Sandisk (SNDK) Q3 2026 Earnings Transcript

Sandisk delivered a major earnings beat, with Q3 revenue of $5.95 billion, up 97% sequentially and 251% year over year, while non-GAAP gross margin expanded to 78.4% and EPS reached $23.41 versus guidance of $4.12 to $14. Management raised Q4 guidance to $7.75 billion-$8.25 billion in revenue and $30-$33 EPS, and announced a new $6 billion buyback after repaying its remaining TLB debt. The call also highlighted rapid AI/data-center demand, with enterprise SSD revenue up 233% sequentially and five multiyear supply agreements covering over a third of fiscal 2027 bits.

Analysis

SNDK just converted a cyclical NAND story into a quasi-contracted cash-flow story, and that is the core re-rate. The market is likely underappreciating how much the new agreement structure de-risks downside: with a meaningful portion of fiscal 2027 supply already spoken for, near-term spot volatility matters less, while pricing power becomes more durable because the company is now negotiating from a position of scarcity plus customer dependence. The financial guarantee layer also changes incentives — customers are not just signaling intent, they are economically bridged into consumption, which should compress the historical drawdown multiple in a NAND downcycle.

The bigger second-order effect is competitive. If SNDK can keep migrating volume into these long-dated agreements, peers with less differentiated product or weaker balance sheets will be forced to chase lower-quality spot demand, effectively ceding the high-value AI/storage mix. That should widen the performance gap versus commodity NAND vendors and could also tighten supply for lower-tier end markets, delaying any broad consumer/PC recovery while the best bits are reserved for data center and premium edge. In other words, the bullish setup is not just better pricing, but better allocation of wafer output.

The main risk is not execution in the next quarter; it is whether the current margin regime proves portable once the initial AI buildout normalizes. Investors may be extrapolating Q4-like economics too far into FY27, when some variable pricing and mix benefits will likely moderate. A second risk is that the market starts treating the buyback as a signal that management sees peak reinvestment behind it — positive for EPS, but potentially a flag that organic growth rates will become the harder variable to sustain once the Stargate and NBM ramps are fully absorbed.

Key Decisions for Investors

• Go long SNDK on any post-print consolidation over the next 1-2 weeks; use pullbacks toward the first 5-7% retracement as entry, targeting a multi-month rerate if RPO disclosures keep compounding.
• Pair trade: long SNDK / short a weaker commodity-memory name or basket (e.g., MU) for 3-6 months; thesis is that contracted revenue plus net-cash/buyback support should outperform more cyclical memory exposure if NAND pricing cools.
• Buy SNDK call spreads 2-4 months out to express continued upside with defined risk; the catalyst path is RPO expansion, Stargate revenue commencement, and incremental NBM wins.
• Trim or hedge if the stock prices in sustained 70%+ operating margins without evidence that FY27 contract economics are expanding; that is where the risk/reward becomes asymmetric to the downside.