Micron Technology Inc (MU) — Q1 2025 Earnings Call Transcript

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.

Hello. Thanks for hosting this call. I've got two questions. On the 100 basis point decline in gross margin in Q2, as you mentioned, it's almost all NAND related mixed, pricing, lower shipments. But qualitatively, what are you anticipating for DRAM pricing? You talked about positive mix impacts, HVM, server DRAM, partially offset by weak commodity pricing. So is blended DRAM pricing going up sequentially in Q2?

Yeah. I think we don't really guide for pricing. 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. The overall set of comments you heard in the prepared remarks and the comments Sanjay and Mark made indicate that the mix improvement continues on a good trajectory. HBM growth continues on a good trajectory. Data center overall revenue trajectory is 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. So those are the positives. We spoke about the near-term issues related to market environment, seasonality, and NAND-related challenges. The mix and data center growth, along with overall DRAM robust results, are what is helping us ensure that the margin for the rest of Q2 is still fairly similar to what we have in FQ1, with only a 100 basis point delta despite some of the headwinds we mentioned.

That's helpful. To follow-up, one of the big Micron-specific inflections has been the strong increase in your enterprise and data center SSD share. You guys have always been strong in SATA. In mainstream PCIe and NVMe, you've gone from 3% to 5% market share in calendar '21 to 10% to 12% exiting last year. In Q3, your share was around 14%, which grew sequentially again in fiscal Q1. So my guess is $1.1 billion to $1.2 billion of data center SSD, which is about 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, and that's impacting Q2. But do you expect sequential re-acceleration in data center SSD in fiscal Q3 and just overall growth in the data center SSD business for fiscal '25 and calendar '25 because AI demand pull for those products is strong?

Yeah. 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've been investing in for five to seven years now, and we have seen a rejuvenation in our product portfolio with industry-leading products. For instance, we have 30 terabyte and 60 terabyte SSDs. The first PCIe Gen5 60 terabyte SSD is on the market, and the 9550 is the fastest SSD in the world. There are lots of amazing products that have enabled us to achieve solid performance, with several quarters of record-setting revenue and market share performance. Looking ahead, beyond the lumpiness you mentioned, we expect a growth in bit shipments restarting in FQ3 for data center SSD and continuing thereafter. The trajectory should grow because we expect the overall data center CapEx from our end customers to remain strong, which should enable us to seize significant opportunity with the portfolio that is the strongest it has ever been in our history.

That's very helpful. Thank you.

Thanks for doing the call back. I have two questions as well. The first question, Mark, I want to go back to 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. Is it 100 basis points? Is it 200? Is it less? When you say constraints, are you implying there isn't the potential or is there a potential to see some gross margin positive drivers into fiscal Q3 relative to fiscal Q2? Even with that underload charges, could we see gross margin flat to up? I'm just trying to understand what you're trying to frame as far as being constrained.

Yeah. So Aaron, we're not guiding third quarter gross margin. We want to make sure that when we provide a number, it's something we have high conviction on. What we do know is that there are some unfavorable items persisting into the third quarter. The NAND challenging market conditions are believed to extend into calendar Q1, which overlaps with our third quarter fiscal; that's still an issue. As I mentioned, in the third quarter, the underload charges will begin to hit, affecting period costs and higher cost inventory. That effect is around 100 basis points impact. There are favorable effects occurring in the second and third quarters and beyond, including continued growth in data centers, mix benefits from higher margin products, and better market conditions in NAND. We believe that these factors will support margin expansion beyond the third quarter. We expect inventories to be up in the near-term due to volume declines in NAND and DRAM in the second quarter.

That's helpful. As a quick follow-up, I know a lot of focus is placed on HBM understandably. But one of the things you've highlighted over the last several quarters is LPDDR5X, especially in the context of the Blackwell product cycle GB200 from NVIDIA. How would you characterize your views on that ramp and where you think Micron's relative share position is within the LPDDR5X data center opportunity?

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. There will be others over time, but we are in the pole position, having driven the innovations that are needed for data centers. Reliability, availability, and serviceability that is built into DDR5 is not available in LP solutions. Micron has made some important innovations in that product category, and we're partnering with customers to seize this significant opportunity. This is an important part of our overall strategy and data center business. When you combine this with the high-capacity DIMM portion of the data center market, we were the first with mono die-based 128 gigabyte DIMMs. The combined momentum of this is driving over 40% sequential growth in data center, and over 400% year-over-year in FQ1, and greater than 50% of our revenue now comes from data center.

Okay. Thank you.

Yes. Hi. I guess the first question is, maybe just some clarity on what may have precipitated some of the cautiousness here and what might have changed over the last couple of months? I know there were some cautious signs back in August, and they receded in the last earnings call. Was it a function of customer inventories turning out to be a bit more than expected, or do you think it was a function of demand?

Yeah. Let me try to address what changed. Over the course of CQ4, we have seen a pushout of the PC refresh cycle that our customers had anticipated. It's not that the refresh cycle won't happen; it will happen in 2025. It's just been delayed. There are numerous drivers supporting this belief and our customers' belief that the refresh cycle is expected to materialize. The average age of PCs has become quite old, and is ripe for an upgrade. As people look to keep their PCs for several years, they will want to future-proof their hardware specifications. Hence, we believe that with modest unit volume growth, the mix improvement from higher memory content in PCs will be a positive driver in 2025. The delay in that upgrade cycle means that our calendar 2024 PC shipment forecast at the unit level is now flat year-over-year, which has been one driver. Inventories that we highlighted in the previous earnings call, which we expected would become healthier by spring, are seeing continued impacts. Some moderation on the data center SSD side is also noted following several quarters of significant buying. Overall, the consumer-oriented segments are where these impacts lie, while the data center demand remains very robust. We expect that by the spring, inventory levels will be much healthier, allowing for shipment growth in the second half of the fiscal year. Overall, DRAM supply remains tight, and NAND continues to be an issue, which we have outlined in the actions taken on supply to align with demand.

That's helpful. Thank you. A follow-up to that: In terms of the lower bit demand assumption overall for Q2, how much of that is DRAM versus NAND? With the mid-teens growth you now have for DRAM for the year, how back-end loaded is that considering what you're seeing in Q1?

Yeah, Chris. Just to be clear, we did say that DRAM bits would be down in Q2, and we're seeing a significant decline in NAND.

Okay.

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

Yes. The DRAM decline in FQ2 is largely driven by the dynamics I described in the consumer-oriented segments. The data center segment is very strong on the DRAM side. We expect sequential increases in volume in both DRAM and NAND starting in FQ3.

Okay. Got it. Helpful clarification. Thank you.

Hi, Mark and Sumit. Thanks for doing the call back. Regarding your SPM capacity, as it increases next year, how much of your conventional DRAM capacity is coming down? Can you size the HBM capacity increase for next year?

Hi, Vijay. It's Manish. We've given data points that as we go towards mature yields in HBM, the trade ratio between HBM and DRAM is at 3. You can sort of do the math on how much of our HBM share will impact conventional DRAM wafer starts. We are focused on maintaining our DRAM bit share through the Atrium transition next year. Our DRAM overall bit growth will align with industry growth numbers discussed for calendar '25.

Got it. And on the CapEx side, can you talk about the split between HBM, conventional DRAM, and NAND?

We're not breaking it out specifically, but overall, HBM is the largest portion of our CapEx for this fiscal year. We’re also focused on facilities construction and some in the back end. This includes significant investments in India and other locations.

The percent spent on NAND is lower than the percentage of business revenue; so it's a much lower spend. The DRAM spend is largely driven by the trade ratio effect and HBM requirements.

That’s excellent. Thanks.

Hi, there. Good afternoon. Thanks for letting us ask a few questions. First, regarding NAND. We've heard of at least one other NAND supplier reducing wafer starts, similar to how you've disclosed today. Do you expect other suppliers to take similar actions or lag like in the past?

It's difficult to predict what others will do. We'd rather not speculate on other activities. We can only speak on how we view the market and the decisive actions we are taking immediately.

Thanks. On the DRAM side, it wasn't that long ago that your DDR5 bit shipments crossed over DDR4. You’re indicating only 10% of your global DRAM sales will be from DDR4 for the next several quarters, which is a lot lower than I expected. Does this shipment mix closely match your internal production mix? Are you holding back inventory, or have you structurally shifted most production to DDR5 and specialty products?

Yes, that's correct. DDR4 and LP4 are in that 10% approximately range of revenue for the remainder of the fiscal year. We do have long life cycle products we support for different segments, such as industrial and automotive, but overall revenue growth means this product mix will remain contained. We've shifted a lot of production to leading-edge products that are constrained. Our mix shift towards RAM is very meaningful.

Great. Thank you.

Hey, Mark. I wanted to come back to your comment on NAND, where you said NAND would account for most of the sequential decline, possibly down 30% plus. Just confirming that, is that correct? And I have a follow-up on the consumer-facing business.

Yeah. It's the majority of revenue decline versus the second quarter sequentially.

Okay. Thank you. Regarding the high inventory levels last quarter, you believed some progress was made through fiscal Q1. Can you quantify how much is left? Are you looking to reduce a week or two in the spring? How much is left on a percentage basis?

Are you talking overall inventories, Quinn? We made progress in the first quarter, with DIO down below 150 days overall. However, we are seeing DIO increase from first to second quarter, along with absolute inventory dollars. This is a volume-driven phenomenon. We expect inventories to draw down through the year, both in NAND and DRAM, but we'll need our leading-edge inventories to supply the market. Overall, we expect DRAM inventories to be below target levels by the end of the fiscal year.

Sorry, Mark. My question was more about inventory at the customers rather than on your balance sheet. It seems you're shipping lower than consumption in Q2. Are you looking to align shipments with consumption in Q3?

Yes. We expect inventory levels at customers to improve by spring. Currently, we're shipping at a pace lower than consumption. We have confidence in demand and related inventory consumption through fiscal '25.

Okay. Thank you.

Thanks. Mark, I want to clarify that I heard you correctly. You said DRAM bit shipments would be in line with the mid-teens for calendar '25, correct? And are you going to under-ship the market? I think you said demand is growing low doubles, but I would think you would under-ship that.

We said that industry bit demand would be mid-teens for DRAM, and we expect to ship in line with the industry.

What about NAND? I mean, NAND demand is up low doubles. Will you under-ship that?

No. For the year, we expect stable bit share in calendar '24 and '25 across both DRAM and NAND. Generally, that is true across both years.

So your NAND shipments are going to be up close to that low double-digit demand, correct?

Yes. Just be mindful of the calendar versus fiscal year skew. That should be factored in.

Got it. Thank you.

I would like to clarify your comments regarding HBM and the volume and pricing agreements being locked in. Are these agreements cancelable, and under what conditions can customers cancel? Are there specific duration limits or restrictions in place?

These agreements relate to the amount of bits and cubes of HBM our customers will purchase, with associated pricing. Lead times lead to firm orders for HBM products. Given the strong demand, more discussions focus on fulfilling more supply. Customers can cancel if the order is not firm and must do so within a reasonable time frame. If it is beyond that timeline, changes can be made, allowing us to manage our supply accordingly.

I had one question on HBM and how to think about the ramp going forward on a quarterly basis. You don't report HBM revenue, but we're estimating you did $800 million to $900 million in revenue in the quarter. I assume it'll increase nicely this quarter. When thinking about starting this quarter, do you believe the revenue trajectory could flatten out from here? If so, what are the constraints? Is it space? Is it TSP capacity? When should we expect your capacity to increase significantly? Is it tied to the Taiwan facility?

We outlined how we would ramp to our market share targets, reaching a $30 billion plus TAM on HBM for calendar '25. While we don't report HBM specifically, the numbers you mentioned for FQ1 are a bit high. There will still be a robust ramp in HBM revenue for us. I'll pass it to Manish for the rest of your question regarding capacity.

We're pleased with the HBM ramp, especially for 8-high, which has exceeded expectations. The process choices we've made are scalable, allowing for a transition to 12-high as capacity grows through calendar year '25. We're gradually adding capacity as we install new equipment and qualify that for production, given that we had limited previous production in HBM3 and HBM2E. We are currently at a low starting point in HBM3E ramp and expect solid revenue growth as we continue through fiscal '25 and into calendar '25.

That’s really helpful. Thank you so much.