The rise of AI in hedge fund management: $12B in AUM

A new study by the National Bureau of Economic Research (NBER) in the United States examines the growth of AI-driven investing and seeks to determine whether the use of such technologies translates into a genuine investment advantage.

More specifically, the study by Shuang Chen, Clemens Sialm, and David X. Xu, published as an NBER working paper in 2026, examines 7,896 U.S. hedge funds over the 2006–2024 period. The researchers combine data from regulatory disclosures by investment advisers, labor-market data, archived descriptions of investment strategies, and Hedge Fund Research data.

The authors employ a relatively strict definition, although one that is broader than generative AI. They define AI-driven investing as a specialized subset of quantitative strategies in which artificial intelligence technologies are used for predictive modeling and the generation of investment signals, either autonomously or as a toolBased on the strict classification resulting from their text analysis, the supporting investment decisions.

Just 89 AI Funds

Based on the strict classification resulting from their text analysis, the researchers identified 89 AI funds, representing just 1.1% of the overall sample.

The study does not provide a complete list of the 89 funds by name, but it presents aggregate data on their performance and investment strategies. One particularly notable finding is the strong concentration of AI funds in macro strategies. In 2024, approximately 60% of AI funds were classified as systematic diversified macro, with investments typically focused on highly liquid assets such as equity indices, commodities, bonds, and currencies.

Overall, more than 80% of the assets managed by the AI funds in the sample were allocated to macro strategies.

From BlackRock to Two Sigma

The study specifically mentions BlackRock’s Absolute Macro Fund, which has been classified as AI-driven since 2021. Another particularly notable example is Two Sigma Investments, which describes its business as quantitative investing and trading and states that it has been using AI in investing for 25 years.

Similarly, Man AHL has formally identified the use of machine learning as a significant investment milestone since 2014, while WorldQuant has built its quantitative asset-management model around data, human talent, and forecasting.

The study also refers to Renaissance Technologies and D.E. Shaw as historical users of such techniques, although it does not establish that they are among the 89 funds identified through the study’s strict text-based classificationThe most significant finding concerns returns. During the earlier part of the period.

Performance Advantage Narrowed After 2017

The most significant finding concerns returns. During the earlier part of the period under review, AI funds generated approximately 6% higher annual returns than non-AI hedge funds on a benchmark-adjusted basis.

However, this advantage narrowed significantly over time. After 2017, the difference in returns became statistically indistinguishable from zero.

This finding is particularly interesting because it cannot simply be explained by early AI adopters gradually losing their lead. Even funds that had begun using AI in the early years of the sample saw a substantial portion of their relative outperformance disappear after 2017.

The picture, however, is not entirely negative. When the researchers compared AI funds with non-AI funds managed by the same investment manager, they identified an advantage of approximately 34.9 to 41.1 basis points per month.

At the same time, AI funds exhibited a lower correlation between their returns and those of other funds within the same strategy category. This finding challenges the view that the use of algorithmic models necessarily leads to more homogeneous investment positions.

In other words, AI appears capable of functioning not only as a tool for alpha generation, but also as a source of portfolio diversification.

AI Funds Manage $12 Billion in Assets

Although the number of funds meeting the study’s strict criteria remains limited, their assets under management are growing.

In 2024, the assets managed by AI funds amounted to approximately $12 billion within the researchers’ sample.

The study’s central conclusion is therefore more nuanced than simply stating that “AI is beating the markets.” Artificial intelligence can serve as a source of alpha and diversification, but it does not provide a permanent investment advantage and does not guarantee higher returns.

As more funds adopt similar technologies, the initial informational and computational advantage may diminish. The true competitive advantage may therefore shift away from the mere use of AI and toward the quality of data, models, infrastructure, and a manager’s ability to translate AI-generated signals into effective investment decisions.