Sandisk Has Gone From Under $50 to Above $1,400 in a Year. Can the Run Continue?

The market is treating SNDK as if it has crossed from cyclical into semi-structural, but the key mechanism is still spot pricing plus mix, not just unit growth. Multi-year commitments should reduce quarter-to-quarter volatility, yet they do not eliminate the core risk: when NAND supply loosens, ASPs can reset faster than contract volumes, and the current margin stack can compress abruptly.

Second-order, the real losers are not just rival memory vendors but any buyer with high storage intensity and weak pricing power. Hyperscalers and server OEMs can likely pass through some of the cost, but PC and handset chains have less room, so a sustained NAND spike eventually shows up as deferred upgrades, lower bill-of-materials flexibility, and softer consumer device attach rates. Within semis, NVDA remains the cleaner AI exposure: storage inflation is only a small slice of AI capex, while SNDK is being valued as if the current scarcity regime can persist indefinitely.

The near-term catalyst path still favors SNDK for one more print if supply remains tight, but the six- to eighteen-month setup is more fragile. The falsifier for the bull case is not a generic growth slowdown; it is either a Q4 guide miss, a sequential gross-margin inflection below the current plateau, or signs that peers are adding capacity faster than demand can absorb it. If that happens, the stock can de-rate violently because the equity is still priced like a quasi-commodity with growth optics.

Contrarian view: consensus may be underestimating how quickly the market can look through one extraordinary year once the next two quarters merely become very good instead of exceptional. The better question is not whether earnings are real today—they are—but whether current earnings are the new base case or the peak of the cycle.