London Stock Exchange speeds up trading data analysis with NVIDIA

London Stock Exchange Group (LSEG) has reduced years of trading data analysis to seconds using NVIDIA processors, based on a benchmark test. This efficiency gain comes from processing its Tick History dataset, comprising 110 trillion rows across 100 million instruments from over 580 exchanges, through NVIDIA’s cuDF software and the Polars data library. The system eliminates duplicate data copies while preserving existing access controls.

The upgrade aligns with LSEG’s Everywhere strategy, embedding its feeds into AI platforms. In December, the group integrated its data with ChatGPT for authorized users, broadening access to real-time and historical market information.

Usage of the Tick History archive is climbing. Customer API requests for Tick History rose 48% in the first half from a year earlier, while traffic on LSEG’s real-time network increased 29%, hitting 26 million messages per second in June. LSEG gave no launch date or pricing for the GPU setup.

Trading firms require historical data without the burden of maintaining separate copies. Ron Lefferts, who has co-led LSEG’s Data & Analytics division since April 2025, explained that users rely on the dataset for three key purposes: long-term strategy testing, execution analysis, including order book reconstruction, and risk evaluations such as market replay and AI-driven research. Execution-quality audits, for instance, help brokers comply with regulatory best-execution requirements.

LSEG already provides cross-market surveillance tools based on a consolidated orderbook of over 40 UK and EU venues, introduced in January. Retail brokers like Capital.com have adopted LSEG’s feeds, signing up in February 2025 for news, market data, and research to serve over 750,000 traders.

High-resolution tick data processing prevents loss of detail that could skew analysis. Ioana Boier, NVIDIA’s global head of capital markets strategy, stated that researchers can now analyze raw data without downsampling, where tick data is condensed into wider intervals, potentially hiding critical patterns.

While LSEG’s Data & Analytics division grew organically by 5.1% in the first half, its slowest growth among four divisions, other units performed better, with Markets expanding 11.9% and Risk Intelligence at 9.7%. Rivals such as KX, which merged with tick data provider OneTick in 2025, have also developed AI trading systems using NVIDIA technology. Meanwhile, CME Group migrated its live futures and options feeds to Google Cloud in 2019, and LSEG began delivering datasets through Databricks in September 2025, including Tick History.