Q1 2025 Earnings Call Transcript

Thank you for standing by, and welcome to Micron's Post Earnings Analyst Call. At this time, all participants are in listen-only mode. After the speakers' presentation, there will be a question-and-answer session. [Operator Instructions] As a reminder, today's program is being recorded. And now, I would like to introduce your host for today's program, Satya Kumar, Corporate Vice President, Investor Relations and Treasury.

Please go ahead, sir.

Yeah. Thank you, Jonathan, and thank you. And welcome to Micron Technology's Fiscal First Quarter 2025 Post Earnings Analyst Call. Joining me today are Sumit Sadana, Micron's Chief Business Officer; Manish Bhatia, EVP of Global Operations; and Mark Murphy, our CFO.

As a reminder, the matters we are discussing today include forward-looking statements regarding market demand and supply, market trends and drivers, and our expected results and guidance and other matters. 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 documents we have filed with the SEC, including our most recent Form 10-Q and upcoming 10-Q for a discussion of risks that may affect our results. Although, we believe that the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee future results, levels of activity, performance and achievements.

We are under no duty to update any of the forward-looking statements to conform these statements to actual results. We can now open the call to Q&A.

Certainly. And our first question for today comes from the line of Harlan Sur from JPMorgan. Your question please.

Hello. Thanks guys for hosting this call back. So I've got two questions. On the 100 basis point decline in gross margin in Q2, as you guys mentioned, it's almost all NAND related mixed, pricing, lower shipments.

But qualitatively, like, what are you guys anticipating for DRAM pricing, where you talked about positive mix impacts, HVM, server DRAM, partially offset by weak commodity pricing. So I guess my question is, blended DRAM pricing up sequentially in Q2?

Yeah. I think, we don't really guide for pricing. I think we have given you a lot of data points that can help you in your modeling, but we are not going to provide guidance on the actual pricing expectation for competitive reasons. And I think the overall set of comments that you heard in the prepared remarks and the comments that Sanjay and Mark made, basically, the mix improvement continues on a good trajectory.

HBM growth continues on a good trajectory. Data center overall revenue trajectory continues to be robust. While we do have some near-term moderation in data center SSD after a lot of growth in the past quarters, the overall data center revenue trajectory remains very robust and we expect that to underpin our performance through fiscal and calendar 2025 and so those are the positives. And of course, we spoke about the near-term issues related to market environment, seasonality, NAND related challenges.

So those are some of the headwinds. The mix and data center growth and overall DRAM robust results is what is helping us ensure that the margin for rest Q2 is still fairly similar to what we have in F Q1 in terms of 100 basis point only delta despite some of these headwinds we mentioned in some of the other areas.

No, that's helpful. And then maybe to follow-up, one of the -- and I've said this before, right, one of the big Micron specific inflections has been the strong increase in your enterprise and data center SSD share, right? I mean, you guys have always been strong in SATA. Mainstream PCIe and NVMe is where you guys have gone from like 3% to 5% market share in calendar '21 to 10% to 12% exiting last year.

I think in Q3, your share was like 14%, and it grew sequentially again in fiscal Q1. So my guess is $1.1 billion, $1.2 billion of data center SSD, that’s like, 65% of your overall NAND business. You guys are now the number two or number three market share leader. I know enterprise SSD can be lumpy as you mentioned, right, that's impacting Q2. But do you guys expect sequential reacceleration in data center SSD in fiscal Q3 and just overall growth in data center SSD business fiscal '25 and calendar '25 just because AI demand pull for those products is just so strong?

Yeah. I think you laid it out really well. We have made tremendous progress in our data center SSD business over the years. This has been an area we had been investing in for five, seven years now, and we have seen a tremendous rejuvenation in our product portfolio with industry-leading products now.

So 30 terabyte SSDs. 60 terabyte SSD we just announced. First PCIe Gen5 60 terabyte SSD on the market. The 9550 fastest SSD in the world. So there are lots of these amazing products that have enabled us to get to a really solid performance, several quarters of record setting revenue and market share performance.

So as we look ahead beyond the lumpiness you mentioned, beyond that, F Q2, we do expect a growth in bit shipments restarting in F Q3 for data center SSD and then continuing from there on into future quarters, that trajectory should grow because we do expect the overall data center CapEx from our end customers to remain strong, and that should enable us to get significant opportunity with the portfolio that we have that is now the strongest it ever has been in our history.

No, that’s very helpful. Thank you.

Thank you. And our next question comes from the line of Aaron Rakers from Wells Fargo. Your question, please.

Yeah. Thanks for doing the call back and I have two questions as well. I guess the first question, Mark, I want to go back to kind of the Q&A in the prepared call. I'm trying to understand, like, you talked about constraints on gross margin into the fiscal third quarter.

Trying to kind of frame the impact that you're thinking about from the underload. Is it 100 basis points? Is it 200? Is it less?

And when you say constraints, are you implying -- are you necessarily implying that there isn't the potential or rather, is there a potential to see some gross margin positive drivers into fiscal 3Q relative to fiscal 2Q? Even with that underload charges, could we see gross margin flat to up? I'm just trying to understand a little bit of what you're trying to kind of frame as far as being constrained?

Yeah. So Aaron, we're not guiding third quarter gross margin. So we're just trying to be mindful to make sure that when we provide a number, it's a number that we have high conviction on. What we do know is that there are some unfavorable items that persist into the third quarter.

The NAND challenging market conditions, we believe, extend into calendar Q1, which overlaps with our third quarter fiscal and so that's still an issue. And then as I mentioned, in the third quarter, the underload charges will begin to hit and those will hit in the form of period costs and then just some higher cost inventory. And that effect is, yeah, you mentioned 100 basis points to 200 basis points, closer to 100 basis points impact. And then there are definitely some favorable effects which continue to occur in second quarter and third quarter and beyond, and that is the most important thing on the call, and that is continued growth in data center, continued mix benefits from higher margin products, HBM, high capacity DIMMs, LPDRAM.

And then as Sumit mentioned, volume growth and better market conditions in NAND, but volume growth in data center SSD is continuing. So all these items will, we believe, beyond the third quarter support margin expansion. And then finally, Aaron, we've been clear that we expect inventories to be up in the near-term because these volume declines in NAND and DRAM in the second quarter. DIO and dollars of inventory will be up.

In DRAM, we're still very tight on leading edge and overall DRAM inventories are fine and to the point where through the end of the year, they will continue to improve from second quarter. And then, at the end of the year, where overall DRAM inventories will be below our target levels. And then NAND, as that market recovers, we expect NAND inventories while not as healthy as DRAM, NAND inventories will improve second to third to fourth quarter.

Yeah. That's helpful. And then, as a quick follow-up, I know a lot of focus is placed on HBM understandably. But one of the things that you guys have highlighted over the last several quarters is LPDDR5X and obviously, getting into the Blackwell product cycle GB200 from NVIDIA.

How would you characterize your views of that ramp and where you think Micron's relative share position is within those -- that LPDDR5x data center opportunity? I'm just curious. Thank you.

Yeah. We are actually the pioneering memory supplier of LP in the data center. If you think about all of the market for LP in the data center worldwide, it is pretty much Micron driving that right now. Over time, there'll be others, but we are in the pole position of having driven the innovations that are needed for a data center to deploy memory that was not purpose built for data center RAS environment.

So reliability, availability, serviceability that is built into DDR5 is not available in an LP solution. And Micron has made some important innovations in that product category. And hence, we are the ones who are really partnering with customers to drive this significant opportunity for us. And this is something that is a really important part of our overall strategy.

It's an important part of our data center business. And when you combine this with the advantages that we have driven in the high capacity DIMM portion of the data center market where we were the first company with mono die-based 128 gigabyte DIMMs (ph), then you look at the overall advantage that we offer to customers with LP in the data center, high-capacity DIMMs in the data center, this -- these two categories for the data center in itself is going to be multi-billions of dollars. HBM separately is multiple billions of dollars. Data center SSD is going to be multiple billions of dollars.

And the combined momentum of that is what you're seeing drove the 40% sequential growth in data center, 400% year-over-year in F Q1, and the greater than 50% of our revenue now that comes from data center.

Okay. Thankyou.

Thank you. And our next question comes from line of Chris Caso from Wolfe Research. Your question, please.

Yes. Hi. I guess the first question is, maybe just some clarity on what may have precipitated some of the cautious -- some of the incremental caution here and what might have changed over the last couple of months? I know that is kind of back and forth a little bit, there were some cautious signs kind of back in August and they kind of received in the last earnings call.

Was it -- I guess the question is, is it a function of the customer inventories turned out to be a bit more than we had expected or do you think it was a function of demand?

Yeah. Let me try to maybe address what changed. So, if you think about over the course of CQ4, we have seen a pushout of the PC refresh cycle that our customers had been anticipating, we had been anticipating. And it's not like the refresh cycle won't happen.

It will happen in 2025. It just been a little bit delayed. There are numerous drivers which we mentioned in our prepared remarks as to why we believe and our customers believe that the refresh cycle will happen. The average age of the PC has become pretty long in the tooth, and it is ripe for an upgrade. And as people look to keep their PCs for a number of years, they will look to ultimately future proof the hardware specs, which is why we believe that even with very modest unit volume growth, the mix improvement that will come from higher memory content in these PCs will be a positive driver in 2025. But the delay in that upgrade cycle means that our calendar '24 PC shipment forecast at the unit level for PCs at our end customers has been reduced and is now very flattish year-over-year in calendar '24, so that has been one driver.

The other is that definitely the inventories that we had highlighted in the last earnings call, that we expected by spring would become healthier at our customers, that inventory and inventory reduction as well as the seasonality of CQ1, those are continuing impacts. Some impact coming from the moderation. You mentioned -- previous discussion we just had a short while ago about lumpiness of demand on the data center SSD side. So some moderation in data center SSD into CQ1 after significant bout of buying over several quarters in calendar '24. So those are the things that have impacted the near-term outlook.

But I just wanted to mention that most of the impact is limited to consumer-oriented segments, and the trajectory of demand in the data center continues to be very robust. Our own data center segment view of the overall revenue trajectory through fiscal '25, calendar '25 remains in a very solid trajectory. You have seen our F Q1 results as well that we highlighted. So it's mainly the consumer-oriented segments and some very temporary moderation of the data center SSD, which we expect will pick up again in a couple of months.

And so if you think about just these consumer-oriented segments, we do expect that by spring time, the inventories will be much healthier. And then we are back to shipment growth because right now, we are shipping to these customers at a rate that is lower than their ship out because they are consuming DRAM and NAND at a faster rate than they're purchasing from us and from the industry. So that's the effect of the inventory that we expect will be in a much healthier place by spring and then a resumption of shipments and growth in shipments for the second half of the fiscal year and then things will be much better.

And overall, DRAM will be continuing to be in a much healthier place. The supply is tight. HBM continues to pressure non-HBM availability of supply. So a large part of the issue is NAND related, and we have outlined the actions we are taking on the supply there to decisively bring our supply and balance with the demand.

That's helpful. Thank you. I mean just a follow-up to that. In terms of -- and I think you've been very clear on what's going on with NAND.

But with DRAM, and I guess, this lower bit demand assumption overall for Q2. How much of that is DRAM versus NAND? And with the mid-teens growth that you now have for DRAM for the year, how back end loaded is that given what you're seeing in Q1? Because, I guess, you really weren't specific in NAND versus DRAM bit demand for -- bit growth for Q1? Yeah.

Yeah, Chris. Just to be clear, we did say that DRAM bits would be down in Q2. We said that NAND would be down meaningfully.

Okay.

Yeah. The majority of revenue decline versus the second quarter will be NAND.

So yeah, I mean, the DRAM decline in F Q2 is largely driven by the dynamics I described on the segment side for consumer oriented segments, right? I mean the data center segment, very strong on the DRAM side, very strong. So it's really the consumer-oriented segments. And we feel that, that's -- we're going to get past a lot of that headwind in the spring timeframe.

So after CQ1, things will start to improve there. So we do expect sequential increases in volume in both DRAM and NAND starting in F Q3. F Q2 is what Mark had mentioned, the sequential decline in DRAM volume, but a meaningful decline sequentially in NAND volume.

Okay. Got it. Helpful clarification. Thank you.

Thank you. And our next question comes from the line of Vijay Rakesh from Mizuho. Your question, please.

Yeah. Hi, Mark and Sumit. Thanks for doing the call back. Just wondering on the -- as your SPM capacity, maybe goes up next year, how much is your conventional DRAM capacity coming down? And can you size the HBM capacity, how much do you think that goes up next year? I have a follow-up.

Hi, Vijay. It's Manish. So I think we've given the data points that -- as we go towards mature yields in HBM, the trade ratio between HBM and DRAM is -- HBM3E and our conventional DRAM is at 3. So that's -- you can sort of do the math at how much of our -- as we grow our HBM share towards our natural bit share, which we expect to achieve sometime next year in calendar '25, you can do the math and determine how much of that impact will be on our conventional DRAM wafer starts because we'll -- essentially for everyone, we're going to have to have three wafer starts or -- in HBM for every one that we would have had for the same number of bits in conventional DRAM.

And we are focused on being able to maintain our DRAM bit share as we go through the Atrium transition here next year. And I think our DRAM overall bit growth will be in line with the industry bit growth numbers -- demand numbers that we had talked about for calendar '25.

Got it. And then on the CapEx side, can you talk to what's the split between HBM and conventional DRAM and NAND? Thanks.

I don't know we're breaking it out specifically, but we have said that when you think about HBM in terms of both the capital that's specific to the HBM product, whether that's in the front-end fabs or in the specific assembly and test process and the clean room needed for new HBM capability, all of that, plus you also have the CapEx to be able to maintain the bid output as you or impacted by that trade ratio, that overall HBM is the largest portion of our CapEx for this fiscal year. Other areas that we're focused on are in facilities in construction and also some in back end. As you know, we have a facility that we're investing in, in India and others, so...

Vijay, NAND is -- clearly, we've mentioned we've cut CapEx in NAND. It's lower. The percent spend on NAND is lower than the percent of the business revenue. So it's a much lower spend. And the DRAM spend, as Manish mentioned, is largely driven by this trade ratio effect and HBM requirements there.

That’s excellent (ph). Thanks.

Thank you. And our next question comes from the line of Brian Chin from Stifel. Your question, please.

Hi, there. Good afternoon. Thanks for letting us ask a few questions. Maybe first on the NAND side.

We've heard of at least one other NAND supply reducing wafer starts, maybe kind of similar as to how you've disclosed today. But back in early 2023, some of the larger suppliers were slow to take actions to reduce wafer inputs. Do you expect maybe some of these other suppliers to follow more quickly this time, similar actions or lag like they did previously?

Difficult for us to project or predict what others would do. We'd rather not speculate other activities. What we can say is, how we look at the market and the decisive actions we are taking immediately. And we feel like these things will at least help Micron's view of supply growth coming from us and our view of demand growth to be aligned, so that's what we are doing.

Okay. Thanks. And then on the DRAM side, I don't think it was that many quarters ago that your DDR5 bit shipments crossed over DDR4. And now you're indicating only 10% or so of your global DRAM sales will be from DDR4 for the next several quarters through fiscal year-end, that's actually a lot lower ratio than I would have expected. Does this shipment mix actually closely match your internal production mix?

Meaning that it's not just you're holding back inventory, but structurally, you've shifted a high majority of your DRAM production to DDR5, HBM specialty, those types of products.

Yeah. I mean that's correct. I mean we do have -- I mean, just for clarity, DDR4 and LP4, right, that's in that 10% approximately range of revenue for the remainder of fiscal year. And we do have long life cycle products that we support for the different segments like industrial, automotive, over time with more defense, aerospace, etc., that will have some of this long cycle effect.

However, as our overall revenue grows, the mix of these products will remain at a very contained level. We have shifted a lot of our mix to these leading edge products, and our leading edge nodes are very constrained. So we are dedicating a lot of capacity to a very fast ramp of HBM, and you're seeing the results from that. And even as we continue to increase our view of calendar '25 time of HBM, we are maintaining our view of getting to our natural share in the second half of calendar '25. So you're seeing that a lot of mix shift is happening.

And I spoke about the LP leadership in the data center, DDR leadership and the growth that is taking place. And even the growth in LP5 on the flagship smartphone side that we are supporting. So yeah, definitely, that shift is very meaningful.

Great. Thank you.

Thank you. And our next question comes from the line of Quinn Bolton from Needham & Company. Your question, please.

Hey, Mark. I just wanted to come back to your comment on the NAND. And I think you said that NAND would account for most of the sequential decline. I mean if that's right, it could be down 30% plus. So I just want to make sure I heard you right that, that was indeed what you said? And then I got a follow-up on the consumer-facing business.

Yeah. I -- it's the majority of the revenue decline versus second quarter sequentially.

Okay. Thank you. And then just -- you talked about the high inventory levels last quarter. It sounds like you thought you might have made some progress through fiscal Q1. Just wondering if you might be able to sort of quantify how much is left?

Can you give us any sense on, are you looking to reduce a week or two weeks in -- through the spring? And from where you started last quarter, I assume most of that inventory reduction is happening in the second quarter that you probably didn't get through that much in fiscal Q1, but just wondering if you might be able to sort of from where you started, how much have you gotten through? And how much is left maybe on a percentage basis, if you don't want to talk on a week's basis?

Are you talking overall inventories, Quinn?

Yeah. Just mostly on the consumer-facing business that seems to be kind of driving most of that sequential decline into fiscal Q2.

Yeah. So we didn't break down inventories by market. We did make progress in the first quarter, as you've noted. DIO was down below 150 days overall.

But we have a volume related decline or deterioration in DIO in the second quarter. So on a -- we'll have DIO go up first to second quarter, we'll have absolute inventory dollars go up. And again, this is a volume driven phenomenon. And then from there, in part due to the continued data center strength, the resumption of data center SSD volume growth and the refresh cycle and content driven growth in consumer and consumer-related markets.

We do see our inventories drawing down through the year, both in NAND and DRAM. Now in DRAM, we will need our leading-edge inventories that we have now just to be able to supply the market we see. So we're going to on our forecast today, and total DRAM inventories below our target levels by end of fiscal [indiscernible].

Sorry, Mark. I guess my question was more inventory at the customers rather than on your balance sheet, sort of, you said your current shipping low consumption in Q2. I guess maybe the other way is, it sounds like you think you start shipping in line with consumption again by Q3?

Yeah. I thought that was covered earlier on that we see inventories getting to a better place in the spring. So -- but as Sumit mentioned, we're shipping in less than they're shipping out or consuming. So we have confidence that both in their demand and related our inventory consumption through fiscal year '25 beyond our second quarter.

Yeah. Okay. Thank you.

Thank you. And our next question comes from the line of Timothy Arcuri from UBS. Your question, please.

Thanks. Mark, I think, I just want to clarify that I heard what you said. So you -- in terms of your DRAM bit shipments, you're sort of saying that they're going to be in line with the mid-teens for calendar '25. Is that right? Number one.

And then number two, are you going to under-ship the market? Because I think you said demand is growing low doubles, which you took that down, but I would think you're going to under-ship that.

We said that industry -- well, we said industry bit supply -- our industry bit demand mid-teens. We gave -- we said with DRAM -- yeah, for DRAM. And when we've said that volumes will be down in the second quarter and expect to ship in-line with the industry.

For the year -- full year?

For the year.

Yeah. Okay. And then what about NAND? I mean NAND demand is up low doubles, but I would guess you're going to under-ship that for the year. Is that fair?

No. I mean, for the year, we expect to have stable bit share in calendar '24 and '25. Across both DRAM and NAND, generally, that is true across '24 and '25. We expect to have relatively stable. Some noise here and there, but relatively stable bit share across both of those years.

Okay. So you'll ship -- your NAND shipments are going to be up like close to that low double-digit demand. Is that fair?

Yes.

All right.

There's always the calendar versus fiscal, right? I just want to be careful because that skews things somewhat, but we have a four month skew between calendar and fiscal year. So that changes things.

Yeah. But I mean, I was talking calendar-to-calendar. So I'm just trying to just say in the calendar year because you're giving us a demand number for calendar year. So I'm saying in that calendar year, you're going to ship in-line roughly for both DRAM and NAND.

Yeah. That's right.

Got it. Okay. All right. I’ll let somebody else ask.

Thank you. [Operator Instructions] Our next question comes from the line of Krish Sankar from TD Cowen. Your question, please.

Hey, guys. This is Hadi for Chris. I would like to clarify your comments regarding HBM and the volume and pricing agreements being locked in. Are these agreements cancelable? If so, I wonder under what conditions can customers cancel? I'm wondering if there are specific like duration limits or restrictions in place? Thank you.

Yeah. I mean, these are agreements that relates to amount of bits and cubes of HBM that our customers are going to purchase 8-high versus 12-high by quarter with pricing associated with them. And once we get into certain relatively lengthy lead times, then we get firm orders because of these lead times that we build with and ship for. And given the environment in the industry, more of the discussion is around chasing more supply because we have an extraordinarily strong product that customers want more of, and we are ramping as fast as possible -- fast as humanly possible to be able to grow our share to the targets that we have outlined.

And so yes, customers can cancel if it is not in that firm time for the purchase order. If it is beyond that time line, yes, customers can make changes, but there's a fairly lengthy time line when based on very long lead times that they have to place from purchase orders for. And customers have been doing that with consistency, and that's what gives us the confidence for a very strong ramp ahead.

Got it. Thank you.

Thank you. And our next question comes from the line of Toshiya Hari from Goldman Sachs. Your question, please.

Hi. Thank you for taking the follow-up here. I had one question on HBM and how to think about the ramp going forward on a quarterly basis. You guys don't report HBM revenue.

So I'm guessing a little bit here. But based on everything you've said over the past couple of quarters, we're sort of estimating you did, I don't know, $800 million, $900 million in revenue in the quarter. I assume it's going to be up again nicely this quarter. So when I think about the starting point this quarter and then the comments you made on calendar '25 TAM and your market share goals, it feels like the trajectory kind of flattens out from here, still up, but flatter.

I'm curious, am I thinking about it the right way? And if so, what are some of the constraints? Is it space? Is it TSP capacity both?

And when should we expect your supply side -- your capacity to increase nicely. Is it tied to the Taiwan facility? Any insights there would be really helpful. Thank you.

I'll give you a couple of opening comments, and then I'll pass it over to Manish to talk about the second half of your question that relates to space and so on. Yes, we did give you the expectation of how we would ramp and how we get to this market share that is roughly equivalent to our DRAM supply share, which is in the low-20s and get there in the calendar second half with a view that there is going to be a $30 billion plus TAM on HBM for calendar '25, right? So that's sort of the outline we have given you. Now you're right, we don't report HBM numbers.

We may choose to provide some data points in the future, but we haven't done so, except for that over $100 million point that we had mentioned when we first crossed that. The numbers that you outlined for F Q1 are a little bit high. But nonetheless, I think there is going to be still a pretty robust ramp occurring on the HBM revenue for us to hit the targets that we have outlined based on the TAM expectation and the growth of TAM on a quarterly basis that we had mentioned. And with that, I'll just pass it on to Manish to answer the second half of your question.

Yeah, Toshiya. So the -- we're pleased with how the 8-high ramp has gone. We've made some comments on the prior call about how that had exceeded our expectations in the last quarter, and we're continuing to have a strong ramp forward here. We think that the process choices we've made are -- this demonstrates that we have a scalable process and that it's going to be able -- we're going to be able to transition it to the 12-high as we move -- transition our capacity towards that ramp, which will become increasingly important as we go through calendar year '25. And really, it's just a ramp.

It's just a ramp of adding capacity across many different dimensions and it's just gradually happening over time as we are installing new equipment and qualifying that new equipment and growing our capacity week-over-week as we move towards that goal. Keep in mind, we had zero -- we had no product in HBM3 and very limited production in HBM2E on a process that was quite a bit different than what we have right now in HBM3. So we really are starting from a very low point as we were beginning this HBM3E ramp and are continuing to grow, and we're adding capacity gradually as we went through this year and continuing into next year.

So as Sumit said, I think we would expect that as we go through the year and as we get to the 12-high being a larger portion of our mix through the year, you will still see solid revenue growth as we're going quarter-over-quarter through the rest of fiscal '25 and even as we go through calendar year '25.

That’s really helpful. Thank you so much.

Thank you. This does conclude the question-and-answer session as well as today's program. Thank you, ladies and gentlemen for your participation. You may now disconnect. Good day.