Sandisk: Unlike Micron, There's Much Higher Risk

Sandisk posted a blowout quarter with $5.95B in revenue, up 97% sequentially, and management guided to $7.75B–$8.25B in next-quarter revenue plus $30–$33 in EPS. The stock still screens attractively at 13x FY27 EPS despite trading at a premium to Micron, reflecting strong AI-driven NAND demand. Key risks remain product concentration and heavy dependence on hyperscaler customers.

Analysis

SNDK is the cleanest leveraged expression of the current AI memory upcycle, but the market is likely underestimating how cyclical this becomes once capacity tightness starts to ease. In the near term, hyperscaler buying can remain price-insensitive because deployment schedules are dictated by rack power and inference rollouts, not controller discipline; that supports multiple quarters of outsized ASPs and mix. The second-order beneficiary is upstream NAND equipment and packaging ecosystems, while the biggest loser is any storage buyer without contract protection, as spot pricing can gap before procurement teams can re-hedge.

The key risk is that SNDK’s earnings power is more exposed to a demand air-pocket than the headline multiple suggests. A concentrated customer base means one or two large hyperscalers shifting inventory policy can compress utilization quickly, and NAND has historically corrected faster than investors expect once lead times normalize. The relevant horizon is 3-9 months, not years: if AI capex pauses or customers re-balance between GPU, HBM, and storage budgets, the earnings delta can reverse sharply even if end-demand remains healthy.

Consensus appears to be treating SNDK as a durable secular compounder rather than a high-beta memory proxy with a superior quarter. That is the likely blind spot: the premium to Micron may be justified in the trough-to-peak window, but the setup is fragile if supply discipline cracks or if rivals prioritize share over margin. WDC is the more interesting contrarian expression because it remains a cleaner way to capture any sympathy rerating without paying for the same concentration risk, though it needs a catalyst beyond sympathy flow to outperform.

For now, the trade is to lean long momentum but hedge duration risk aggressively; this is a better earnings-trend trade than a core fundamental compounder. If management confirms backlog visibility and supply stays tight, upside can persist for another 1-2 quarters, but the first sign of order normalization should be treated as a fast de-risking event.

Key Decisions for Investors

• Long SNDK into the next earnings cycle, but size as a tactical momentum position only; target a 20-30% upside continuation if guidance holds, with a hard exit on any commentary about lead-time normalization or customer inventory digestion.
• Buy SNDK put spreads 2-4 months out to hedge a post-print air pocket; structure for a 2:1 or better payoff if the market starts discounting peak margins before the next quarter.
• Pair trade: long SNDK / short WDC for 1-2 quarters if you want pure AI-memory beta with less execution noise; the spread should work if investors keep paying up for direct exposure, but cut if WDC gets a product-cycle catalyst.
• For more contrarian exposure, consider shorting the basket of low-quality storage suppliers on any further multiple expansion in SNDK; the trade works best if channel checks show any shift from shortage to allocation.
• Set a risk trigger on hyperscaler capex commentary over the next 30-60 days: any sign of budget reallocation away from storage should prompt de-grossing SNDK immediately.