Morgan Stanley updated its research note on Sandisk (SNDK) on June 22, raising the target price sharply from $1,100 to $1,750 while maintaining an overweight rating. The core argument is that AI inference demand is fundamentally reshaping the NAND market structure. Cloud data-center customers are price-insensitive, granting Sandisk ongoing pricing power. Combined with the New Business Model (NBM) long-term agreements that lock in gross margins, the company's future profitability is now largely foreseeable.
AI Inference Reshapes Demand, Cloud Business Surges
After 64% sequential growth in Q4, Sandisk's cloud business accelerated to 233% in Q1. Behind this lies a structural shift in demand: cloud providers are paying premiums for KV cache and context-window storage used in AI inference. Morgan Stanley's analysis shows cloud revenue now accounts for a high share of Sandisk's Q1 sales, almost entirely driven by TLC (triple-level cell) NAND, as customers prioritize density and performance. These customers do not behave like consumer buyers; they sign long-term contracts with fixed pricing, providing real gross-margin support.
NBM Agreements Lock in High Margins, Show Earnings Resilience
Sandisk's NBM agreements have already locked in over one-third of fiscal 2027 bit shipments. These are typically 3-to-5-year contracts with fixed-price or collar structures. Critically, even at floor prices, these contracts can sustain approximately 80% gross margins. For context, Sandisk's gross margin was only 30.3% in FY25, is expected to rise to 69.2% in FY26e, and hit 86.7% in FY27e—improvement that appears durable. Morgan Stanley believes the company could eventually bring 70%–80% of shipments under NBM coverage, providing a strong earnings buffer. The ability to maintain ~80% margin at floor prices means the company can remain highly profitable even during price wars.
Supply Tightness Supports Pricing Power; Strong Earnings Outlook
The current NAND supply tightness is likely to persist. AI data-center expansions are still accelerating, and storage demand remains far from saturated. Sandisk's long-term contracts lock in customers, hedging much of the cyclical risk. Morgan Stanley expects ASPs to continue rising through calendar 2026 and possibly into mid-2027. With 40%–50% of revenue coming from North America and data centers now the largest end market, pricing leverage rests with suppliers amid tight supply and high customer loyalty. The company targets 15%–19% bit growth, achieved mainly through technology node transitions (density and process improvements) rather than capacity expansion. From FY25 to FY27, revenue is projected to grow from $7.355 billion to $48.826 billion (about 6.6x), and EPS from $2.74 to $14.73. Growth quality matters more than speed—high-margin cloud business is the driver. The company has announced a $6 billion share buyback, with management believing the current stock is among the cheapest in the semiconductor sector.
Catalysts and Risks Coexist
Upside catalysts: higher-than-expected penetration of data-center eSSDs, edge AI applications boosting NAND content growth, and potential returns from advanced technology investments like HBF (High Bandwidth Flash). Downside risks: lower-than-expected industry growth, increased competitor capital expenditure, Sandisk losing share in data centers, and Chinese NAND maker YMTC continuing to gain market share. Morgan Stanley's bullish case rests on three pillars: structural demand shift from AI inference, gross-margin protection via NBM contracts, and sustained NAND supply tightness. The target price of $1,750 (up from $1,100) corresponds to ~28x FY27 PE. While the forecast will be revised based on earnings and customer feedback, the logic framework is more instructive than the specific price target.