A new paradigm of winning Alpha is achieved! The prototype of AI fund manager emerged on Wall Street, J.P. Morgan's AI intelligent body outperforms the classic 60/40 investment.

With more and more investors relying on cutting-edge artificial intelligence tools to assist in various major investment decisions from stock selection to risk management, Wall Street financial giant JPMorgan Chase & Co. is testing a more ambitious question: whether artificial intelligence models can autonomously allocate funds and achieve alpha excess returns beyond market benchmarks. A blockbuster research report from JPMorgan shows that researchers have built a series of AI-driven investment agent workflows (focused on autonomous investing and position adjustments), with these AI agent subsystems autonomously adjusting positions and transitioning investment themes between stocks and bond assets based on changing market conditions. The research report shows that the top-performing systems outperform traditional 60/40 investment portfolios by 0.7 percentage points annually, with lower volatility, and have also outperformed JPMorgan's own rule-based market regime investment model in backtesting over the past twenty years. However, JPMorgan's strategy team emphasizes that these results are based on historical simulations, not live investment evaluations. JPMorgan warns not to see them as a significant basis for artificial intelligence to consistently outperform the market, and advises not to accept investment portfolio answers from artificial intelligence uncritically. Nevertheless, this research also highlights that AI agents are evolving from "assisting analytical tools" to intelligent investment infrastructure capable of market state recognition (economic growth/inflation cycles), asset rotation, risk control, and portfolio optimization. This also implies a shift in the competition dimension of future alpha returns, from purely relying on human experience to a mixed investment system of "human macro judgment + continuous calculation optimization by AI agents". AI moves from assisting research to autonomous decision-making: Intelligent investment agents are reshaping the trillion-dollar asset management industry The early results undoubtedly encourage investors. The financial giant's researchers have built a series of AI-driven investment agents that dynamically adjust between stocks and bonds based on the changing market environment. According to the team led by strategist Thomas Salopek, the best-performing system in historical backtesting over the past 20 years outperformed the traditional 60/40 investment portfolio (60% stocks, 40% bonds) by 0.7 percentage points annually, with lower volatility, and also beat JPMorgan's own rule-based market regime and cycle models. There is an important limitation to these results. The research is based on historical simulation, not actual fund investments, and JPMorgan also warns that it should not be seen as proof that artificial intelligence can consistently outperform the market. Nevertheless, this signals future directions as the rapid expansion of automated trading trends in the field shows no signs of slowing down. "AI agents can be set with a process that allows them to make important real-time investment decisions autonomously in an uncertain environment, achieving excess alpha performance relative to reasonable benchmarks." These strategists wrote in a report on Thursday, stating that this work is the institution's first attempt to build an artificial intelligence system for identifying market investment regimes and cycle states. The JPMorgan AI agents in the backtesting performed better than the traditional 60/40 portfolio by 0.7 percentage points on an annualized basis, while lowering volatility. All eight tested AI agent investment workflows outperformed the 60/40 portfolio in risk-adjusted returns, indicating that AI is evolving from an "information processing tool" to an "investment decision infrastructure". The so-called "alpha" is defined as investment returns significantly exceeding "beta returns" - referring to synchronous investment returns data that far exceed those achieved by benchmark stock index trackers. Returns achieved by benchmark indices are also known as "beta returns". The birth of the next generation of Wall Street infrastructure: AI agents may become the new engine of asset allocation For ordinary retail investors, how can they achieve alpha returns in the future? AI agents may become one of the strongest automated execution assistants for investors. JPMorgan's research is not a proof that AI can consistently beat the market, but it is the first demonstration that AI agents have the potential to identify market states, dynamically allocate assets, control risks, and assist with capital decisions in a manner approaching professional investment processes. AI agents that autonomously perform various tedious and complex tasks are likely to be the ultimate trend in AI applications for the next decade. The emergence of AI agents indicates that artificial intelligence is evolving from an information-assisting tool to a highly intelligent productivity tool, which is why the valuations of Anthropic, the parent company of Claude Cowork, have surpassed 1 trillion US dollars and exceeded OpenAI. This experiment showcases an important development direction for Wall Street's application of artificial intelligence in the next phase, or an early glimpse into the next phase of artificial intelligence adoption pattern on Wall Street. Over the past two years, banks have been applying large language models to research analysis, code development, and internal investment tools. Now, they are further testing: can these systems evolve from assisting AI investment analysts' work to actively participating in one of the financial industry's most important decisions - allocating capital between different markets. As these research results are released, more and more academic studies are beginning to focus on a question: what changes will occur in the market if all investors start relying on similar artificial intelligence models for investment decisions. While this technology may allow investors to increase speed and obtain more comprehensive information, researchers warn that it may also lead to more crowded trades, making the market more susceptible to manipulation by increasingly large long or short AI forces, and amplifying risks during market pressure stages when a large number of AI agents systems reach similar investment conclusions. It is worth noting that JPMorgan's strategy team also acknowledges these risks. "We strongly caution against accepting the results given by artificial intelligence uncritically, as these results may essentially be overconfident answers based on in-sample data," they wrote. "AI agent-style AI tools need to be built on a thoughtful asset allocation process foundation, rather than naively assuming that intelligent agents can be a source of domain knowledge." However, these findings further enrich a growing body of evidence that artificial intelligence is performing increasingly complex investment tasks. The JPMorgan team used intelligent agents driven by OpenAI and Anthropic models to design an AI agent operational system, which can divide the market into four classic states based on economic growth and inflation environments: Goldilocks, re-inflation, stagflation, and risk aversion. Subsequently, these AI agents were tasked with determining asset allocation strategies based on different market environments - for example, increasing stock allocations during periods of strong economic growth and increasing fixed income asset allocations in deteriorating economic outlooks. All eight AI agent subsystems tested outperformed the traditional Wall Street 60/40 investment portfolio in risk-adjusted performance. They also beat JPMorgan's existing rule-based market regime and cycle models, indicating that this cutting-edge AI technology can significantly enhance investment returns based on a classic framework that has been used to guide asset allocation decisions. "We are excited about the development potential of intelligent agent artificial intelligence, although we remain cautious and will not completely delegate asset allocation decisions to AI agent-style operational systems," Salopek and colleagues wrote.