Cathie Wood Goes Bargain Hunting as AI Stocks Decline: 3 Stocks She Just Bought

Cathie Wood bought three AI-linked stocks on recent weakness: Space Exploration Technologies, Palantir Technologies, and Cerebras Systems. The article emphasizes dip-buying into newly public or recently re-rated names, including SpaceX after a drop of more than 20% from its peak and Cerebras after a 68% first-day IPO surge and subsequent pullback. The piece is primarily a stock-picking and positioning update rather than a fundamental catalyst, so likely market impact is limited.

Analysis

The setup is less about these names in isolation and more about the market's willingness to finance duration again. A soft risk tape plus visible insider-style accumulation from a high-profile allocator can compress the liquidity discount on pre- or post-IPO AI names, but that effect is fragile: these are still sentiment-led flows, not cash-flow-led re-ratings. The second-order winner is the entire private-markets complex, because any sign that large-cap public AI appetite remains strong improves late-stage private markups and lowers the cost of capital for adjacent venture-backed compute, data, and tooling names.

Palantir is the cleanest fundamental beneficiary because its commercial expansion gives it a path to translate AI enthusiasm into recurring spend rather than one-off beta. The key market inefficiency is that investors often underwrite it as a pure multiple story, but the real optionality is budget reallocation inside enterprise software: if AIP proves sticky, it can pull share from legacy analytics, SI, and workflow vendors over 2-4 quarters. Nvidia and AMD are the collateral beneficiaries, but the alpha is increasingly harder there; any upside from the theme is more likely to be captured through ecosystem suppliers or a relative-value long/short than through outright ownership.

The contrarian view is that the recent dip-buying behavior is exactly the wrong signal to chase after a sharp run: it may mark a short-term liquidity bid rather than durable institutional accumulation. For the newer AI IPOs, the failure mode is a window of re-underwriting where revenue growth stays impressive but the market stops paying for it because secondary supply, lockup expiries, or capex intensity dominate the narrative. That makes the next several weeks a tradeable positioning event, but the next several months depend on whether these companies can convert “AI promise” into visible margin expansion and operating leverage.

Key Decisions for Investors

• Long PLTR vs. long-only software basket: initiate a 3-6 month relative-value long PLTR / short legacy analytics-name basket. Target 15-20% relative upside if AIP adoption keeps inflecting; cut if commercial growth or net retention decelerates for one quarter.
• Avoid chasing NVDA outright here; use it as a hedge against thematic exposure. If buying AI beta, prefer a PLTR/NVDA pair: long PLTR, short NVDA on a 1:1 dollar basis to capture application-layer multiple expansion while reducing hardware valuation risk.
• Speculative long CBRS via call spreads, 2-4 months out, sized small. The risk/reward is asymmetric if the market re-rates post-IPO AI infrastructure winners, but the downside is high if second-quarter lockup/supply overwhelms demand.
• Short a basket of non-differentiated AI IPOs or adjacent high-multiple infrastructure names against PLTR. The market is likely to reward names with demonstrable enterprise adoption and punish those whose story still depends on future proof points.