MCP连接器与基于使用量的数据消费

UK | Brokers, Asset Mgrs & Exchanges Jef U: The Ins and Outs of MCP Connectors &Usage-Based EXPERT STUDIES GUEST SPEAKER Damian Sasso We are in a rapid adoption phase of AI-driven data consumption, with theModel Context Protocol (MCP) serving as an important new distributionchannel. A key medium-term investment debate for LSEG and peers centreson whether incremental growth in data and feeds can (more than) offsetpotential pressures on desktop revenues. Clearer pricing frameworks andevidence of revenue substitution will ultimately be critical in determiningoutcomes post adoption. Group Product Manager – PolyAILimitedFormer Group Director,Collaboration Services – LondonStock Exchange Group plc HOSTED BY Tom Mills Research Analyst, DiversifiedFinancials This report contains an edited version of the transcript. MCP is an enabler of AI-based distribution.MCP is an open, model-agnostic standarddesigned to allow LLMs to consume structured data safely and consistently. In practice, it actsas a connective layer between proprietary datasets and AI-driven applications, enabling dataand workflows to be accessed within model-native interfaces (rather than traditional terminalsor APIs), while retaining user authentication, permissioning, and commercial controls. For datavendors, MCP primarily serves as an enabler of AI-based distribution rather than a standaloneproduct. The competitive advantage sits in the data.In an AI-enabled distribution model, competitiveadvantage does not sit with MCP itself but with the underlying data assets and the taxonomyand delivery frameworks that make that data usable and trusted at scale. Proprietary real-timefeeds, reference data, symbology, and long-standing historical datasets retain value becausethey are deeply embedded in client systems and aligned with regulatory and operationalrequirements. Consistent taxonomy across datasets is particularly critical in an AI context, asit allows LLMs to generate reliable outputs without re-engineering client data infrastructure. Data leakage risks appear overstated in the near-term. Concerns that AI models canreconstruct proprietary datasets through repeated querying appear overstated in the nearterm. MCP-based architectures continue to rely on entitlements, authentication, and usagetracking analogous to existing real-time feeds and API-based distribution models. Structural changes in data consumption.AI agents imply a gradual shift toward consumption-based real-time data usage rather than seat-based entitlements. This raises unresolvedquestions around how exchanges and data licensors classify model-driven applications:whether they are treated as extensions of desktop usage or as proprietary applicationsrequiring separate commercial terms. Weighing up the net revenue impact. AI-driven distribution may expand the addressablemarket, particularly across the long tail of financial services users—such as smaller assetmanagers, private capital and family offices—that were never heavy desktop buyers. However,this potential upside must be weighed against the risk of revenue dilution at Tier 1 andTier 2 institutions if AI-mediated distribution leads to desktop substitution without a clearcommercial offset. Nevertheless, trading-focused desktop applications appear relativelydefensible. Tom Mills * | Equity Analyst Pricing strategy: adoption first, monetisation later.LSEG's current strategy prioritisesadoption and long-term enterprise engagement over immediate usage-based charging. Whileusage-based pricing is a logical end-state, both technology and customer behaviour remain tooearly-stage to set optimal pricing frameworks. Recently signed long-term access agreementsimprove near-term revenue visibility and customer lock-in, but could create future tension Surinder Thind, CFA ^ | Equity Analyst Laura Gris Trillo, CFA * | Equity Analyst Jefferies University ‡ | Equity Research Team if AI adoption drives usage materially beyond initial assumptions, potentially leading torenegotiation of terms or the introduction of usage caps. Jefferies University: Insights from Experts Jefferies University: Insights from Experts Jefferies University is an executive-level education program that connects independent experts withinstitutional investors on matters critical to making informed investment choices. Through this globalprogram, Jefferies’ clients have the opportunity to meet thought-leaders, innovators and scholars ona wide range of subjects that may be relevant to building investment portfolios. Topics can be tailoredto suit your needs and delivered in different formats including private or group conference calls, in-person presentations or via video-conference as well as at our numerous conferences and summitsheld around the world. Jefferies University: Exponential Series Jefferies University courses go beyond investment matters. OurExponential Seriesalso delves intoissues that can impact investors on a personal level. Join us for sessions tha