JPMorgan AI Agents Beat 60/40 Portfolio by 0.7 Point in 20-Year Backtests
Updated
Updated · Bloomberg · Jul 9
JPMorgan AI Agents Beat 60/40 Portfolio by 0.7 Point in 20-Year Backtests
1 articles · Updated · Bloomberg · Jul 9
Summary
JPMorgan’s best AI investing agent outperformed a traditional 60/40 stock-bond portfolio by 0.7 percentage point a year in backtests covering the past two decades.
The system was designed to allocate money autonomously, shifting between stocks and bonds as market conditions changed rather than following a fixed mix.
Lower volatility strengthened the result, and the top agent also beat JPMorgan’s own rules-based market regime model, according to strategists led by Thomas Salopek.
The test reflects a broader push by investors to use AI not just for stock picking or risk management, but for full portfolio allocation.
As AI investing agents beat old benchmarks, are they creating a new, more dangerous market bubble centered on big tech?
With AI now autonomously managing funds, is the era of the star human portfolio manager officially over?
Who is responsible when autonomous AI agents, managing billions, inevitably make a costly, market-shaking mistake?
JPMorgan’s AI Agents Beat the 60/40 Portfolio: A 0.7% Edge and the Future of Automated Investing
Overview
JPMorgan Chase, the largest U.S. bank by assets, is making aggressive investments in artificial intelligence, with a nearly $20 billion annual technology budget and a strong push to be an early mover in AI adoption. Under CEO Jamie Dimon’s leadership, AI is being integrated across nearly every function of the bank, including its 319,000 employees. As a result, JPMorgan’s advanced AI agents have reportedly outperformed the traditional 60/40 investment portfolio by 0.7 percentage points over a 20-year period in internal backtests. This highlights how deep AI integration can drive long-term investment performance and reshape financial strategies.