After Surging 3,110%, Has Sandisk Already Had Its "Nvidia Moment"?

Sandisk (SNDK) is accelerating with FY revenue of $20.2B (+175% YoY) and a data center business up 437% to $5.2B, including Q4 data center revenue nearly doubling sequentially to $2.9B. The company secured New Business Model agreements with 8 customers that provide a $93.9B floor and lifted remaining performance obligations from $59.8B to $91.1B (after two post-quarter deals), with terms averaging 4+ years. Despite the stock’s ~30x run and +500% in 2026 (after a ~33% pullback from June peak), the article argues valuation remains relatively cheap vs Nvidia, citing ~20x trailing P/E and ~7x forward P/E.

Analysis

The real beneficiary is not just SNDK but the downstream layer of the AI stack: storage, cache, and inference plumbing. That shifts bargaining power away from pure compute vendors and toward suppliers with long-duration allocation contracts, which should compress revenue volatility and support a higher multiple than legacy memory names. The second-order loser is any enterprise storage incumbent still anchored to HDD economics; if hyperscalers keep re-architecting data paths around lower-latency flash, replacement cycles can migrate faster than consensus expects.

Near term, the stock can keep levitating if management proves the contracted backlog converts into margins rather than just volume. The key watch item over the next 1-3 months is whether ASP discipline survives as supply tightens; if not, the market will quickly reclassify this as a cyclical earnings spike, not a durable rerating story. The main falsifier is a guide-down in data-center mix or a sequential slowdown in order intake after the next hyperscaler budget reset.

The contrarian miss is that investors may be treating these agreements like software-style annuities when they are really structured supply contracts with embedded cyclicality. That still matters because memory can re-price violently once the industry moves from shortage to balance, so the upside is likely more duration-driven than unlimited. Over 6-18 months, the upside case is a sustained scarcity regime; the downside case is that the market has already priced in the good news before the margin inflection fully shows up.

Key Decisions for Investors

• Long SNDK on pullbacks, but size modestly: use the next 5-10% drawdown or post-earnings volatility to build, targeting a 3-6 month rerate if backlog conversion stays intact.
• Pair trade: long SNDK / short WDC for a cleaner 'flash displacing legacy storage' expression; thesis breaks if HDD pricing tightens or WDC shows unexpected enterprise share gains.
• For risk-controlled upside, buy a 3-6 month SNDK call spread instead of common stock; the name has already moved sharply, so convexity is preferable to chasing spot.
• Use NVDA as a hedge, not a short: if hyperscaler capex rotates back toward compute, SNDK relative outperformance should pause; a SNDK/NVDA pair is only attractive if storage spend is clearly accelerating faster than accelerator spend.
• Watch for a guide-up in data-center gross margin and backlog conversion at the next print; absent that, take profits on any further multiple expansion because the market may already be discounting a quasi-annuity story.