Billionaire Ken Griffin Sells Sandisk Stock and Buys a Quantum Stock Up 1,900% Since Early 2023

In Q3 Citadel Advisors (Ken Griffin) cut its Sandisk position by ~97% (selling nearly 2 million shares) and initiated a small stake in D‑Wave Quantum. Sandisk benefits from an unprecedented NAND supply shortage driven by AI—non‑GAAP earnings fell 33% in Q1 but the company expects a ~160% bounce in Q2 and Wall Street projects adjusted earnings to grow ~259% annually to FY June 2027, yet the stock trades at ~170x earnings and faces cyclical downside risk. D‑Wave, an early commercial quantum-annealer vendor, has strong growth momentum (stock up ~1,900% since 2023) but trades at ~347x sales despite analysts projecting ~70% CAGR to 2027 and a relatively small addressable market (~$4bn by 2030), making its valuation speculative; Citadel’s moves appear tactical and small-sized rather than long-term conviction.

Analysis

Market structure: Short-term winners are integrated NAND suppliers (SNDK, WDC) and AI datacenter beneficiaries (NVDA) as constrained NAND supply supports pricing and margins through 2026; losers include downstream OEMs facing inventory shocks if memory pricing reverses and momentum-exposed quantum names (QBTS) if sentiment cools. Citadel’s SNDK trim signals institutional positioning shifting from cyclicality risk to idiosyncratic/spec momentum trades (QBTS), implying market breadth may narrow into mega-cap AI names while smaller caps see higher volatility.

Risk assessment: Tail risks include a rapid capex/production ramp by Kioxia/Chinese entrants that could collapse NAND ASPs >40% within 12–18 months, and for QBTS, technical setbacks or regulatory export controls that could devalue equity by >70% over years. Immediate (days–weeks) risk is elevated IV and headline sensitivity; medium (3–12 months) hinge on spot NAND prices and SNDK guidance; long-term (2–5 years) depends on quantum commercialization and AI-driven durable demand.

Trade implications: Favor de-risking memory cyclicality and favor AI infrastructure exposure. Use option structures to express views: buy protective puts on SNDK or sell premium against small QBTS momentum longs. Rebalance size: cap speculative quantum exposure to <0.5% portfolio and keep memory exposure hedged through 6–12 month derivatives.

Contrarian angles: Consensus underestimates geopolitical bottlenecks that could extend the memory supercycle beyond 2026, which would keep SNDK elevated — but that’s a binary risk. Conversely, QBTS could become an M&A/partnering target for hyperscalers before revenue scales; however, current 347x sales multiples imply any positive technical proof points will be quickly priced, so nimble, size-capped momentum plays are preferable to buy-and-hold.

Key Decisions for Investors

• Trim SNDK equity exposure to ≤1% of portfolio within 2 weeks; if retaining any exposure, purchase a 6–9 month put spread (buy 30% OTM puts, sell 15% OTM puts) sized to protect remaining position—target cost ≤2% of notional to cap downside if NAND ASPs collapse.
• Establish a tactical long QBTS momentum sleeve no larger than 0.25–0.5% of portfolio using 3–6 month 25% OTM calls (or outright equity) with a hard stop-loss at −40% from entry and time-based exit at 6 months unless fundamental progress is announced.
• Pair trade for 6–12 months: go long NVDA (2% portfolio) vs short SNDK (1% portfolio) to express secular AI demand outperformance versus cyclical memory risk; trim both sides on quarterly spot NAND price prints or SNDK guidance changes >±20%.
• Reduce broader memory/semi cyclicals weight by 3–5% and redeploy into semiconductor equipment/AI infrastructure leaders (allocate to NVDA or traded ETF exposure) over the next 1–3 months to capture secular capex while avoiding peak-memory exposure.