Sandisk Is Up More Than 1,700% in a Year and Still 33% Off Its Peak. History Says This is What Happens Next.

Sandisk shares have returned 4,334% since their February 2025 relisting, though the stock at $1,552 is 33% below its June 25 peak of $2,335. Fiscal Q4 revenue rose 51% sequentially to $8.97B, while full-year revenue increased 175% to $20.25B and data-center revenue climbed 437% year over year, supported by memory pricing and AI-driven demand. The company guides to $10.5B of fiscal Q1 2027 revenue, up 17% sequentially, and has $93.9B in contracted backlog, although only $16.5B is guaranteed; at 7x forward earnings, the article views the post-peak selloff as an attractive entry point.

Analysis

SNDK’s apparent value multiple is only compelling if NAND contract pricing and utilization remain near current levels; the market is discounting the normal memory-cycle outcome of price erosion rather than treating reported forward earnings as durable. The key underwriting issue is not headline backlog but its enforceability, cancellation provisions, volume flexibility, and whether commitments are tied to hyperscaler capex that can be deferred. A modest reversal in NAND pricing would have disproportionate EPS impact because recent margin expansion is largely operating leverage on price rather than unit growth.

The 33% retracement is unlikely to reset positioning after a >40x move: retail momentum ownership, concentrated gains, and a narrow set of data-center buyers leave SNDK vulnerable to another air pocket around earnings or any customer-capex caution. Over the next 1-3 months, NAND spot and contract-price data, management’s gross-margin bridge, and the conversion of contracted orders into shipments matter more than revenue guidance. Over 6-18 months, rising supply from competitors and hyperscalers’ effort to qualify alternate NAND sources are the structural risks; NVDA is a less direct beneficiary than memory suppliers because higher storage costs can modestly pressure system-level AI infrastructure economics.

Contrarianly, consensus may be underestimating how quickly an AI-storage bottleneck can persist if enterprise inference workloads increase data retention and retrieval intensity. But the asymmetry is unfavorable for chasing common equity: a low stated P/E does not protect against a cyclical earnings reset, while a single-quarter pricing or backlog-conversion miss can drive material multiple compression. WDC should have limited direct sensitivity absent retained economic exposure, but may trade sympathetically and offers no clean substitute for SNDK exposure.

Key Decisions for Investors

• Do not initiate a directional SNDK common-equity long before the next earnings release; require evidence that gross margin is stable or expanding despite sequential pricing normalization and that guaranteed backlog converts to shipments. A guide-down in gross margin or customer-volume commitments would falsify the long case.
• For a tactical long, use a defined-risk SNDK call spread dated 3-6 months after the next earnings print rather than stock: enter only if post-results implied volatility falls and management quantifies backlog conversion/cancellation terms. Target a 2:1 payoff profile; premium paid is the maximum loss.
• Initiate a small SNDK / MU relative-value short-spread watchlist, not an immediate trade. Short SNDK versus long MU if SNDK’s premium to MU on normalized mid-cycle earnings expands further while NAND spot pricing rolls over; close if SNDK demonstrates sustained contract-price floors or materially better margin resilience.
• Monitor monthly NAND contract-price benchmarks, hyperscaler capex revisions, and SNDK’s reported data-center mix. A sustained 10%+ decline in NAND contract pricing or a major customer deferral would justify reducing any residual SNDK exposure quickly, given operating-leverage and liquidity risk.