JPMorgan Commits Hedge Fund to AI in Technology Arms Race

(Bloomberg) -- JPMorgan Chase & Co. is pushing its use of artificial intelligence beyond investment banking and into hedge funds.

The bank's asset management arm is planning a strategy to invest in emerging and established machine-learning statistical-arbitrage hedge funds, according to a person familiar with the matter. The vehicle -- dubbed for now the Machine Learning Fund Ltd. -- will operate within JPMorgan's $15 billion fund-of-hedge funds business, according to June 20 regulatory filings.

JPMorgan has spent billions in the technology arms race with rivals like Morgan Stanley, deploying AI and machine-learning in investment banking and tapping industry experts to help shape strategy. Last year, the company started an equity data science unit within its asset management business to explore how AI can improve investment decisions.

Robots That Learn Are the Hottest Weapon in the Investing Arms Race

The alternatives business has stumbled on a few recent hedge fund bets. It invested $200 million with Duane Park Capital Management, which closed recently after only two years in business. The group also pulled a big investment from quant firm Mana Partners, which struggled almost from the get-go.

JPMorgan's machine-learning strategy comes as the use of AI among hedge fund managers accelerates in an effort to stand out in a crowded field. With machine learning, a subset of AI, computers improve their ability to find trading signals based on data sampled over time, without needing much human intervention.

Hedge funds deploy AI in many ways, from basic research to sophisticated live trading, and it's unclear what type of managers the bank is targeting. A Barclayhedge study last year found that two-thirds of managers it surveyed use machine learning in idea generation, more than half deploy the technology for portfolio construction, and 25% utilize it for trade execution.

AI Hedge Fund Voleon Doubles Assets to $5 Billion After Stumble

Overall, 56% of the 55 hedge funds surveyed in May 2018 said they utilized a machine-learning approach in their investment processes. That compares with about 20% in a Barclayhedge survey from August 2017.

AI and machine-learning hedge funds have lost money this year, according to a Eurekahedge index of about 14 funds that use the technology. But over the longer-term, they've outperformed other computer-driven funds.

The index has had annualized gains of about 7% over the last five years, about three times the returns of Eurekahedge's CTA/Managed Futures hedge fund index. Hedge funds, broadly, posted annualized gains of about 3.6% over the period.

A spokeswoman for New York-based JPMorgan declined to comment.

--With assistance from Michelle F. Davis.

To contact the reporter on this story: Katia Porzecanski in New York at kporzecansk1@bloomberg.net

To contact the editors responsible for this story: Alan Mirabella at amirabella@bloomberg.net, Vincent Bielski, Josh Friedman

©2019 Bloomberg L.P.