
SanDisk surged 6.31% to $1,335.13 after reporting Q3 FY2026 revenue of $5.95B, a 26% beat versus estimates, with datacenter revenue up 645% year over year to $1.467B and gross margin expanding to 78.4% from 22.5%. Management guided Q4 EPS to $30-$33 versus the $22.01 estimate, while multiple analysts raised price targets, including Bernstein to $1,700, reinforcing the bullish rerating. The company also highlighted over $42B in multi-year AI infrastructure supply agreements, underscoring strong demand and pricing power amid tight NAND supply.
The key second-order effect is that SNDK’s print is not just a one-off upside surprise; it is evidence that AI capex is migrating from compute scarcity to storage scarcity. If hyperscalers are locking multi-year commitments, the margin structure for the entire NAND stack can stay elevated longer than consensus expects, because pricing discipline now has a contractual anchor rather than just spot-market tightness. That means the market may be underestimating duration: this is less a quarter-by-quarter re-rating and more a multi-year operating regime change. The biggest winners beyond SNDK are likely the suppliers closest to constrained upstream nodes and the memory equipment ecosystem, while the losers are downstream OEMs and consumer electronics assemblers that rely on opportunistic NAND pricing. A sustained memory upcycle also pressures storage-heavy server OEM margins unless they can reprice quickly, which tends to lag by 1-2 quarters. In equities, the hidden bull case is not just “AI spending,” but improved visibility, which lowers perceived earnings risk and can justify higher multiples across the group. The main risk is that consensus extrapolates a peak gross margin environment into perpetuity. If capacity additions or inventory normalization arrive in late 2026, the air pocket can be sharp because the stock has likely pulled forward several years of optimism. Another reversal trigger is customer pushback: if AI buyers start optimizing for cost per workload and shift architecture toward more compute-per-dollar efficiency, storage intensity could decelerate faster than revenue models imply. Contrarianly, the move may be under-discounting the breadth of beneficiaries outside SNDK itself. The cleanest expression may be a relative-value long in the strongest memory name versus short a storage/OEM name exposed to input-cost inflation, rather than chasing outright momentum. Near term, this is a momentum tape, but over 6-12 months the better trade is to own the supply-constrained upstream and fade any downstream margin compression.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request DemoOverall Sentiment
extremely positive
Sentiment Score
0.92
Ticker Sentiment