Oct 29, 2026 earnings call — AI-driven NAND growth and NBMs
Reported revenue of $11 billion and announced an 8-for-1 stock split to be distributed November 19.
Revenue $11.0B (est $10.7B) · EPS $49.51 (est $46.67) · Next-quarter revenue outlook $12.2B–$12.6B
Key quotes
"BiCS8 is the backbone of our current production portfolio and the industry benchmark for performance and density across TLC and QLC."
"We had shared with you that about 50% of our bits are under NBM for '27 and about 2/3 for '28."
"During inference, each generated token must reference the context that precedes it, creating a growing KV cache that serves as the model's working memory."
Analyst Q&A 12
customer deals reserving NAND supply
Q (Benjamin Reitzes · Melius): Will you sign many more NBMs given capacity limits? Also, remind us what bits are under contract for the next few years and if that changed.
A: Management said the NBM pipeline is very strong and growing. About half the bits are under NBM for this fiscal year and about 2/3 for next year, and 95% of NBMs are with investment-grade companies that generated about $130 billion in operating cash flow. They will remain patient and selective and expect the percent of the portfolio committed to rise over the next quarters, and they are open to offering NBMs to any customer.
"95% of our NBMs are with companies that are investment grade, and they generated about $130 billion in operating cash flow." — Luis Visoso
why data center share rose
Q (Joseph Moore · Morgan Stanley): Why did data center participation increase so much? Is it BiCS8, market changes, or something else? And could data center become half of bits?
A: Management said the large data center growth reflects multiple things: the right NAND node (BiCS8), clean CMOS from CBA, product portfolio including TLC and ramping QLC, and more refined AI inference architectures driving nuanced NAND use. NBMs enable deeper strategic conversations and repeat business, and they expect mix to move substantially higher through calendar year '27.
"So we got the right product at the right time based on the right technology." — David V. Goeckeler
pricing outlook and new NAND tiers
Q (Mark Newman · Bernstein): How will the mix of price vs volume drive guidance and pricing next quarter? Also, any update on HBF and how new NAND types fit in data centers?
A: Luis said roughly 60% of recent revenue growth came from volume and 40% from pricing, and they expect both to contribute next quarter with modest pricing improvement. David said HBF customer conversations are robust and ongoing, and that as AI inference scales architects are refining designs and pulling NAND into more parts of the architecture for cost and endurance reasons.
"we do expect some modest pricing as we put in our prepared remarks." — Luis Visoso
node transition and allocation limits
Q (David Gibson · MST Financial): Will you accelerate BiCS node transitions to grow bits faster? And are there limits to allocating more supply to data center?
A: Management said they have no plans to speed node transitions and remain committed to the mid- to high teens bit growth they discussed, because node moves need planning and tool purchases. On allocation, they said product mix from the fab limits short-term moves and it's a constant fitting exercise; Luis added wafers take about 120 days to produce, so short-term changes are hard.
"No. I mean we're committed to the bit growth we talked about, mid- to high teens." — David V. Goeckeler
NBM pipeline, supply and deposits
Q (Christopher Muse · Cantor Fitzgerald): You said NBM pipeline exceeds supply through the end of the decade. How does that change view of industry bit supply and your CapEx plans? Also, can deposits be used for buybacks?
A: David said the process is working and they now have a robust pipeline through 2030 and beyond that gives more visibility, but negotiations and commitments are still underway before supply changes. Luis said most guarantees are held by third-party financial institutions and not on their balance sheet; only prepayments are on the balance sheet and are a small amount, and there is no legal restriction on them.
"But as we explained at Analyst Day, the vast majority are financial guarantees, which are held by third-party financial institutions." — Luis Visoso
enterprise SSD content and share goals
Q (Karl Ackerman · BNP Paribas): How will enterprise SSD content per server progress in 2027 with agentic AI growth? And do you see achieving a low- to mid-teens enterprise SSD share in 2027?
A: Management said they see around 60 to 90 gigabytes per CPU in those inference and orchestration applications, and they will do further TAM analysis on server demand. On share targets, they avoid specific share guidance; they expect mix to move higher into data center but will manage the portfolio to maximize value and watch NBMs and quarter-by-quarter signals.
"I mean we know at a pretty granular level the amount of demand that we expect per CPU, something 60 to 90 gigabytes per CPU is what we see in those kind of applications." — David V. Goeckeler
node roadmap and SLC demand
Q (Vijay Rakesh · Mizuho): Will BiCS8/9/10 transitions help meet 2027 demand? And thoughts on SLC demand tightening the NAND market next year?
A: Management said BiCS9 is important for faster interface speeds and CBA flexibility, BiCS10 is still a ways out while they continue ramping BiCS8. On SLC, it's early to say; they are talking to customers signaling SLC interest, and SLC use would reduce available QLC/TLC supply depending on deployment, so they are watching it.
"BiCS10 is still a little ways out, right?" — David V. Goeckeler
seasonality and buyback plan
Q (James Schneider · Goldman Sachs): Will fiscal Q1 (March quarter) be up, down, or flat sequentially? And how will you execute buybacks as authorization runs down?
A: Luis said historically Q3 (the March quarter) is normally their weakest sequential quarter and advised being conservative modeling sequential growth from Q2 to Q3; last quarter bits were down 19% sequentially. David said the company will return 100% of excess cash to shareholders, meet the Board quarterly, and will continue to execute buybacks appropriately; they have already bought back about 4% of market cap since starting.
"So if you look at historically, right, our Q3 is normally our weakest sequential quarter." — Luis Visoso
TLC vs QLC and mobile growth
Q (Kinney Chin · TD Cowen): What's the current demand split between TLC and QLC for CSP customers, and will that shift later this decade with NBMs? Also, can mobile bits still grow next year given unit declines?
A: Management said they are around a 70-30 split of TLC to QLC today and are seeing QLC reaccelerate in some use cases; their Stargate QLC product is ramping and well received. For mobile, they expect exabytes to grow next year even with slightly declining unit counts, driven by higher content per device; they will manage portfolio flexibly and update mix over time.
"We're still like at that 70-30 number on TLC to QLC." — David V. Goeckeler
who are the AI customers and receivables change
Q (Wamsi Mohan · BofA): When you say leading AI companies could indicate demand upside, are you referring to AI labs or neoclouds? Any credit concerns from neoclouds? And is the receivables pickup due to NBM payment terms?
A: Luis said they have not seen credit issues and demand is strengthening; their largest customers are investment-grade and generate strong cash. On who they sell to, they sell to hyperscalers, model makers and others. For receivables, he said the pickup was due to a change in customer mix, which affected linearity, not NBM payment terms.
"No. We basically saw a little bit of a change in customer mix, and that has a little bit of an impact on our receivables, a little bit on linearity, but it's mostly customer mix." — Luis Visoso
capital structure and HBF milestones
Q (Amit Daryanani · Evercore): Does a more predictable NBM model change thoughts on capital structure or willingness to take on leverage? And for HBF, what are the milestones from sampling to revenue ramp?
A: David said they plan to stay debt-free with a strong balance sheet, invest in internal innovation like HBF and 3D Matrix Memory, and return 100% of excess cash to shareholders. For HBF the key near-term milestone is getting the die back, taping out the controller, assembling the product and putting a device in customers' hands so customers can test integration and then give feedback that drives next steps.
"But the big step is to actually put a product in their hands." — David V. Goeckeler
SLC impact and JV flexibility
Q (Mehdi Hosseini · SIG): Is growing SLC demand hurting TLC growth or is it additive? And can the JV add SLC capacity short-term to meet demand while HBF commercializes?
A: David said SLC is not cannibalizing other NAND products but expanding the TAM into use cases needing different endurance and performance, and their cells can run in SLC, MLC or TLC modes. On capacity, he said they think about supplying wafers in the modes customers need and it's early in customer discussions, so they will monitor how much SLC is deployed since a little SLC can replace a lot of QLC.
"No, I don't think it's a cannibalization of other NAND products." — David V. Goeckeler