AI and Machine Learning in Quant Finance 2026: What Compensation at Citadel and Jane Street Reveals About Hiring
Why the quantitative developer median has moved above the quantitative researcher median, and what that ordering signals about where hedge funds are putting their money this cycle.
A working read on the state of machine learning across hedge funds, prop shops, and high-frequency desks, written for people who already know the mechanics and want the figures that move decisions.
Roughly 47% of mid-to-large hedge funds had at least one generative AI system running in production by the first quarter of 2026, and the firms doing it cluster in the quant and multi-strategy tier above $5bn in assets. The version of this story you keep seeing says AI has arrived.
The number underneath it says AI has already become ordinary at the top of the industry, which changes what the edge actually is. Once production machine learning is standard equipment across the upper tier, the advantage shifts to how cheaply and how quickly a desk can iterate, and that is an engineering and data problem before it is a research one.
Adoption has saturated the top of the industry
Breadth is close to total at scale. More than 70% of global hedge funds now run a machine-learning component somewhere in the trading pipeline, and around 18% rely on AI for more than half of their signal generation.
In 2025, over 35% of new fund launches described themselves as AI-driven or AI-enhanced, which tells you where allocator attention and marketing spend are pointing as the next fundraising cycle opens.
The research base feeding all of this is growing at a pace that should worry anyone relying on a single durable signal. Academic output on language models in finance rose 594% between 2023 and 2025, climbing from 36 papers to 250 in the leading machine-learning and NLP venues.
The techniques behind next year’s signal stack are being published in the open right now, which shortens the useful life of any one research idea and raises the premium on the funds that can read, test, and deploy faster than their competitors.
The gap that matters is no longer between funds that use machine learning and funds that do not, because almost everyone at scale does. The gap is between the funds that have rebuilt parts of the investment process around AI and the funds that have only pointed a chatbot at their earnings memos, and that distance is widening every quarter.
Morgan Stanley‘s 2026 outlook describes a phase in which generative models reshape how alpha gets produced and scaled, and the capital flows are consistent with that view, with the strongest adopters concentrated in systematic and multi-strategy firms.
