JPMorgan Chase tested AI agents that autonomously allocate money between stocks and bonds by reading market conditions. Over two decades of historical backtests, the best-performing system beat a traditional 60/40 portfolio by 0.7 percentage point per year with lower volatility. Although the results are promising, JPMorgan cautions they reflect simulations, not live trading, and should not be viewed as evidence that AI can reliably beat markets.
Summaries like this, in your inbox every morning.
Sign up free →What happened
JPMorgan researchers built AI-powered investing agents that automatically shift between stocks and bonds based on market conditions. In backtests over the past two decades, the best-performing system outperformed a traditional 60/40 portfolio by 0.7 percentage point a year with lower volatility, and also beat JPMorgan's own rules-based market regime model.
Why it matters
This marks Wall Street's next phase of AI adoption—moving beyond assisting workers to making one of the industry's most consequential decisions: how to allocate capital across markets. However, JPMorgan warns the results are based on historical simulations rather than live investing and should not be treated as proof that AI can consistently outperform markets.
What to watch
The strategists described this as the firm's first attempt to build an AI system for identifying market regimes. The findings come as academic research raises questions about what happens if too many firms turn to similar AI models, which could produce more crowded trades, make markets easier to manipulate, and amplify stress periods if firms reach similar conclusions.
JPMorgan's experiment reflects a broader shift in how Wall Street is deploying artificial intelligence. Over the past two years, banks have embedded large language models into research, coding, and internal investing tools—primarily to assist human workers. This backtest represents a step further: autonomous AI agents making capital allocation decisions with minimal human intervention. The 0.7 percentage point annual outperformance, paired with lower volatility than the 60/40 benchmark, suggests the technology can detect and respond to market regime changes faster than traditional rules-based systems.
However, the bank's own caveats are significant. Backtesting—even over two decades of historical data—cannot account for how markets behave when the AI model itself becomes a market participant, or when many firms adopt similar strategies simultaneously. The strategists acknowledged these risks directly: crowded trades, easier manipulation, and stress amplification if consensus coalesces around the same AI conclusions. This tension—between individual outperformance in a backtest and potential systemic fragility if the strategy becomes widespread—may define the next chapter of AI in finance.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
