BMLL and SIGMA AI partner on real-time market data analytics

BMLL and SIGMA AI have announced a partnership to combine BMLL’s historical order book datasets with SIGMA AI’s real-time analytics engine, creating a service intended to compare live market activity with historical trading patterns. The companies said the product will support execution intelligence, benchmarking, market impact scoring and market regime classification for trading firms and market infrastructure providers.

BMLL provides harmonised historical Level 3, Level 2 and Level 1 data and analytics for capital markets. SIGMA AI provides real-time intelligence tools for financial markets through a streaming platform it calls “Quant in the Cloud”.

Under the collaboration, BMLL’s data will be used to establish historical reference points for volatility, liquidity, spread behaviour and auction dynamics. SIGMA AI’s platform will apply continuous computation, machine learning inference and natural-language delivery to live market data, according to the companies.

The planned service is designed to flag unusual activity by checking current trading conditions against historical benchmarks derived from BMLL data. The companies said this approach could allow users to identify anomalies, classify market regimes and receive alerts as market conditions change.

Historical data as a real-time reference layer

Order book data records the supply and demand visible in a market at different price levels. Level 1 data typically captures the best bid and offer, Level 2 adds depth across price levels, and Level 3 includes individual order-level information. BMLL said its historical datasets cover these layers and can be used to define baseline measures across venues, instruments and time periods.

In the proposed model, those baselines become a reference layer for SIGMA AI’s live analytics. For example, a current spread, liquidity condition or auction behaviour can be assessed against prior observations rather than viewed in isolation. The companies said the aim is to bridge historical reporting and real-time decision support.

The anticipated user base includes exchanges, market data providers, buy-side institutions using both low-touch and high-touch trading workflows, sell-side brokers, platform vendors, market makers and quantitative firms, according to the announcement.

Paul Humphrey, chief executive of BMLL, said SIGMA AI had joined the BMLL Activate Data Credits Programme. He said real-time analytics and decision support were “a natural application layer” on top of high-quality historical order book data, adding that the partnership reflected BMLL’s focus on enabling AI-driven workflows using its data.

Andy Simpson, founder and chief executive of SIGMA AI, said the firms were building a production-ready analytics layer that compares live market conditions with historical context derived from BMLL. He said the product would initially focus on Saudi Arabia, the UK and European instruments, with a later expansion to the US.

Data credits programme supports development

The partnership is anchored in SIGMA AI’s participation in BMLL Activate, a data credits programme designed for partners building analytics and data-driven products on top of BMLL’s historical market data.

According to BMLL, selected participants receive a defined allowance of data credits that can be used for agreed access to the BMLL Data Lab or BMLL Data Feed. The access covers specified datasets, instruments, venues and historical time windows.

The programme is intended to reduce or remove upfront data licence costs during the build phase, while giving participants a route to commercial rollout if a use case proves viable, BMLL said. No financial terms for the BMLL and SIGMA AI partnership were included in the announcement.

This story draws on original reporting from Finextra Research.