one question here. There's a few questions from the audience, but one that's got the most votes here, so I'll have to ask you is, the supply chain reports have indicated potential for large inventory levels at large cloud vendors. Are you getting any indications of that? And what is your visibility into their demand levels for second half '20? We talked a bit about the visibility of demand over second half '20, but we didn't talk about the first part of the question, which is inventory levels becoming greater at the cloud vendors.

As I said before, I think the cloud demand has been strong during the first half of this calendar year. And we see cloud demand continuing to be healthy as well. Again, a lot of the workloads related to work-from-home economy, requiring greater demand from cloud infrastructure, and it has put some strains for that, and that has worked well for memory and storage requirements as well. Just like we have pointed out before that we have ourselves carried extra inventory of some raw materials, et cetera, it is possible that some of the customers may carry extra inventory of memory and storage solutions as well. But I would like to point out that nothing, at least, to our knowledge, that is that unusual. We may not have perfect visibility ever into our customers' inventories, and COVID environment may make that uncertainty more challenging. However, nothing that unusual that we see here with respect to any inventory considerations. And again, I think, the important thing is that cloud is still in the sweet spot of its growth cycle. And just like even the 2018-2019 inventory that was built up, it was worked down to normal levels pre-COVID at a reasonably rapid pace and along the lines of the projections that we had made regarding how that inventory will get normalized. Same thing here that if anything, the demand trends for cloud are only continuing to increase. Having said all of this, I mean, we all must recognize that COVID has brought about a certain level of uncertainty in the markets and does create challenges with respect to visibility as well, but we continue to work closely with our customers. And I do see that once we get past this pandemic, I think, the growth trends in cloud will continue as well as growth trends in markets, such as smartphone and other consumer and automotive markets will resume post the pandemic. So we are focused on making sure we work closely with customers and understanding their requirements and bringing the right products to them to address our well-diversified markets, which vary from cloud to PC to smartphone to automotive and industrial and consumer markets.

Got it. So we have -- we're running out of time. If anyone has any final quick votes on the last question, do that now because we're running out. There's one other question I want to ask you, Sanjay, which is a question we're asking to all CEOs at this conference, which is, as you think through and beyond this current pandemic, how do you expect your priorities to shift, especially as they relate to cutting costs or increasing levels of investment?

So for us, our focus, of course, is on product leadership. And we have made investments over the course of last 3 years, and we have produced strong results in terms of continuing to bring leadership products in the marketplace. I mentioned about 1z DRAM technology being a leader in the industry in terms of its launch and its shipment. I talked about high-value solutions with NAND, now increasing to 70% of our revenue mix versus in 2018, a projection of 80% goal for 2021 that we had mentioned. We are well on our way to there. So our -- as we navigate through the current challenges of the COVID pandemic, our focus is absolutely making sure that we are making the right investments for the long-term strength of the company on the side of product leadership, high-value solutions and making sure that we are managing our investments in OpEx as well as CapEx to provide strong ROI. And with respect to OpEx, we want to make sure we are investing in R&D related to the long-term technology capabilities, both -- I mean in all 3 technologies in NAND, DRAM as well as in 3D XPoint, which we see as a technology, which will give us a unique portfolio and opportunities to grow in the future years. And on the CapEx side, our focus is on supply to be aligned -- with supply growth for us to be aligned with our expectations on the demand growth. So these are really the focus areas for us as we look at emerging stronger out of this pandemic.

So one, just final question here, which got quite a few votes. What are your DRAM and NAND supply expectations from China? And I believe we're talking here about Chinese competitors over the medium to long term?

I think if you look at Chinese competitors, as I've said before, barriers to entry are pretty high because it's not just about producing the wafers. It's really having the products with the right quality, with the right cost structure, with the right ramp-up in production capabilities and having the right performance and right power to meet the customers' expectations as well. However, we do not ignore any competition. We don't ignore any competition coming from China as well, even though the barriers to entry are high. We always -- when we look at industry supply bit growth projections, we do assign a certain percentage of supply bit growth coming from these new potential entrants from China. However, what we have seen is that year-over-year, those actually do continue to shift out even though we keep assigning a certain level of supply growth projection capability to them. And that just points to the difficulties of this technology and making it meaningful for any broad adoption in the end market applications. And I think it's important that technology nodes are leading-edge technology nodes. I think when you hear about on the DRAM side from China, I think those are lagging edge nodes. So I don't think I worry too much about, frankly, on that front. On the NAND side, Chinese competitor has talked about 128 layer of nodes. But again, with NAND, as you know, it's not just about having a technology node, you need to have controller and the firmware capability and all the reliability and all the applications to go with it. So we remain focused on the strong technology and product road map we have. And as we continue to execute on that, we believe we will do fine, and Micron has competed over 4 decades with scores of companies. And as you can see, that we have tremendous technology, product and manufacturing capability and a strong customer ecosystem to continue to address our market opportunities ahead. That's what we are focused on.

Great. Thank you very much, Sanjay. As we're out of time, I'd just like to direct investors on the call again to our Procensus poll. If you're viewing the video, it should be a link on the left of the screen. If you click on that, there's just a few questions, easy, multiple choice questions, and you will see the results of that as soon as you finish. Thanks again, Sanjay, and your team for joining us at this conference. Hope you join us again next year. We're also hopefully upping our numbers again, next year would be great as well. Thanks very much.

Thank you, Mark. Thank you, everybody.