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

Sandisk has returned 4,334% since its February 2025 relisting but is 33% below its June 2026 peak, trading at $1,552 per share. Fiscal Q4 revenue rose 51% sequentially to $8.97 billion, while full-year revenue increased 175% to $20.25 billion and data-center revenue surged 437%. Management forecasts $10.5 billion of fiscal Q1 2027 revenue, up 17% sequentially, supported by a $93.9 billion contracted-revenue backlog, though only $16.5 billion is guaranteed. The article argues the post-peak pullback leaves the stock inexpensive at 22x trailing earnings and 7x forward earnings, despite risks around hyperscaler contract execution and profit-taking.

Analysis

SNDK’s apparent valuation discount is better read as peak-cycle skepticism than a durable value signal. NAND pricing has historically mean-reverted sharply once customers complete inventory builds or capacity returns; because recent earnings leverage is predominantly price-driven, even a modest ASP reversal can compress both EPS and the multiple simultaneously. The key underwriting question is not headline backlog, but the enforceable take-or-pay portion, cancellation provisions, pricing reset clauses, and the extent to which hyperscaler commitments overlap with already-signaled procurement plans.

The near-term setup is vulnerable to crowded momentum deleveraging: a further 15-25% decline can occur without any fundamental revision if systematic holders reduce exposure after a broken parabolic chart. Over 1-3 months, the relevant catalyst is whether next-quarter guidance converts contracted demand into realized shipments while gross margin holds; an upside revision would force shorts to cover, but merely meeting elevated expectations is unlikely to re-rate the shares. Over 6-18 months, incremental NAND wafer supply from Samsung, SK Hynix and Kioxia is the principal threat, with Micron (MU) the cleaner listed proxy for a broader memory upcycle.

Contrarian view: the market may be underestimating the durability of enterprise SSD content growth, but it may also be over-crediting AI infrastructure spend as uniquely favorable to SNDK. AI capex primarily benefits accelerators and high-performance storage architectures; a transition toward internally designed storage, alternative suppliers, or lower-cost QLC deployments could preserve unit demand while eroding SNDK’s realized pricing. WDC should not be treated as a direct sympathy trade absent confirmation of remaining commercial agreements or equity ownership following the separation.

Key Decisions for Investors

• Do not initiate a directional SNDK long solely on the reported forward P/E. Require verification of backlog enforceability, customer concentration, and ASP/gross-margin assumptions before adding exposure; treat any disclosure that guaranteed commitments are not rising as a bearish catalyst.
• For a 1-3 month relative-value expression, consider long MU / short SNDK in equal dollar beta-adjusted size after SNDK fails to reclaim its 50-day moving average. The trade captures broader memory-cycle upside while hedging SNDK-specific execution, valuation-quality, and momentum-unwind risk; exit if SNDK raises margin guidance while MU does not.
• For existing SNDK holders, reduce into sharp rallies ahead of the next earnings release and retain only a defined core position. A break below the prior post-peak support level without a corresponding industry NAND-price deterioration would signal positioning-driven downside and argues against averaging down.
• Monitor monthly NAND contract-price data, MU commentary, and announced wafer-capacity additions from Samsung/SK Hynix/Kioxia. A sustained sequential decline in contract pricing or any customer deferral is the thesis falsifier for a long; conversely, two quarters of shipment-led growth with stable gross margin would justify reassessing the peak-cycle short case.