An industry analysis published Monday by Sify documents a structural shift in how hedge funds use AI: the technology is moving from back-office automation and research assistance into active portfolio management and trade decision-making. The shift is backed by performance data from multiple funds that have already made the transition.
The Numbers
Point72 Asset Management’s Turion fund, which launched in October 2024 focused on AI semiconductor and hardware stocks, returned approximately 14% in its first three months and grew to nearly $1.5 billion in assets, according to Reuters. By November 2025, the fund had posted roughly 30% gains for the year, per Bloomberg.
Bridgewater Associates launched a $2 billion fund in July 2024 that uses machine learning as its primary decision-making basis, according to Bloomberg. The fund incorporates models from OpenAI, Anthropic, and Perplexity alongside proprietary technology Bridgewater has developed over more than a decade. Co-Chief Investment Officer Greg Jensen leads the effort through what the firm calls its Artificial Investment Associate Labs project, per Hedgeweek.
Sydney-based Minotaur Capital went further: it has no human analysts at all. Its proprietary system, called Taurient, processes nearly 5,000 news articles per day. In its first six months, the fund returned 13.7% compared to 6.7% for the MSCI All-Country World Index, according to Hedgeweek, citing Bloomberg data.
Scale of Adoption
The individual cases reflect a broader trend. An Alternative Investment Management Association (AIMA) survey from February 2024 found that 86% of hedge fund managers had granted staff access to generative AI tools. That survey covered 157 hedge fund managers controlling approximately $783 billion in aggregate assets. A separate AIMA finding showed 67% of firms were already using AI specifically for generating investment ideas, according to data from BarclayHedge and the AIMA report.
From Copilot to Decision-Maker
The key distinction in 2026 is the role AI plays. Earlier deployments focused on research acceleration: scanning earnings calls, summarizing sell-side reports, flagging anomalies. Point72’s partnership with AI platforms to process earnings calls in real time, identifying linguistic patterns and sentiment shifts that human analysts might miss, represents that first wave.
Bridgewater’s and Minotaur’s approaches represent the second: AI systems that make or heavily inform the actual allocation decisions. Bridgewater’s fund combines LLMs with reasoning tools designed to understand causal relationships in markets. Minotaur eliminated human analysts from the workflow entirely.
The question for the agent ecosystem is whether this pattern, autonomous AI systems operating in high-stakes financial environments, will accelerate or slow the push for agent accountability frameworks. These funds are generating real returns with real capital. The governance infrastructure that regulators and compliance teams will demand for autonomous financial agents may end up setting the template for agent oversight across every other industry.