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

Sandisk (now a stand-alone NAND/SSD specialist post the Western Digital spin) reported FY revenue of $20.2B, up 175% YoY, with the data-center segment surging 437% YoY to $5.2B and nearly doubling sequentially to $2.9B in Q4. The key underwriting catalyst is ~$93.9B in minimum committed revenue via New Business Model agreements with eight data-center/edge customers, supported by $91.1B remaining performance obligations after two post-quarter deals (weighted average term >4 years). Despite a ~30x stock run over the past year (and >500% in 2026), the article argues valuation remains cheap versus Nvidia’s early AI breakout (P/E ~20; forward P/E ~7), suggesting further multiple expansion if the mix keeps shifting toward data center.

Analysis

The real mechanism here is not "AI demand" generically; it is the migration of storage from a volatile component line item into a semi-contracted infrastructure layer. That should compress earnings volatility and, over time, justify a higher quality multiple versus the typical NAND basket. The closest second-order winners are memory-adjacent equipment names and controller/IP suppliers that benefit if SNDK’s customers extend the storage buildout; the losers are any storage vendors still living on spot pricing and inventory turns.

Near term, however, the setup is more crowded than the story implies. A 30x move means the market is already paying for some version of the backlog narrative, so the next leg likely depends on conversion into cash flow and margin stability, not just headline revenue growth. The key risk is that the new contract structure can reduce downside but also cap upside if NAND pricing tightens faster than expected; that would leave investors owning a "safer" business at a less exceptional return profile.

The contrarian miss is assuming RPO is the same as durable value creation. What matters over the next 1-3 quarters is realized gross margin, working-capital discipline, and whether hyperscalers keep rebalancing spend toward storage after the current GPU-heavy wave. If that mix stalls, SNDK can de-rate fast because the stock is still being treated like a cyclical semiconductor name in practice, even if the narrative says otherwise.

Key Decisions for Investors

• Tactical long SNDK only on pullbacks of 10%+ or after an earnings reset; target a 3-6 month hold, but require confirmation that data-center mix and gross margin both continue higher. Risk/reward improves if forward EV/EBITDA stays below the memory peer group.
• Pair trade: long SNDK / short WDC for 1-3 months to express the view that contracted AI storage exposure deserves a higher multiple than legacy storage exposure. Falsify if AI-related storage demand broadens into HDD/nearline faster than expected.
• Buy a modest SNDK call spread into the next quarterly print only if implied volatility cheapens after the recent run; use it as a convex way to play continued backlog conversion while defining risk.
• Set a watch item on any sequential slowdown in SNDK data-center growth or gross margin compression; if either cracks, the rerating thesis is likely over-extended and profits should be reduced.