David Tepper filed Appaloosa Management's Q2 2026 13F on August 14th, disclosing a $4.2 billion equity portfolio that moved sharply away from the AI names that anchored his book through 2024 and early 2025. The firm reduced or eliminated stakes in six of its ten largest technology holdings, including complete exits from two positions that had each exceeded $300 million as recently as Q4 2025. No replacement mega-cap technology stake appeared in the filing. The portfolio now shows broader sector distribution and smaller position sizes across its top twenty names.
Appaloosa cut its Nvidia stake by 68%, trimming the position to approximately $180 million from $562 million in the prior quarter. The fund eliminated its Microsoft position entirely, a holding that represented $340 million in Q1. Meta, previously a top-five position at $410 million, was reduced by 54% to $189 million. Amazon dropped 31% to $278 million. Alphabet holdings fell 42% across both share classes. The only significant AI-infrastructure add was a new $95 million stake in Broadcom, first disclosed this quarter. Energy and financials saw modest increases, but no sector received the capital that flowed out of technology at a scale that would suggest a new thesis.
The move matters because Tepper built credibility as an AI-infrastructure bull when it was expensive to be one. He added to Nvidia in Q3 2023 when the stock traded at $450 pre-split, and he defended those positions publicly while other macro traders called tops. His concentration in the group peaked in Q4 2024, when six AI-related names accounted for 63% of disclosed long equity. That positioning worked: Appaloosa's flagship fund returned an estimated 34% in 2024, outperforming the S&P 500 by 10 percentage points according to investor letters reviewed by Bloomberg. When a manager with that kind of carry thins the book, allocators notice. The filing does not explain whether the rotation reflects valuation discipline, a macro view on rate paths and liquidity, or simply profit-taking after a two-year run. What it does show is that Tepper is not adding to AI exposure at current levels, and he is not rotating into a new identifiable theme at size.
Allocators should watch three things. First, whether Tepper's public commentary in September or October—he typically appears on CNBC during earnings season—offers a valuation framework or a liquidity view that explains the rotation. Second, whether the Q3 filing in mid-November shows reinvestment into technology at lower levels, or continued distribution into other sectors. Third, whether other Tiger cubs and 2000s-era hedge funds follow similar rotation patterns in their August 14th filings, which would suggest a cohort move rather than an idiosyncratic call. Tepper has been early before. He was also early exiting financials in Q1 2007, a trade that looked premature until it didn't.
Appaloosa manages approximately $6.8 billion in AUM as of mid-2026, down from a peak of $20 billion in 2011, and the 13F captures roughly 62% of that capital assuming typical long-short construction. The disclosed equity book is now the most diversified it has been since Q2 2022.