A PostgreSQL Database for Every Agent

YugabyteDB 2026.1 and YugabyteDB AMP introduce a serverless, scale-to-zero PostgreSQL database tailored for AI agents, offering each agent its own isolated database at a minimal cost. This platform efficiently manages bursty agent workloads, ensuring predictable costs with Resource Governance and scaling seamlessly from a prototype to global deployment without requiring rewrites. The update includes features like in-database RAG, vector search, and the MAGE graph engine, along with enhanced tools such as database branching and YugabyteDB Connection Manager. These innovations are designed to meet the growing demands of AI applications, providing a comprehensive data infrastructure that supports the entire lifecycle of agent operations, from provisioning and scaling to teardown, through MCP calls. By consolidating multiple functionalities into one platform, including native vector search and multi-tenant graph capabilities, YugabyteDB 2026.1 aims to streamline the complex data requirements of AI-driven applications while maintaining operational efficiency and scalability.