Sovereign AI and Data Factory

Sovereign AI and Data Factory is a fully integrated hardware and software platform co-designed by EDB and Supermicro®. It enables organizations to deploy production-grade Postgres® and AI workloads in their own data centers without external dependencies.

This engineered system is optimized for sovereignty, security, and scale. It's preassembled, lifecycle-managed, and ready to run advanced Postgres, vector, and AI pipelines as soon as it's installed.

Why choose a hardware-based system?

Cloud-first isn’t always an option. Sovereign Data and AI Factory exists for environments where:

• Data and model control are legally or operationally required
• Cloud GPU access is inconsistent, oversubscribed, or restricted
• Sovereign AI demands zero external inference or data movement
• Unified observability across Postgres, vector, and AI is critical
• Lifecycle reliability matters across hardware and software

This system removes friction from modern workloads by tightly coupling the hardware, orchestration platform, and supported runtimes so you can focus on outcomes, not infrastructure.

What’s inside

Delivered as a rack-mounted system into your data center, the platform includes:

• Validated Supermicro servers
• Lifecycle-managed Kubernetes and Hybrid Manager
• Built-in HA Postgres orchestration
• pgvector, embedding pipelines, and model-serving capabilities
• Optional GPU-based AI inference and agent runtime
• Full monitoring, patching, and version upgrade automation

All configuration is handled before delivering access. Once installed, use the Hybrid Manager web interface to begin deploying Postgres clusters, RAG pipelines, and agentic applications immediately.

Supported workloads

You can run a wide variety of sovereign or offline-capable workloads:

• High-availability Postgres databases
• Hybrid internal DBaaS environments
• Vector search and embedding pipelines
• Retrieval-augmented generation (RAG) using local LLMs
• On-premises AI inference with GPU acceleration
• Internal chatbots, copilots, and agentic orchestration

See Common workloads and use cases for examples.

Hardware configurations

We offer two validated configurations:

• Core Compute — For high-performance Postgres and vector workloads.
• Advanced AI — Adds GPU-enabled nodes for inference and agent-based systems.

Both configurations are expandable and fully supported. See System specifications for details.

Explore capabilities

• Concepts and architecture
• Glossary and key terms
• What to expect during setup
• Common workloads and use cases
• FAQ

To learn more about lifecycle operations, see Using Hybrid Manager.