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

Hi. Good afternoon, everyone. This is Mark Newman from Bernstein. And next session up, we've got David Zinsner, sorry, CFO of Micron; and also with Farhan Ahmad, who's the Head of IR for Micron Technology. So I think you all know who Micron is. We're going to have a brief introduction from Dave on how business is going. And then we'll go straight to Q&A. Reminder, this is interactive Q&A. We've got the Pigeonhole link. If you don't know where that is, you should reach out to our corporate access team. The Pigeonhole link will let you add questions. I'm monitoring that right now, and I will try to get to your questions if you put any in. And you can also vote on any questions that are already submitted. And so with that, Dave, hope you're doing well. And would you like to give a brief -- kind of few opening comments on how business is going in the last few months and maybe outlook.

Okay. Great. Thanks, by the way, Mark, for having us. Appreciate the time. Just always want to start with the preamble here that I might make forward-looking statements. And they also may have risk factors. So anybody that wants to look at those risk factors, they're filed in the most recent 10-K. Also, I might make comments around non-GAAP numbers. So there's a reconciliation of GAAP to non-GAAP on the Investor Relations website. So thanks, everyone, for joining. I hope everybody is staying safe and healthy. Just a quick update on how the quarter is going. So we are a little bit more than 10 weeks through the quarter. We still have a little less than 3 weeks to go. And obviously, those 3 weeks tend to be pretty important weeks for Micron. That being said, I think through the first 10-plus weeks of the quarter, we feel really good about how the business is going. It's tracking really well. Markets are tracking well. We seem to be executing particularly well, operationally. So very pleased with how the quarter is shaping up. I think as we talked about on the call also, we're incredibly enthusiastic about how calendar '21 is likely to shape up for us. A lot of tailwinds to the business. Mobile should be really strong, given 5G. The cloud business should -- which is healthy today, should actually see good growth next year. Graphics is doing well for the quarter. We would expect graphics to do well through the year. Automotive is already starting to see a recovery. And I think any of the markets that have seen an impact from COVID, like automotive, like consumer, should start to see recovery, in some cases, already are seeing recovery, should see recovery in '21. So we're pleased, obviously, really pleased with how healthy things are for the quarter, and incredibly enthusiastic about how next year should develop in particular since it looks like a couple of vaccines will be available. We should start to see good GDP growth next year, recovering off of this year. That should also be helpful. We have had an impact from our industrial business. That should obviously be in a much better situation next year. So yes, so again, still a couple of weeks to go. So can't claim victory just yet, but pretty satisfied with how things have gone so far.

Great. Well, the good news is you've just broken through $60 today. Now currently trading at $61 as I see. So it's nearly, I think, investors starting to kind of get behind it. Yes, it's been -- for the stock, it's been a fairly rough year, I think, compared to other tech stocks, right? And the reason for that has been some of the worries around demands and pricing did great for a few quarters and then has, unfortunately, weakened second half of this calendar year. Clearly, we're all hoping for this recovery in the demand, which will lead to stabilization recovery and prices. I know that's what we are forecasting.

But I'd like to get from you comments on how you're seeing that demand picture play out. Maybe if we could first talk about data centers, that's been one of the areas that many companies have pointed out, that has been the area of weakness from some -- a bit of an inventory build-out, it sounds like, from some of the cloud guys and on-premise's data centers also been particularly weak. Are you seeing that come back? And are you seeing it come back in this current quarter? If not, what are you thinking about that in terms of when that recovers? Yes, some of the...

So obviously, data center is somewhat a tale of 2 businesses. On the cloud side, as COVID started to hit and as people started to move towards work from home and e-learning and so forth, cloud had a really big pickup and benefited a lot from that shift. And obviously, when you have a good couple of quarters, that is going to tend to kind of settle out. But I would call the cloud business pretty healthy and continues to be healthy. And like I said, I think calendar '21 is shaping up for a really good year for cloud. On the enterprise side, obviously, that's been a bit more impacted from COVID as, certainly, there's been less investment in that space and certain customer -- certain end customers have been more impacted by COVID and, of course, are investing less in that business. It might take a while for that, obviously, to recover. But the likelihood is that it will, as people move more back to working at the location of their business. And on top of that if some of these businesses that have been impacted that invest a lot on the data center side for enterprise. The likelihood is that will -- that -- which is a headwind today will turn into a tailwind in '21.

So this current quarter compared to the previous quarter, has the cloud -- the cloud side improved Q-on-Q?

So as far as quarter-over-quarter, we did expect it -- particularly because the fourth quarter was -- had an extra week in it, we did expect that not to be -- not to incrementally increase. But I would still characterize it as pretty healthy, that business.

Okay. And -- but still not quite at the level of where it was earlier on in the calendar year, right?

That's fair, I think.

So when do you think it's going to get back to that level? Are you confident it's going to get back to that level, first half of calendar year '21?

Yes. I mean it's hard to say exactly the timing, but I feel pretty good about '21 in general for cloud. I think that was -- this -- if anything, the cloud business does tend to have a little bit of lumpiness to it sometimes. But the market is very healthy. All of the cloud customers that we've heard from have talked fairly positively about their CapEx investments, both for this year and next year. So we would expect that to be a very healthy market for us.

I mean I think, one of the issues, I think, analyzing this cloud market, or I'd say, data center market in general, is how big that cloud segment is, how fast it's growing and through the cycle, it's growing very, very fast, right? That sounds pretty much liable. But it tends to be quite lumpy and quite difficult to predict. That's the problem with it because it's so opaque. And one of the particular things is we don't really have a fairly accurate assessment of the inventory level. We hear numbers from expert network checks, a few numbers here and there, while, we don't really have an active assessment. I mean how does Micron get comfortable with that, like planning around? It's a bit different than when you're looking at the PC or mobile market, where you can see how many phones and how many PCs are being made and sold. And you have a rough idea of the inventory. Look, the cloud side, considering that's just far more opaque.

Yes. I mean it's a little, obviously, a little more challenging. But we do talk to those customers and get a good sense for what their annual consumption is going to look like. And we do cross-reference that with the expectations around capital spending, both from what the customers are saying, but also what third parties are suggesting might be the investments necessary. I mean one thing you can rely on is that content will increase over time in that space, for sure. And the -- their business profile is going to show growth. There might be, quarter-to-quarter, perturbations in that business because of timing of rollouts of when data centers are going to get rolled out. But I don't -- I think in general, it's a very positive enough into the right business. I think a lot -- maybe a lot is -- our concern around that has been around, there was a pretty significant inventory build through the course of 2017 and 2018. And I think there's just general concerns around, "hey, could that happen again." And I think, first of all, our read on the inventory levels is that it's a much, much different situation today. There has not been that kind of build of inventory. And I think that's partially -- we know that just from our discussions with the customers and so forth. But I also think that was perhaps earlier on in the maturity of purchasing decisions. And I think, over time, both from the supplier and the customer perspective, things have gotten more mature in that market. And I would expect it to behave much differently than it did back in that 2017, 2018 time frame.

Yes and it makes sense. Are they giving you -- are they giving, well, your sales team or management team, any data around the inventory level? And do you believe those numbers they're sharing?

Yes. I mean, of course, we get inputs from the customers, but we have our own kind of models on the inventory levels of those customers, based on our own kind of understanding of the market. And we have a whole market intelligence team that spends a lot of time kind of analyzing that. But at the end of the day, the best indication is how demand looks like, and that gives you a really good sense of where things are most likely. And I think, right now, based on how we see things unfolding for the quarter, looks like it's in fairly good shape.

Right, right. And demand, right? This isn't easy. It's now a problem for data centers. So then same question for mobile, really. I mean I think the obvious strength there is this migration to 5G, which Micron has talked quite a bit about. So maybe touch on that, the content growth from 5G is obviously, pretty significant. But the other area, of course, is the units. I think this month -- this year, there was quite a bit of weakness in the first half of the year. How are you seeing that playing out going into next year?

Yes. So on a unit per sell hit, the unit perspective, before I talk about 5G content. On the end units for mobile phones, we feel pretty good about next year. We expect growth to probably be double-digit growth year-over-year in the mobile space at the end unit level. Of course, we should do much better than that because of the first question you asked, which was the content. If you look at a 4G phone, it was running somewhere in the 3 to 4 gigs of DRAM. As we migrate to 5G, if you want or if a mobile customer wants a 5G experience for their end customer, our best read on that is it takes at least 6 gigs of DRAM. We've seen SKUs at 8. We've seen them at 12. We've even seen one at 16. So this is a pretty content-rich story from a DRAM perspective in mobile as we transition to 5G. 5G has done, I think, despite COVID, much, much better than I think I had ever anticipated. I mean, I think early on, people were -- once COVID hit, were suggesting that maybe 5G handset units would be somewhere like $150 million, I think, was one of the numbers I heard. I think it's going to be well north of $200 million for calendar '20. It's going to probably be more than double that in calendar '21. And given our content, increase is there, that's going to be a great story. And another thing we've heard sometimes is that, "hey, well, with the faster download speeds, is that actually a headwind on the NAND front." But we just haven't seen that. We have seen content increases on the NAND side as well with 5G. And I think it makes sense. You can download faster. You're going to download things faster and more. And then you're going to want to store those things, particularly in periods of time where you lose connectivity. And then you have a lot of whatever you downloaded on your device. And so if anything, we think that the content increases on the NAND front in mobile as well. So this should be a huge, huge tailwind for us next year and for multiple years, as 5G becomes a bigger and bigger percentage of the overall mobile units.

Yes, it looks like the -- I mean, the whole features of faster downloads being negative to NAND, it theoretically makes sense. But maybe we're not there yet. Maybe that's something to worry about later on because yes, if you can download more you're more -- if you can download more faster, you are likely to want to keep it as a local backup because when you're on a plane or in the tunnel, when you're in the train or something, the 5G connection doesn't work so well. So we're likely still going to want files locally saved.

Correct.

And so maybe we move on a bit to the supply side. So what are you seeing there on the supply side, maybe you can comment a bit about what Micron is doing on supply. I think you've been pretty consistent that you're keeping capacity flat for both DRAM and NAND, I believe. Maybe you can just comment on that. And then move on to their competitors. Any worries about what competitors are doing on the supply side, particularly NAND Flash?

Right. So we talked about in our fiscal '20 year that we pretty dramatically brought CapEx down, both in DRAM and NAND pack. On the equipment side, we brought both of them down more than 40%, year-over-year. And that was in an effort to really carefully -- partially to carefully manage the bid output. The other reason was we were investing a relatively limited amount in our first-generation NAND and so that also helped. I think as we go into fiscal '21, if you look at DRAM, we'll be roughly similar levels in terms of front-end CapEx spend, with fiscal '20. So again, trying to be very prudent around how much we invest in that business to make sure we continue to maintain good balance. If we look at the growth rate for DRAM next year from a demand perspective, we think the likelihood is that the growth is somewhere in the 20% range. We're going to be well below that. The industry by the way, we think, will be below that as well, but we will definitely be well below that just to carefully manage that supply. On the NAND front, as I mentioned, we had brought down also, in the end, CapEx investment quite a bit in '20. We will increase that CapEx in '21 but that's more a function of transitioning from first generation replacement gate to second generation replacement gate. We just announced 176 layer NAND last week, which is our second generation replacement gate. So we'll put a full investment in that this year. So that will certainly increase our bid output. But we'd expect that to be below the industry and below demand. So again, we're maintaining pretty healthy balance there. And managing things tightly. As far as the industry goes, I mentioned, I think the industry will undersupply demand in '21 on the DRAM front. On the NAND front, it's -- that's one that we got to monitor and be cautious about. Early on, there looks like there's some healthy investment there. And our hope and expectation is that as we progress through next year that, that we'll still be in a good -- or that we'll be in a good supply/demand balance. But we're -- that's obviously one we're more cautious about.

Yes. I agree. And I think the DRAM side looks like, on our numbers, it looks like it's significant on the supply next year on DRAM. NAND flash, tight but not as obviously tight compared to DRAM. And so I think that the NAND issue is that capacity is being added already from like the Samsung right now. Whereas DRAM, the overall capacity is still flattish to slightly down to the industry. So the DRAM situation looks great. But I think the NAND side is where there are worries, given a bit of capacity coming online from Samsung and we don't know how much exactly, like how much and how fast they're going to add. They've started to add some. And I don't think even they've decided yet how much they're going to add, which is, we know is a bit of a concern. And also what NPC going to do is another question as well.

Right. Well, I'm not going to pontificate about what the other players in the market might or might not do. I would just say, we are definitely maintaining good discipline in the end. And I would say the other thing about NAND for us is this 176 layer NAND has a great cost structure. It will be really good for us from a cost perspective next year. That, I think, makes us feel good about our position in NAND. Also QLC, which we said is now moving the needle for us, also helps us from a cost perspective. So we feel really good about the cost structure. We've been moving our mix to higher-value products, which have better margin dynamics to them as well. That also, I think, helps us from a profitability perspective. What the other players may or may not do is obviously difficult for me to comment on or call on. As it relates to emerging competitors, we remain paranoid about emerging Chinese competitors, no question about that. As we remain paranoid about all our potential competitors or competitors. But I would just say this stuff is really hard to do. It's -- as we -- and we have proven that sometimes. One thing to get the NAND right, it's a whole other thing to get it right in the system with firmware and with a controller that works appropriately. The Qual hurdles are unbelievably difficult in this business with the customers. They have high expectations and do not want to see their quality impacted by lack of quality from the suppliers. So I think for us, it's, maintain our focus on that. Keep the competitive edge. Super happy that we were first to market with the 176 layer, which I think is indicative of our ability to be ahead. And I think you do that in the technology space, generally, it works out for you.

Do you think there's a chance that these Chinese companies get added to this entity list, like Fujian Jinhua was. Remember, Fujian Jinhua was added to the entity list for essentially stealing life being from Micron, right? I'm sure you're aware of that situation. And it didn't work out very well for them is what I understand because they can't get equipment. No, I don't know that there is such an obvious case of stealing technology for CXMT, the DRAM player and YMTC, the NAND player. But, I'm sure there's probably likely some IP infringement here and there. So I'm just wondering, is there potentially a chance that one or possibly both of those may get added to the entity list in the future because of that IP infringement potential that they most likely have.

Yes. Well, as it relates to IP, obviously, we are super focused and we will do everything necessary to protect our IP, which we have obviously done most recently. But as it relates to government actions, we have no say in what the government may or may not do. So it's hard for me to comment on what they may or may not do. I'm sure they actively look at these type of things and make decisions based on what they think is the right thing to do. And by the way, that was the true of the case you mentioned that did happen. We certainly didn't have any say on that. That was the government's decision, and we support that.

But, yes, sure. And I understand you don't know what the government is going to do. However, I haven't seen any, so far, any litigation against YMTC or CXMT, whereas there was litigation from Micron against Fujian Jinhua. Is that -- because it's -- they're too small that they haven't really shown up on the radar yet? Or is it that their IP infringement is less obvious?

Well, we certainly had specific evidence around the prior case. And so we'll defend our IP when we have that evidence. If evidence comes up later, that's -- where we believe our IP has been infringed, you can expect us to aggressively defend that IP.

Right. Well, I think, I mean the Fujian Jinhua case was like, not just IP infringement, there is actually IP theft as well, wasn't there?

Correct.

You don't need to have seen IP theft to have really -- like you're -- if someone is infringing on your IP, you can still have litigation on infringing patents. That's a much lower bar that presumably they're infringing, if not Microns, then some of the memory companies, but we haven't seen any litigation so far. So I'm just wondering what are the reason for that is, and it could just be that they're just too small, they haven't really shown up on the radar screen because they're not really shipping any significant volumes there.

Right. No, I can't postulate a reason, to be honest with you, Mark.

Okay. Well, there was a couple of questions. I had a question on tech. You kind of touched a bit on it. And there's a question here in, from an investor. Can you share latest updates in detail about the node transition, what's the progress and when is it expected for introduction to the market? Yes, so you touched on NAND, maybe talk a little bit more about the timing of NAND and then let's talk about DRAM in terms of technology, if that's okay?

Yes. So the second generation NAND 176 layer is in production. It's ramping as we speak. It does take time for that all to flow through meaningfully into the cost side. So I think for the first couple of quarters, NAND's cost per bit is probably flattish, maybe down a little bit. But we really don't see meaningful cost declines until the back half of our fiscal year. That's where we really start to see the attraction around the 176 layer really take shape on our cost front. We talked about on the earnings call that we thought cost declines in NAND would be in the kind of low to mid-teens percent, and that's accounting for mix. Obviously, it's higher when you take out the fact that we're mixing up as well. And the fact that 176 came when it did supported that timing, I mean, we certainly understood when we were going to roll it out. So we built that into our forecast for what we thought cost declines would look like for NAND.

So low to mid-teens for FY '21, and that's all related to 176 because the first gen had negligible cost decline, right?

Yes.

So these are ones...

Yes.

And then when is...

And then I just want to make sure I'm clear about it, though, is the low to mid-teens is also a function of the fact that high-value solutions continue to be part -- more and more a part of our portfolio. And so that tends to weigh down the cost per bit for us. I mean, if you strip that out, obviously, the cost -- year-over-year cost declines are going to be more than low to mid-teens. But since we do have that mix element in there, we tend to talk about it with the mix impacts as well.