London's algorithmic trading sector delivered its most profitable year on record in 2025, with leading firms including Man Group, Marshall Wace, and Winton collectively generating estimated net returns exceeding £14 billion. The surge has been driven by the deployment of transformer-based AI models that can process vast quantities of unstructured data — from satellite imagery and shipping manifests to social media sentiment and central bank communications — to identify trading opportunities invisible to traditional quantitative strategies. Man Group's AHL division reported a 38 per cent return on its flagship fund, its best performance since 2008.

The profits have triggered an intense talent war across the City, with senior AI researchers commanding packages of £800,000 to £1.5 million at top-tier systematic funds. "We're competing directly with DeepMind, OpenAI, and the big tech labs for the same pool of PhD-level machine learning talent," said Luke Ellis, Man Group's chief executive. "The difference is that in finance, these researchers can see the direct impact of their work within days, not years." Graduate programmes at quant firms have expanded significantly, with Citadel Securities and Jane Street both opening new London offices in 2025, adding over 200 AI-focused roles between them.

Regulators and market structure experts have raised questions about the systemic risks posed by increasingly sophisticated AI-driven trading. The Bank of England's Financial Policy Committee noted in its November stability report that the growing correlation between AI-powered strategies could amplify market dislocations during stress events. "If multiple firms are training models on similar data sets and reaching similar conclusions, the potential for herding behaviour increases materially," warned Sir Jon Cunliffe, the Bank's former deputy governor for financial stability. The FCA has signalled that it will introduce enhanced reporting requirements for AI-driven trading strategies by 2026.