LSEG has launched Model-as-a-Service (MaaS), expanding its marketplace strategy beyond data distribution into the hosting, commercialisation and integration of financial models. At launch, Societe Generale has joined as a provider, making fixed income, FX, ESG and equities analytics available through the platform.
The move positions LSEG not just as a data vendor, but as an infrastructure layer for distributing and operationalising models across the institutional market.
From Data Vendor to Model Platform
MaaS enables institutions to host, distribute and consume models through what LSEG describes as a secure, governed marketplace. For providers, it offers a route to commercialise proprietary analytics without building their own distribution and compliance frameworks.“We’re seeing rapid acceleration in innovation and transformation across modelling and workflows, and our goal is to be an enabler of that shift,” Aysegul Erdem, Head of Modelling Solutions, LSEG Institutions, tells TradingTech Insight. “Institutions increasingly need enterprise-grade integration that delivers consistency across the organisation, while still allowing them to select best-in-class models not only from banks and independent providers, but also from buy-side institutions, fintechs, and even research groups as more of these organisations commercialise their modelling IP. At the same time, AI-enabled workflows will fundamentally reshape how these models and insights are accessed and consumed. Through our Models-as-a-Service offering build in partnership with Microsoft and supported by our broader AI partner ecosystem, this is exactly the transformation we aim to deliver.”
As workflows become increasingly API-driven and AI-enabled, the value chain is shifting from raw datasets to deployable analytical components. LSEG is seeking to occupy that layer by combining its data assets, infrastructure and cloud partnerships.
“Given our strength in data, we see a significant opportunity to act as an incubator for innovation,” adds Erdem. “The traditional silos between datasets, markets, and analytics are breaking down, and we want to stay ahead of that convergence, helping clients unlock more sophisticated, connected insights.”
Societe Generale’s participation represents an early example of the sell side productising quant IP within a managed marketplace structure, with its models delivered alongside LSEG analytics through a single integrated environment.
Model Integration
“One point that has resonated strongly with clients is the ability to use multiple models simultaneously,” notes Erdem. “For example, they can combine analytics from one bank with models from another, run comparisons, and even create agents that orchestrate different components together. That level of composability simply doesn’t exist today and it’s powerful. It also lowers barriers to entry: in many markets, hiring specialist quants for every asset class is challenging, especially in disparate or emerging markets. If clients can access best-in-class expertise through aggregated models and integrate them via AI, it opens up new opportunities and significantly broadens access.”Institutions with their own compute environments can integrate via API-to-API connections and execute models locally, bearing the associated compute costs. Alternatively, they can host models on LSEG’s platform, with execution supported through LSEG infrastructure in partnership with Microsoft Azure.
Commercial arrangements reflect differences in model complexity, from lightweight analytics to compute-intensive strategies. For trading technology teams, this raises practical considerations around execution topology, performance management and cost allocation.
Intellectual Property and Control
Intellectual property protection is central to the proposition. LSEG positions itself as a neutral intermediary capable of hosting proprietary models without requiring direct code or data sharing between competing institutions.
“IP protection is fundamental for us,” points out Erdem. “Many providers choose to work with us specifically because of how seriously we take IP considerations. We treat this with the highest level of care from both legal and commercial standpoints. For example, a bank may not be willing to share data or code with a competitor, but they are comfortable working with us as an independent party. The same applies in private markets and other sensitive domains.”
The platform allows providers to determine the level of transparency made available to users.
“In terms of transparency, the delivery of the models are configurable,” explains Erdem. “A model provider can decide how much to expose. They can allow clients to adjust parameters and incorporate their own views, or they can limit exposure and keep it more closed. It depends entirely on the provider’s preference. Our infrastructure supports either approach.”
The marketplace structure also formalises model distribution, potentially simplifying contractual and compliance processes compared with bespoke bilateral arrangements.
Marketplace Economics
MaaS incorporates revenue sharing between LSEG and model providers. End clients subscribe to the infrastructure and to specific models, with consulting services available where bespoke integration work is required.
“We are providing a facility for commercialising models and model-building components from our partners and fully integrating them with our data offering,” says Erdem. Building models internally alongside data is one challenge; commercialising and distributing them at scale introduces additional infrastructure and commercial complexities. By creating this ecosystem, we enable a faster route to market for our partners and deliver greater choice and flexibility for our clients.”
The approach reflects the emergence of platform-style economics within financial modelling, where providers contribute analytics and institutions consume them through a shared operational environment.
Embedding Models into AI Workflows
Integration with Microsoft’s ecosystem, including Model Context Protocol (MCP) connectors and availability within Copilot Studio, underpins the AI strategy. The objective is not simply model distribution, but the ability to embed models within enterprise AI workflows.
“We are very excited about MCP and are doing significant work around it,” notes Erdem. “If the models are already available within the marketplace, integration becomes much easier, creating composability across providers and workflows. With AI, you can dynamically generate context-aware workflows, build agents, trigger alerts, and orchestrate multiple components together. Our goal is to support highly customised workflows while maintaining consistency across trading, risk, and other core functions.”
This architecture allows firms to compare models, combine analytics from multiple providers and integrate them into trading, risk and portfolio workflows.
Expanding the Ecosystem
While the launch initially centres on sell-side providers, LSEG’s ambition is broader. The marketplace is intended to aggregate analytics across complex domains such as private markets and commodities, while also supporting model-building components in addition to finished models.
By leveraging LSEG’s infrastructure and client network, specialist providers can accelerate time to market and focus on developing analytics rather than building distribution capabilities.
The implications are particularly significant for the evolving alternative data landscape. Alternative data strategies are increasingly delivered as signals, factors and predictive models rather than raw datasets. A governed marketplace capable of hosting and distributing those analytical components could materially reshape how such analytics are commercialised.
A Structural Shift
The launch reflects a broader shift in capital markets infrastructure. Data is increasingly becoming a modular input into analytical processes, while models themselves are emerging as commercial products that can be distributed and consumed through shared platforms. Distribution is moving towards API-native marketplaces, while AI-driven workflows are becoming the layer through which these analytics are orchestrated across trading, risk and investment functions.
If successful, Model-as-a-Service could represent an early step towards a financial model ecosystem in which analytics from multiple providers are discoverable, deployable and combinable within a common operational framework.
Subscribe to our newsletter
