Redpanda has introduced new capabilities in its Agentic Data Plane (ADP), including a centralized AI Gateway, AI observability via OpenTelemetry, built-in AI agents, and unified authentication and authorization. These features create a unified governance layer that enables enterprises to securely connect AI agents and Model Context Protocol (MCP) servers to live enterprise data, providing full visibility, control, and trust at scale.
Enterprises transitioning AI agents from experimentation to production face significant hurdles in safely managing access to sensitive data and systems. Redpanda’s ADP serves as a central control plane that routes all agent interactions through a single governance layer, reducing the complexity of configuring policies at individual data sources and enabling secure, scalable operations without compromising real-time performance.
The AI Gateway acts as a unified access point between applications, AI models, and MCP services. It enforces policies, applies token budgets and spending limits, optimizes usage with deferred tool loading, and maintains an admin-controlled registry of approved MCP servers for rapid, YAML-based deployment.
Security is anchored in OIDC authentication and granular authorization policies applied to every request from users, services, or agents. Observability captures complete request-response flows for inspection in the Redpanda console or export to external tools, supporting compliance, incident analysis, and continuous improvement of agent behavior.
Built on Redpanda’s low-latency streaming engine, ADP delivers continuous context updates and event-driven execution. With over 300 connectors via Redpanda Connect, enterprises can integrate data from databases, SaaS platforms, streams, and lakes without data movement or architectural changes, preserving existing investments while enabling real-time agentic workflows.
These capabilities are now available to select Design Partners on Bring Your Own Cloud (BYOC) deployments in AWS, with planned expansions to GCP, Azure, and self-hosted environments.
Redpanda is pioneering the agentic data plane - a new category in AI infrastructure that makes it simple and secure to connect AI agents with enterprise data and systems. Built on a multi-modal data streaming engine, Redpanda empowers agentic applications that reason and act in real-time with speed, autonomy, and precision. Global leaders including Activision, NYSE, Cisco, Moody's, Texas Instruments, Vodafone and 2 of the top 5 banks in the U.S. rely on Redpanda to process hundreds of terabytes of data a day.