Sandisk Stock Is Down 45% Over the Past Four Months. Is It a Bargain or a Value Trap?

Sandisk reported fiscal 2027 Q1 sales of $11.0 billion, up 378% year over year, and EPS of $52.60 before its announced eight-for-one stock split; data centers are driving growth. The stock was up 444% over the year as of Nov. 4 but down 47% since the start of July, amid concerns that NAND memory profits may be near a cyclical peak. Eight multi-year new business model agreements aim to lock in revenue, while the article notes the risk that low valuation multiples reflect unsustainable peak earnings.

Analysis

The key valuation question is not whether AI needs NAND, but whether Sandisk can preserve pricing and utilization as customers convert current demand into supply commitments. Multi-year agreements could dampen spot-cycle volatility, yet they do not establish attractive economics unless terms protect price, volumes, and returns on capacity; they may instead lock Sandisk into lower-margin supply if the cycle turns. Verify whether agreements are take-or-pay or otherwise volume-binding, how pricing resets, and whether customer concentration is rising.

The market may be discounting a familiar memory-cycle pattern: strong pricing invites capacity additions, then supply arrives after demand momentum fades. Samsung Electronics, SK Hynix, and Micron are relevant competitive supply checks; their investment and wafer-allocation decisions could matter more to normalized NAND earnings than near-term AI spending headlines. Also, AI infrastructure demand is not uniform across memory types, so do not extrapolate HBM or DRAM tightness directly to NAND pricing.

Near term, expect sentiment to hinge on guidance and evidence that contract economics support margins—not the stock split or trailing multiples. Over 1–3 months, watch NAND pricing, inventory, and disclosed NBM terms. Over 6–18 months, capacity response and the pace of AI data-center buildouts are the larger swing factors. The contrarian opportunity is that investors may be treating all current earnings as peak-cycle; the counterpoint is that low multiples offer little protection if normalized earnings reset sharply. No valuation conclusion is secure without normalized earnings and contract-level detail.

Key Decisions for Investors

• If seeking exposure, consider a small, staged SNDK position rather than buying solely on the low trailing multiple; add only if subsequent guidance and NAND pricing support sustained profitability.
• For a lower-beta expression, pair a modest SNDK long with a semiconductor-sector hedge. The thesis is company-specific NAND execution, not a broad call that every AI-memory category will remain tight.
• Treat improving contract disclosures as a catalyst only after verifying price-reset mechanics, minimum volumes, customer concentration, and capacity commitments; absent those details, keep it on watch rather than underwriting revenue visibility.
• Falsify the long thesis if NAND pricing or utilization weakens alongside margin/guidance deterioration, or if competitors' capacity additions outpace demand. A recovery in pricing and evidence of durable contract economics would support adding exposure.