AI is Revolutionising the Stock Market

[CrimsonFury]

The Financial Times published a deep look at how AI is fundamentally reshaping stock market structure, moving well beyond the narrative of AI stocks being good investments and into the more complex territory of AI as a trading and market-making tool. Hedge funds running AI-driven systematic strategies have been delivering strong numbers. D.E. Shaw's Oculus fund returned 36.1 percent in 2024 and Citadel's Tactical Trading arm posted 22.3 percent, while across the industry hedge funds delivered a record 543 billion dollars in investor gains in 2025. The piece examines how machine learning models processing earnings calls, news feeds, and alternative data are executing strategies at speeds and scales that human analysts simply cannot match.

The FT piece does not shy away from the systemic risks this creates. When multiple AI systems train on the same historical data, they tend to converge on similar strategies. That creates herding behaviour that could amplify volatility rather than dampen it, with flash crash risk being the obvious concern. Regulators at the SEC and CFTC are grappling with a landscape where agentic AI platforms can execute complex multi-step strategies without human intervention, pushing the definition of market manipulation and intent into genuinely novel legal territory. The piece also notes that the shift is changing what jobs exist in finance. The demand is for data engineers and model training specialists rather than traditional analysts, which represents a structural change in how Wall Street talent looks.

B31F1E09 5Aae 4Cad Af15 97Adb15Dba70www.ft.com

[Estuary59]

The herding risk is the one that worries me most. If every major fund's AI is trained on the same data and converges on the same trades, you get a system that is fragile in a very specific and hard to predict way

[BigDog92]

D.E. Shaw and Citadel have been doing this for decades. What is new is not that quants use AI, it is that the capability has jumped and the cost has dropped enough that second-tier funds can now access similar tools

[Sharon_77]

The legal question around agentic AI and market manipulation is genuinely unsolved. How do you prove intent when the system executing the trade has no intent in any meaningful sense

[GradientPiston]

Flash crashes are already happening with some regularity. If AI trading systems become more correlated not less, the magnitude of those events is going to increase

[Sophie92]

The job market shift is real and it is accelerating. I know people at mid-sized hedge funds who have been told their analyst roles are being restructured around AI tooling. It is not hypothetical

[BankHolidayBlues87]

Alternative data is where the real competitive edge is. NLP on earnings calls has been around for years. The funds reading satellite imagery of car parks and shipping container counts are operating in a different league

[RayOfLight99]

The SEC needs to grapple with this urgently. Traditional market surveillance was designed to catch human traders. An AI executing thousands of correlated micro-trades looks nothing like what the rules were written to detect

[CyberRider56]

I am sceptical that the returns being attributed to AI are entirely separable from the AI stock rally itself. If your systematic fund was long tech and AI names in 2024 you did well regardless of strategy sophistication

[Coastal Estuary]

At some point the alpha from these strategies gets competed away as more funds use similar approaches. The question is how long the edge lasts before it becomes table stakes

[Cass_9]

The 543 billion in hedge fund gains in 2025 is the number that jumps out. That is an extraordinary wealth transfer from retail and passive investors to systematic funds. The distributional implications are significant

[202694]

AI is bad for asking for prices and should i invest. it is good at explaining terms.