Cathie Wood’s ARK buys Amazon stock, sells AMD and Rocket Lab

ARK ETF daily trades for Friday, April 24, 2026 showed a major purchase of 4.02 million shares of X-Energy Inc worth $92.5M and 280,450 shares of Amazon worth $71.5M, while trimming AMD by 215,643 shares worth $65.8M. The fund also sold Rocket Lab, Teradyne, Caterpillar, and Iridium, indicating active portfolio rotation rather than a fundamental company-specific catalyst. Overall, the article is a flow-driven update on Cathie Wood’s positioning with limited immediate market impact.

Analysis

The flow signal is more interesting than the headlines: the combination of AMZN add and AMD/TER trims looks like a rotation away from high-beta “AI picks-and-shovels” toward platform exposure and away from names whose upside is now more execution-sensitive. In the near term, that should help AMZN sentiment because it reinforces the idea that the market is willing to pay for durable cloud/compute optionality even as it becomes more selective on semiconductor and test-equipment names.

AMD is the clearest second-order loser because the sell decision reads less like a one-day rebalance and more like a view that the incremental trade is no longer in the chip vendor, but in the infrastructure and end-demand beneficiaries. If that interpretation is right, peers with similar valuation-to-growth mismatch can underperform for several weeks as active managers de-risk crowded AI beta. TER is exposed to the same narrative through capex cyclicality: if ordering momentum cools even modestly, test and validation spend tends to lag by 1-2 quarters, so the market may be discounting a softer second half before consensus catches up.

Contrarianly, the setup is not uniformly bearish for semis. A visible seller can create a temporary overhang, but if AMD guidance or channel checks improve, the stock can rip back quickly because positioning is likely still too dependent on passive AI optimism. The bigger miss may be that AMZN’s incremental buying is less about e-commerce and more about a broader “compute owner” trade; if that is the real thesis, capital should continue gravitating to the infrastructure layer with pricing power, while weaker second-tier hardware names get repriced on slower payback periods.

Key Decisions for Investors

• Short AMD for a 2-6 week tactical window; use a tight stop if it reclaims recent relative-strength highs, because the trade is vulnerable to any positive AI capex surprise. Best risk/reward is a partial short rather than full-size because the catalyst is flow-driven, not fundamental deterioration.
• Long AMZN on pullbacks over the next 1-3 weeks; the more important thesis is cloud/compute optionality, which should support multiple expansion if market leadership narrows. Favor call spreads over stock if implied vol remains elevated.