Down 23%, Should You Buy the Dip on Sandisk Stock?

Sandisk's data-center business expanded from 12% to 38% of bit volume by fiscal 2026, helping drive fiscal Q4 revenue growth of 372% year over year and adjusted EPS growth of 68% sequentially to $39.25. The company has secured nearly $94 billion of future revenue through eight customer agreements averaging more than four years, which management expects to cover over 50% of fiscal 2027 bit volume and roughly two-thirds in fiscal 2028. Management forecasts mid-to-high-teens revenue growth from fiscal 2028 through 2030 and an approximately 50% free-cash-flow margin; however, uncontracted NAND volume remains vulnerable to demand and supply-cycle volatility.

Analysis

The key investable question is whether SNDK’s contracted-volume model changes its earnings multiple or merely delays the NAND cycle. If agreements contain meaningful take-or-pay terms, price floors, and customer prepayments, SNDK deserves to trade closer to enterprise infrastructure peers than historical commodity-memory valuations. If they are principally framework agreements with volume flexibility, the market is likely underestimating residual exposure to spot NAND pricing and the apparent earnings visibility is overstated.

Near term, the stock’s extreme prior appreciation makes it vulnerable to a modest NAND pricing deceleration even if fundamentals remain strong; a 10x forward P/E is not necessarily cheap when consensus EPS embeds peak-cycle pricing. The more important 1-3 month catalyst is disclosure around contract enforceability, committed versus forecasted bits, capex intensity, and the mix of enterprise SSDs versus lower-value NAND components. A positive read-through would pressure Kioxia and Micron (MU) to secure similar contracts, while capacity additions by Samsung Electronics or SK hynix would be the principal 6-18 month threat to industry margins.

Contrarian view: AI storage demand may prove less durable than compute demand because storage architectures can improve utilization through deduplication, tiering, and higher-capacity drives. Hyperscalers also possess disproportionate purchasing power; long-duration agreements can cap upside during a supply squeeze while leaving suppliers exposed if customer deployment plans slip. The thesis is falsified by a sequential decline in enterprise ASPs, rising inventory days, contract volume below management’s implied mix, or capex growing materially faster than revenue.

Key Decisions for Investors

• Do not chase SNDK on the pullback; establish a 50% starter only after the next earnings release confirms sequential enterprise ASP stability and discloses binding committed volume versus non-binding pipeline. Add only if forward estimates hold while the stock remains below 10-12x normalized forward EPS.
• Use a defined-risk bullish structure rather than outright stock ahead of contract-detail disclosure: buy 6-9 month SNDK call spreads, financed selectively with out-of-the-money puts only if implied volatility is elevated. Target a 2:1 payoff; exit if enterprise revenue growth decelerates materially or management raises capacity spending without matching contracted demand.
• Monitor a relative-value pair: long SNDK / short MU only if SNDK demonstrates superior contract coverage and lower spot-price sensitivity. Reverse or avoid the pair if MU reports tighter HBM/DRAM supply translating into stronger earnings revisions than NAND, since AI memory budgets may favor DRAM over storage.
• Set alerts for Samsung and SK hynix NAND capacity announcements, NAND spot-price declines, and SNDK inventory growth. Any two occurring together would signal cycle risk is reasserting itself and warrants cutting long exposure before consensus estimates reset.