Lakehouse architecture

What is a sovereign AI lakehouse?

A sovereign AI lakehouse is a data lakehouse (open tables plus compute) that runs entirely within a chosen jurisdiction, including its control plane, metadata, logs and AI models. It lets organizations use modern analytics and AI on regulated data without sending that data, or anything derived from it, outside their borders.

Published . Last updated

On this page

What is a lakehouse?

A lakehouse stores data in open table formats such as Apache Iceberg on low-cost object storage, then runs warehouse-grade SQL, Spark and machine learning directly on those tables. It replaced the old pattern of a data lake for raw files plus a separate warehouse for reporting.

What makes a lakehouse sovereign?

Three things.

  1. Everything stays in the jurisdiction. Not only tables, but catalog metadata, query history, lineage, logs, embeddings and AI prompts.
  2. The operator is under the right law. A platform operated by a company subject to foreign law can be compelled to hand over data, wherever it is stored.
  3. You can run it yourself. On-premises, private cloud or air-gapped, when rules require it.

Why does AI make sovereignty harder?

AI multiplies the places data goes. A single question to an agent can touch a prompt log, an embedding store, a model endpoint and a tool call. If any of those sit outside your jurisdiction, data has left. A sovereign AI lakehouse keeps models and agent infrastructure inside the same boundary as the data.

Is a cloud region in my country enough?

Often not. A region gives you residency. It does not change who operates the platform or which laws apply to that company. Many Canadian regulators and procurement teams now ask about both.

How does DataNature implement it?

DataNature runs the full stack inside your deployment: Connect, StreamHouse, Lakehouse, Lakebase, Data Engineering, DataNature AI, BI, Atlas and the Semantic Layer, plus AI360 Studio, DocuContext and Voir. Tables are open Delta or Iceberg. Models can be self-hosted. In Canada, it is operated by a Canadian company on Canadian infrastructure.

Who needs one?

Banks, insurers, telecoms, health systems and governments first. Increasingly, any enterprise whose customers ask where their data goes when it meets AI.

What is inside a sovereign AI lakehouse, technically?

Sovereignty is a property of every layer, so it helps to name them. In DataNature, the stack that runs in your jurisdiction includes:

  • Storage: Apache Iceberg tables on object storage you choose.
  • Compute by type: StreamHouse for the real-time hot path, Spark 4 for batch transformation, Trino for interactive SQL, DN Warehouse for low-latency serving, Lakebase (PostgreSQL) for operational state.
  • Governance: Atlas for catalog, row filters, column masks, lineage and audit.
  • Identity and secrets: Keycloak for single sign-on, HashiCorp Vault for credentials and per-tenant encryption keys.
  • AI: experiment tracking and a model registry, a Milvus vector store, DocuContext for documents, an AI gateway to your approved models, and Voir agents with guardrails enforced in code.
  • Operations: Kubernetes, Terraform and CI/CD, so the environment can be rebuilt anywhere.

The control plane is part of that list, which is the main difference from managed platforms that keep orchestration and metadata in the vendor's cloud. Read the architecture deep dive for each component.

How does DataNature Atlas keep this secure and compliant?

Governance is part of the architecture, not a separate tool bolted on later. DataNature Atlas is the platform's governance module:

  • Knows your data: a catalog with automated PII detection and classification across Bronze, Silver and Gold.
  • Controls access: role-based grants plus attribute-based policies, row filters and column masks (null, redact, partial, hash, year only).
  • Governs AI: Voir agents are scoped principals; masks, row filters and PII guardrails apply to every tool call and answer.
  • Proves it: lineage from source to metric to decision, and an audit log of who or what accessed which data.
  • Stays local: Atlas runs inside your deployment, so policies, lineage and logs never leave your jurisdiction.

These controls support GDPR, PIPEDA, Quebec Law 25, CCPA, HIPAA and sector rules such as OSFI E-23. Read DataNature Atlas in depth or see security and compliance.

See it on your own data.

Book a demo

Demos and quotes are handled by Dlytica, the company behind DataNature.