tapebrief
SNDK · Q2 2026
Preliminary brief— based on press release only. Full analysis including management tone and Q&A will be added when the transcript is available.

SNDK · Q3 2026 Earnings

Sandisk

Reported April 30, 2026

30-second summary

Sandisk's Q3 FY2026 broke every dimension of the prior guide: revenue $5.95B (+251% YoY, +97% QoQ) cleared the $4.80B high end by $1.15B, non-GAAP EPS of $23.41 was 67% above the $14.00 high end, and non-GAAP gross margin of 78.4% printed 1,140bps above the 67.0% high end. The story is no longer cyclical recovery — three Q3-signed contracts carry ~$42B minimum contractual revenue; five signed agreements in aggregate carry >$11B in financial guarantees, with two additional NBMs signed in Q4 to date. Management guided Q4 FY2026 to $7.75–8.25B revenue (+308–334% YoY off the $1.90B Q4 FY2025 base) with gross margin stepping further to 79–81%, and announced a $6B share repurchase authorization. The structural-reset thesis from Q2 is now backed by signed paper.

Headline numbers

EPS

Q3 FY2026

$23.41

Revenue

Q3 FY2026

$5.95B

+251.0% YoY

Gross margin

Q3 FY2026

78.4%

Free cash flow

Q3 FY2026

$2.99B

Operating margin

Q3 FY2026

69.1%

Key financials

Q3 FY2026
MetricQ3 FY2026YoYQ2 FY2026QoQ
Revenue$5.95B+251.0%$3.02B+96.7%
EPS$23.41—$6.20+277.6%
Gross margin78.4%—50.9%+2750bps
Operating margin69.1%—35.2%+3390bps
Free cash flow$2.99B—$0.98B+205.4%

Guidance

Massive guidance beat in Q3 FY2026 with 24–35% revenue and 67–79% EPS upside; Q4 guidance raised to $7.75–8.25B revenue (308–334% YoY) and $30–33 EPS, reflecting structural margin expansion from new datacenter-led business models and pricing power.

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

MetricPeriodPrior guideActualΔResult
RevenueQ3 FY2026$4.40 billion to $4.80 billion$5.95 billion+$1.15 billion to +$1.55 billion above guide (24–35% beat)Beat
Non-GAAP EPSQ3 FY2026$12.00 to $14.00$23.41+$9.41 to +$11.41 above guide (67–79% beat)Beat
Non-GAAP Gross MarginQ3 FY202665.0% to 67.0%78.4%+11.4 to +13.4 points above guideBeat

New guidance

MetricPeriodGuideYoY
RevenueQ4 FY2026$7.75 billion to $8.25 billion+308–334% YoY
Non-GAAP EPSQ4 FY2026$30.00 to $33.00—
Non-GAAP Gross MarginQ4 FY202679.0% to 81.0%—
Non-GAAP Operating ExpensesQ4 FY2026$480 million to $500 million—
Non-GAAP Tax ExpenseQ4 FY2026$775 million to $875 million—
Diluted Shares OutstandingQ4 FY2026~158 million—

Product revenue

Q3 FY2026
SegmentQ3 FY2026YoY
Datacenter$1.467B+233.0%
Edge$3.663B+118.0%
Consumer$0.82B-10.0%
Datacenter Revenue Growth YoY645%—

Management tone

Cyclical recovery → BiCS8 ramp story → structural margin reset claim → signed multi-year contracts with ~$42B minimum revenue (three Q3 contracts) and >$11B financial guarantees (five agreements in aggregate).

Three quarters ago management was selling a pricing-led margin step-up; two quarters ago they relabeled it a "structural reset"; this quarter they put paper behind the claim. The disclosure that "the three contracts signed during the quarter provide minimum contractual revenue of approximately $42 billion" and that "the five agreements signed so far include financial guarantees that exceed $11 billion" is the most consequential single datapoint in any Sandisk print since the WD separation. Last quarter the language was "we are engaged in discussions with customers to evolve from quarterly negotiations towards multi-year agreements" — explicitly discussion-stage. This quarter it is executed agreements with quantified financial guarantees. That converts the structural-margin thesis from management assertion into contractually-backed forward revenue.

The value-capture framing crossed a rhetorical threshold this quarter. From the call: "We're finally for the first time in decades in this business, getting to the point where the value of our technology is getting recognized... quite frankly, the value of our technology has been recognized in the market. It's just other people have been collecting that value." A quarter ago management was arguing that datacenter mix earns a different multiple; this quarter they are explicitly arguing that the entire industry has historically transferred NAND value to downstream players, and Sandisk is now repositioning the contract structure to retain it. The implication for sustainable margins is much more aggressive than the Q2 framing.

The end-market repositioning hardened into declarative infrastructure language. Q1 framing: "Calendar year 26 will be the first time that data center market is the largest market in NAND." Q2 framing: datacenter is "highly strategic product that's part of a very sophisticated AI architecture." Q3: "NAND has become a critical component of the underlying infrastructure" for AI, with management calling it "the only economically viable solution" to meet AI capacity and latency requirements. This is the third consecutive escalation in how management positions NAND's role in AI infrastructure, and it is now backed by 645% YoY datacenter growth and signed hyperscaler contracts.

The cyclicality-elimination claim is the bullish-prose extreme. From management: "This marks a fundamental evolution of our business... durable, structurally higher earnings and a significantly more predictable and less cyclical business." Two quarters ago the company was selling margin expansion; one quarter ago it was selling a margin reset; this quarter it is selling the elimination of the boom-bust pattern that has defined NAND for forty years. The five-business-model disclosure ("these five new business models account for over a third of our bids in fiscal year 2027") is the operational mechanism — if a third of FY27 bids are under fixed-margin recurring structures, the historical NAND cycle genuinely is partially broken for this company. The credibility test is whether Q4 prints inside the 79–81% gross margin guide and whether the contract base expands further next quarter.

Recurring themes management leaned on this quarter:

New business models (NBMs) creating recurring revenue and demand certaintyAI-driven NAND demand acceleration, particularly for inference workloads and KVCacheData center emergence as primary growth engine (233% sequential growth)Technology differentiation commanding pricing power and customer preferenceElimination of cyclicality through long-term contracted commitmentsCapital-efficient growth via nodal transitions requiring minimal incremental CapEx

Risks management surfaced:

Multi-year contracts may not hold or customers may fail to perform obligations (though mitigated by financial guarantees)KVCache opportunity sizing remains complex with multiple variables (concurrent sessions, input tokens, cache hit ratios, storage duration)Supply-demand balance dependent on achieving nodal transitions and capacity expansionMarket dynamics changing rapidly, making forecasting difficultPC and smartphone unit declines creating headwinds in edge market

Answers to last quarter's watch list

Q3 FY2026 gross margin delivery against 65–67%. Delivered 78.4%, 1,140bps above the high end of guide. The pricing-led margin model has now been validated three quarters in a row, each by a larger margin than the last. The Q4 guide of 79–81% extends the trajectory further. The "structural reset" framing is no longer in question; the question is now where the ceiling is and how durable the contract-backed pricing actually is.
Resolved positively
Datacenter dollars crossing $700M. Datacenter printed $1.467B — more than double the $700M threshold — with +233% QoQ growth, far above the +40% bull threshold. The "datacenter becomes NAND's largest end-market in CY26" claim is operationally validated for Sandisk, and management guided mid-70s% datacenter growth for calendar 2026.
Resolved positively
Multi-year customer agreements moving from "discussion" to "signed." Three multi-year partnerships were signed in Q3 carrying ~$42B minimum contractual revenue, with two additional NBMs signed in Q4 to date — five in aggregate, backed by >$11B in financial guarantees. Across all five, these agreements account for over a third of FY2027 bids. This was the single most important narrative dependency a quarter ago and it has resolved as decisively as it could.
Resolved positively
FY2026 full-year guide. Still not provided. Fifth consecutive quarter without one. Given the company now claims ~$42B of minimum contractual revenue on the three Q3 contracts alone and "demand visibility through CY27," the continued absence of a FY-level envelope is conspicuous to the point of being a disclosure choice rather than a forecasting limitation.
Resolved negatively
Stargate (BiCS8 QLC) revenue contribution. Not separately quantified on the print. Management indicated Stargate begins shipping for revenue in Q4. The datacenter segment growing to $1.467B (+233% QoQ) was driven primarily by TLC enterprise SSD; Stargate-specific revenue was not broken out.
Continue monitoring
Tax rate normalization pace. Q4 tax guide of $775–875M on implied pre-tax income roughly inferred from EPS and share count suggests the rate is normalizing further. The absolute dollar step is large but consistent with the revenue and margin scaling, not a rate spike. Effective rate trajectory remains a material EPS-leverage variable but is not eroding faster than expected.
Continue monitoring

What to watch into next quarter

Q4 FY2026 gross margin delivery against 79–81%. Three consecutive massive margin beats give credibility, but 79%+ from a NAND supplier is unprecedented territory. A print inside or above guide cements the contract-backed pricing thesis; a print below 79% would be the first crack in the structural-reset narrative since it was introduced.

Datacenter dollars crossing $2.5B and YoY growth above +500%. Q3's $1.467B grew +233% QoQ off Q2's $440M. The bull case for "fastest growing market" and mid-70s% CY26 datacenter growth needs another sequential leg. Below $2.0B in Q4 datacenter revenue would force a re-examination of the demand trajectory.

FY2026 full-year guide. A fifth consecutive omission while disclosing ~$42B of contractual revenue on three contracts is now an active governance question. If Q4 prints without an FY envelope, investors should treat the asymmetry between contract-base disclosure and forecast-envelope withholding as deliberate.

NBM agreement count and remaining FY2027 bid conversion. Five agreements signed (three in Q3, two so far in Q4), with NBMs representing "over a third" of FY2027 bids. Watch whether the count crosses ten signed agreements by end of Q4 and what share of FY2027 bids has converted — the durability of the structural-margin story scales directly with this metric.

Tax expense and effective rate. Q4 tax guide $775–875M on a midpoint EPS of $31.50 × ~158M shares implies pre-tax income near $5.7B and an effective rate near ~14%. Watch whether Q1 FY2027 tax guidance steps materially higher as Malaysia loss carryforwards exhaust — this is the principal hidden risk to EPS leverage.

Pace of repurchases under the newly authorized $6B program. With $3.7B cash, zero debt, ~$3B quarterly FCF, and a Board-authorized $6B buyback effective immediately with no expiration, the operative question is execution pace. Watch whether Q4 shows actual repurchases against the $6B authorization, and how aggressively management deploys cash given the visibility from signed NBMs.

Sources

  1. Sandisk Q3 FY2026 press release (Exhibit 99.1), filed via SEC EDGAR: https://www.sec.gov/Archives/edgar/data/2023554/000162828026028879/sndkq3-26ex991xpressrelease.htm
  2. Sandisk Q3 FY2026 earnings call transcript (prepared remarks and Q&A)
  3. Sandisk Q2 FY2026 press release and earnings call (for prior-quarter guidance baseline and tone comparison)

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.