**Yes, AMD remains worth considering for many investors as a long-term AI play, but Cathie Wood's recent sale isn't a strong "sell" signal—it's mostly profit-taking after a massive rally.**

On April 24, 2026, ARK Invest sold ~215,643 AMD shares (worth roughly $66–75 million at ~$348/share), while buying a similar amount of Amazon. This followed AMD's ~70% surge in the prior month (including a ~14% jump that day, boosted by strong Intel earnings signaling broader chip demand). ARK still holds a substantial position—AMD ranks as one of its top holdings (around 4–5% in ARKK recently, with total ARK exposure over $800 million post-sale). Wood/ARK has trimmed AMD multiple times in 2026 amid rebalancing, not as a outright rejection of the company.

### Why the rally happened

- **AI momentum**: AMD's data center segment (driven by Instinct MI300/MI350 GPUs) has shown strong growth. Hyperscalers like Microsoft, Meta, and Oracle are using or testing AMD accelerators for cost/diversification reasons versus Nvidia. Recent customer wins and roadmap visibility (MI350 series in 2025, MI400 in 2026 with advanced HBM4 memory) support expectations of double-digit AI market share gains.

- **Broader tailwinds**: Intel's results highlighted CPU demand tied to AI workloads; AMD continues gaining server CPU share from Intel while pushing Ryzen AI in PCs.

- **Stock performance**: AMD has been extremely strong—YTD returns over 50%, 1-year ~230–250%, with the stock hitting all-time highs near $350 before pulling back. It's volatile and high-beta.

### Bull case for buying AMD now

- **Growth runway**: Analysts project robust data center revenue expansion (e.g., 30%+ overall revenue growth in 2026 in some models), fueled by AI training/inference demand. MI350 offers competitive specs (high memory bandwidth), and MI400 is positioned as a step up. Long-term targets from Lisa Su include scaling data center revenue significantly.

- **Valuation and analyst view**: Forward P/E is elevated (often 40x+ or higher depending on estimates), but many see it as reasonable for the growth. Wall Street consensus is **Moderate Buy/Strong Buy** (dozens of Buy ratings, few Holds, no Sells). Price targets vary widely—median around $270–300, with highs to $380—implying mixed near-term upside from recent ~$320–340 levels but optimism on execution.

- **Ecosystem progress**: ROCm software is improving (though still trails Nvidia's CUDA in maturity and utilization). AMD benefits from any hyperscaler push for multi-vendor supply.

### Risks and bear case

- **Nvidia dominance**: Nvidia still holds ~80–90%+ of the AI accelerator market. AMD's real-world performance (e.g., utilization rates) often lags due to software/ecosystem gaps, even if raw specs look competitive on paper.

- **Execution and competition**: Delivering on the aggressive yearly roadmap (MI350 → MI400) while scaling production and winning meaningful share is key. China export restrictions add uncertainty. Custom ASICs from hyperscalers (Amazon, Google, Meta) could pressure merchant GPU demand.

- **Valuation and volatility**: After the recent run-up, the stock isn't "cheap." Pullbacks are common in semis. Cathie's sale (and any rotation out of hot semis) can contribute to short-term pressure, as seen in the ~3–4% dip post-trade.

- **Macro**: AI capex enthusiasm could moderate if returns disappoint or recession risks rise.

**Bottom line**: Cathie Wood's trade looks like classic ARK behavior—selling into strength to reallocate (here toward Amazon and other themes like nuclear energy) while keeping meaningful exposure. It doesn't scream "the thesis is broken." AMD has real AI tailwinds and a credible #2 position, but it's not Nvidia and carries higher risk of execution slips or valuation compression.

If you're bullish on sustained AI infrastructure buildout and believe AMD can take 10–20%+ share over time, it can make sense as part of a diversified tech portfolio (perhaps alongside or versus NVDA). Consider your time horizon, risk tolerance, and position sizing—semis move fast. This isn't personalized advice; do your own research or consult an advisor, and watch upcoming earnings/roadmap updates for fresh data.