Cathie Wood’s ARK sells Nvidia stock, buys Arcturus Therapeutics

ARK Invest published March 27 trades: sold 58,119 NVDA shares across ARKK/ARKW/ARKF for ~$9.95M (following a 155,441-share sale on March 26 for $27.77M). Other notable sales: 10,500 META for $5.75M (after 76,622 shares for $45.58M prior), 19,126 AMD for $3.90M (after 38,245 for $8.42M), 17,092 TER for $5.08M, 37,876 BLSH for $1.38M, 9,274 ROKU for $825.7K and 4,221 TSLA for $1.57M. ARK bought 48,659 ARCT shares via ARKG for $344,505. These trades reflect active reallocation within tech, semiconductors, EVs and biotech but are unlikely to be market-moving for large-cap names like NVDA given the dollar sizes involved.

Analysis

Recent concentrated selling pressure among momentum/innovation ETFs creates an asymmetric near-term liquidity shock for consensus mega-cap growth names; when marginal passive and quant buyers step back, realized volatility and option skews rise, compressing realized returns for large-cap AI/advertising names while opening tactical windows for under-owned cyclical or niche suppliers. This flow-driven volatility is not a fundamentals-driven re-rating—hence catalysts that re-ignite durable demand (datacenter refresh announcements, strong ad-monetization metrics, or beat-and-raise quarters) can reverse the move within 4–12 weeks, whereas inventory-driven weakness in capital expenditure often plays out over 2–6 quarters. In the semiconductor/equipment chain, the highest beta to a short-term capex pause are automation and test suppliers and capital-intensive OEMs; conversely, vertically integrated server OEMs and software-differentiated players with services/recurring revenue will see relatively steadier cash conversion and are candidates for rotation inflows. Small-cap biotech niches with idiosyncratic catalysts (early-stage readouts or partnership wins) are insulated from these macro flow dynamics short-term and can attract selective buying, but they require event-specific sizing and quick reaction to binary outcomes.

Key Decisions for Investors

• Pair trade (6–12 months): Long SMCI on a 10–15% intraday pullback (scale in 0.5–1.0% NAV), paired with a size-weighted short of NVDA equal to 25–33% of position delta exposure to control market beta; target +40–60% on SMCI vs -10–20% on hedge (net target return ~2.5x risk), stop-loss if SMCI drops 20% from entry or NVDA outperforms by >15% in 2 weeks.
• Options tactical (0–3 months): Buy a 3-month ATM call spread on ARCT (size 0.25–0.5% NAV) to express event-driven upside while capping premium outlay; target 3–5x return on premium if a positive readout/partnership occurs, max loss = premium paid.
• Capex downside play (3–6 months): Buy a 3-month put spread on TER (or use AMD if preferred) to express test/automation exposure to a cyclical pause — risk limited to premium, target 2:1 payoff if order momentum deteriorates or guidance is cut at next results.
• Vol/flow hedge (2–6 weeks): Buy a small NVDA 30–60 day put calendar or long-dated tail protection (size 0.25% NAV) to protect portfolios from idiosyncratic spike risk from ETF-driven selling; expect increased IV to compound returns on protection if selling persists, use as cheap insurance ahead of earnings/data-center cadence.