SanDisk: NAND allocation scarcity turns a commodity memory maker into a contracted annuity

Stevie AI on SanDisk Corp (SNDK-USA | sandiskcorps)

9/20/2026

Summary

SanDisk has re-rated from a cyclical NAND flash supplier into something closer to a contracted capacity provider, as Named Bit Model (NBM) agreements with eight strategic datacenter customers now lock in 50%+ of FY2027 supply and two-thirds of FY2028 supply under multiyear terms backed by $16.5B in financial guarantees. This structural shift—datacenter bit mix rising from 12% a year ago to 38% exiting FY2026 and guided past 50% in FY2027—converts a historically boom-bust commodity business into one with visible, floor-protected pricing even as the broader NAND market triples toward $300B in CY2026 and approaches $500B in CY2027 on management's own estimates. The financial inflection has already begun: revenue swung from $7.4B in FY2025 (with a $-1.6B net loss, EPS -$11.32) to $20.2B in FY2026 with net income of $11.4B and EPS of $73.76 — a swing from deeply negative to strongly profitable in a single year, driven by NAND allocation tightness and the early ramp of NBM contracts. Guided Q1 FY2027 revenue of $10.3-10.8B (non-GAAP gross margin 83-85%, EPS $44-46) implies the business is compounding sequentially even before the full-year NBM mix shift is reflected. Our four-year forecast has revenue climbing to $43.5B in FY2027 and peaking growth-wise before decelerating to $64.2B by FY2030, with EPS rising from $134.45 (FY2027) to $279.89 (FY2030) as elevated margins normalize post-2027 but aggressive buybacks (executed against the $1791.82 reference price) continue shrinking the share count. We apply a 13x forward P/E to FY2027 EPS — a premium to typical cyclical-memory multiples (historically 6-10x) but a discount to high-growth semiconductor multiples, reflecting the durability the NBM contracts provide against a still-real commodity/cyclicality risk — yielding price targets that imply meaningful upside from the current $1791.82 reference price in the near term, moderating as the multiple compresses with decelerating growth into FY2029-30.

Thesis

1. **NBM contracts fundamentally change the earnings quality of a commodity business.** Historically, NAND suppliers like SanDisk have been pure price-takers subject to brutal multi-year boom-bust cycles as supply gluts crushed pricing. The Named Bit Model changes this by locking in multiyear volume and pricing commitments with hyperscale datacenter customers, complete with floors and ceilings. Management has explicitly stated it expects attractive margins even at floor pricing, and with eight customers representing over half of FY2027 supply and two-thirds of FY2028 supply, the revenue base is becoming structurally more predictable than at any point in SanDisk's history as a standalone entity. 2. **Datacenter mix shift is the single largest driver of margin durability.** Datacenter bit mix has moved from 12% a year ago to 38% exiting FY2026, with management guiding to over 50% in FY2027 and continued outpacing of the broader market in CY2027. Since datacenter/NBM pricing is contractually elevated relative to historical spot/edge pricing, this mix shift alone pushes blended average selling prices higher even without unit growth, which is why non-GAAP gross margins are guided to 83-85% for FY2027, translating to an approximate 78-80% GAAP-equivalent margin — a level unprecedented for a NAND supplier historically operating with mid-cycle margins in the 20-40% range. 3. **Vertical integration and BiCS technology provide the underlying cost structure to support NBM commitments.** The Kioxia joint venture gives SanDisk access to some of the world's largest NAND fab complexes, and its BiCS (Bit Cost Scaling) architecture—described by management as the industry gold standard—combined with proprietary controller design gives it cost-per-bit economics competitive with SK Hynix and Micron. This matters because the NBM model only works if SanDisk can profitably deliver bits at floor pricing; the technology roadmap (BiCS 8 and BiCS 10 ramps) is the mechanism that protects margin even as capex-to-revenue is guided down to approximately 6%, a historically low reinvestment rate that will support free cash flow generation ($15.9B in FY2027 alone). 4. **The market is mispricing SanDisk as a cyclical memory stock rather than a contracted-revenue technology supplier.** Traditional NAND/memory names trade at single-digit to low-double-digit forward P/E multiples because the market prices in mean-reversion of margins and pricing. But with 50-67% of near-term supply locked under multiyear contracts with floor pricing and financial guarantees, and management guiding bits to remain on allocation beyond calendar 2027 (i.e., demand outstripping supply industry-wide), the downside cyclicality that justifies a sub-10x multiple is structurally reduced for at least the FY2027-2028 window. This is the core mispricing: the stock's growth and margin profile is now more comparable to a specialty semiconductor supplier with revenue visibility than a spot-priced commodity producer. 5. **Aggressive buybacks amplify EPS growth independent of operating performance.** With the buyback program executed against a fixed $1791.82 reference price and FCF of $8-16B annually through the forecast period, share count reduction becomes a persistent EPS tailwind. This explains why EPS growth (from $134.45 in FY2027 to $279.89 in FY2030, more than doubling) outpaces net income growth (from $27.5B to $36.6B, up about 33%) over the same period — nearly half of the EPS CAGR is coming from capital return rather than operating leverage, a dynamic the market has not fully modeled given SanDisk's brief history as an independent, buyback-capable entity. 6. **Catalysts are near-term and observable.** The August 2026 Investor Day already provided expanded disclosure on BiCS 10, HBF development, and NBM margin durability assumptions. Each of the next four quarterly earnings releases (Q1-Q4 FY2027) will provide direct evidence on whether the guided $10.3-10.8B Q1 revenue and 83-85% non-GAAP gross margin are holding, and whether bit growth is tracking the guided mid-teens sellable-bit trajectory. Given the NAND market's move toward $500B in CY2027 per management's own estimate, each print is a live test of the NBM thesis versus the market's skepticism, and a string of confirming quarters should be the catalyst for multiple re-rating.

Risks

1. **NBM customer concentration is a genuine tail risk, not a modeling footnote.** Eight strategic customers account for over 50% of FY2027 supply and two-thirds of FY2028 supply. While $16.5B in financial guarantees and collateral mitigate the downside, a material disruption at even one or two of these hyperscale customers — whether from a shift in their own capex plans, an internal NAND sourcing strategy change, or financial distress — could impair a very large share of forward revenue almost overnight, given how concentrated the counterparty base is relative to the size of the commitments. 2. **NAND is still fundamentally a cyclical commodity, and floors can be renegotiated or breached in practice even if not on paper.** Management's own narrative concedes that margin normalization is expected post-FY2027 as NAND supply catches up with demand; if oversupply arrives earlier or more severely than modeled (e.g., aggressive capacity additions from SK Hynix, Micron, or Chinese suppliers such as YMTC), pricing power in both NBM and non-NBM channels could erode faster than the fixed/floor mechanics anticipate, especially at contract renewal points beyond the current multiyear windows. 3. **The $1791.82 buyback reference price embeds significant execution and capital allocation risk.** If the stock price diverges materially from this reference level, or if free cash flow generation undershoots (FCF is already guided to decline from $15.9B in FY2027 to $8.1B in FY2030 as capex needs for BiCS 8/BiCS 10 persist and margins normalize), the buyback-driven share count reduction that underpins much of the EPS growth thesis could slow, directly compressing the EPS trajectory investors are underwriting. 4. **Margin assumptions embedded in the forecast are unusually elevated relative to history and could compress faster than modeled.** The 83-85% non-GAAP gross margin guidance for FY2027, an extraordinary level for a NAND manufacturer, depends on tight allocation conditions persisting through calendar 2027 as management expects. Any acceleration in industry capacity (BiCS competitors' own node transitions, unexpected yield improvements industry-wide) could pull forward the margin normalization that is currently modeled as a post-FY2027 event, compressing net income and EPS well below the FY2028-2030 targets. 5. **Bit growth guidance of mid-teens (versus a structural mid-to-high-teens potential) reflects a deliberate inventory build to support NBM commitments, introducing working capital and obsolescence risk.** Operating at elevated inventory days to support customer commitments and account for higher component costs ties up capital and creates exposure if NBM demand assumptions prove optimistic — inventory that was built to fulfill contracted bits could become a drag on margins and cash flow if any customer scales back commitments. 6. **Technology transition risk remains material given the capital intensity and complexity of BiCS 8/BiCS 10 ramps.** Even with capex-to-revenue guided down to approximately 6%, any execution stumble in nodal transitions (yield issues, delayed qualification with hyperscale customers, or a competitor achieving a structural cost-per-bit advantage) could impair SanDisk's ability to deliver bits profitably at the floor pricing embedded in NBM contracts, undermining the entire durability thesis the market is being asked to underwrite.

📈 Price Targets