FTSE Russell launches AI-driven thematic index series with MarketPsych

FTSE Russell has launched a thematic index series whose theme classifications come from text analytics. Developed with MarketPsych — described in the announcement as a leader in sentiment-driven analytics and behavioural-economics tools — the series feeds LSEG datasets of news, transcripts and financial data through the firm's AI-enabled framework, which scores them for context, meaning and relevance, before FTSE Russell's rules-based index construction decides what the portfolio holds, as ETF Express reported.

Five pillars organise the series: Technology & Innovation, Society & Demographics, Geopolitical Realignment, Natural Resources & Infrastructure, and Health & Wellbeing. FTSE Russell says the reach runs from artificial intelligence and quantum computing to new energy, space exploration and ageing populations, calling it the next generation of its thematic solutions for global investors seeking systematic, forward-looking access to themes.

Gerald Toledano, FTSE Russell's global head of equities and multi-assets, said the launch builds on the long-standing LSEG partnership with MarketPsych and gives clients a systematic, investable route into evolving themes. Richard Peterson, MarketPsych's chief executive, said the partnership extends the reach of his firm's AI and natural-language-processing models to a growing global investor base and helps address longstanding demand for thematic exposure.

The pitch rests on a simple split: MarketPsych decides which companies belong to which theme, while FTSE Russell decides how they are weighted, capped and screened, and how quickly membership turns over. The claim to adaptiveness sits in the first half — a classification built from the current text flow re-rates a theme as the language around it shifts, rather than waiting for a methodology committee to convene. This publication has argued that the label hides the rules, and the value-fund case is the proof: a construction decision separated a 21% year from an 11% one while the category label stayed the same.

Buyers of this series will not be getting 'AI' as an exposure; they will be getting a rules-based portfolio whose constituent list was assembled upstream by a text model and licensed under a theme name that will do most of the marketing.

What is not yet in evidence is a fund: the coverage names no ETF issuer, no product and no fee, which is the point at which a thematic series either becomes an allocation or stays a licensing catalogue. Also unstated is the refresh cadence for the thematic dataset — the detail to watch, because a classifier tracking the news cycle will rotate a theme's membership faster than most allocators re-underwrite the position.