Micron Technology, Inc. (MU) 2023 UBS Global Technology Conference Transcript

Micron Technology, Inc. (NASDAQ:MU) 2023 UBS Global Technology Conference November 28, 2023 9:35 AM ET

Company Participants

Sanjay Mehrotra – President & Chief Executive Officer

Mark Murphy – Chief Financial Officer

Conference Call Participants

Tim Arcuri – UBS

Tim Arcuri

Hi. Thank you. I’m Tim Arcuri. I’m the semiconductor analyst here at UBS. Very pleased to have Micron kick off the conference with us. Our first year having this conference. Very pleased to have both Sanjay and Mark here. And so we saw the 8-K this morning just now. So maybe that’s a good way to start off the conversation. So maybe you can run through how the quarter played out. I saw revenue was quite a bit better, earnings toward the high end, gross margin approaching breakeven. Just talk through some of the dynamics behind that.

Sanjay Mehrotra

So I can get started here. First of all, Tim, thank you for having us here. Great to see you all. And of course, during the course of the presentation, we’ll be making certain forward-looking statements. There are risks and uncertainties associated with those. Our risk factors are listed in our SEC filings that we make from time-to-time, please do refer to those.

So last update that we had provided was at the time of our earnings call at the end of September in the Q4 earnings call. And in that update, we have said that, industry environment was improving, inventories were improving and that we were seeing pricing bottoming out. In fact, pricing is starting to increase. We had also guided to increasing revenues and overall improved outlook for the year. And we had guided at that time to $4.4 billion in revenue, plus/minus $200 million, and we had guided to gross margins at minus 4%, plus/minus 200 basis points.

This morning, as you just noted, Tim, we did file 8-K and also did a press release updating the quarter. What we said is that revenue for the quarter would be approaching $4.7 billion, so higher than the high end of the guidance we had provided. We also said that the gross margins would be approaching 0%, minus 0.5% to 0% and again, significantly higher than the range that we had provided at the end of September.

We said that operating expenses will be $990 million approximately, and that is higher than what we had guided before. And those higher expenses are driven by timing of certain R&D program expenses timing of certain asset sales as well as higher incentive accruals based on the improved financial outlook for the full year basis as well. It is important to note that, we do expect that in fiscal second quarter, our operating expenses will go down. And we guided based on this to an EPS for Q1 to a loss of approximately $1 and that is at the high end of the range that we had guided to at the end of September.

And of course, I do want to point out here that our quarter ends on Thursday, November 30, and therefore, there can always be certain adjustments and which will be typical factors, including post quarter end close factors. Now, we do see that the pricing outlook is positive for the remainder of the year as well. And we, of course, remain extremely focused on managing our supply and with the improved pricing outlook that we are seeing, we do expect that in FQ2, our gross margin percentage would be positive. And with our focus on supply discipline as well as our focus on managing our business, optimizing for financial results, we do remain quite excited about the opportunities that are ahead for our business, particularly for the memory and storage industry.

Our technology, our product portfolio, our manufacturing excellence and customer relationships position us well to address the consumer market recovery to participate in that consumer market recovery as well as exciting growth opportunities in automotive and industrial. And certainly, AI driving content growth across all end market segments and certainly including data center that we are excited about.

We are excited about our HBM3E product, and I’m sure we’ll talk more about that later in the discussion. We have talked earlier that, that HBM3E product is in qualification with a major customer. And post the qualification, we expect to begin production ramp of the product in early part of calendar year 2024 and that will contribute meaningfully towards revenue in the latter part of our fiscal year. So exciting opportunities ahead. And I want to open with that and take further questions from you.

Question-and-Answer Session

Q – Tim Arcuri

Perfect. Perfect. Thank you. You’ve said that 2025 is going to be a record year for the industry. It sounds like you feel just as confident about that as you ever have. Can you just talk maybe about some of the supply/demand dynamics this cycle that might be a little different coming out of the downturn that give you the confidence that 2025 will be a record year?

Sanjay Mehrotra

Yes. We have said that for a while that 2025, we believe will be a record year for the industry. 2024 will be the year of recovery. I think, it’s really, really important that the focus on supply management is maintained. Micron is extremely focused on its CapEx management, OpEx management, supply management, even though pricing has increased and has good outlook, we want to make sure that the supply is controlled because industry, as you can see, is still far from the normal kind of profitability levels.

So the discipline and focus is critically important, and this will be our mantra as we drive the business moving forward as well. So 2024 will be the year of recovery with the pricing improvements continuing to progress through the years as well as with a focus on demand and supply management. 2025, we fully expect it to be a record year for the industry and to get to that record year in terms of revenue for the industry in 2025, it really just takes nominal levels of price increasing, nominal levels of business growth, nothing exceptional that is expected from it.

And in terms of the dynamics, of course, the most important dynamic is to continue to control the supply and Micron is going to remain focused on it. Other important dynamic that we have talked about before is for DRAM, for HBM. We have discussed before that HBM, while it takes small percentage of bids in the industry, it requires a significantly higher percentage of wafers. And HBM is in a high-growth mode, HBM requires more wafers because HBM die is more than two times bigger than a standard die. And also the packaging stack is much more complex. Therefore, its mature yields across the industry would be lower than the maturities of assembly and test for standard products.

So the combination of the two requires you to produce more than two times wafers to produce the same number of bids at the same — number of bits in the same technology node. And this is a factor that will affect the industry demand supply dynamics as well. The leading edge nodes are already in short supply, and the HBM factor leaves less of the leading-edge nodes for the non-HBN fab products as well. So these are some of the factors that will play an important role. Of course, with continued focus on supply discipline as we approach 2025, and as we look at 2025 to be a record revenue year for the industry.

And of course, the demand trends, we see those demand trends continue to demand drivers across end market segments continue to be healthy as smartphone and PCs content growth continuing to go higher and of course, AI servers that require significantly more memory, 6x to 8x DRAM content and 2x to 3x more NAND content. And with the AI revolution and yet in the very early stages, memory is a key enabler and a memory like HBM is a key enabler for high-performance systems that are needed for these generative AI kind of applications. So all of those factors, AI will continue to be a key growth driver across our end markets as we look ahead at the 2025 demand picture.

Tim Arcuri

Thank you. One thing that I hear from folks that they get a little concern that once the industry starts to improve, that this latent capacity will begin to come back online. And I think you spoke to some of the reasons why there’s probably not as much latent capacity as what you think there is because of some of these structural things happening. But can you talk about that? Is that a valid concern that there is all this latent wafer capacity out there to actually bring back online as things improve?

Sanjay Mehrotra

Yeah. That’s a very important question, and we did address it in our earnings call in September. As part of CapEx management, what we have focused on, of course, as part of supply management that we have been running underutilization in the fabs. We have talked about underutilization approaching 30% for DRAM and NAND in our fabs today. And we have said that we have taken that underutilized equipment and diverted it towards ramp-up of the new technology nodes, such as 1-beta in DRAM as well as 232-layer in NAND.

And what that does is that the new technology nodes typically have longer cycle time because they have increased process complexity. So as you move the equipment from the legacy nodes, towards leading-edge node, you are achieving a structural reduction in wafer capacity, wafer production capacity as well. So what we have said is that as we look at today, approximately 30% underutilization in the fabs by the end of the fiscal year, most of that underutilization will be eliminated, except it will continue in the legacy nodes, of course.

However, with the CapEx discipline that we have had managing our CapEx at lower levels and yet targeting ramp-up of new nodes, structural wafer capacity will come down meaningfully and that structural wafer capacity, this CapEx management across the industry and use of legacy equipment toward leading-edge node, we believe is happening in the rest of the industry as well, leading to structural wafer capacity reduction. So that is another factor that bodes well for the longer-term industry demand supply balance trends.

Important thing here is that as this underutilization switches to the structural reduction in wafer capacity, there is no longer a light switch to turn on and say all of a sudden that, okay, underutilization is going to turn into utilization in the fab and now you get increase capacity. There is no more that high level of underutilization that’s going to be there as we progress through the year in the industry because equipment is being used towards leading-edge nodes. So there’s no longer that light switch to just quickly turn on and off the wafer production.

And as you know, that to add any new capacity, it takes a significant lead time, it takes lead time for equipment, it takes new CapEx. And we will, of course, remain extremely disciplined in terms of managing that CapEx, managing that supply growth and therefore, any changes to structural capacity layer have a much longer lead time versus turning on or off the utilization?

Q – Tim Arcuri

Great. I want to ask a question on inventory, maybe to Mark, and then we can come back to HBM. So Mark, I know that the targets 120 days. If we use your COGS and we add the $1 billion that’s more of a strategic inventory that you built, you probably have $1 billion to $1.5 billion worth of excess inventory. Can you talk about how you’re working that down and sort of how you think about whether this could sort of short circuit the price increase, I think, in addition to the fact that there’s latent bit supply, there is a concern that, well, you have a lot of inventory also on your balance sheet. So can you talk about how that’s going to come down and the timing on that?

Mark Murphy

Sure. And similar to what we covered on the last earnings call, we do intentionally have some inventories associated with the NAND build, the replacement for — or floating gate, for example. We have some other pockets of inventory that we hold. Excluding that, as you mentioned, we would expect to work down about $1 billion, $1.5 billion of excess. We don’t believe that we’ve been tracking this for some time and that inventories have moved as we’ve expected. It’s been a helpful fact to have inventories and the build because it allowed us to flex on CapEx and do these capacity-related decisions that Sanjay went through. So we don’t see that adding to any sort of price pressure or anything.

So we do expect by the second half of fiscal 2024 to be down within a few weeks of our target level. And we’ll continue to manage inventories aggressively. It’s one of the indicators we’re watching for in the industry. We’re, of course, looking at price and price indications going forward. Those have — as we released this morning, those have exceeded our expectations. So we’re pleased today to state that we believe definitively that the gross margin will be positive in the second quarter. So we will continue to make capacity decisions as a function of what we see on our inventory projections and what we see in the pricing environment.

Q – Tim Arcuri

Great. Thanks, Mark. So let’s talk about HBM. We could talk all — the whole session about HBM. There’s a lot of people skeptical that you can catch up on HBM. I’m not skeptical at all. You bet on HMC. So there’s a natural thing where you’re behind on HBM. But on 3E now, you’ve got the core technical acumen to catch up on 3E. So can you talk about why we should be confident that you can catch up on HBM, some of the qualifications that you cited? And maybe what’s your share target is in HBM?

Sanjay Mehrotra

So you’re right to note that Micron had, in the past, invested in HMC technology. And that, of course, gave us tremendous amount of packaging expertise, the market moved to HBM. And Micron had HBM2 product, but at limited levels. And we chose to leapfrog to HBM3E with industry-leading specifications. And now we have industry-leading HBM3E product that has a 10% to 15% better performance than any other competitive HBM3E product offering that is being contemplated out there.

It has more than 25% better power as well, while having 10% to 15% better performance, and that combination is extremely important in AI applications. So we have a great product, well received by our customer base. It is, as I mentioned earlier, in qualification stages. Post the qualification, we will begin to ramp up production in early calendar year 2024 time frame, and that production will be ramping through the course of the year, second half being more heavily weighted in terms of revenue shipments versus the first half.

In our fiscal year, we do — as we have said before, we plan to have meaningful revenue from HBM3E product. And in 2025 time frame, we should have our HBM 3E share that would be in line with our DRAM industry share. We are confident in our product is proceeding well. Of course, qualifications have to complete and we are very much focused on preparing for production ramp-up.

Tim Arcuri

Great. And can you talk also just about your position in NAND. We’ve talked about DRAM. But can you talk about NAND now? Obviously, there’s another memory company that’s thinking about spinning off an and asset, you used to run that business. So can you talk about your position? And do you feel like you have enough scale. Do you feel like you have to build scale? How do you see your position in the NAND market? And also how it dovetails with your position in DRAM evolving over time?

Sanjay Mehrotra

While our share in NAND is approximately 12 percentage points, we have a strong position with respect to technology. We have industry’s most advanced technology 232-layer that’s in high-volume production, applied across multiple end market product segments. And of course, our NAND, not only do we lead well, I mean, almost all of our NAND today, as we have said before, is in our 176-layer and 232-layer NAND all of our products.

We have strengthened our product portfolio over the course of time. We have strong position in mobile NAND products, managed NAND products. These are multi-chip products that involve DRAM as well as NAND in them. So having DRAM and NAND is a core plus for us, and that has enabled us to have a strong position.

A few years ago, Micron really did not have any position in managed NAND solutions. And today, we have a strong share position with managed NAND solutions in mobile. As we discussed in our September earnings call, calendar Q2 gave us the highest share that we have had in client SSD as well as in data center SSD, so strong momentum in both clients as well as in data center SSDs with the NVMe SSD portfolio that we now have, leveraging our 176- and now 232-layer NAND technology. So our technology is well positioned, our products across the end market segments from mobile to client to data center SSDs are now well positioned. We certainly target to be gaining more share in data center SSDs. Now that the market is turning as well as our product portfolio is well positioned, Micron also has had a strong position with QLC. We have been industry leaders with QLC technology and we last quarter had reported record QLC shipments for the company as well.

So strong overall technology, product position, strong customer relationships, our customers value us having both DRAM as well as NAND solutions. And of course, as I mentioned, some of the products such as mobile products require DRAM as well as NAND in them. So we are pleased with our position. And of course, NAND and DRAM, they have synergies, our manufacturing synergies, procurement synergies as well as sales synergies.

Tim Arcuri

So I guess the message is that sure you’d like to have more scale, but you don’t feel like you need to have more scale?

Sanjay Mehrotra

We are pleased with where we are. And as we have always said, our goal is to not be chasing share, but instead really increasing the mix of the products and our sales towards what we consider higher values products and solutions. And this is the strategy we have been successfully executing to over the course of last few years. And with the results we are showing, as I’ve discussed, client SSD as well as data SSD reported — third-party reported market share in FY Q2, I think it shows that we are gaining momentum in terms of improving the product mix. And that’s what our focus is to have our revenue share outperform our bit shares for NAND.

Tim Arcuri

Great. Thank you. So let’s talk about China for just a couple of minutes. Lots to unpack there. You sized the CAC band at low double-digit per percent of your revenue. But I think you did meet with the Commerce Minister in Beijing earlier this month, and I think you’ve committed to investing, I think, $600 million in a Xeon packaging site. So there’s a lot of cross currents there in China. So can you just talk — and then also, we have Swisher and we have CXMT who are ordering equipment handover fit. So can you talk about just the dynamic in China and sort of whether we’ve set a lowest bar possible in China, and it’s possible that you could get back into China.

Sanjay Mehrotra

So China is an important market, clearly for us. And yes, we have been impacted by CSC decision regarding use of our products, particularly in the data center and the networking markets in China, and that impact has showed up in our F Q4 results. And of course, as part of our F Q1 guidance that we have provided to you as well. As we have said before, we continue to work on mitigating the revenue loss as a results of the CAC decision with shoring up our share in other parts of the market that are not impacted by CAC decision in China as well as with our international customers.

We have said before that we are doing a very good job in terms of executing to that strategy of maintaining our global bit share, both in DRAM and NAND by making sure that we are able to increase that share with other customers and make up for the share losses that are occurring in China. So it takes a while for that strategy to fully play out, but we are very encouraged by our engagement and by our response from our customers on that front. I believe our team is doing a very good job in that regard.

We, of course, remain fully engaged, fully committed to China with respect to supporting our customers there who can purchase our products, who value our technology, who value our product, to value our quality as well as delivery and manufacturing capabilities. So engaging with the customers is important. Our operations in China, assembly and test operations, you noted that we have announced investments to increase the mix of the products that we will be making in China to support the needs of the customers in China as well as other multinational customers operating out of china. And that investment requires Chinese approvals, and we are in the process of gaining those approvals. So it remain fully committed to our operations, to our customers and to improving our relations with the government as well. I have now visited China twice last visit that you just mentioned took place few weeks ago, twice this year, and we are continuing to make good progress with our meetings with the government officials there as well.

Tim Arcuri

Great. Thanks. And how much of a factor are the local Chinese DRAM and NAND makers? How much of a factor are they in the market? Do you see them from a competitive perspective? Do you see CXMT, do you see in YMTC. And are you, at all, concerned about just the sheer dollars that they’re spending. You can look at what AMAT’s reporting and what KLA is reporting and what LAM is reporting, and you can add the numbers up and you can get to a pretty big number for what China is spending in memory right now and in particular, DRAM. Does that concern you? How do you kind of think about all that?

Sanjay Mehrotra

So I think various geopolitical considerations, export control regulations that are out there. It can have some impact on some of this, what can be made in China or cannot be made in China. But clearly, we do always have a sign that there will be some part of the memory market that will be supported, particularly within China, by the Chinese memory manufacturers. Our focus, of course, remains on continuing to advance our technology road map and you see us as clear leaders with having 1-beta technology in DRAM as well as 232-layer NAND in production. And that’s our focus.

I think that’s what we remain totally focused on. Of course, Chinese suppliers will have also some share of the market, particularly in China, but important focus going forward will be on leading-edge technologies on leading-edge products such as HBM, such as data center SSDs. And we believe that our — as long as we continue to execute to our strategy well, we are well positioned to address the market growth that is ahead of us for both memory and storage.

Tim Arcuri

So based upon what you see, you see the same dollars that they’re spending, you’re not worried that at some point, they become disruptive on the supply side?

Sanjay Mehrotra

I think the name of the game is going to be to continue to be in the forefront of technology. And that’s what Micron’s focus is, of course, staying at the forefront of technology while managing our CapEx while carefully managing our supply growth as well.

Tim Arcuri

Great. Mark, I wanted to ask you a question about costs. They were down quite a bit in fiscal Q4 on mix. Fiscal Q1 sound — cost sounded like they were going to be flat and maybe even a bit off that tough comp last quarter. Can you talk about how costs — I mean, the quarter’s isn’t and over yet, but can you talk about how cost came in and maybe how we should think about the cost curve throughout the rest of this fiscal year and into calendar 2024.

Mark Murphy

Sure. I think at the outset, what I’ll say is that because Micron’s on the leading node we have a cost advantage in the sense that we’re getting greater density, and that helps us maintain our competitiveness. The whole industry is being impacted by the fact that utilization levels are down. And then as Sanjay mentioned, as tools are redeployed to the leading edge, the wafer starts are going down and that in and of itself, you get both utilization effects and then eventually, that transitions into lower wafer starts, and that’s going to weigh on cost downs going forward.

And then the trend that you’re going to see increasingly is that effect kind of wanes and volumes eventually increase in the future, an increase of HBM mix will come in. And that, as Sanjay mentioned, that’s more silicon, that’s more cost. Now of course, with HBM, there’s more price and you get paid for that on both the direct cost and the capital risk. So that effect will dampen cost downs as well. And again, that’s an industry-wide phenomenon. So we expect cost downs to be dampened through 2024. And then beyond that, improvements fab efficiency improves and HBM mix stabilizes.

Q – Tim Arcuri

And you would still say that down mid — let’s say, low teens is probably the right longer-term cost curve in DRAM? Or do you think you’ve been getting to low teens? Are we in a world where DRAM cost downs are going to be sub-10?

Mark Murphy

Yes. I think what we’ve said is high single digit for DRAM. And I think we — again, what’s important is that we’re competitive on technology, and these effects of utilization and HBM will just be clear when we, for example, talk at earnings about it in more detail.

Tim Arcuri

Great. Well, we’re just about out of time. This is a great way to kick off the conference. Thank you, Sanjay and Mark.

Sanjay Mehrotra

Thank you.