Data Centers Now Deliver a Third of Sandisk's Revenue -- $2.98 Billion in a Single Quarter

Sandisk’s fiscal Q4 (ended July 3) datacenter revenue surged to $2.98B (~33% of $8.97B total), up from just $213M a year earlier, and pricing drove performance with ~two-thirds of sequential revenue growth from higher pricing. Sandisk signed 10 New Business Model agreements (8 customers) with minimum $93.9B in expected revenue at floor pricing, supported by $16.5B in customer cash deposits/instruments. Profitability swung sharply higher with gross margin rising to 84.6% (from 26.2% a year ago) and quarterly net income of $6.9B; full-year free cash flow improved to $11.5B from a $120M outflow, and guidance for Q1 FY2027 calls for revenue of $10.3B–$10.8B with gross margin 83%–85%.

Analysis

SNDK is starting to look less like a pure spot-memory beta and more like a quasi-contracted supply platform, which should compress the drawdown profile in the next pricing downturn. That matters most for relative value: peers with heavier exposure to open-market NAND pricing should see more earnings volatility and lower multiple support, while SNDK can argue for a structural premium if the contract book really covers a majority of shipments by FY28.

The near-term setup is still about margin durability, not revenue growth. The market is likely underestimating how much of the current earnings power is still a function of elevated pricing, so the stock can remain choppy if spot NAND rolls over before the contracted mix fully dominates. The key catalyst window is the next 1-3 quarters: if management keeps guide intact while gross margin stays above the low-80s, the bear case of an imminent commodity reset weakens; if guidance starts to reflect ceiling effects or slower datacenter mix conversion, the rerating can stall quickly.

Second-order, the customer deposits and longer-dated supply commitments improve SNDK's working-capital and financing flexibility, which could support capex and inventory positioning through a downturn. The flip side is that ceilings on variable pricing cap upside in a continued boom, so this is not a clean secular growth story; it is a volatility-reduction story with a still-cyclical end market. The contrarian miss is that consensus is probably too focused on peak earnings optics and not enough on the fact that floors plus prepayments can materially change the left tail, even if they do not fully eliminate the cycle.

Key Decisions for Investors

• Long SNDK on weakness, sized as a medium-term relative-value position; thesis is lower earnings volatility and higher floor on FY27/FY28 earnings than the market is giving credit for. Best entry is a post-earnings or post-rally pullback of 5-8% where the multiple is still below the broader semiconductor group.
• Pair trade: long SNDK / short WDC for 1-3 months if NAND pricing shows any sign of peaking. The relative thesis is that SNDK's contracted mix should protect margins better, while WDC remains more exposed to spot-driven compression and narrative de-rating.
• If you want cleaner sector expression, use SNDK long against SMH or SOXX short into any memory-sentiment squeeze. Risk/reward favors the pair if investors rotate into cyclical semis broadly and the market misprices SNDK as just another commodity memory name.
• Do not chase at elevated multiples; instead, watch for confirmation that FY27 gross margin stays above ~82% and contract coverage keeps rising. If margins slip below that band or guide is cut, the thesis is invalidated and the stock should be treated as a high-beta cyclical again.
• Alert item: monitor spot NAND pricing and hyperscaler capex commentary over the next 1-2 quarters. A sharp spot decline or capex pause would be the fastest way to prove the contract floor is less protective than it looks.