Investors enamoured with the potential of artificial intelligence (AI) to revolutionise the investment world may want to pause and consider the performance data of the Eurekahedge AI Hedge Fund Index. This index tracks hedge funds that leverage AI and machine learning in their trading strategies. However, its performance over the past 15 years offers a sobering reality check: from December 2009 through July 2024, the index delivered a 9.8% annualised return, trailing the 13.7% return of the S&P 500 over the same period.
One might expect that as AI technologies become more sophisticated, they will increasingly outperform traditional investment strategies. However, the data from the Eurekahedge AI Hedge Fund Index tells a different story. Interestingly, the AI hedge funds performed better in the earlier years of the sample, casting doubt on the narrative that AI’s learning capabilities and access to more market data would lead to better returns over time.
While the AI index does show better performance on a risk-adjusted basis—with 57% lower volatility than the S&P 500—this is not enough to bridge the performance gap. Even over the last four and a half years, where AI’s potential should have theoretically shone brighter, the index still lags behind the S&P 500, even when adjusting for risk.
This underperformance is not entirely surprising, given William Sharpe’s “arithmetic of active management.” Sharpe, a Nobel laureate, argued that after accounting for costs, the return on the average actively managed dollar would inherently be less than that of the overall market. This concept, grounded in basic arithmetic, suggests that active managers, including those using AI, face significant hurdles in consistently outperforming the market.
Sharpe’s reasoning is simple: the combined portfolio of all active managers must, by definition, equal the market itself. Therefore, after subtracting management fees and transaction costs, the net return of these managers will inevitably be lower than that of the market.
AI and machine learning technologies are not cheap, and their costs further eat into returns. Additionally, as more active managers adopt AI, the competition intensifies, making it increasingly difficult for any AI-driven strategy to outperform. Lawrence Tint, former U.S. CEO of BGI (which created iShares), echoes this sentiment, noting that even if some AI managers can initially beat the market, their job becomes progressively harder as AI use becomes more widespread. The fewer non-AI managers left to take the other side of trades, the less likely AI strategies are to find market inefficiencies to exploit.
The allure of AI in investing is strong. Still, the performance of the Eurekahedge AI Hedge Fund Index serves as a reminder that technology alone is not a guaranteed path to outsized returns. Investors should be cautious about expecting AI-driven funds to consistently outperform the broader market, especially given the inherent costs and the competitive landscape. As AI becomes more ubiquitous in trading, its advantages may diminish, making traditional index funds a more reliable choice for long-term investors.