Published Date : 30/05/2025
Wall Street runs on information, and investors rarely have to look too far for market-moving data. Everything from trade policies to earnings reports provides clues about the current and future health of corporate America and the U.S. economy. One of the most telling data releases occurred on May 15, when institutional investors overseeing at least $100 million had to file Form 13F with the Securities and Exchange Commission. These quarterly filings detail the stock, ETF, and (select) options transactions of top-tier money managers.
Though Warren Buffett is the most well-known asset manager, he's far from the only billionaire investor known for outsized returns. Another notable figure is David Tepper of Appaloosa Management. As of the end of March, Tepper was overseeing close to $8.4 billion in assets, and his first-quarter trading activity was particularly noteworthy for his net-selling of artificial intelligence (AI) stocks.
David Tepper's Appaloosa was a big-time seller of AI stocks. While he did make some purchases, such as opening a new position in Broadcom and adding shares to existing positions in Meta Platforms and Taiwan Semiconductor Manufacturing, the amount of selling was significantly more pronounced. Tepper sold shares of five prominent AI stocks in the March-ended quarter:
- Advanced Micro Devices (AMD): 1,200,000 shares sold (exited position)
- Intel: 1,000,000 shares sold (exited position)
- Lam Research: 750,000 shares sold (60% reduction)
- Nvidia: 380,001 shares sold (56% reduction)
- Microsoft: 460,000 shares sold (47% reduction)
This selling activity could be seen as simple profit-taking. Excluding Intel, whose stock has underperformed, shares of Nvidia, Microsoft, AMD, and Lam Research have all seen significant gains as the AI revolution has taken shape. Given Tepper's average holding period of around 29 months, he has shown a willingness to take profits when his positions are in the green.
However, the worry is that there may be more than just benign profit-taking behind Tepper's aggressive net selling of AI stocks. On one hand, demand for AI-graphics processing units (GPUs) and AI solutions has been exceptionally strong. Nvidia's Hopper and Blackwell GPUs hold a significant market share lead in AI-accelerated data centers, with AMD ramping up production of its Instinct AI-accelerating chips. Intel's CPUs are also playing a role in the rapid expansion of AI-data center infrastructure.
While these companies appear well-positioned to benefit from the ongoing build-out of enterprise data centers, the economics of supply and demand might disagree. For Nvidia, nothing has been more important to its rapid sales and profit growth than AI-GPU scarcity. Demand overwhelming the supply of GPUs has allowed Nvidia to place a premium on its products. However, with AMD and other competitors ramping up production, and many of Nvidia's largest customers developing their own AI chips, AI-GPU scarcity is expected to decrease in the coming quarters.
This reduction in scarcity is expected to weigh on Nvidia's pricing power, as well as AMD and possibly Intel. It could also impact the pricing power for semiconductor equipment companies like Lam Research, whose equipment is crucial for packaging high-bandwidth memory in AI-accelerated data centers.
Another potential concern for AI stocks that may have influenced Tepper's decision is the historical pattern of next-big-thing technologies and innovations. Since the advent of the internet in the mid-1990s, every major technological trend has experienced a bubble-bursting event early in its expansion phase. These bubbles form because investors often overestimate how quickly a new technology will gain utility and widespread adoption.
Even with strong demand for AI hardware, as evidenced by Nvidia's rapid sales growth, most businesses deploying AI solutions haven't yet figured out how to optimize them or generate a profit. Every game-changing technology needs time to mature, and AI is far from reaching that point.
If the AI bubble were to burst, we could see widespread weakness among the prominent AI stocks Tepper exited or reduced. Nvidia, which generated more than 90% of its net sales from its data center segment in the fiscal fourth quarter, would likely be hit the hardest. Even Microsoft, with its high-margin cloud infrastructure service platform Azure, which incorporates generative AI solutions, would not be immune. While its high-margin software sales with Office and Windows would help insulate the stock, growth for Azure could slow during an AI bubble-bursting event.
Given the historical pattern of technological bubbles and the current dynamics in the AI market, billionaire David Tepper may be allowing these odds to guide his investment strategy. For investors, this move could be a signal to re-evaluate their positions in AI stocks and consider the potential risks and rewards in the coming quarters.
Q: Why did David Tepper sell off AI stocks?
A: David Tepper may have sold off AI stocks for profit-taking, but also due to concerns about the potential for a bubble burst in the AI market and the decreasing scarcity of AI GPUs.
Q: Which AI stocks did Tepper sell?
A: Tepper sold shares of Advanced Micro Devices (AMD), Intel, Lam Research, Nvidia, and Microsoft.
Q: What is the significance of AI GPU scarcity?
A: AI GPU scarcity has been crucial for Nvidia's rapid sales and profit growth, allowing them to charge premium prices. However, with increasing competition and internal chip development by large customers, this scarcity is expected to decrease.
Q: How might an AI bubble burst affect the market?
A: An AI bubble burst could lead to widespread weakness among AI stocks, particularly affecting companies like Nvidia and Microsoft, which have significant exposure to the AI market.
Q: What is the historical pattern of next-big-thing technologies?
A: Historically, every major technological trend has experienced a bubble-bursting event early in its expansion phase, as investors often overestimate the speed of adoption and utility of new technologies.