AI agents are moving rapidly from demos to decision-making. That shift changes the stakes considerably, and it’s surfacing a question enterprises keep arriving at from different directions: 

When AI is making mission-critical decisions, acting on your data, touching your operational systems—how much of that process is truly yours?

At EDB, we’ve been building toward an answer from the data layer for a long time. Open, portable, governed infrastructure—Postgres at the center—isn’t a new position for us. But the rest of the industry is catching up fast. 

Which is why we expect that sovereignty will be a key topic of discussion at Red Hat Summit this year, alongside AI, security, and infrastructure as a core thread running through the program in Atlanta this May. We’re featured in 2 sessions that get into the specifics of what building sovereign, production-grade AI actually looks like, technically and organizationally.

Why EDB and Red Hat: Open Source, All the Way Down

The EDB and Red Hat collaboration is built on a shared premise: enterprises shouldn’t have to choose between capability and control. Both companies are committed to open source not as a licensing model, but as a design philosophy that keeps customers portable, infrastructure auditable, and vendor lock-in off the table.

On the technical side, the integration runs deep. EDB Postgres AI on Red Hat OpenShift gives organizations a unified data layer for both structured data and vector embeddings—meaning the same Postgres instance that runs your operational database can also serve as the knowledge base for your AI applications. The pgvector extension handles vector storage and semantic search directly in the database, eliminating the complexity of a separate vector store and keeping your data where you already govern it. Red Hat OpenShift AI provides the platform layer: model lifecycle management, MLOps and LLMOps tooling, and the hybrid cloud flexibility to run workloads wherever they need to run.

What that combination makes possible is AI that’s grounded in your data, running on infrastructure you own, governed by policies you set—across clouds and on-premises environments, without rearchitecting every time the market shifts. That’s what sovereignty looks like in practice, and it’s what we are showing up to Red Hat Summit to talk about.

EDB Sessions at Red Hat Summit 2026

Building Scalable RAG Architectures with EDB, OpenShift, and OpenShift AI

Wednesday, May 13 |  11:40 AM–12:00 PM EDT  |  Developer Zone

Add to your schedule »

Postgres has always been a serious database. What’s changed is that it’s now a serious AI data layer, too. This session, led by Red Hat Technical Director ofHybrid Platforms, Natale Vinto, is a live demonstration of what that looks like end to end.

The architecture is RAG—Retrieval-Augmented Generation—built on EDB Postgres for Kubernetes on Red Hat OpenShift, integrated with Red Hat OpenShift AI. Postgres serves as the vector store and knowledge base: ingesting data, indexing it, and making it retrievable at the speed and governance level production workloads require. No separate vector database. No proprietary pipeline. Red Hat OpenShift orchestrates the workloads; Red Hat OpenShift AI handles model serving at scale.

The demo runs the full workflow, from data ingestion and vector indexing through to model inference, on an open, hybrid cloud stack you actually own. That’s the point.

Don’t Forget to Breathe: The Human Story Your AI Demo Is Missing

Wednesday, May 13 |  2:35–2:55 PM EDT  |  Expo Hall — Discovery Theater 1

Add to your schedule »

A few hours later, EDB Machine Learning Engineer Bilge Ince joins Red Hat’s Shane Heroux for something a little different.

Most AI demos don’t fail because the technology doesn’t work. They fail because the story doesn’t land. Ince and Heroux work through both directions: starting from a known problem, where the demo practically writes itself, and starting from a solution looking for a problem, which is the harder, messier path most teams actually face. They’ll show examples of demos that work, and explain exactly why they do—and why most don’t.

If you’re the person who has to translate what your stack does into something that compels a room that doesn’t think in infrastructure terms, catch this session.

Find Us in Atlanta

Red Hat Summit runs May 11–14 in Atlanta. If you’re attending, come find us at one of these sessions. And if you want more background on what EDB and Red Hat are building together, this is a good place to start.