Bull of the Day: Sandisk (SNDK)

Sandisk, spun off from Western Digital in Feb 2025, reported a blockbuster start to FY26 with revenue roughly $2.3–2.38 billion and a 37% EPS beat while non-GAAP gross margin expanded to ~29.8–29.9%. Management guided Q2 EPS of $3.00–$3.40 and revenue $2.55–$2.65 billion with margins targeting 41–43%, and raised full-year revenue to about $10.45 billion (+42% y/y), underpinning a 680% stock rally since August. Analyst estimates have more than doubled across near- and longer-term horizons amid tight NAND supply and structural AI-driven demand — including Citi and others quantifying substantial incremental SSD/NAND needs tied to NVIDIA's upcoming Vera Rubin platform — supporting a bullish investment thesis for hyperscaler and AI-edge exposure.

Analysis

Market Structure: Winners are pure-play high-density NAND suppliers (SNDK, MU) and hyperscalers (AMZN, MSFT) that pay premiums for allocated capacity; losers include legacy HDD vendors (WDC/Seagate) and smaller SSD OEMs unable to qualify BiCS8. Tight supply and multi-quarter allocations—bolstered by Citi’s Vera Rubin estimate (34.6M TB NAND in 2026; 115.2M TB in 2027 = 2.8%/9.3% of demand)—create near-term pricing power and elevated gross margins (management guiding 41%–43%). Cross-asset: higher cash flows and capex lift semiconductor equipment and HY credit spreads compress, options IV on SNDK will stay rich, and cyclical commodity/capex inflation could modestly pressure yields.

Risk Assessment: Tail risks include rapid Chinese/IDM capacity additions or an aggressive Micron ramp that flips to oversupply (high-impact, 12–24 months), a major design win failure at NVDA, or export-control shocks interrupting fabs. Immediate (days–weeks): sentiment-driven 20–40% pullbacks possible; short-term (months): earnings/guidance cadence will reprice multiples; long-term (2027+) structural demand depends on sustained ICMS server shipments and controller availability (Stargate). Hidden dependencies: controller/fab yield constraints and software/firmware qualification cycles create chokepoints that can cap growth even with demand.

Trade Implications: Allocate size via staging and options to manage blow-ups: prefer a core long SNDK (phased 2–4% portfolio) held through FY26, augmented with 6–12 month call spreads to cap cost; pair-trade SNDK long vs WDC short to express NAND vs HDD divergence. If IV spikes, sell near-term covered calls against newly bought shares or buy protective puts 25–30% OTM; overweight semiconductor suppliers and equipment suppliers while reducing long-duration consumer cyclicals if yields rise.

Contrarian Angles: Consensus underestimates speed of a supply response and overprices permanent margin expansion—680% rally implies >50% of upside is multiple expansion. Historical memory cycles (DRAM 2016–18) show sharp mean reversion once capex accelerates; watch for early signs of supplier capital commitments (announced wafer starts rising >15% YoY) as a trigger to pare positions. Unintended consequence: hyperscalers could push for internal sourcing or stricter price clauses, compressing ASPs despite volume growth.

Key Decisions for Investors

• Establish a phased long position in SNDK equal to 2–3% of portfolio now (limit buys $360–$395), add to reach 4–5% only on confirmed pullback of 15–25% (target add zone $300–$335). Hold through FY26 earnings cadence; take 50% profits on 50% price appreciation or cut to breakeven if SNDK misses guidance by >10%.
• Implement a 6–12 month asymmetric options trade: buy SNDK Sep/Dec 2026 call spread (e.g., buy 400 / sell 600) size = 1% portfolio notional to capture upside while capping premium; simultaneously buy 30% OTM Sep/Dec 2026 puts as tail protection if downside flows emerge.
• Execute a relative-value pair: long SNDK (1% notional) vs short WDC (0.75% notional) to isolate NAND upside vs HDD secular decline; rebalance if spread narrows >25% or if WDC announces major NAND-related capex shifts.
• Rotate 3–5% from consumer discretionary into semiconductor suppliers (MU, select equipment names) over next 4–8 weeks; reduce exposure to PC-tilted names if SNDK gross margin guidance >40% becomes consensus (suggests price elasticity falls).
• Monitor three triggers over 30–90 days before increasing sizing: (1) NVDA Vera Rubin confirmed shipments >20k by mid-2026, (2) supplier wafer-start announcements up >15% YoY, (3) hyperscaler multi-year contracts announced with volume and price disclosure—if none materialize, reduce SNDK exposure by half.