SanDisk (SNDK)

Investment Review After Q3 FY2026 Earnings (2026.04.30)

TL;DR

SanDisk is a pure-play NAND flash and SSD maker that spun off from Western Digital in Feb 2025. Q3 FY2026 revenue exploded to $5.95B (+251% YoY, +97% QoQ), crushing the $4.4-4.8B guide. Non-GAAP gross margin hit 78.4% and EPS came in at $23.41 vs $12-14 guide. Stock currently $1,097 at 18x forward P/E(ttm). The company signed five multi-year customer supply contracts (NBMs) covering over a third of FY27 bits with $42B in committed RPO and $11B in financial guarantees, and authorized a $6B share buyback after fully repaying debt. Key risk: NAND has historically been the most cyclical semi market - 78% gross margins are unprecedented and the durability of the new customer contract model is the central investment question.

Summary

Finding	Assessment
5-Year Stock Return	Stock currently $1,097, up 2,189% from $48 at Feb 2025 spin-off (14 months ago). TTM OP $5.47B (Q3 FY2026), $11.0B implied for Q4 FY2026E. P/E 18x. Adj FCF $2.96B in Q3 (50% margin). Net cash $3.74B (zero debt).
Business Model	Pure-play NAND flash/SSD company. Flash Ventures JV with Kioxia (~80% capacity, extended through Dec 2030). $1B Nanya stake for DRAM. Q3 FY2026 mix: Edge 62%, Data Center 25%, Consumer 14%. Five multi-year NBM contracts ($42B RPO).
Competitive Moat	NARROW Kioxia JV + BiCS8 tech + SanDisk consumer brand. NBM contracts ($42B RPO, $11B+ guarantees) may broaden moat if held through cycle. NAND is base commodity. Samsung/SK Hynix retain scale advantage; pricing durability the central question.
Past 5-YR Drivers	AI-driven NAND supercycle, BiCS8 technology ramp, supply discipline (mid-high teens bit growth), data center mix shift, multi-year NBM contract framework, WDC separation unlocking pure-play focus.
Recent Stock Move	Stock currently $1,097, +73% QoQ from $635 at Q3 close. Most explosive coverage stock at +2,189% since spin-off. Driven by Q3 beat, NBM contract announcements, and $6B share buyback.
Revenue Trend	ACCELERATING Rev YoY: 251% (Q3 2026) → 320% (Q4 2026E).
Margin Trend	FLAT OM%: 70% (Q3 2026) → 70% (Q4 2026E).
TTM OP Trajectory	DECELERATING TTM OP QoQ: 332% (Q3 2026) → 101% (Q4 2026E). TTM OP $5,472M.
Key Opportunities	Five NBM contracts cover >33% of FY27 bits; $42B RPO; data center +233% QoQ to $1.47B; QLC Stargate ramp in Q4; CY26 data center exabyte growth raised to mid-70%; Q4 guide $7.75-8.25B / 79-81% GM / $30-33 EPS.
Key Risks	NAND cyclicality - 78% GM is unprecedented; NBM variable-pricing components could compress in later years; customer concentration in hyperscalers (5 contracts cover >33% bits); Samsung/SK Hynix capacity decisions; newly public with no downcycle track record.
Catalysts	Q4 FY2026 actuals (guide $7.75-8.25B / 79-81% GM / $30-33 EPS); additional NBM signings pushing coverage above 50%; QLC Stargate revenue ramp; HBF NAND sample late CY26; $6B buyback execution.
P/E(ttm) Valuation	COMPRESSED at 18x on peak-cycle TTM OP. Counter-cyclical pattern: low P/E at peak earnings. If NBMs sustain durable model, 18x is cheap on $80-100 EPS run-rate. If cycle turns, earnings reset 40-60% and P/E expands as denominator collapses. Market pricing "show me the durability."

Q3 FY2026 Key Data

Revenue $5.95B (+251% YoY, +97% QoQ). Massive beat vs $4.4-4.8B guide. Non-GAAP GM 78.4% (vs 65-67% guide), OM 70.9%, EPS $23.41 (vs $12-14 guide). Adjusted FCF $2.96B at 50% margin.
Data center revenue +233% QoQ to $1.47B. Edge +118% to $3.66B; Consumer -10% to $820M (seasonal). Data center now 25% of mix; QLC Stargate begins shipping for revenue in Q4. Bit shipments down high-teens QoQ as inventory built for Q4 ramp.
Q4 FY2026 guided at $7.75-8.25B revenue, GM 79-81%, EPS $30-33. ~158M diluted shares. CapEx 4% of rev ($240M Q3). Cash $3.74B, paid off remaining $650M TLB - now zero debt and net cash positive.

Key Takeaways from Earnings Call

Five multi-year "New Business Model" supply contracts signed (3 in Q3, 2 in Q4 to date) - cover over a third of FY27 bits with $42B minimum RPO from the three Q3 contracts alone, and $11B+ in financial guarantees managed by third-party institutions ($400M prepayments already on Q3 balance sheet). Mgmt expects coverage to push above 50% over time. Pricing is fixed in near-term contract years and increasingly variable later, letting SNDK capture upside if NAND prices keep rising.
AI-driven NAND demand structurally accelerating - CY26 data center exabyte growth raised to mid-70% (from 60s three months ago, 40s six months ago, 20s nine months ago). KB cache, RAG, agentic AI workloads make NAND "the only economically viable" capacity tier for inference at scale. QLC Stargate ramp + TLC enterprise SSD strength position SNDK for sustained data center share gains.
Capital allocation pivoting to returns: $6B share repurchase authorized with no expiration, debt fully retired, net cash $3.7B. Kioxia JV extended through Dec 2030 plus $1B Nanya investment for DRAM supply secure long-term capacity. Mgmt: "we believe our margins are sustainable" - the structural reset thesis underpins the buyback decision.

Key Segments and Revenue Mix (Q3 FY2026)

Segment	Revenue	% of Total	Key Details
Edge (PC + Smartphone)	$3,663M	62%	+118% QoQ in Q3 FY2026; premium device mix shift; on-device AI driving higher capacity per unit
Data Center	$1,467M	25%	+233% QoQ in Q3 FY2026; TLC enterprise SSD demand from AI inference; QLC Stargate ramping in Q4
Consumer	$820M	14%	-10% QoQ (seasonal); premium portable SSD launch; "Space to Hold More" brand campaign
Total Q3 FY2026	$5,950M	100%	+251% YoY, +97% QoQ. Non-GAAP GM 78.4%. Five multi-year NBM contracts ($42B RPO).

Business Model

SUPPLIERS

Kioxia (Flash Ventures JV partner)
Silicon wafer suppliers
Equipment makers (Applied, ASML, Lam)
Packaging & test providers

Wafers / Manufacturing →

SANDISK

NAND flash design & R&D
BiCS8 3D NAND technology
Enterprise SSD & consumer storage
Flash Ventures JV (~80% capacity)
Mid-to-high teens bit growth plan

SSDs / Flash Storage → $ Revenue

CUSTOMERS

Hyperscale cloud (AWS, Google, Meta)
PC/mobile OEMs (Dell, HP, Apple)
Consumer retail (SanDisk brand)
Gaming & automotive OEMs
Enterprise data center builders

Moat Assessment: NARROW

Flash Ventures JV with Kioxia provides world-class 3D NAND manufacturing at scale with shared economics. The 25+ year partnership (extended through 2034) gives SanDisk access to ~80% of Kioxia’s manufacturing capacity with shared R&D costs and coordinated technology transitions. This reduces capital intensity vs. building alone. However, the JV also means SanDisk depends on a partner that is also a competitor.
SanDisk brand is the #1 consumer flash storage brand globally, creating channel and retail pricing power. Consumer revenue of $2.3B with premium product innovation (Extreme Fit, FIFA licensing, gaming) provides a diversified base. But consumer flash is more commoditized than enterprise SSD, with Samsung and others competing aggressively on price.
Multi-year NBM contracts may broaden the moat if they hold through cycle. Five signed agreements ($42B RPO, $11B financial guarantees, >33% of FY27 bits) introduce contractual revenue visibility unprecedented in NAND history. Pricing has fixed and variable components. If NBMs sustain through a NAND price reversal, SNDK earns a structural-revenue valuation premium. If customers walk or variable terms compress margins in later years, the moat reverts to commodity NAND with Samsung/SK Hynix scale advantage.

Stock vs. Earnings — Since IPO Performance

Stock currently $1,097, up 2,189% from $48 at the Feb 2025 spin-off — the most explosive post-IPO move in our coverage. SanDisk separated from Western Digital as a pure-play NAND flash/SSD company and immediately caught the AI-driven NAND supercycle. The stock went from $48 to $1,097 in 14 months as revenue went from $1.7B (Q3 FY2025) to $5.95B (Q3 FY2026) and non-GAAP gross margins expanded from 23% to 78%.
TTM OP swung from -$480M (Q4 FY2024) to +$5.47B (Q3 FY2026 actual), a transformational earnings inflection. Q4 FY2026 TTM OP projected at $11.0B based on company guidance. Non-GAAP operating margin expanded from negative to 71% in 7 quarters, driven by AI-driven pricing power, BiCS8 technology leadership, and supply discipline (mid-to-high teens bit growth, no capacity adds).
Free cash flow turned massively positive: $2.96B in Q3 FY2026 alone (50% margin). SanDisk paid off the remaining $650M term loan in Q3, ending with zero debt and $3.74B cash. The Flash Ventures JV with Kioxia (extended through Dec 2030) plus a $1B Nanya investment for DRAM supply provide manufacturing scale with light CapEx intensity (4% of Q3 revenue).

Key Drivers of Operating Profit Growth (Since Separation)

AI-driven NAND supercycle created a structural supply-demand imbalance, driving unprecedented pricing power. Data center exabyte growth forecasts raised 4 times: from 20s to 40s to 60s to mid-70% for CY2026. AI inference, KB cache, and agentic workloads make NAND "the only economically viable" capacity tier. Data center became SNDK's fastest-growing market at +233% QoQ in Q3 FY2026.
BiCS8 technology ramp positions SanDisk with the industry’s best-performing NAND at the perfect moment. The 2-terabit die with excellent QLC performance is driving strong customer adoption. PCIe Gen 5 TLC drives qualified at 2 hyperscalers with more in pipeline. BiCS8 QLC “Stargate” product expected to ship for revenue in coming quarters, providing additional data center tailwind.
Disciplined supply management and mix shift drove gross margin from 23% (Q3 FY2025) to 78% (Q3 FY2026). SanDisk maintained mid-to-high teens bit growth without capacity additions, letting pricing reflect demand. Five multi-year NBM contracts (>$42B RPO, $11B+ guarantees) lock in committed supply at premium pricing - mgmt frames this as a "fundamental reshaping" of the cyclicality narrative.

Quarterly Key Metrics

USD	Q3'25	Q4'25	Q1'26	Q2'26	Q3'26	Q4'26E
Quarter End	Mar 2025	Jun 2025	Sep 2025	Dec 2025	Mar 2026	Jun 2026
Stock Price	$48	$45	$112	$237	$635	$1,097
Stock QoQ	—	-5%	+147%	+112%	+168%	+73%
Rev YoY	-1%	+8%	+23%	+61%	+251%	+320%
OM%	-2%	3%	8%	35%	70%	70%
R40 (Rev YoY + OM%)	-3%	+11%	+31%	+96%	+321%	+390%
TTM OP ($M)	641	507	386	1,266	5,472	11,009
TTM OP QoQ	-16%	-21%	-24%	+228%	+332%	+101%
FCF/OP ttm	-53%	-24%	134%	114%	82%	86%
Debt/EBITDA	2.2x	2.4x	1.9x	0.4x	0.0x	0.0x
P/E(ttm)	21x	35x	186x	47x	22x	18x

Recent Stock Performance & Drivers

Stock currently $1,097, up 73% QoQ from $635 at Q3 FY2026 close. SNDK has returned +2,189% since Feb 2025 spin-off - the most explosive coverage stock. The post-Q3 surge reflects the unprecedented Q3 actuals (251% YoY rev, 78% GM, $23.41 EPS) plus the announcement of five multi-year NBM contracts and a $6B share buyback. P/E(ttm) compressed to 18x.
TTM OP went from $507M (Q4 FY2025) to $5.47B (Q3 FY2026 actual), with $11.0B implied for Q4 FY2026E. Non-GAAP operating margin reached 71% in Q3 - data center +233% QoQ, edge +118% QoQ. Q4 guide $7.75-8.25B revenue, 79-81% GM, $30-33 EPS implies the ramp is not done. Current TTM OP $5,472M.
Capital pivot: zero debt, $3.74B cash, $6B share buyback authorized. Paid off remaining $650M term loan in Q3. Generated $2.96B adjusted FCF (50% margin). Extended Kioxia JV through Dec 2030; invested $1B in Nanya for DRAM supply. Capital returns phase has begun.

Opportunities

Five signed multi-year NBM contracts cover >33% of FY27 bits with $42B minimum RPO and $11B+ in financial guarantees. Three signed in Q3, two in Q4 to date. Mgmt expects coverage to push above 50% over coming quarters. Pricing fixed in near years, variable later. Each new signing adds RPO and reduces cyclicality - a meaningful re-rating catalyst if execution holds.
Data center accelerated +233% QoQ to $1.47B and is now SNDK's fastest-growing segment. CY26 industry data center exabyte growth raised to mid-70% (from 20s nine months ago). KB cache + agentic AI workloads make NAND the only viable inference capacity tier. QLC Stargate begins shipping for revenue in Q4 - a second product cycle leg on top of TLC enterprise SSD strength.
Q4 FY2026 guidance is unprecedented: $7.75-8.25B revenue (+30-39% QoQ on top of +97%), 79-81% GM, $30-33 EPS. Implies a roughly $30B annualized revenue run-rate at ~70% OM. CapEx light at 4% of rev (JV structure). $6B share buyback compounds at higher EPS levels.

Challenges

78% gross margin and 71% OM are unprecedented in NAND history. Even with NBM contracts, sustainability through a cycle bottom is the central question. Samsung, SK Hynix, and Micron retain capacity expansion optionality. The market is implicitly skeptical: SNDK trades at 18x forward P/E vs WDC/MU at similar multiples on much lower margin assumptions - the durability premium has not yet been priced in.
NBM contract structure is opaque: customer names, durations, target margins, and variable-pricing mechanics are undisclosed. Mgmt confirms "fixed in near years, variable later" - which means SNDK gives back upside if NAND prices rise further, AND could see margin compression if prices fall in later contract years. $11B in financial guarantees is meaningful but smaller than $42B RPO would suggest.
Customer concentration risk: 5 NBM contracts already cover >33% of FY27 bits, suggesting a handful of hyperscalers drive the structural narrative. Any one customer reducing AI infrastructure spend, switching to alternative storage tiers, or developing in-house NAND solutions could unwind the thesis. Edge and Consumer share is declining as data center takes mix - reducing the diversification cushion.

Next 12 Months

Q4 FY2026 guide $7.75-8.25B / 79-81% GM / $30-33 EPS — if achieved, FY26 EPS approaches $60+ vs FY25 loss. A beat above $8.25B with GM >81% would reinforce the structural narrative; any softness or guidance reset for FY27 would raise peak-cycle concerns and could trigger a sharp correction.
NBM coverage expansion: mgmt targets >50% of FY27 bits under multi-year contracts vs >33% today. Each new signing adds RPO + financial guarantees and reduces cyclicality. Disclosure of customer names, target margin range, or aggregate contract count would be a material re-rating catalyst.
QLC Stargate ramp at hyperscalers in Q4-Q1 FY27 is the second product-cycle leg. First revenue ships in Q4 FY26. High-Bandwidth Flash (HBF) NAND product sample late CY26 with system shipping early-to-mid CY27. Both extend SNDK's data center growth runway beyond TLC.

Next 12-24 Months

NAND supply-demand through CY2027 will determine if this is a structural shift or a cyclical peak. If demand growth (mid-70% data center) continues to exceed supply growth (mid-to-high teens bit growth) and NBMs hold, SNDK FY27 EPS could be $80-100+. If AI capex pauses or competitors add capacity, EPS could reset to $30-40 and the multiple expands as earnings collapse.
Samsung, SK Hynix, and Micron capacity discipline is the biggest external risk. Industry restraint supports the thesis; any signal of accelerated NAND capacity expansion would pressure pricing despite NBMs (variable-pricing components would still take some hit). NBM commitments cushion but do not eliminate this risk.
SNDK can re-rate from cyclical commodity P/E (10-15x peak) toward durable-earnings multiple (20-25x) if NBM coverage exceeds 50% by FY27 end. The $6B buyback compounds at higher EPS levels. If structural-revenue narrative holds, today's 18x P/E on $5.47B TTM OP looks cheap; if it cracks, peak earnings get reset and the multiple expands as the denominator collapses.

Core Metrics Scorecard

	Q3'25	Q4'25	Q1'26	Q2'26	Q3'26	Q4'26E
Quarter End	Mar25	Jun25	Sep25	Dec25	Mar26	Jun26
R40 > 40%				✅	✅	✅
R40 accelerate		✅	✅	✅	✅	✅
TTM OP QoQ > 10%				✅	✅	✅
TTM OP QoQ accelerate				✅	✅
TTM FCF/OP > 60%				✅	✅	✅
Debt/EBITDA < 3.5	✅	✅	✅	✅	✅	✅
Total passes	1	2	2	6	6	5
Stock QoQ	—	-5%	+147%	+112%	+168%	+73%
P/E(ttm)	21x	35x	186x	47x	22x	18x

SNDK earned 6 passes in Q2'26 and Q3'26, the maximum on our scorecard. Total passes went from 1 (Q3'25) to 6 (Q2'26 and Q3'26), correlating with the +1,689% to +2,189% stock move. R40, TTM OP QoQ, and FCF/OP all crossed their thresholds simultaneously - the hallmark of a cyclical inflection capturing maximum momentum. Q4'26E shows 5 passes (TTM OP QoQ accelerate fails as growth rate slows from +332% to +101%).
R40 at 321% (Q3'26 actual) and 390% (Q4'26E) is the highest in our entire coverage. The combination of 251% YoY revenue growth and 70% OM% reflects extraordinary NAND cycle dynamics. R40 this extreme typically signals peak-cycle conditions - the sustainability question is whether NBM contracts can convert peak-cycle earnings into durable earnings.
P/E(ttm) compressed from 186x (Q1'26) to 18x (Q4'26E) as TTM OP went from $386M to $11.0B. This counter-cyclical pattern (low P/E at peak earnings) mirrors MU and WDC. If NBMs sustain durable earnings power, 18x is cheap on $80-100 EPS run-rate. If the cycle turns, earnings reset 40-60% and P/E expands to 35-50x on depressed earnings.

Chart: R40 vs Stock Price (Since IPO)

Chart: Scorecard Total Passes vs Stock Price (Since IPO)

Source: SEC EDGAR annual and quarterly filings, company earnings conference calls and presentations.

Disclaimer: This report is for educational purposes only. NO investment advice.