Sandisk has gained more than 600% in 2026. It still has room to run, Rosenblatt says

Rosenblatt initiated Sandisk with a buy rating and $2,400 price target, implying 36% upside from Monday's close, citing NAND flash's growing role in AI infrastructure. Sandisk shares have already surged 644% year to date as AI-model growth boosts demand for high-density, high-performance enterprise storage. The firm expects Sandisk's BiCS8 and BiCS10 3D NAND platforms to preserve a performance and bit-cost advantage, while 24 of 28 covering analysts hold buy or strong-buy ratings.

Analysis

The investable question is not whether AI lifts NAND demand, but whether SNDK can convert a higher-value enterprise mix into durable gross-margin expansion before industry supply responds. A NAND upcycle typically attracts aggressive wafer-output additions from Samsung, Kioxia and SK hynix; if pricing remains disciplined for the next 2-3 quarters, SNDK's operating leverage should be unusually high, but a renewed capacity race would rapidly reassert commodity economics. The key verification points are enterprise SSD mix, realized price-per-bit relative to cost-per-bit, and management's capex/wafer-start discipline—not benchmark claims around node density.

The stock's 644% run and near-unanimous positive sell-side positioning create unfavorable short-term asymmetry: incremental upgrades are less important than any evidence that AI storage demand is being satisfied through inventory builds or that hyperscalers are substituting DRAM/HBM, networking, or proprietary storage architectures for incremental NAND content. Over 1-3 months, monthly NAND contract-price data and the next earnings guide should determine whether estimates continue rising. Over 6-18 months, SNDK's advantage is most credible if it wins qualified enterprise designs and sustains margin gains while competitors' supply additions stay delayed.

A less obvious beneficiary is equipment and controller exposure rather than NAND beta alone: sustained technology migration supports memory-capex suppliers such as LRCX and AMAT, while enterprise-storage vendors such as PSTG could capture higher system-level spending without bearing NAND spot-price risk. Conversely, a NAND price spike can pressure OEM storage margins at DELL and HPE before they reprice contracts. The contrarian view is that the market may already be capitalizing a multi-year AI scarcity premium into a historically cyclical memory asset; absent evidence of structural pricing power, upside from here likely requires earnings revisions rather than further multiple expansion.

Key Decisions for Investors

• Do not chase SNDK after the parabolic move; establish only on a post-earnings pullback if management raises enterprise SSD revenue/margin expectations and confirms constrained industry supply. Size as a tactical 1-3 month long, with a stop on a guide-down or evidence of sequential NAND contract-price declines; target risk/reward should be at least 2:1 rather than underwriting the published price target.
• Prefer a relative-value expression: long SNDK / short a broad semiconductor ETF such as SOXX for 3-6 months, contingent on SNDK demonstrating enterprise mix expansion above industry NAND growth. This isolates company-specific margin upside while reducing AI-beta and rate-sensitive multiple risk; exit if gross-margin guidance fails to improve despite higher revenue.
• Monitor Samsung Electronics (005930 KS), SK hynix (000660 KS), and Kioxia capacity announcements as a falsification signal. Material wafer-start expansion or a return to inventory-led channel demand would favor reducing SNDK exposure and could create a tactical short setup after the next earnings window.
• For lower-volatility AI-infrastructure exposure, evaluate LRCX or AMAT only if memory makers translate improved pricing into higher 2027 capex guidance; without that confirmation, equipment upside is speculative and no position is warranted.