Cathie Wood, the CEO of Ark Invest, continues to divide opinions among investors. Advocates laud her for her daring approach to growth stocks within Ark’s exchange-traded funds (ETFs), while detractors argue that she neglects valuation metrics, prematurely sells successful stocks, and clings too long to underperformers. Notably, Wood’s Ark fund liquidated a majority of its Nvidia shares during 2022 and early 2023, just as the AI boom propelled Nvidia’s stock to unprecedented heights. Over the last three years, Nvidia’s stock has skyrocketed by over 950%, whereas Wood’s flagship Ark Innovation ETF has only climbed approximately 120%. As the past year progressed, Wood has gradually reestablished a position in Nvidia, albeit at significantly elevated prices.
This trade may indicate that investors should be cautious when considering Wood’s forecasts. Nevertheless, her insights continue to be closely monitored by many growth-centric investors. Ark’s recently released “Big Ideas 2026” report highlights Wood’s perspective on the intensified rivalry Nvidia is likely to encounter from AMD in the AI sector this year. This begs the question: can AMD actually compete with Nvidia, or is Wood downplaying the latter’s decisive competitive advantages?
Nvidia has evolved from primarily selling discrete GPUs for high-end gaming PCs to a robust player in the data center arena, introducing increasingly powerful GPUs tailored for machine learning and AI applications. Unlike traditional CPUs, which are designed for sequential processing, GPUs can efficiently execute parallel tasks, allowing for the rapid handling of substantial data operations.
As the AI landscape expanded, Nvidia fortified its pioneering edge with successive innovations, rolling out advanced GPU architectures such as Turing, Ampere, Hopper, and the upcoming Blackwell. Moreover, Nvidia has secured its data center clientele by integrating them into its CUDA (Compute Unified Device Architecture) platform—a proprietary programming system specifically optimized for its chips. This ecosystem creates a substantial barrier for competing AI chip manufacturers.
According to Carbon Credits, Nvidia commanded an impressive 92% of the discrete GPU market in 2025, while AMD only held an 8% share. Leading AI firms, including Microsoft, OpenAI, Alphabet’s Google, and Meta Platforms, rely heavily on Nvidia GPUs to fuel their cutting-edge AI technologies.
While the cost of Nvidia’s data center GPUs is hefty, starting at around $25,000 for an individual H100 chip, the company’s established reputation maintains its pricing power against lower-cost challengers. As competition in the AI software sphere intensifies, it is likely that companies will favor Nvidia’s trusted products over cheaper or lesser-known alternatives.
Conversely, Wood posits that AMD holds potential due to its substantial revenue derived from data center operations, which includes Epyc CPUs and Instinct GPUs designed for servers. AMD’s data center segment has consistently experienced double-digit revenue growth, even amid regulatory challenges in China and a slowing enterprise spending environment.
AMD markets its Epyc CPUs and Instinct GPUs as more affordable counterparts to Intel’s Xeon CPUs and Nvidia’s premium GPUs. For instance, the Instinct MI300X, which competes with Nvidia’s H100, is priced around $15,000. Major clients such as Microsoft, OpenAI, and Meta are already integrating AMD’s solutions.
However, the reality may not be a direct competition between the two manufacturers; both Nvidia and AMD could thrive without necessarily undermining each other. Furthermore, Nvidia’s extensive advantages, particularly its CUDA ecosystem, likely positions it as the primary choice for high-end AI applications.
Additionally, AMD offers a diversified array of products beyond just GPUs, including CPUs and embedded chips, which could dilute its focus on data center investments and hinder its ability to match Nvidia’s scale and technological prowess.
Both Nvidia and AMD might represent attractive investment opportunities in the AI landscape. While AMD’s data center growth is notable, it may not pose a significant threat to Nvidia in the immediate term. Industry projections estimate that the global AI market will expand at a staggering compound annual growth rate (CAGR) of 30.6% from 2026 to 2033. Forecasts for Nvidia project a CAGR of 47% in revenue and 45% in earnings per share from fiscal 2025 to fiscal 2028. Currently, Nvidia is viewed as fairly valued at 26 times its expected earnings for next year, indicating ample room for growth, regardless of Cathie Wood’s nuanced perspective.
