Cathie Wood Goes Bargain Hunting: 3 Stocks She Bought Last Week

Cathie Wood’s Ark Invest bought more Nvidia, MercadoLibre and AMD during Thursday’s sell-off, when the stocks fell 9%, 12% and 17%, respectively, in the market’s biggest single-day decline in more than 19 months. The article presents the pullback as a potential buying opportunity: Nvidia was down 23% from its recent high and traded below 12 times next year’s profit target, while AMD had more than doubled in 2026 and was expected to grow faster than Nvidia next year. MercadoLibre reported 45% revenue growth in Q3 (41% currency-neutral), but margin pressure and rising credit provisions were cited as risks; its shares had fallen nearly half from their peak two summers ago.

Analysis

Wood’s dip-buying is a sentiment datapoint, not evidence of institutional demand large enough to change earnings. The more useful distinction is earnings durability: NVDA and AMD share exposure to data-center spending, but AMD’s higher forward multiple leaves less room for execution slippage even if its growth forecast is faster. NVDA’s upcoming report is a near-term catalyst; the key test is whether customer spending and forward guidance support the implied profit ramp, not whether it beats a headline estimate. Over 6–18 months, power availability, permitting resistance and hyperscaler custom silicon could constrain or redirect accelerator demand, limiting the assumption that overseas buildout simply offsets U.S. friction.

MELI’s issue is different: shipping subsidies can defend marketplace share while diluting operating leverage, and faster growth in credit-loss provisions than in the portfolio warrants scrutiny of loan vintages and delinquency trends. A low historical multiple, if realized, is not a floor when earnings quality is weakening. The contrarian risk is that the sell-off already prices in slowing growth, making an outright short vulnerable to stabilization; verify credit cohorts before taking directional exposure. Near term, NVDA earnings can reset the AI complex; over 1–3 months, guidance revisions and lending-quality disclosures matter more than dip-buying headlines.

Key Decisions for Investors

• Avoid adding to NVDA ahead of the report solely on the drawdown or reported fund purchases. Reassess after guidance: persistent capex commitments and supported forward growth would validate the dip; weaker demand visibility or margin commentary would falsify the long case.
• Consider a small, dollar-neutral NVDA-long/AMD-short relative-value position only after NVDA’s event risk passes, reflecting the valuation and execution-risk gap rather than a view that AMD lacks AI exposure. Exit or reassess if AMD demonstrates sustained accelerator share gains and upward estimate revisions, or if NVDA guidance weakens.
• Keep MELI on watch rather than shorting into a large decline. Require stabilization in operating margin and evidence from delinquency, nonperforming-loan and credit-vintage data before buying; worsening credit metrics alongside further margin compression would support avoiding or hedging the shares.
• Track data-center power and permitting developments and hyperscaler capex commentary over the next 1–3 months; these are potential demand constraints for both chip names, while custom-silicon adoption is a longer-term substitution risk.