SNDK · Q3 2026 Earnings

Cyclical recovery (Q4 FY2025) → Supply-constrained pricing cycle (Q1 FY2026) → Structural industry reset (Q2 FY2026) → Contracted-revenue enterprise (Q3 FY2026).

Last quarter management asserted that "we believe NAND is becoming a more durable, structurally attractive industry with higher average returns." This quarter they put a number on it: the three Q3-signed NBM agreements carry approximately $42B of minimum contractual revenue, and the five agreements signed to date carry aggregate financial guarantees of more than $11B. The quote that captures the shift is from this call: "I actually don't...probably never expect to collect those [financial guarantees] because I think our customers are extremely serious about needing this product...within weeks we're having a conversation about how we increase the amount of product we can get to them." What was directional rhetoric in Q2 is a balance-sheet construct in Q3 — customers have posted collateral. This converts the "structural reset" thesis from a margin claim to a contractual one.

The value-capture framing has crossed a line that management teams almost never cross. From this call: "for the first time in decades in this business, we're getting to the point where the value of our technology is getting recognized...other people have been collecting that value and it hasn't been the producers." That is management explicitly indicting the historical NAND value chain and stating that Sandisk is now the party capturing the rent. Two quarters ago the language was "demand improving"; this quarter it is "we are taking back value that was being collected by others." That is a permanent narrative break, not a quarterly tone shift.

The Datacenter framing has completed its multi-quarter arc. Q1 FY2026: "highest-engagement growth pillar." Q2 FY2026: "expected to become the largest market for NAND in 2026." Q3 FY2026: "Data Center has become our fastest growing market, and the workloads driving that demand including inference, reasoning, and agentic systems, represent a structural and durable shift." Combined with management's CY2026 datacenter growth forecast stepping from mid-20s% (two quarters ago) to mid-40s% (last quarter) to mid-to-high 70s% (this quarter), the trajectory of forecast revisions itself is the signal. Two consecutive near-doublings of a forward forecast is either a regime change or a loss of calibration; with $42B of minimum contracted revenue against the three Q3 NBMs alone, the burden of proof has shifted to the bears.

The supply-allocation posture is now explicitly anti-cyclical. From the call: "we're not going to wait until next year and see what the market gives us...we're going to supply the customers that we have agreements with...we're allocating supply to the highest value opportunities." The -10% QoQ Consumer print is the visible manifestation of the allocation choice even as the segment is still +44% YoY — Sandisk is voluntarily letting the cyclical/seasonal pieces of the portfolio give way to redirect bits to contracted Datacenter customers. That is the inverse of every prior NAND down-cycle behavior pattern, where producers chased volume into weak end markets.

Capital allocation has also crossed a line: with net cash achieved (zero debt, $3.735B cash), the board authorized a $6B share buyback effective immediately with no expiration. That removes the "what will they do with the cash" overhang and reframes the question as one of pace.

What is still hedged: the FY2026 framework remains absent for the fourth consecutive quarter, even with three of four quarters now reported and Q4 guided. Management is willing to make extraordinary structural claims but unwilling to commit to a full-year revenue or EPS number. With $42B of minimum contracted revenue against the Q3 NBMs alone, the continued absence of an FY guide is conspicuous — either there is internal disagreement about how much of the contracted revenue lands in FY2026 vs. FY2027+, or management is preserving optionality to keep beating quarterly numbers without resetting expectations.