David Goeckeler: Yes. So on the first question, the business with our customers is planned pretty far in advance. So there wouldn’t be a situation where something would happen intra-quarter and that would drive a big share shift. The reality is we’ve got great products, and they’re very much resonating with our customers and we can deliver them at scale. And they have bring best-in-class TCO and clearly customers are adopting those at a significant rate. So — like is it sustainable? We continue to bring great products to market. That’s what we plan to do. We’re very confident in our road map on HDD and we’ll continue to bring the best TCO solutions to our customers. On bit growth, we do expect flat bit growth into the calendar Q2, but we’ll see a pickup in bit growth in the second half of the year.
Amit Daryanani: Got it. Thank you.
David Goeckeler: Thank you.
Operator: Our next question comes from Harlan Sur from JPMorgan. Please go ahead with your question.
Harlan Sur: Good afternoon. Nice job on the quarterly execution. Another question on enterprise is so you guys have been really smart on how you are allocating flash bits, right, with a strong focus on profitability. So as you reallocate more bits towards eSSD in the second half, is the profitability profile of enterprise SSD portfolio expected to be accretive to the overall flash business and your shares peaked previously sort of in that sort of high single-digit percentage range in enterprise. Just given a more competitive portfolio, like what type of share is the team targeting kind of mid- to longer term?
David Goeckeler: Okay. Thanks, Harlan, your questions are very related. So we saw a pickup in enterprise SSD in the March quarter. It’s still — quite honestly, it’s still relatively small numbers, but it’s growing quite well. So we wouldn’t have supplied those bits if it wasn’t the right thing to do from a portfolio strategy point of view. We’ll see when we get to the second half, what pricing looks like, that versus other options we have, and then we’ll decide how much supply we put into those products. And you’re really getting into core of our portfolio strategy, which is to have a lot of optionality. We have a lot of optionality across client SSD, across gaming, now across enterprise SSD, across mobile, across consumer, obviously, which is a big business for us.
And then based on what we see going into the quarter. And then very importantly, what happens during the quarter, how do we allocate our supply to get the best return? And clearly, we’re in an environment right now where things got better throughout the quarter. So as we go through the quarter, we find more opportunity to mix and get more profitability, and that’s what happened in the March quarter, and you saw the results of having that agility into the business. So I really don’t want to call a share number or anything like that because it tends to distort, what we want to do is maximize profitability, not maximize share in any particular market. We want to maximize where we get the most return for our supply.
Harlan Sur: I appreciate that. And then maybe a question on BiCS 8. I know you’re not calling out any timing yet, but you have had it sort of in preproduction for quite some time. How are the early yields on this technology? And I guess, more importantly, like can the team still drive mid-teens percentage annualized type cost down with the new bonded the raise technology?
David Goeckeler: Yes. So what I’ll say about yields is we’re very confident in the technology. I mean, we feel very, very good about it. It’s a major advancement in the architecture OptiNAND from an industry perspective to the CBA architecture. And it’s — the development has gone well. We feel very good about it. We can productize it when we need it. Again, this gets into a larger conversation about the dynamics of the market and when is the supply needed, and we’re going to be very, very disciplined about going through any transition or putting any CapEx to the market until we see the profitability that we want to get. So we feel very good about BiCS 8. There was a second part of the question.
Harlan Sur: On the cost downs.
David Goeckeler: Cost downs.
Wissam Jabre: Yes, maybe I’ll take that, Harlan. Yes. On the cost down, we’re still anticipating the mid-teens percentage year-on-year cost downs. So there’s no change there.
Harlan Sur: Perfect. Thank you.
David Goeckeler: Thanks, Harlan.
Wissam Jabre: Thank you.
Operator: Our next question comes from Carlos Colorado from UBS. Please go ahead with your question.
Carlos Colorado: Thanks for taking my question. So I have — the first one is about nearline. You are going outperforming your competition by a lot. So what are the underlying reasons in your opinion for this? And do we have to expect this to normalize over time? And do you think this can be sustained? And I have a follow-up. Thanks.
David Goeckeler: Yes, the performance of the HD business is driven by the product, right? It’s pretty straightforward. Products are — they’re great products. This architecture that we built on ePMR, OptiNAND, UltraSMR, customers are really committed to SMR. They deliver the best TCO in the market. We can produce them at scale and that’s what leads to the performance.
Carlos Colorado: Okay. Thanks and the follow-up is, you mentioned that AI has revenues are of SSD sales. You have a perfect vantage point to see if AI is driving applications that traditionally were HDD is that now being transferred to SSD some of those applications, or is the classic question on cannibalization from one to the other. Is that — is AI change in that scenario? Thanks.
David Goeckeler: We do not see any cannibalization clearly, HDD plays a big role in the AI storage life cycle as well as the whole ingest phase because all of the big data lakes and all of the raw data sets, those are all going to be stored on HDD. It’s just the economics of where you store that data and how do you access that data. It’s all that part of the AI pipeline, if you will, is going to be HDD. Now you have all of these other new use cases around training and inference, and those are all going to be SSDs. So it’s really about growth as opposed to substitution. And that’s what’s so exciting about this. And obviously, once you get the models trained, then the models are going to turn out more data, which is going to be stored on HDD.
So you got this virtuous cycle going. So it’s kind of literally rising tide lifts all boats. It’s not a substitution game. Clearly, there’s a lot of new use cases being developed around AI, like the whole training infrastructures that are being built, that’s what’s driving these very high-capacity storage-based enterprise SSDs that we’re seeing demand for. So hopefully, that helps.
Operator: Our next question comes from Krish Sankar from TD Cowen. Please go ahead with your question.
Krish Sankar: Thanks for taking my question. I have two of them. First one on flash for Dave. You spoke about the BiCS 6 hyperscale qualifying it. My understanding was a BiCS 6 was kind of more like a sub node and BiCS 8 is going to be the bigger one. I’m just kind of curious to get to your enterprise SSD market share target. Do you really need BiCS 8, or can you achieve it with BiCS 6? And then I have a follow-up.
David Goeckeler: You’re right. BiCS 6 is when we say stub node, it’s we’re not going to take the whole portfolio to BiCS 6. So we have a big portfolio, and we’re choosing which products to take the BiCS 6. And clearly, we’re taking the products that require QLC and the kind of things you’re talking about. So we feel good about our node a plan in the fab being able to supply what we need in these markets.
Krish Sankar: Got it. Got it. And then, Dave, on the hard drive side, I think you said in the past that you can get to 40 terabytes of the ePMR technology. I’m just kind of curious with obviously a competitor like trying to ramp up HAMR, and it took them a while like a few years to even get the 3 terabytes per disk in R&D to [indiscernible]. Can you give us an update on your HAMR road map or the status of your HAMR technology, how we think about 30, 40 terabytes plus.
David Goeckeler: So we’ve been working on HAMR for quite some time. We understand HAMR extremely well. We understand all the issues with HAMR, and what it takes to get it qualified. Clearly, we’re doing that all behind the scenes because we have a product portfolio with the best TCO we can offer in the market today, and we can do that all the way up to 40 terabytes. And 40 terabytes is where the economics flip over and you get the 4 terabyte per platter or 40 per unit, where essentially the capacity increase will cancel out the increase in costs you have to put in the unit to get the economics to work on margin, right? That’s kind of a complicated — a lot to say in 1 sentence. But our portfolio is very focused on the right product with the right cost at the right time.
The right time for HAMR is at 40 terabytes. And we’ve got a lot of development going on that product. We have for a long time. We, quite frankly, don’t need to do it in public because we have another portfolio that’s selling extremely well, which we’ve talked about throughout this process. But have a lot of confidence in our HAMR development. And quite frankly, our customers know exactly what we’re doing, and where we’re at and what our plans are, and they’re comfortable with that as well.
Krish Sankar: Thanks, David.
Operator: And our next question comes from Tom O’Malley from Barclays. Please go ahead with your question.
Tom O’Malley: Thanks for taking my question. I’m going to do one on the CFO side real quick on OpEx. So big step up in the June quarter, and you’re talking about some special projects. How should we think about that progressing? Is that investments that are going to stick around for the next couple of quarters? Or should that reset back to kind of the lower base you’ve been running out? You’ve just seen OpEx move from kind of the 550 to 660 over the past year, obviously, revenue increasing as well, but any color there on what that investment is for and if you see a step down after that?
Wissam Jabre: Yes. Sure, Tom. So let me first start by saying that the way we think of OpEx is we don’t see OpEx increasing faster than revenue. So we’re still very focused on that cost discipline and OpEx discipline. When it comes to this quarter, we’re expecting some increase. The increase is almost 50-50 driven by variable comp as the financial outlook has improved much faster than anticipated. So there’s a bit of increase there. But also, as you mentioned, there’s some project-specific R&D investments that also — that we have sort of a direct correlation and line of sight to revenue. I would say for the next couple of quarters, the range that we’ve guided for Q4 is a reasonable range. I know it’s too early to talk about fiscal year ’25. But for modelling purposes, we can use the same type of numbers for now.
Tom O’Malley: Helpful. And if I look at your cost guidance for the year, a couple kind of with what you’re looking at for June of ’24, when I’m looking at gross margins, it seems like you need to have a pretty significant step-up in HDD gross margins. Are you planning for all of that the underutilization to come out of the model in the June quarter? And if any remains, can you let us know how much you’re expecting?
Wissam Jabre: So for this most recent — for Q3, we had a little bit — and we disclosed, we talked about those. But as you can see, the numbers are becoming less and less significant. And so for the June quarter, there’s still a little bit, but it’s not really very significant for us to talk about on this call.
Tom O’Malley: Thank you.
Wissam Jabre: Thanks, Tom.
David Goeckeler: Thanks, Tom.
Operator: Our next question comes from Vijay Rakesh from Mizuho. Please go ahead with your question.
Vijay Rakesh: Yes, hi. David, Wissam. Just a quick question on the Flash side. Dave, when you look at the profitability, as you mentioned, how does the BiCS 6 compare to — if you look at some of the competitive NAND out there, either in terms of die size or cost per gig versus some of the peers?
Wissam Jabre: So that’s a very complicated question. I mean we can go into it in detail offline. We obviously do tons of work, and I appreciate your question that it’s a multifaceted issue. It’s die size, it’s memory hole density, it’s all kinds of very complicated thing goes into producing a NAND product. Look, we think the product compares extremely favorable. We think it leads the market. Again, for the last — looking back many years, we have been able to produce bits at one third less CapEx than the industry average, and we expect BiCS 8 to continue that leadership in the market. So we feel very, very good about the product, about its performance. Again, when you build, this is like kind of one of the magic of wafer bonding.
