TRG buys BST, buy side tests AI for research, and more

This summer, with wildfires burning across the globe, what’s burning for financial firms, and how will they contain it or put it out?

Announced this week

Substantive, Aiera survey: Buy-side testing AI to consume research

Euronext-owned information provider Substantive Research has released a survey that looks at 35 of the largest asset managers’ use of AI, in partnership with Aiera, a content delivery platform. The results found that: 77% of respondents have generative AI platforms like Claude and ChatGPT in place enterprise-wide; broker data licensing restrictions are the biggest barrier to adopting research/data feeds, accounting for 69%; and 37% are taking four to six months to approve and adopt these models.

The buy side is reviewing how AI can accelerate its research consumption, and is actively monitoring how AI can improve research discovery, accelerate insights and improve analyst productivity.

LSEG data now available for AI through MCP

The London Stock Exchange Group will make its data and analytics available through Model ML via its MCP connector, enabling access to its content for financial workflows. This allows users to access LSEG data, including pricing, estimates, fundamentals, reference data, ownership, macroeconomic indicators, ESG data, news, forecasts and analytical models and apply AI to them. The initiative is governed by LSEG Everywhere, the exchange’s AI and data strategy that utilizes MCP and partners with AWS, Microsoft, Google Cloud, Claude, ChatGPT, Snowflake and Databricks.

TRG Screen buys BST America

Market data and subscription cost management provider TRG Screen has acquired New York-based data advisory and consulting firm BST America. The deal expands BST’s offering into the strategic and consultancy space.

The deal allows TRG to support clients at every stage of their evolution and data maturity, to make better market data-related decisions, and to “further expand and enrich the domain expertise our customers rely on, making proven best practices even more accessible,” said Nadine Scott, chief customer strategy officer at TRG Screen.

Bondwave enhances Effi oversight, TQA platform

Fixed-income trading technology vendor BondWave has released the latest version of its Effi platform, which includes expanded trade oversight and transaction quality analysis abilities to give clients more control over trade surveillance and better visibility into execution quality. 

The new release adds 14 new configurable attributes, bringing its total to 23, such as execution yield, base price, and reviews such as minimum denomination, increments, muni limited offering, and sales credit/markup variance.

Users can also view information directly from their trade file, searchable by security ID and date, allowing them to research calculations and providing greater transparency.

ETrading releases report on bond tape environment

A new report from Etrading Software’s ETS Connect Plus examines how UK fixed-income professionals are responding to the UK bond consolidated tape, which went live in late June. Almost all those surveyed expect the bond tape to be positive for their firm and the market, and 100% are actively consolidating—or planning to consolidate—their fixed-income data providers as a result of the tape. Of the 100 buy-side and sell-side firms surveyed, 74% will consume the tape via a third-party, and 65% cite integration complexity with existing systems and workflows as their most significant challenge.

BridgePort pilot crypto standard for settlement in trading industry

BridgePort, a middleware layer for the institutional crypto industry, has begun piloting its Digital Asset Master Agreement (DAMA), a legal standard for off-exchange settlement between trading firms, custodians and exchanges, which participants can use as a common starting point for establishing tri-party settlement agreements.

BridgePort is now forming a working group comprising trading firms, custodians, and exchanges to pilot DAMA across live off-exchange settlement relationships.

FIX taps Atomic Wire to update standards authoring

FIX Trading Community has adopted a new automated development workflow for FIX based on the Orchestra Domain Specific Language from Atomic Wire Technology. Rather than manually maintaining repository artefacts, the Orchestra DSL enables proposed changes to be expressed in a structured, machine-readable form. These changes can then be automatically validated, combined and used to generate updated repository artefacts, providing greater consistency, traceability and confidence throughout the standards development lifecycle.

In addition, the process automatically generates derived items, such as the FIXML standard, schemas and implementation reports, to ensure uniformity across published accounts.

What you might have missed from us

BlackRock aims to transform data quality checks with AI

WatersTechnology editor-in-chief Anthony Malakian presents an interview with BlackRock’s managing director of the firm’s AI platform engineering team Dan Romuald Mbanga, who—after more than a decade with Amazon Web Services and Google—is using AI as a weapon to combat bad data and improve quality controls.

A third of banks do not maintain logs for GenAI models

Sibling publication Risk.net outlines how firms are responding to AI in its inaugural Model Risk Management study, showing that only one respondent—out of 44 surveyed—actually stores and reviews a complete record of the prompts and outputs for AI models currently in production, while 35% already run at least one agentic AI system in production, 23% have a structured pilot or proof-of-concept underway, and a further 28% are actively weighing deployment.

Ethereum rebuild casts doubt on banks’ QC prep

Asia editor Wei-Shen Wong discusses quantum computing, financial firms’ experiments with it, and the recent slowdown in testing and adoption. The reason? Sure, cost and trying to find a home for this technology today is debatable. But on the horizon is Q-Day, when quantum computing is predicted to become powerful enough to break today’s standard encryption methods for credit cards, blockchains, passkeys, and more. Can Ethereum help, and how?

Few banks formally evaluate GenAI human-in-the-loop controls

Only one in five banks use tools to formally monitor the effectiveness of human-in-the-loop controls on AI and large LLMs. The numbers include four of six G-Sibs saying they use key performance indicators, a further six assess it periodically, five use human judgment without systematic measurement, while 14 banks have not assessed human-in-the-loop effectiveness at all.

In other news

Datacenter builders are racing to offload stakes worth billions, The Wall Street Journal

US datacenters running AI computing are marketing stakes in their operating companies as investors seek control of infrastructure supporting the growth of AI, The Wall Street Journal’s Anissa Gardizy reports. Bankers are working with a number of datacenter providers to present parts of their businesses—collectively valued at tens of billions of dollars—to private equity firms. A controlling interest can give a buyer access to operational capacity, development sites, client relationships, and future expansion.

This begs the question whether any datacenters specifically serving financial clients will follow suit (or already have?) and how financial services companies—for whom proprietary data is something to be held confidentially, or where possible, commercialized and monetized on their terms—will react if they do.