SNDKMay 25, 2026 at 2:05 PM UTCSemiconductors & Semiconductor Equipment

Sandisk: Blowout AI-Driven Earnings Mask Unproven De-Risking

Read source article

What happened

A Seeking Alpha article highlights Sandisk's robust AI demand and exceptional earnings beats, with data center and edge revenues surging 645% and 295% YoY respectively, assigning a Buy rating. However, the DeepValue master report reveals that Q3 FY26 revenue growth was overwhelmingly driven by ASP per GB (+248%) rather than volume (flat exabytes YoY), indicating peak-cycle pricing. Contract indicators like $511M in customer advances and $41.6B in remaining performance obligations are promising, but the 10-Q warns that multi-year agreements may not fully offset lost revenue if customers underperform. The company's fixed-cost exposure through Kioxia ventures and quarterly pricing resets in at least one NBM contract undermine the narrative of structural earnings durability. At $1,187 and a 39x P/E, the stock prices in a de-cyclical shift that lacks empirical confirmation from the latest SEC filings.

Implication

Near-term, Sandisk's Q4 FY26 guidance of $7.75–$8.25B revenue and $30–$33 non-GAAP EPS may be achievable given continued AI demand, creating positive sentiment. However, the reliance on ASP inflation and the filing's explicit 'no assurance' language on contract performance suggests the market is overpricing stability. A reversion in NAND pricing could expose the fixed-cost downside from Flash Ventures and cause earnings to contract faster than anticipated. The $6B buyback provides a floor but also increases per-share sensitivity to the cycle. Investors should wait for at least one more quarter of NBM disclosure continuity and volume-driven growth before adding positions at current levels.

Thesis delta

The bullish narrative from the Seeking Alpha article—that Sandisk is a de-risked AI play—is contradicted by the DeepValue report's evidence that near-term results are pricing-driven and that contract guarantees carry material limitations. The thesis shifts from 'buy on AI momentum' to 'wait for proof that earnings durability is not a function of peak pricing.' The market's confidence in de-cyclicality is premature, and the risk/reward favors skepticism until NBM contracts are stress-tested.

Confidence

medium