Cathie Wood Goes Bargain Hunting: 3 Stocks She Just Bought

Ark Invest's Cathie Wood added to Nvidia, SpaceX and Airbnb, emphasizing growth opportunities despite their differing share-price setups. Nvidia authorized a record $150B buyback and has $235B of total repurchase capacity, while revenue growth accelerated from 56% in Q2 FY2026 to 106% in Q2 FY2027; current-quarter guidance calls for 89% growth. SpaceX remains unprofitable but is expected to turn profitable next year after shares fell more than one-third from their debut peak, while Airbnb is projected to deliver 16% revenue growth this year, its fastest pace in three years.

Analysis

NVDA's authorization is not an incremental earnings catalyst until cash deployment accelerates; at its current scale, the economic effect on EPS is modest relative to quarterly estimate revisions driven by hyperscaler capex. The more important signal is capital-allocation: a large unspent authorization can support drawdowns and reduce float, but it also raises the hurdle for organic AI infrastructure returns. Over the next 1-3 months, the stock remains most sensitive to Blackwell/rack-scale shipment timing, cloud-provider capex commentary, and any evidence that inference workloads are shifting toward lower-cost custom silicon from GOOGL, AMZN, and hyperscalers.

ABNB has a cleaner idiosyncratic setup than the AI complex: accelerating growth would support a multiple re-rating only if it comes with stable take rates and marketing efficiency rather than discounting. The key underappreciated risk is regulatory fragmentation: city-level restrictions can create nonlinear supply losses in high-value urban markets, while simultaneously benefiting hotel operators and online travel peers with more conventional inventory. A softer consumer environment would likely pressure booking lead times and ADR before it appears in headline revenue growth, making next-quarter gross booking value and nights/booked guidance the relevant confirmation points.

SPCX is a high-beta duration asset rather than a near-term fundamental trade. A path to profitability can compress its financing risk, but launch cadence, Starlink subscriber net additions, spectrum policy, and satellite replacement capex determine whether growth converts into free cash flow. The contrarian view is that the post-listing rebound may already price the visible growth trajectory; without independently verified segment margins and cash-burn data, buying strength is less attractive than waiting for a financing or execution-driven dislocation.

Key Decisions for Investors

• Maintain NVDA as a core long but do not add solely on the repurchase headline; add on a 8-12% pullback or after next hyperscaler capex read-through confirms 2027 demand. Thesis fails if management guides data-center growth materially below consensus or gross margin falls on rack-scale mix; use a 10% downside stop from entry.
• Express relative AI risk with long NVDA / short SMH only if NVDA's forward EPS revisions continue to outpace the semiconductor basket over the next 4-8 weeks. Target 8-12% relative outperformance; exit if custom-ASIC commentary from AMZN/GOOGL drives a sustained NVDA estimate-cut cycle.
• Accumulate ABNB over a 3-6 month horizon only after checking whether growth is accompanied by expanding adjusted EBITDA margin and stable urban supply. A long ABNB / short EXPE pair is preferable to an outright consumer-discretionary long; target 10-15% relative return, with regulatory supply restrictions or GBV deceleration below guidance as falsifiers.
• Keep SPCX on watch rather than initiating after the rebound. Require disclosure confirming positive contribution margins, a credible free-cash-flow bridge, and manageable satellite replenishment capex; otherwise treat a 20%+ pullback around earnings or financing updates as the only attractive entry window.