Bridgewater Launches $2 Billion Fund Managed by Machine Learning

(Bloomberg) — Bridgewater Associates has launched a fund that uses machine learning as a primary basis for its decision-making.

Bloomberg’s Most Read

The vehicle debuted with nearly $2 billion in capital from more than half a dozen clients and began trading on Monday, according to people familiar with the matter who requested anonymity to discuss the strategy.

The hedge fund giant, led by CEO Nir Bar Dea, told investors it is leaning on its own proprietary technology, which it has been building for more than a decade. It’s the result of a broader venture led by co-chief investment officer Greg Jensen, and the new fund will also be expanded to include models developed by OpenAI, Anthropic and Perplexity, among others, the people said.

The new fund will be managed by Jensen. Westport, Conn.-based Bridgewater has been testing the strategy since late last year with a small portion of its flagship Pure Alpha fund — about $100 million — to make sure the technology works, the people said.

Bridgewater declined to comment on the fund.

Bar Dea, 42, has been transforming Bridgewater since founder Ray Dalio ceded control in late 2022. The fund’s launch is the latest step in a years-long transition that also included a major management overhaul. Meanwhile, the Pure Alpha fund is up 14.4% this year through June 26 after more than a decade of largely lackluster returns, including a 7.6% loss in 2023, people familiar with the matter said.

Bridgewater’s assets under management total $108 billion.

The push for machine learning is a “good manifestation of us taking the flag and putting it on top of the mountain,” Bar Dea said in an interview, although he declined to provide details about the new fund. “This is perhaps the most significant and purest manifestation of the moment in which we find ourselves.”

It also has the potential to change Bridgewater’s hiring strategy and staffing composition to include more data scientists, said Jensen, 49, who has been thinking about how machine learning could impact the hedge fund’s investments since 2012. That This year, Bridgewater hired David Ferrucci, who led the team of engineers responsible for Watson at International Business Machines Corp. He left the hedge fund firm in 2021 but remains a consultant.

The story continues

Jensen, who has worked at Bridgewater since 1996, said he committed his own money to OpenAI’s first round of funding nearly a decade ago and, years later, wrote one of the first checks to Anthropic.

Statistician Jasjeet Sekhon, a professor at Yale University, was hired by Bridgewater as the initiative’s chief scientist in 2018. Early last year, the company formed a division called Artificial Investment Associate Labs, or AIA. The venture combined technologies including large language models, machine learning data models and reasoning tools.

Through the venture, Bridgewater has built a way to use AI to understand casual relationships in markets and is using it to generate returns. AIA technology is the key decision maker in Bridgewater’s newest strategy.

‘Giant leap’

“The big leap here is using machine intelligence to generate alpha – that’s a leap,” said Jensen. “If this were a side hobby for people who normally have the responsibility of Pure Alpha, they wouldn’t be able to have the focus necessary to make this giant leap that we’re taking.”

Jensen, who majored in economics and applied mathematics at Dartmouth College — and won a gold bracelet at the 2022 World Series of Poker — discussed the strategy’s limitations.

Jensen said Bridgewater humans will still oversee a range of functions, including risk management, data acquisition and trade execution to ensure the investment process is complete. He said a popular question from investors has been: “How do you keep the machine from getting out of control?”

Large language models “have the hallucination problem,” he said. “They don’t know what greed is, what fear is, what the likely cause-and-effect relationships are.”

Tests using AIA systems included the question of how asset prices would be affected if Donald Trump won the November election and raised tariffs on Chinese goods. Bridgewater also experimented with the AIA machine learning process to calculate the impact on bond prices under the Federal Reserve’s quantitative tightening process.

“You will have an intelligence that can read all the newspapers in the world,” said Jensen. “Machines are better at finding patterns over time and across countries.”

(Updates with Ferrucci in ninth paragraph, reasoning tools in 11th.)

Bloomberg Businessweek Most Read

©2024 Bloomberg LP