ICEIntercontinental Exchange, Inc.

“In March alone, our Servicing business processed approximately 4 billion API and web services calls up nearly 20% year-over-year driven by increased use of our AI and Business Intelligence tools, a signal that our infrastructure is becoming more embedded in client operations, not less.”

“What was previously a 46 touch-point process spanning 10 days, now requires just 6 touch points over 2 days.”

“at our ICE Experience Conference in March, we unveiled AI-powered Voice and Chat Agents for Mortgage Servicing to handle routine borrower inquiries, execute common loan management actions and help servicers manage fluctuating call volumes. We also launched 16 exception-based automation agents for complex servicing workflows, including escrow management, investor reporting and disaster-related processes.”

“ICE now offers an AI model control protocol server, or MCP server located in our data center and available on the ICE proprietary Cloud to ease access to ICE's nonproprietary data. And we are actively engaged with major AI model vendors to explore the development of additional server protocols and topology for further access to and the protection of ICE's proprietary data.”

“Internally, we're already deploying AI in production across our organization. Teams are using it to undertake code writing, enhanced pricing workflows, accelerate index calculations, support client interaction and earlier identified loan servicing issues. These are not experiments.”

“The nonproprietary data in ICE's MCP server recently launched and is being offered under existing license agreements to some of our customers to see if this type of delivery has benefits versus traditional data connectivity methods.”

“Artificial intelligence fits squarely within that strategy. It accelerates the way regulated workflows are processed by embedding intelligence directly into our systems of record, preserving governance and audibility and while improving speed and insight. Importantly, as automation increases, value shifts towards workflow outcomes rather than seat pricing. We recognized that early on that pricing on workflow outcomes would be the preferred pricing model. As AI is incorporated into these workflows, that pricing model remains durable and stands to benefit us.”

“As these data sets scale, the ways in which clients use our data continue to expand, whether powering automated workflows, AI models or real-time decision-making, every use case requires high-quality proprietary inputs and we believe ICE controls the most comprehensive and institutionally trusted data sets across these markets.”

“Private global data center network connecting over 750 data sources and 150 trading venues across 24 countries is a physical infrastructure asset that cannot be replicated quickly or cheaply, and it continues to benefit from secular demand trends, including higher messaging activity and AI-driven demand for capacity.”

“certainly, the appetite for data across the entire segment with some of the comments that Jeff made around AI and things that we're doing that, there are 3 real components that we have in the portfolio, a very large segment of proprietary data, that is very well versed in the regulatory community.”

“Our data team is actively transforming our proprietary and nonproprietary financial, market and commodity data into AI-ready formats. And they themselves are internally integrating AI technologies into ICE to enhance data utility, extraction and analysis for our clients.”

“Because this is a highly regulated market that requires a deep understanding of risk, audit and governance before deployment, we embed AI directly into the systems of record, reinforcing ICE's role as a neutral trusted platform that does not compete with its customers.”

“Importantly, customers are embedding our proprietary and secure real-time data for inference in their workflows and not simply consuming it to train models and then move on. As these use cases deepen demand for ICE's proprietary data increases rather than decreases.”

“Our AI capabilities are deployed within that same framework, not outside of it, which means the governance auditability and controls our customers rely on extend fully to every automation we deliver.”

“you heard us talk about we built an MCP server that is there for AI, whether or not that becomes an Oracle for reference data, which will be in demand for people trading on chain.”

“invest in things like Polymarket and OKX and a few other AI-related investments that we've been making.”

“the growing reliance on proprietary and institutional-grade data by AI systems and human decision-makers alike”