Cathie Wood’s ARK buys CoreWeave, sells Bullish stock

ARK disclosed a mixed set of daily trades for April 28, 2026, led by a $18.2M purchase of CoreWeave (162,306 shares) and a $14.2M buy of Alphabet (40,656 shares). The firm also added to Kratos ($6.2M), Intellia ($7.8M), and Spotify ($5.7M), while trimming Bullish ($1.1M), Intercontinental Exchange ($1.7M), and Roku ($4.2M). The report reflects routine portfolio repositioning with a clear tilt toward technology and biotech rather than a material macro catalyst.

Analysis

The more interesting signal is not the individual names but the factor bundle: ARK is rotating toward high-beta, AI-adjacent compute, defense tech, and platform software while trimming financial infrastructure and crypto-exposure proxies. That mix suggests a preference for businesses where near-term product cycles can re-rate multiples faster than fundamentals alone would justify, which usually works best when liquidity is improving and risk appetite is stabilizing. The second-order effect is that any sustained follow-through could tighten supply-demand in smaller names with limited free float, especially in CRWV, KTOS, and NTLA, where discretionary ownership can move the tape more than fundamentals in the next 1-3 months.

CRWV is the clearest momentum expression, but it also carries the highest sequencing risk: if AI capex enthusiasm pauses, the stock can de-rate violently because expectations are already pulling forward multi-quarter demand. By contrast, GOOG is the lower-volatility way to express the same AI monetization theme; if large-cap tech leadership broadens, the market may reward balance-sheet quality over pure narrative exposure. KTOS looks like the most asymmetric defense-tech trade because it benefits from the same strategic spend cycle as primes, but without the same analyst saturation, so incremental flows can matter disproportionately.

The sells matter too: trimming ICE and BLSH looks like a de-emphasis on toll-collector and crypto-beta exposure at a moment when macro conditions are making speculative assets less forgiving. ROKU is the weak-link here: if ad budgets soften again, the stock can underperform even in a benign tape because it lacks the durability of SPOT’s subscription/engagement model. NTLA remains a long-duration catalyst story, but biotech accumulation only works if data-readout timing stays within the market’s patience window; otherwise, carry and dilution risk can overwhelm sentiment.

Key Decisions for Investors

• Long KTOS vs short ROKU for 4-8 weeks: defense-tech flow plus budget resilience versus ad-beta vulnerability; target 1.5x upside/downside skew if risk appetite stays stable.
• Buy GOOG on pullbacks for a 3-6 month horizon: preferred AI monetization exposure with lower drawdown risk than CRWV; use CRWV only as a tactical momentum long, not a core position.
• Pair long SPOT / short ICE for 1-2 months: consumer recurring revenue and engagement should be less rate-sensitive than transaction/toll-collector exposure if liquidity stays choppy.
• Avoid chasing CRWV after this print unless it consolidates for 5-10 sessions; if entered, size small and use tight stops because the trade is flow-driven and can unwind fast.
• Keep NTLA as an event-driven long only if there is a clear catalyst window within the next 1-2 quarters; otherwise prefer options over equity to limit dilution and binary trial risk.