EDB Postgres® AI Q1 2026 Release Highlights

The sovereign safe harbor for the agentic era

Only 13% of enterprises have successfully moved agentic AI projects into production. The bottleneck is infrastructure, not ambition. Our survey of global executives confirms the goal: a unified, sovereign platform for transactional, analytical, and AI workloads under their direct control.

Traditional data stacks, built by stitching together disparate legacy and cloud platforms, have become a liability. In the high-volume, continuous workflows required by AI agents, these fragmented systems introduce sovereignty risks, inconsistent performance, and unpredictable consumption costs.

EDB Postgres® AI is the escape — a sovereign safe harbor providing petabyte-scale performance and 58% cost savings without vendor lock-in. This quarter’s platform updates are generally available today, and prove that you don’t have to sacrifice cost for scale or sovereignty. Read on to learn more.

Predictable performance and cost for petabyte scale analytics

Enhancing control and observability for warehouse analytics

We are doubling down on making Postgres the best engine for analytics in this new agentic era. To do that, we put EDB Postgres AI for WarehousePG to the test. In a new study, McKnight Consulting Group evaluated our massively parallel processing (MPP) solution against cloud-only data warehouses like Snowflake and Databricks:

• Cost efficiency: Up to 58% TCO savings compared to cloud data warehouses with a predictable cost model that eliminates consumption-based pricing and hidden fees.
• Predictability: Even when many employees or AI agents use the system simultaneously, WarehousePG remains fast and reliable. It offers up to 52% more consistent performance for high-concurrency workloads than cloud-only systems, which experience greater slowdown under the same pressure.
• Sovereign control: Unlike cloud-locked solutions, EDB Postgres AI gives you unparalleled deployment flexibility — in the cloud, on-prem, air gapped, or hybrid.

The results confirm that petabyte-scale analytics and AI readiness do not require proprietary platforms or cloud lock-in. To dive deeper, read the full benchmark report here.

To streamline operations for these MPP workloads, the new WarehousePG Enterprise Manager (WEM) is now available. WEM replaces complex command-line tools with a visual interface for comprehensive management and monitoring of WarehousePG clusters.

By integrating real-time telemetry, AI-assisted development, and interactive configuration, WEM streamlines complex distributed database operations into a single pane of glass for auditing cluster health, tuning SQL performance, and managing security in high-performance environments. This fusion of massive scale and simplified management allows EDB Postgres AI for WarehousePG to set a new standard for operational excellence in the agentic era.

Redefining the boundaries of Postgres performance

In addition to enhancing warehouse analytics, we continue to invest in accelerating analytics for operational data. Last month, at NVIDIA GTC, we unveiled deep integration between our PGAA extension and NVIDIA RAPIDS for Apache Spark. By leveraging Spark’s horizontal scalability and offloading compute-intensive analytical workloads to NVIDIA GPUs, enterprises can eliminate the traditional CPU bottlenecks that often stall complex queries in high-concurrency environments. This GPU acceleration allows customers to achieve up to 91x faster analytical queries on live operational data, providing the massive throughput and sub-second latency required for AI agents to process, interpret, and act in real time. For a deeper dive, read our technical blog on achieving predictable performance at scale.

Building autonomous agents with EDB Postgres AI

According to Gartner, enterprise spending on AI-capable database systems is set to triple by 2028. In their recent report, Innovation Insight: Database Management Systems for Enterprise AI Agents, Gartner recognizes EDB for its ability to fulfill the three vital roles of an agentic database: acting as the long-term memory, the primary knowledge source, and the reliable engine for task execution.

This quarter we are ushering in the next generation of EDB Postgres AI to make it even easier to build agents directly on top of your data. These updates are specifically designed to help your team close the gap between prototype and production by providing the integrated tooling needed to activate your data and orchestrate agent behavior.

Getting data AI-ready with an upgraded Vector Engine

Effective agentic workflows depend entirely on the quality and speed of the underlying knowledge base. This quarter, we’ve significantly enhanced the platform’s built-in Vector Engine by introducing VectorChord as a high-performance companion to pgvector.

This update provides a tailored indexing strategy for every workload. While pgvector remains the standard-bearer for most RAG applications, VectorChord serves as a seamless, drop-in upgrade for enterprises scaling into the billions of vectors, providing 100x faster indexing and the sub-second latency required for high-concurrency multi-agent systems.

Enhancing Agent Studio, AI Pipelines, and Model Serving capabilities

EDB Postgres AI’s Agent Studio, now integrated with Langflow, offers a visual drag-and-drop interface to build, test, and deploy agents. Users can build from scratch, use open-source starter flows, or create custom reusable templates. To bridge the gap between agent orchestration and your data, we’ve introduced the EDB MCP Server, Database, and Embedding components within Langflow. This utilizes the Model Context Protocol (MCP) — an open standard for secure tool and data connectivity — to allow agents to interact directly with your Postgres databases as tools, eliminating custom integration hurdles.

We’ve also introduced a number of enhancements to AI Pipelines. You can now create multi-step pipelines with simple point-and-click configuration, unifying data prep and embeddings so you build once and maintain once. For SQL users, the AIDB extension has also been enhanced with advanced PDF parsing and in-database LLM capabilities. These updates enable the native ingestion of complex documents and the summarization of unstructured data into actionable insights—all with just a few lines of SQL.

To ensure you have the right model for every use case, our model serving capability now allows you to connect to a provider like Hugging Face without manual glue code. By connecting your model of choice directly to the platform, you reduce data movement, improve security, and maintain full sovereign flexibility across your AI stack.

For more details on the latest AI Factory enhancements, read the dedicated blog.

Managing your AI-ready estate at scale

The agentic era changes how you interact with your own infrastructure, not just what you build on top of it. We are simplifying the way you manage your data and infrastructure with language and industry-standard automation.

Streamline operations with natural language 

EDB Postgres AI now includes a native chatbot that lets you manage the platform in natural language — no scripting required. You can streamline tasks including:

• Create or edit migration projects.
• Modify user roles and permissions.
• Get performance and health recommendations directly.

This moves manual operations to interactive dialogue, a meaningful step toward a database that participates in its own administration, and freeing up your team for more strategic work.

We’ve also added autonomous storage scaling, which expands capacity in response to workload demands, within guardrails you define, such as maintenance windows or maximum storage caps. No real-time intervention and no over-provisioning required to keep your apps up and running. For workloads that can’t tolerate a moment of downtime, our Hybrid Manager now supports three-location deployments for 99.999% uptime, ensuring you never depend on a single region or data center for your business continuity.

The foundation for automation: strategic partnership with Red Hat Ansible Automation Platform

We are proud to announce that EDB and Red Hat have jointly validated a reference architecture that provides a resilient data layer for Red Hat Ansible Automation Platform (AAP). For mission-critical automation platforms, a resilient foundation is not merely a best practice—it’s essential to protecting business processes from interruption.

Together, EDB Postgres AI and AAP deliver:

• High Availability: Multi-AZ resilience with automated failover in under 30 seconds.
• Disaster Recovery: Robust cross-region protection for uninterrupted operations.
• Seamless Deployment: Full-stack, Ansible-native integration.

Every AAP release is tested against EDB Postgres AI, ensuring a certified, production-ready foundation for enterprise automation at scale. Learn more about this partnership here.

Ready to see the data behind the performance?

Don’t just take our word for it—the McKnight Consulting benchmark report covers the full methodology, workload configurations, and results across cost, concurrency, and sovereignty. Read the full benchmark report here.

To go deeper on any specific capability, please see our documentation or reach out to our team.