Cathie Wood’s ARK sells AMD stock, buys Shopify and GeneDx

ARK reported heavy portfolio rotation on May 5, 2026, led by a $32.6M purchase of 255,804 Shopify shares and a $16.7M purchase of 245,914 GeneDx shares. On the sell side, it reduced exposure to AMD by 45,917 shares worth $15.7M and trimmed CoreWeave by 98,880 shares worth $12.4M, while also selling smaller stakes in Teradyne, Twist Bioscience, and 10x Genomics. The trades suggest ongoing repositioning within tech and biotech rather than a broad directional market call.

Analysis

The biggest signal here is not the individual names, but the direction of capital inside ARK’s book: they are rotating out of semis/test equipment and select hyper-scaling infrastructure names into software and genomics where the perceived payoff is more idiosyncratic and less tied to the AI capex cycle. That matters because the sell list clusters around beneficiaries of the prior hardware spend wave, which often precedes a period of multiple compression once investors start asking whether end-demand is keeping up with supply. In other words, this looks less like a stock-picking change and more like a late-cycle factor rotation away from crowded “picks-and-shovels” exposure.

AMD and TER are the most important tells for near-term sentiment. If an aggressive growth allocator is trimming semis while the broader AI trade is still institutionally crowded, the second-order read-through is not just lower upside for those names; it is potential de-rating across adjacent beneficiaries with similar narrative torque but weaker earnings revision support. CRWV is especially vulnerable because the market is still price-sensitive on anything tied to AI capacity buildout, and repeated distribution can pressure both relative performance and new issue demand for the ecosystem.

On the buy side, SHOP and NTLA suggest ARK is leaning into names where operating leverage is longer-dated and less dependent on this quarter’s margins. That creates a cleaner path for momentum if macro uncertainty persists, but also a sharper downside if rates rise or risk appetite fades because these are duration assets in disguise. WGS fits the same pattern: genomics gets a bid when investors want exposure to differentiated data assets rather than commoditized tools, but the follow-through depends on evidence of commercial conversion rather than thesis alone.

The contrarian angle is that this may be less bullish for SHOP/NTLA than it appears and more a sign that ARK is forced to chase what has recently worked in its own basket. If so, the better trade is to fade the highest-beta hardware names on rallies rather than chase the buys, because the market is likely to reward visible earnings and punish “story” names until the next growth scare resets positioning.

Key Decisions for Investors

• Short-term: fade rips in AMD over the next 1-3 weeks; use strength above recent intraday highs to establish a tactical short or buy puts, targeting a 5-8% downside if semis continue to de-rate on weak incremental demand signals.
• Pair trade: long SHOP / short a basket of AI hardware proxies (AMD, TER) for 1-2 months; thesis is that software sentiment can stay supported while hardware multiple expansion stalls, with the pair benefiting from factor rotation rather than broad beta.
• Add NTLA on pullbacks over the next 2-4 weeks only if volume confirms; the position works best as a catalyst-driven trade into data/events, with upside optionality meaningfully larger than downside if risk appetite remains intact.
• Avoid chasing CRWV into strength; consider a hedged short via put spreads for 1-2 months because repeated selling into a crowded AI infrastructure name often precedes a sharper drawdown than fundamentals alone would imply.
• Relative-value: long WGS vs TXG for a 1-3 month horizon; the market appears to prefer names with clearer data monetization narratives, while commoditized tools remain vulnerable to margin skepticism.