How NVIDIA technology will speed up market data analysis
The collaboration uses NVIDIA cuDF, a tool for processing large datasets on GPUs, alongside Polars, a data-processing framework. Benchmark analysis indicates that years of Tick History data can be analysed in seconds, compared with processing times of several hours under existing approaches. This could help firms examine individual trades and price quotes without reducing the level of detail in the data.
The system also avoids creating another complete copy of the underlying dataset. Existing licences, permissions and access controls remain in place, allowing financial institutions to work with governed market information while managing data access. The initiative forms part of LSEG’s Everywhere AI strategy, which aims to bring its financial data closer to the computing systems used for research, market analysis and artificial intelligence applications.
Rising demand for historical market data
LSEG Tick History contains approximately 110 trillion rows of data covering 100 million financial instruments, more than 30 years of history and over 580 exchanges, trading venues and contributors.
The partnership comes as demand for market information and faster analysis continues to rise. During the first half of 2026, traffic across LSEG’s real-time network increased by 29% compared with the corresponding period in 2025. The network reached a peak of 26 million messages per second in June 2026. Meanwhile, customer application programming interface (API) requests for Tick History data rose by 48% over the same period.
These figures highlight the growing need to process large volumes of financial information efficiently. As equity markets move towards near-24-hour trading, firms are also looking beyond traditional systems that load market data after closing and run analysis overnight. Faster processing could allow them to study historical conditions continuously throughout the trading day.