Tepper's Appaloosa Sells AI Stocks

David Tepper of Appaloosa Management, a billionaire fund manager, oversaw close to $8.4 billion in assets as of March. A notable detail of Tepper's first-quarter trading was the net selling of artificial intelligence (AI) stocks.

Tepper's Appaloosa did some buying in the AI space, opening a new position totaling 130,000 shares in Broadcom, an AI-networking specialist. During the first quarter, he also added 60,000 shares of Meta Platforms and 20,000 shares of Taiwan Semiconductor Manufacturing to existing positions. However, the amount of selling Tepper undertook in the AI arena differed greatly relative to buying. Appaloosa's 13F revealed that Tepper sold shares of five prominent artificial intelligence stocks in the quarter that ended in March.

Possible Reasons for Selling

This selling activity may represent simple profit-taking. Shares of Nvidia, Microsoft, AMD, and Lam Research have increased as the AI revolution has taken shape (excluding Intel, whose stock has underperformed). With Appaloosa averaging a holding period of around 29 months, Tepper has shown a willingness to sell when his positions are profitable.

There may be more than just profit-taking or stop losses (in Intel's case) behind Tepper's net selling of AI stocks.

AI Market Dynamics

Demand for AI-graphics processing units (GPUs) and AI solutions has been exceptionally strong. Nvidia's Hopper (H100) and Blackwell GPUs hold a significant market share lead in AI-accelerated data centers, and AMD is increasing production of its Instinct AI-accelerating chips. Intel's central processing units (CPUs) also contribute to the expansion of AI-data center infrastructure.

While these companies seem well-positioned, the economics of supply and demand might present challenges. AI-GPU scarcity has been crucial to Nvidia's sales and profit growth, allowing the company to place a premium on Hopper and Blackwell GPUs. However, with AMD and other competitors increasing their production, and many of Nvidia's largest customers developing AI chips internally, AI-GPU scarcity is expected to decrease in the coming quarters. This could affect the pricing power of Nvidia, AMD, and possibly Intel. It could also affect the pricing power for semiconductor equipment companies like Lam Research, whose equipment is important in packaging the high-bandwidth memory needed in AI-accelerated data centers.

Historical Context

Another factor that may have led David Tepper to reduce Appaloosa's holdings in AI stocks is the history of next-big-thing technologies. Since the advent of the internet in the mid-1990s, these trends have experienced a bubble-bursting event early in their expansion phase. Bubbles form because investors overestimate how quickly a new technology will gain adoption. Despite strong demand for AI hardware, many businesses deploying AI solutions have not yet optimized those solutions or generated a profit on their AI investments. AI needs time to mature.

If the AI bubble were to burst, weakness could be seen among the five AI stocks Tepper exited or reduced in the March-ended quarter. Nvidia, which generated over 90% of its net sales from its data center segment in its fiscal fourth quarter, would likely be affected the most. Microsoft would also be impacted; growth for Azure, its cloud infrastructure service platform that incorporates generative AI solutions, would likely slow during an AI bubble-bursting event. David Tepper may be considering these historical trends in his investment strategy.