Cathie Wood’s ARK buys Palantir stock, sells AMD shares

ARK disclosed net portfolio shifts on April 11, 2026, buying 85,485 shares of Palantir for about $11.15 million while selling 44,446 shares of AMD worth $10.52 million. It also reduced its Strata Critical Medical position by 75,389 shares, or $305,325. The trades signal continued conviction in PLTR and lighter exposure to AMD and SRTA, but the article is primarily a flow update rather than a fundamental catalyst.

Analysis

ARK’s flow is a sentiment signal, not a fundamental one, but it matters because it is concentrated in names where incremental demand can still move implied positioning. The more interesting read is that the buyer is leaning into a software/data platform with high narrative elasticity while trimming a cyclical semis winner, suggesting a preference for assets with cleaner AI monetization optionality and less dependence on capex-cycle timing. That matters for relative performance: if AI breadth rotates away from hardware beta, the market is likely to reward “software as decision infrastructure” more than compute suppliers over the next 1-3 months.

PLTR is the clearest beneficiary because any added demand from a visible growth sponsor reinforces a loop of passive ownership, retail follow-through, and higher multiple stability. The second-order effect is on adjacent AI software names: if investors start treating PLTR as the preferred public-market AI operating system, smaller platform vendors and legacy analytics software names may underperform as capital consolidates into the perceived category leader. Conversely, AMD’s reduction is not a call on the chip cycle per se; it can pressure sentiment across AI hardware multiples if others interpret it as a signal that near-term upside in accelerators is more crowded than consensus assumes.

The contrarian risk is that this flow is late-cycle performance-chasing rather than insight. If the next earnings prints show accelerating enterprise AI monetization and stable inference demand, trimming semis too early could be a mistake; compute remains the gating factor for model deployment over the next several quarters. SRTAW looks more like a forced de-risking of a thinly traded position than a real thesis call, so I would not extrapolate it beyond a liquidity constraint lens.

Near term, this is a positioning trade with a 2-8 week horizon, not a fundamental catalyst. The fastest reversal would come from AMD guidance that re-anchors AI semiconductor growth expectations or a PLTR print that shows any slowdown in commercial deal velocity; either would unwind the sentiment leg quickly. Absent that, the flow backdrop should keep PLTR better bid relative to AI hardware peers, especially if market breadth narrows and investors pay up for “proof” over “promise.”

Key Decisions for Investors

• Long PLTR / short AMD for 2-6 weeks as a relative-value expression on AI software over hardware breadth; target a 5-8% spread capture, stop if AMD regains relative strength on guidance or channel checks.
• Buy PLTR on pullbacks rather than strength-chasing; use a staggered entry over 3-5 trading sessions and prefer call spreads over outright stock to cap valuation risk if flow-driven momentum fades.
• Sell/underweight AMD into any pre-earnings strength over the next 1-2 months; the risk/reward is asymmetrically worse if the market is already pricing uninterrupted AI capex acceleration.