Cathie Wood's Biggest Buy Last Week Was SpaceX. She Sold Almost Exactly That Much Palantir the Same Day.

ARK bought 205,031 SpaceX shares worth about $27.5M (its largest disclosed buy of the week) while selling 156,110 Palantir shares for about $27.2M on the same day, showing a roughly $27M in/$27M out rotation. The note frames SpaceX as trading near its $135 IPO price (after a prior run to $225.64) while Palantir—up ~40% since its August surge—trades at ~154x earnings despite rapid growth and an increased FY revenue outlook (~$8.15B). Net impact is limited, but the trades highlight ARK’s relative-value positioning between AI/space infrastructure and expensive AI software valuation.

Analysis

This reads less like a “buy/sell AI” statement and more like a duration trade inside the theme. The marginal message is that capital is being pulled from the highest-multiple public software winner into a private infrastructure asset with a nearer-term path to monetization, which is bearish for crowded AI application names and mildly supportive for compute/infra proxies like NVDA where revenue is still tied to real capex rather than narrative premium.

For PLTR, the risk is not operational collapse; it is multiple fragility. At ~150x earnings, the stock is priced for a long runway of 50%+ growth and expanding margins, so even a modest deceleration or guide-reset can trigger a 20-30% drawdown in weeks as the market reprices terminal assumptions. The recent trim pattern matters because it suggests sophisticated holders are willing to sell into strength before the market tests whether growth can stay at this pace for multiple quarters.

SPCX is a different beast: near its reference price, it has anchoring support, but the equity still behaves like a long-duration private asset whose value depends on sustained capital-market appetite. The second-order risk is that “private winner” enthusiasm can reverse if financing conditions tighten or if the next lockup/secondary print comes at a lower implied multiple. Over 6-18 months, the key question is whether the company proves it can turn growth into durable free cash flow; until then, the valuation remains hostage to sentiment rather than fundamentals.

Key Decisions for Investors

• Short PLTR via 1-3 month put spreads or a covered call overwrite on existing longs; risk/reward is attractive if growth merely normalizes, with a potential 20-30% downside on multiple compression versus limited premium outlay.
• Use NVDA as the liquid beneficiary proxy for AI-capex reallocation: add on 5-10% pullbacks over the next 1-2 months if the market continues to favor infrastructure over software; thesis breaks if hyperscaler capex rolls over.
• Stay neutral on SPCX directly; if access is only through ARK funds, treat any weakness in ARKK/ARKW as a better entry than chasing the private-name mark near reference price. Add only if there is evidence of sustained post-IPO demand, not on one-day flows.
• Relative-value pair for 1-3 months: long NVDA / short PLTR to express ‘AI picks-and-shovels over AI software multiple risk.’ Best entry is after a PLTR squeeze or on any negative AI sentiment day; exit if PLTR keeps accelerating revenue growth above 50% with no margin giveback.
• Watch for a PLTR re-rating trigger: if the next earnings report shows guidance raised but growth still slows, that is the clearest falsifier for the bull case and a catalyst to press shorts; if growth re-accelerates above current expectations, cover quickly.