Since late 2022, the evolution of artificial intelligence (AI) has been garnering the attention and capital of investors. Empowering software and systems with the capacity to make split-second decisions and become more effective over time at their assigned tasks is a game changer for most industries around the globe.
Wall Street's largest financial institutions and leading money managers aren't oblivious to the sky-high addressable opportunity AI brings to the table.
Less than two weeks ago, on Nov. 14, institutional investors with at least $100 million in assets under management filed Form 13F with the Securities and Exchange Commission. This required quarterly filing details which stocks, exchange-traded funds, and select options that Wall Street's savviest fund managers bought and sold in the latest quarter (in this case, the quarter ended in September).
Image source: Getty Images.
Unsurprisingly, billionaire money managers have been busy bees in the AI arena. Over the last two quarters, Appaloosa's billionaire boss, David Tepper, has more than sextupled his fund's position in graphics processing unit (GPU) goliath Nvidia (NVDA 0.23%). All the while, he's completely exited Appaloosa's stake in a hypergrowth AI stock that was, until recently, flirting with a trillion-dollar valuation.
Billionaire David Tepper shifts course with Nvidia
What makes Tepper's recent buying activity of Nvidia stock so noteworthy is that he had been a persistent seller of shares between June 30, 2023, and March 31, 2025. During this period, 13Fs show that his fund's stake fell from a stock-split-adjusted 10.2 million shares to only 300,000 shares.
However, Appaloosa's billionaire investor purchased 1.45 million shares of Wall Street's AI darling during the June-ended quarter and added 150,000 more shares in the quarter ended in September. The 1.9 million shares held, as of Sept. 30, represent a 533% increase in just six months.
The leading catalyst fueling Tepper's optimism might be the short-lived crash that occurred on Wall Street in early April. In the days following the reveal of President Donald Trump's tariff and trade policy, all three of Wall Street's major stock indexes entered into correction territory or a bear market. This created brief price dislocations in high-growth stocks, allowing billionaire money managers like David Tepper to pounce.
NASDAQ: NVDA
Key Data Points
Appaloosa's chief investor was likely also enamored with Nvidia's first-mover advantage in the AI space. Its Hopper (H100), Blackwell, and Blackwell Ultra GPUs have enjoyed exceptional demand, which has led to otherworldly pricing power for Nvidia and a gross margin that consistently tops 70%.
Building on this point, Nvidia's CUDA software platform has been the company's anchor. CUDA is the software platform developers use to build and train large language models, as well as to maximize the compute capacity of their Nvidia GPUs. CUDA is ensuring that GPU buyers remain loyal to Nvidia's ecosystem of products and services.
But it can also be argued that Appaloosa's fourth-largest holding is far from a slam-dunk investment at a $4.3 trillion valuation.
Although Nvidia's fiscal third-quarter operating results highlighted plenty of demand for AI infrastructure, it's hard to ignore the reality that external and internal competition are ramping up. External competitors can offset the AI-GPU scarcity that's fueled Nvidia's exceptional pricing power. Meanwhile, the internal development of AI-GPUs by some of Nvidia's top customers, based on net sales, could cost the company valuable data center real estate and delay future upgrade cycles.
Nvidia's valuation is a question mark, too. While hindsight reveals that a price dislocation existed in early April, Nvidia's price-to-sales ratio of 23.5, as of the closing bell on Nov. 21, remains well above historical norms. Even though this is the most optimistic David Tepper has been on Nvidia stock in quite some time, headwinds do exist.
Image source: Getty Images.
Hypergrowth AI stock Oracle gets shown the door
On the other end of the spectrum, billionaire David Tepper ushered shares of integrated cloud applications and cloud infrastructure services provider Oracle (ORCL +11.36%) to the door.
During the first quarter of 2024, Appaloosa's stake in Oracle peaked at 2.3 million shares. However, Tepper was a seller in each of the subsequent six quarters. Appaloosa's 13Fs show that 550,000 shares of Oracle stock were dumped during the second quarter, with the remaining 150,000 shares jettisoned in the latest quarter.
The most logical reason for Tepper to sell Oracle stock is to lock in profits. When Appaloosa's stake peaked in the first quarter of 2024, shares of Oracle vacillated between $100 and $125. During the third quarter of 2025, they briefly popped to $345, giving the company a nearly $1 trillion valuation. For six quarters, Tepper had ample reason to cash in his fund's chips and lock in supercharged gains.
Oracle's outperformance during the September-ended quarter was sparked by growth guidance that completely floored Wall Street and investors. The company's remaining performance obligation (RPO) -- the backlog of future revenue based on contracts signed -- as of the end of the fiscal first quarter (ended Aug. 31) surged 359% from the prior-year period to $455 billion!
NYSE: ORCL
Key Data Points
What was even more pronounced was CEO Safra Catz's growth forecast for on-demand cloud-computing service Oracle Cloud Infrastructure. This high-margin operating segment is projected to grow its sales by:
- 77% in fiscal year (FY) 2026 to $18 billion
- 78% in FY 2027 to $32 billion
- 128% in FY 2028 to $73 billion
- 56% in FY 2029 to $114 billion
- 26% in FY 2030 to $144 billion
But there might be more to Tepper's exit than just a desire to cash in Appaloosa's chips.
For instance, Oracle has missed Wall Street's earnings per share (EPS) estimate in three of the last four quarters. While these "misses" have been relatively small ($0.01 or $0.02 per share), they signal the possibility that Oracle may fail to live up to its now-lofty expectations.
Furthermore, Oracle, along with Nvidia, are both exposed to the historical headwinds that come with next-big-thing technologies. Since the proliferation of the internet in the mid-1990s, there hasn't been a hyped innovation or game-changing technology that's avoided an eventual bubble-bursting event.
While Oracle's RPO paints an optimistic picture, most businesses have yet to optimize their AI solutions and/or aren't close to generating a positive return on their AI investments. In other words, the risk of an AI bubble forming and bursting remains quite high.
