Sandisk - Earnings Call - Q4 2025

Executive Summary

• Q4 2025 revenue was $1.901B, up 12% q/q and 8% y/y, with non-GAAP EPS of $0.29; both exceeded the prior quarter’s guidance, driven by better-than-expected bit growth and early pricing actions.
• Gross margin (non-GAAP) improved to 26.4% from 22.7% q/q; excluding $51M underutilization and $42M fab start-up costs, non-GAAP GM would have been 31.3%, indicating embedded margin tailwinds as BiCS8 ramps and start-up costs step down over the next two quarters.
• Q1 2026 outlook issued: revenue $2.10–$2.20B and non-GAAP EPS $0.70–$0.90; subsequent Q1 2026 actuals significantly beat with $2.308B revenue and $1.22 non-GAAP EPS, reinforcing demand strength and pricing power.
• Strategic catalysts: ramp of BiCS8, enterprise SSD traction (QLC, PCIe Gen5/6), and High Bandwidth Flash (HBF) roadmap with ecosystem partnership (SK Hynix), positioning SNDK for AI-driven demand in data lakes and inference workloads.

What Went Well and What Went Wrong

What Went Well

• Q4 revenue and non-GAAP EPS exceeded guidance; “Sandisk delivered strong results this quarter, with revenue and non-GAAP EPS exceeding our guidance” — CEO David Goeckeler.
• Mix/pricing tailwinds: implemented price increases and managed supply to undersupply, supporting margins; non-GAAP GM up 370 bps q/q and would be ~31.3% absent underutilization/start-up costs.
• Strategic progress in datacenter: bits shipped to datacenter >12% of total; unveiling 256TB NVMe enterprise SSD on Ultra QLC, with Tier-1 customer qualifications underway, aligning with AI data lake needs.

What Went Wrong

• Non-GAAP GM remains below prior year due to nodal transition costs and start-up impacts; Q4 non-GAAP GM 26.4% vs 36.4% y/y (down 10 ppt).
• GAAP profitability pressured: Q4 GAAP net loss of $23M; FY25 GAAP results include earlier goodwill impairment ($1.83B in Q3), highlighting volatility during separation and transition.
• Underutilization and start-up costs still a headwind in Q4/Q1: ~$51M underutilization and ~$42M start-up costs in Q4; guiding ~$60M start-up costs in Q1 before step-downs in Q2/Q3.

Transcript

Speaker 2

Good afternoon and welcome to SanDisk's fourth quarter fiscal year 2025 earnings conference call. All participants are in listen-only mode. Should you need assistance, please signal a conference specialist by pressing STAR, then zero on your telephone keypad. After today's presentation, there'll be an opportunity to ask questions. To ask a question, you may press STAR, then one on your telephone keypad. To withdraw your question, please press STAR, then two. Please note this event is being recorded. I would now like to turn the conference over to Ivan Donaldson, Vice President of Investor Relations. Please go ahead.

Speaker 3

Before we begin, please note that today's discussion will contain forward-looking statements based on management's current assumptions and expectations, which are subject to various risks and uncertainties. These forward-looking statements include expectations for our technology and product portfolio, our business plans and performance, market trends and opportunities, and our future financial results. We assume no obligation to update these statements. Please refer to the registration statement on S-1/A and other filings with the SEC for more information on the risks and uncertainties that could cause actual results to differ materially from expectations. We will also make reference to non-GAAP financial measures today. Reconciliations between the non-GAAP and comparable GAAP financial measures are included in written materials posted in the Investor Relations section of our website. I'll now turn the call over to David.

Speaker 0

Thanks, Ivan. Good afternoon and thank you for joining SanDisk's fourth quarter fiscal year 2025 earnings call. SanDisk delivered another strong quarter with revenue and non-GAAP earnings per share exceeding guidance. Revenue for the fourth quarter was $1.9 billion, up 12% over the prior quarter, with both BIT shipments and ASPs up mid-single digits. Non-GAAP earnings for the quarter were $0.29 per share. During the quarter, we reduced our inventory from 150 to 135 days and reduced our net debt to $368 million. We remain on track toward reaching our goal of becoming net cash positive. In our fiscal fourth quarter, we estimate that overall demand exceeded supply, which we anticipate to continue through calendar year 2026. Data center demand remains strong, supported by robust capital investment from leading cloud providers. In the client end market, growth was primarily driven by rising average capacity across mobile and PC markets.

Overall, we close fiscal 2025 strong, and while we are optimistic about the future for fiscal 2026, we will continue to manage the business prudently. Fiscal year 2026 marks a pivotal transition as Bixate becomes our prominent node. We've made strong progress ramping up this industry-leading node into high-volume manufacturing. Bixate delivers best-in-class performance, density, and power efficiency capabilities that are critical across a wide range of end markets, including hyperscale data centers, PCs, mobile, and gaming. While the nodal transition brings above-average capital intensity and below-average cost reductions in the near term, we expect this to be a year of significant financial improvement, with both expanding margins and cash generation, as macro headwinds subside and the demand and supply dynamics remain favorable. Our leadership with Bixate reflects years of focused investment and disciplined engineering execution, and we are well positioned to continue and extend that lead.

Moving on to key business highlights. In the fourth quarter, data center represented over 12% of our total BIT shipments, a meaningful milestone as we scale in this critical part of the market. We continue to make steady progress in both storage enterprise SSDs intended for fast AI data lakes and compute enterprise SSDs intended for compute-heavy applications. We are proud to announce a 256-terabyte NVMe enterprise SSD powered by our industry-leading Ultra QLC platform, which we unveiled at the Future of Memory and Storage conference. With this product, we are setting a new benchmark in performance, capacity, and efficiency for AI-driven workloads. This breakthrough reflects our deep commitment to helping customers accelerate their most demanding data challenges with scalable next-generation flash solutions. Our plan is to qualify our high-capacity Ultra QLC platform at several major Tier 1 customers by the end of fiscal year 2026.

On the compute enterprise SSD front, we are encouraged by the progress we have made in qualifying solutions with key customers, including an ongoing qualification with the second major hyperscale cloud provider and other customers using the NVIDIA GB300. Customer feedback has been consistently positive, particularly regarding TLC's performance and efficiency in high-throughput read-intensive applications. Expansion into the data center space requires sustained effort, given its long qualification cycles. We are progressing in line with our expectations in enterprise SSD, and our strategy is anchored in the adoption of QLC-based SSDs and PCIe Gen 5 and Gen 6 solutions with our Ultra QLC platform, the technology designed to support the scale and density of fast data lakes needed for AI and other data-intensive workloads. Cloud infrastructure spending remains significant, with industry analysts estimating major U.S. hyperscale CapEx growing 47% year-over-year to $368 billion, alongside rising investments in Asia and Europe.

Our full-stack portfolio, from advanced NAND components to high-capacity SSDs, is well-aligned with the evolving needs of cloud and AI infrastructure customers. With Bixate scaling, continued traction in PCIe Gen 5 and QLC, and accelerating global AI workloads, we are confident in our ability to drive sustained growth in this end market and earn our fair and growing share in this fast-evolving space. In client, Bixate SSDs are now qualified across all major PC OEMs, with additional qualifications underway. As the market transitions towards QLC, our differentiated CMOS bonded array architecture and system-level capabilities give us a competitive advantage. In addition, in mobile, we secured shipment approvals from key customers for our next-generation storage solutions, targeting the premium portion of the market. Our continued innovation in flash-based storage is driving positive customer response across channels. In consumer, we continue to strengthen the SanDisk brand through differentiated product innovation.

Our Bixate node technology is showing healthy consumption, supported by solid channel sell-through and seasonal strength so far this calendar year. Meanwhile, our new SanDisk USB4 portable SSD is recognized as one of the fastest solutions on the market, delivering exceptional performance and reliability for content creators, gamers, and everyday users. We are deepening strategic partnerships with industry leaders, including Nintendo and Xbox. Our co-branded SanDisk Creator microSD Express card for the Nintendo Switch 2 is gaining adoption, and the new C50 expansion card for Xbox reinforces our leadership in gaming storage. We are also expanding into emerging high-growth markets with targeted solutions, including a new high-performance USB drive designed for DJs and creative professionals. These efforts reflect our commitment to product innovation, which continues to drive positive customer response across high-value, growing consumer use cases.

Looking beyond our current product portfolio, we are also investing in innovations that will redefine the future of memory and storage technology. During our Investor Day in February, we unveiled our high-bandwidth flash (HBF) memory technology, which we trademarked as HBF. We are proud to announce that just last week at FMS, HBF received the Best of Show award in the Most Innovative Technology category. We have made significant progress advancing this groundbreaking technology over the last six months. First, we established a technical advisory board, including industry experts and senior technical leaders from within and outside of SanDisk. Second, we formed an ecosystem partnership with SK hynix to standardize HBF technology specifications. Third, we announced that we expect to have HBF technology available by the second half of calendar year 2026 and product samples, including the controller device, in the first half of 2027.

We are very optimistic about this technology and the opportunity it is expected to create as AI continues to proliferate. With that, I'll turn it over to Luis for our financial results and guidance.

Speaker 1

Thank you, David. Let's start by diving deeper into the quarter financials. Revenue for the fourth quarter was $1,901 million, up 12% quarter-over-quarter and 8% year-over-year. Sequentially, BIT shipments and average selling prices were up mid-single digits. Our performance was above our guidance range of $1,750 million to $1,850 million, primarily from better-than-expected BIT shipments growth. For fiscal year 2025, revenue was $7,355 million, up 10% from fiscal year 2024. Cloud revenue for the fourth quarter was $213 million, up 8% sequentially and 25% year-over-year. Client revenue was $1,103 million, up 19% sequentially and 3% year-over-year. Consumer revenue was $585 million, up 2% quarter-over-quarter and 12% year-over-year. Non-GAAP gross margin for the fourth quarter was 26.4%, within our guidance range of 25.5% to 27% and up 370 basis points from the prior quarter.

Our non-GAAP gross margin for the fourth quarter includes $51 million in underutilization charges and $42 million in fab startup costs. Excluding these charges, our non-GAAP gross margin would have been 31.3%. Non-GAAP operating expenses for the fourth quarter were $402 million, in line with our guidance range of $395 to $405 million. Fourth quarter non-GAAP earnings per share were $0.29, above our guidance range from a loss of $0.10 to a profit of $0.15. The beat was primarily from the additional revenue flowing through to the bottom line. Key GAAP to non-GAAP reconciliation items include stock-based compensation, which represents 2.6% of revenue, $17 million in separation charges, and $16 million in restructuring charges, as we reduced our workforce by approximately 200 employees. Moving on to the balance sheet, we closed the quarter with $1.5 billion in cash and cash equivalents.

During the quarter, we reduced our inventory days from 150 to 135, as demand exceeded supply in line with our strategy. During the quarter, we also made our first quarterly $5 million term loan B payment and prepaid an additional $95 million, reducing our long-term debt to $1.8 billion. For perspective, we prepaid an additional $100 million of the term loan B after the fourth quarter closed. These prepayments reflect our confidence in our future cash flow expectations, which includes funding our Bixate investment. We closed the fourth quarter with 147 million fully diluted shares. Moving on to free cash flow. During the quarter, we generated $77 million in adjusted free cash flow. This included $94 million from operations and $28 million cash received from our activities related to Flash Ventures, partially offset by $45 million invested in our backend operations and offices.

The $28 million received from our operations related to Flash Ventures includes $343 million in gross capital spending, with $109 million funded through depreciation as part of our cost of goods sold and $262 million funded from external sources, mainly subsidies and equipment leasing from the joint venture. Before moving on to guidance, I want to reinforce that our goal is to create value for our customers and shareholders. We will continue to innovate to drive performance, density, and power efficiency while managing supply in line with demand. This includes adjusting wafer starts, underutilizing fabs when needed to align with demand, and planning capacity based on expected demand growth and node productivity. In the fiscal fourth quarter, we also began implementing price increases. Several of our products are on allocation, and we expect mid-single-digit undersupply for the fiscal 2026, supporting further price increases.

We anticipate our BIT shipments growth in fiscal 2026 to be consistent with broader demand growth. We will continue to adjust supply based on market conditions, with an emphasis on driving healthy and sustainable profitability levels. With that, let's move on to the first quarter guide. We expect revenue for the first quarter to be $2,100 million to $2,200 million. The midpoint of the revenue guide is up 13% growth quarter-over-quarter. We expect revenue growth to come from BIT shipments growth and higher average selling prices, with similar contributions from both drivers. We expect non-GAAP gross margin for the first quarter to be between 28.5% and 29.5%, which includes $10 million to $15 million in underutilization charges and approximately $60 million in fab startup costs. At midpoint, non-GAAP gross margins are expected to increase by another 260 basis points quarter-over-quarter, mostly from higher pricing.

We expect fab startup costs to reduce significantly as of the third quarter. We expect non-GAAP operating expenses for the first quarter to be between $415 million and $430 million. Non-GAAP interest and other income and expenses to be between $40 million and $45 million, and non-GAAP taxes to be between $35 million and $40 million. Higher operating expenses compared to the prior quarter are due to an additional week in the quarter and one-time cost associated with qualification samples for our expanding product portfolio. Combined, the impact of these two costs is between $20 million and $25 million. As a result, we expect non-GAAP EPS to be between $0.70 and $0.90 based on 148 million fully diluted shares. Free cash flow for the quarter is expected to be positive. This includes an increase of gross capital spending to support our Bixate program and a reduction in our inventory days.

As a result, we expect our net debt to continue to decline. With that, let me turn the call back to David.

Speaker 0

Thanks, Luis. In conclusion, as we enter fiscal year 2026, we see momentum in our product portfolio, an improving supply and demand environment, and early benefits from our pricing actions. We are evolving from a model driving elasticity-based TAM growth to one focused on creating value from our innovation in combination with disciplined capacity management. Ultimately, the steps we are taking today are not just supporting quarterly financial improvements. They are laying the groundwork for a stronger, more resilient SanDisk, built for long-term success and value creation for our shareholders, as well as enabling us to drive value for customers long term. With that, let's open the call for questions.

Speaker 2

Thank you. We will now begin the question-and-answer session. To ask a question, you may press STAR, then one on your telephone keypad. If you are using a speakerphone, please pick up your handset before pressing the keys. If at any time your question has been addressed and you would like to withdraw your question, please press STAR, then two. At this time, we will pause momentarily to assemble our roster. Our first question comes from CJ Muse of Tantor Fitzgerald. Please go ahead.

Yeah, good afternoon. Thank you for taking the question. I guess I was hoping you could spend a little bit of time giving more depth on gross margins. Obviously, great guide, 260 bps higher, but I think less than expectations. If you could kind of walk through underutilization, cost downs, mix, anything that could be helpful. Thanks so much.

Speaker 1

Yeah, thank you, CJ. As you saw, underutilization is moving lower and lower. Now we are forecasting it to be between $10 million and $15 million in the quarter, which is significantly down from what we had in Q4. Really, the biggest impact on gross margin this quarter is startup costs, right? Startup costs are somewhere around $60 million, which continues to impact us in the quarter. If you look at both together, underutilization and startup costs for the quarter, it's about 300 basis points. The good news is they will go away, right? These costs are coming down. They will be reduced next quarter and will be significantly reduced in Q3. We'll continue to see those 300 basis points kind of flow into our gross margin as we go forward.

Very helpful. Thanks.

Speaker 2

Our next question comes from Joe Moore of Morgan Stanley. Please go ahead.

Great, thank you. I wonder if you could give us a little bit more color on the high-bandwidth flash partnership. I guess, you know, what's the thought process behind partnering and how do you extract the most value working with SK hynix on that?

Speaker 0

Hey, Joe. Good to hear from you. Yeah, we're very excited about this. We've been working on this technology for a while. I think we are all looking at the AI architecture and understanding what are good ways to optimize it. We view it as a memory-bound problem. We've been looking for how we can bring flash to this equation with our scalability roadmap. As we talked about at our Investor Day, we think we've found a way to design flash where we can overcome the bandwidth or achieve the bandwidth requirements that we need. We've been driving this forward. What we really want to do is drive this as an industry standard.

When SK hynix contacted us and had the same goal of saying, hey, let's see how we can take this technology and make it a standard, I think anything in this industry that has this broad of applicability, you know, this is a technology that can play from the edge, so PCs, smartphones, all the way into the cloud for inference. We think it's a new paradigm for how inference is driven. The ability, all customers are going to look for that as an industry standard. We think that's how we move this the fastest and how we drive it forward. We've got a controller to design and all the interfaces to standardize around that. Doing it with an industry partner, we think, is the right way to go to move this along and drive adoption as quickly as possible.

Quite frankly, we think the industry is going to need this kind of capacity in the memory architecture to drive inference at scale. That's a little more background for you.

Great. Your confidence in this kind of being commercially viable next year from a sampling perspective, how speculative should we view these investments as being at this point?

I mean, Joe, what we said is next year we'll be sampling the die for the NAND, and then early 2027, we'll have the controller that goes along with that. We've been building NAND die for 25 years now, so we have a lot of history in that. It is a development project, so there's always uncertainty and unknowns, but we know how to manage through those. I think one of the things folks have been asking us since our Investor Day was to put some timelines around this and give us a better idea of how this technology is going to evolve. We've seen a lot of progress since then over the last five months, so we feel very comfortable putting those dates out, and we've got teams working aggressively towards those milestones.

Great, thank you.

Thanks, Joe. Appreciate it.

Speaker 2

Our next question comes from Aaron Rakers of Wells Fargo. Please go ahead.

Yeah, thanks for taking the question, guys. I guess I want to go back to CJ's question a little bit, appreciating the inputs in the gross margin. I know, Luis, you had mentioned that the startup cost of $60 million looks like it's up from the $42 million this last quarter. You did suggest that that should decline significantly going forward. Any kind of glide path of how we should think about that $60 million as we look out over the next couple of quarters? Also tied to the gross margin, if I back those items out, I adjust, it looks like your rate of cost down relative to the Bixate ramp. I'm just curious, that seems to be a bit of a headwind. When do we start to see maybe that cost curve start to normalize and you start to see the benefits of that cost down from Bixate?

Thank you.

Speaker 1

Yeah, thank you for the question. What you should expect is, call it two steps as we go down from the startup cost, a significant step down next quarter, and then even a further step down in Q3, in our fiscal Q3. Expect two steps to get very close to minimal impacts on startup costs. That's kind of what you should expect there. Underutilization, expect unless the dynamics change in the market, it should be pretty close to zero going forward. As you saw, our inventory levels are coming down. I also mentioned that we're seeing some of our products are actually on allocation because we're seeing a good market. As David mentioned, we see an undersupply situation there. We're feeling good there. In terms of ongoing savings, at the end of the day, the way you bring cost per gigabyte down is through node transitions, right?

The node transition that we're going through is as we implement Bixate. We mentioned that in Q4, we closed about 7% of our BIT shipments were Bixate. We will be somewhere between 40% and 50% by the end of fiscal year 2026. That's when you will start to see the impact and the benefit of that transition throughout the year as we continue to change our mix from Bixate 6 towards Bixate. Hopefully, that helps, Aaron.

Yeah, that does. Real quickly, you mentioned growing BIT growth consistent with the broader demand of the market. Can you just remind us of what that growth is, the broader market demand?

Speaker 0

Aaron, we see 2025, we see demand low double digits, very low double digits with supply under that. We see that moving up to mid to high double digits in calendar year 2026, but we still see demand below that. We mentioned in the prepared remarks that we continue to see this undersupplied market through the end of 2026. I think the last quarter played out as we saw it back at our Investor Day. We're optimistic on the second half. As we go into 2026, we still see, when you look at the whole year, production demand, we see more demand than supply. We feel very good about the market. Just to comment on your first part of your question, Luis did a great job with it. We're clearly in a state where we have some headwinds on the cost side.

We're going through a fab startup, which is a pretty big episodic event, which we drive through. We've been on BICS 5 for quite some time. It was still the predominant node last quarter. It's been a fantastic node for us, the highest yielding node in the history of BICS. Now we're going to go through a year here where we start with Bixate at mid-single digits % of the portfolio and roughly end with it as half. We're going to have that be a tailwind to the business as we mix into a more efficient node. We're going to have the startup costs, as Luis talked about, step down over the next couple of quarters. We're going to turn what have been headwinds in the business into tailwinds, and all of that against the backdrop of what we see as a favorable supply and demand environment.

Thanks, David.

Thanks, Aaron. Appreciate it.

Speaker 2

Our next question comes from Wamsie Mulhane of Bank of America. Please go ahead.

Yes, thank you so much. How are you thinking about the growth in client in the second half of the calendar year, given some of the dynamics of PC demand pulled forward and maybe some in smartphones as well that seem to have happened here in the first half?

Speaker 0

Hey, Wamsie. Yeah, we've seen very consistent behavior out of these customers, right? I mean, we've been seeing consistent demand. Maybe there was some pull ahead early in the tariff days, but the market has absorbed that. As we look into the second half, we see pretty consistent demand from the signals our customers are sending us. The overall backdrop is we see an undersupplied market, and we see consistent demand from those customers. We also believe that their inventory levels have been normalized. We don't see a lot of big distortions on the market here as we move through the second half.

You also just mentioned tariffs. Can you talk about how you're thinking strategically about tariffs on semis, particularly given that your fab footprint is kind of all Japan? How are you thinking about navigating that, given some of the comments we've heard lately?

Yeah, we're staying, obviously, we're staying very, very close to that. I mean, I think the NAND business in general is a very dynamic business. It's something you've got to stay very close to on a day-to-day basis. Tariffs are another thing that are part of that equation. We stay very close to that. We have conversations with all the right folks we need to to understand, get the best view possible of where this is all going to land. It is pretty dynamic right now. I think the big issue for us, of course, is the 232 tariffs that we still have yet to see. When we see those, we'll have more to say about how do we respond to it. We're very confident in our ability to navigate this whole situation with our global footprint.

OK, thanks, Dave.

Thanks, Wamsie. Appreciate it.

Speaker 2

Our next question comes from Vijay Rakesh of Mizuho. Please go ahead.

Yeah, hi, David and Luis. Just a quick question on the data center side. Obviously, it looks like you're ramping your 256-terabyte QLC SSD because your target family, I guess. Can you talk to how the roadmap looks? You obviously have a second hyperscaler you're qualifying. What's the target that you're trying to get to by the end of fiscal 2026 in terms of the number of hyperscalers or your mix of data center revenues? I think you mentioned 12% of BITs. How do you see that going?

Speaker 0

Yeah, I mean, our goal, you know, we've been pretty consistent. This is an area where we want to increase our optionality and be able to mix up at least to our fair share of the market, if not more. It's a market where there's long development cycles. We've got, you know, we have a previous generation of products in the market that we've been driving over the last several years that have been performing. Now we have the next generation of products coming out on the compute side. We've been talking about that for two or three quarters. We've got good traction with one hyperscaler. We're qualifying a second right now. To your point, we were very happy to announce at FMS the actual, you know, in our booth, we had the 256-terabyte StarGrade platform drive. People could come and see it and touch it and see the performance.

That product is just now going into the first customer's hands. From there, you go into a qualification process, which would be 6 to 12 months. It's an evolving story. We feel very good about where we're at. We think, you know, with our new development, we're moving on to the front foot as far as the availability of our products at the right capacity points as the market transitions. As we go through 2026, I think we'll see a ramping set of qualifications first. We'll start to see the revenue ramp behind that.

Got it. Just a quick follow-up for Luis. I think you mentioned to enhance the previous question of 300 BITs underutilization. Should we expect that to completely come off by your fiscal Q3? Over the next two quarters, I guess you should see a pricing tailwind to margins and that underutilization coming down. Is that how we should look at it? Thanks.

Speaker 1

Yeah, just to make sure you have the right numbers, right? Underutilization was about $50 million in Q4. What I'm saying is going to be between $10 million and $15 million in Q1. I would expect it to be very close to zero thereafter. Obviously, we're going to adjust, as I said, based on supply and demand because we believe that, you know, that's where we need to be in a supply and demand balance or slightly undersupplied. We'll continue to manage towards that. The number that comes down over time is startup costs, right? Q4 was about $42 million. Q1, we're guiding somewhere around $60 million. As I mentioned, expect this to come down in two steps, in Q2 and Q3, being minimal.

Got it. Thank you, Dave.

Does that help?

Yeah, very much. Thanks.

Speaker 0

Thank you.

Speaker 2

Our next question comes from Carl Ackerman of BMO Capital Markets. Please go ahead.

Yes, thank you. I have two, if I may. Could you talk about the average capacity trends you are seeing across flagship mobile in the second half of this calendar year, as well as content growth of conventional Windows 11 PCs and AI-enabled PCs that support your view for demand to outstrip supply into fiscal 2026?

Speaker 0

Yeah, as we look at 2025, we see average capacity in smartphones up high single digits. In 2026, a little tick above that. PCs, we see average capacity in 2025 up mid-single digits. As we go into 2026, up mid to high single digits. We going into 2026, we see higher average capacity per unit across both smartphones and PCs.

Very helpful, David. For my quick follow-up, given the comments that you made about within enterprise SSD, should we expect your data center SSD business to grow more than twice the demand you expect overall in fiscal 2026, given the cloud investment and visibility they may be giving you today? Thank you.

Yeah, we're not going to forecast for the whole year. Carl, we feel really good about where the portfolio is and the customers are pulling it in. I think you may have caught in Luis's comments, one of the things that's happening this quarter is we've got a little extra OpEx that we're going to spend because the number of units we need for qualifications is higher than we expected, which is an investment we'll make at any time. The qualification process in this market and these size drives is very, you know, this is a new platform. It's a significant process. The good news of that is once you get through it, then there's significant consumption on the other side.

We'll give you updates as we go throughout the fiscal year and then beyond of kind of how we're doing on those qualifications and what we expect in revenue growth. We certainly are confident in the product. We're seeing very good response from customers. We look forward to getting to the point where we have the optionality to mix higher into this part of the market, depending on what the market gives us. I think, as you know, we already have a great portfolio in consumer. We have a great portfolio in gaming. We have a really strong position in client. We're going to fill out this part of the portfolio. We'll be in a really good spot that no matter what the market presents us, we'll be able to mix into the best return for our shareholders.

Very clear. Thanks.

Thank you, Carl.

Speaker 2

Our next question comes from Krish Sankar of TD Cowen. Please go ahead.

Hey, guys. This is Eddie for Krish. One of your main peers is reported to be exiting the China NAND market, specifically in mobile. I wonder what's your exposure to China today and what end markets you have biggest exposure to. If you have any views to share about the competitive dynamics in China, that would be helpful.

Speaker 0

Yeah, I mean, we play across the wide range of markets, the global markets. I don't tend to think about them about one country at a time. You know, I think everybody has to decide what the mix of their portfolio is going to be. It doesn't surprise me that people make decisions of markets to enter and exit. We see China as an attractive market for us. We have a lot of partners there. Obviously, it's a good consumer market for us as well. I think that's something one of our peers decided to do. For us, this is an important part of the market.

That's helpful, Dave. A quick follow-up on the June quarter, you guys guided BITs to be flat but reported upside mid-single digits. I wonder where the upside came from, what end market?

Speaker 1

Yeah, really, it was across the board. We saw good growth. Client was good. We saw good growth across the portfolio, actually. We are very happy with how things trended. You can also see that reflected in our inventory days, right? As I mentioned, we're getting much tighter and to a point where sometimes our customers want products we don't have. We saw good performance across the board versus our expectations.

Thank you, Luis. Thank you, Dave.

Speaker 0

Thank you.

Speaker 2

Our next question comes from Stephen Fox of Fox Advisors. Please go ahead.

Hi, good afternoon. I was just wondering if you could talk a little bit more about the startup costs in the quarter, what they're related to, how quickly you get the return, and if it's any different than the expectation was 90 days ago in terms of the $60 million for the fiscal Q1. Thanks.

Speaker 1

Yeah, no, the plan hasn't changed. No, the plan is very consistent with what we had. Let me explain a little bit, right? As you go from one node to another, you know, new nodes require more tools, more steps, more space. You build a little bit more clean room for that. Once you build the space, you need to actually clean the clean room. You need to get everything kind of ongoing. Sometimes you need to relocate tools because we reutilize most of our old tools in new nodes. That all adds up in extra costs, some energy costs while you're not producing BITs. That's kind of what we classify as startup costs. Most of this or a big portion of this cost goes away, right? Again, as I said, you relocate tools and that doesn't stay with you.

Some of these costs stay with you, like energy consumption, but they will go to inventory with your BITs. It's a very different handling of those costs. Hopefully, that clarifies.

Yeah, that's helpful. Thank you.

Speaker 0

Thanks, Stephen.

Speaker 2

Once again, if you have a question, please press STAR, then one. Our next question comes from Asia Merchant of Citigroup. Please go ahead.

Oh, great. Good evening, everyone. Thank you for taking my question. If I may, at the analyst day, I think you guys talked about perhaps mid-teens, kind of a CAGR as you think about maybe it was through cycles. I'm just wondering, given the high SSD, like the Ultra QLC that you're talking about, how does that affect the way you're thinking about the NAND market and the BIT growth that we should be expecting? Is this something, as you're coming out of a market that saw some perhaps some pull forwards in the first half of this year, how we should think about the demand upside, if Ultra QLC or high SSDs materialize into significant AI workloads? Thank you.

Speaker 0

Yeah, so first of all, you know, the numbers, when we talk about the growth numbers, it's kind of all in. We kind of go by each segment. We look at, obviously, the big three, smartphone, PC, and data center, and then kind of aggregate that all up into an overall number where I talk about we're either undersupplied or oversupplied. In the data center market, you know, we continue to see strong growth. I think we just saw a couple of weeks ago with the hyperscalers reported that CapEx took another step up. There's an enormous amount of investment in AI. I think that, you know, our customers are pulling us forward on the roadmap on these very high-density enterprise SSDs as they build out data lakes.

We think we're going to, we've been investing in this Stargate platform for, you know, it takes years to build a platform like that. We feel like, you know, we're now at the point where we can put that product in customers' hands. As we go through 2026, we'll get through qualifications and then the actual selling and volume. All those, the demand for all of those products and the demand for those workloads and all those CapEx numbers are factored into the numbers we put out for how we see the supply and demand for the market.

If I may, you have a pretty aggressive Chinese competitor in this market. I know you talked a little bit about undersupply. It takes a few months, maybe a couple of quarters for production adjustments to affect supply. Is there anything you can share on, when you look at the competitive moves that are happening, how you're thinking about that particular competitor in terms of affecting a nice tight supply-demand environment that you're talking about?

Yeah, we factor all the players in the market. All of the competitors are in our forecast for market demand. They all have, they all play in different markets, different portfolios to the questions earlier, different places they can play or want to play. We have that all factored into our numbers, and that's where we come up with this. I think we've believed, going back to our Investor Day, that we were going into an undersupplied market. We've seen it play out that way, and we continue to see that going forward, factoring in all the investments, all the nodal transitions that are going on on the supply side. To your point, all of the different use cases, we've talked about it here, whether it's data center and AI workloads, or it's PCs, number of units, average content per unit, same thing for smartphones.

That's where we add it all up, and we come up with its undersupplied market. In the second half, we see about a 5% undersupplied market. I think we'll get all of 2026 more dialed in as we get closer to it.

Thank you.

Thank you.

Speaker 2

Our final question comes from Mark Miller of The Benchmark Co. Please go ahead.

Thank you for the question. I was just wondering if you could give some more color on where you're seeing allocations.

Speaker 1

Yeah, thank you for the question, Mark. We're seeing allocations across many of our products, particularly as we transition from Bixate node to Bixate node, there are certain products where we're just short on both of them, on Bixate node and on Bixate node. We're just managing that with our customers in terms of timings and how do we maximize value for them and for us. It's a little bit of some of our products on Bixate node and some on Bixate node.

Are you seeing any major share changes in the data center market?

No, I wouldn't say that.

Speaker 0

I mean, look, we've seen some, you've seen some episodic, you know, BIT shipment numbers from different players in the market. Share is going to move around in this market where you got to look over, as you guys well know, you got to look over longer periods of time than one quarter to the next quarter, look year over year. I think the good news of all of that is we've seen those BIT shipments. The market has absorbed those numbers, and we still have this favorable supply-demand balance, which is the big picture that we look at. We're encouraged by that going forward.

Thank you.

That's great. Thank you, Mark. That's a great place to wrap up. We appreciate all the questions, everybody's participation. We'll see you during the quarter. Thank you very much.

Thank you.

Speaker 2

This concludes today's question and answer session and conference call. You may disconnect your lines. Thank you for participating and have a pleasant day.