Cathie Wood, founder of Ark Investment Management and a prominent advocate for disruptive technology stocks, has made a notable reversal in her holdings of two leading AI chip giants, drawing wide attention across the investment community.
Wood’s flagship Ark Innovation ETF (ARKK) has delivered mixed performance. As of late April, the fund has surged 49% over the past 12 months, yet posted an average loss of 8.7% across the past five years. Every move Wood makes tends to set market trends, and her latest adjustment has left investors debating whether to follow her lead.
Recently, Wood has reduced her positions in Advanced Micro Devices (AMD) while ramping up holdings in its rival Nvidia (NVDA). Even with the sell-off, AMD remains the fifth-largest holding in Ark’s portfolio. The stock has quadrupled in value over the past year, pushing its forward valuation to 35.5 times earnings based on 2027 analyst forecasts. The steep run-up in valuation is widely seen as the key reason behind Wood’s profit-taking.
AMD boasts solid dual growth drivers in the AI sector. Its graphics chips are well-suited for AI inference workloads, helping the company capture greater market share. It has secured two major GPU partnership deals each valued at roughly $100 billion, and is expected to supply next-generation chips to AI firm Anthropic, strengthening its edge in the inference market.
Beyond inference, the rise of agentic AI is lifting demand for central processing units. The industry sees a dramatic drop in the typical GPU-to-CPU ratio for agentic AI applications. As a top player in data center CPUs, AMD has consistently chipped away at Intel’s market share. The related market is projected to reach as high as $200 billion in the coming years. Given AMD’s total revenue stood at below $35 billion in 2025, the company is poised for robust long-term expansion.
Nvidia, by contrast, stands out with its attractive valuation and explosive growth, which explains Wood’s aggressive buying. Its forward price-to-earnings ratio based on fiscal 2028 estimates is just 16 times. The firm reported an 85% year-over-year revenue jump to $81.6 billion last quarter, and guided for a 95% revenue increase to $91 billion in the second quarter.
As the undisputed leader in AI model training, Nvidia is also expanding into inference and agentic AI. It has developed in-house ARM-based CPUs and added language processing units through acquisitions. By combining different hardware products, the company delivers tailored solutions for the full inference process. Supported by its comprehensive networking lineup, Nvidia now offers end-to-end server systems covering AI training, inference and agentic AI scenarios.
Between the two chip leaders, Nvidia appears the more compelling pick at present thanks to its valuation profile and strong growth momentum. Meanwhile, analysts note investors can hold both stocks. Both companies own solid growth prospects, allowing investors to capitalize on the booming AI chip industry.
