For years, the realms of application development and analytics have existed in silos, creating a significant barrier for developers. The reliance on fragile ETL pipelines to transfer data from operational PostgreSQL instances to data lakes has not only hindered efficiency but also resulted in a data tax characterized by duplicated storage and a persistent disconnect between real-time operations and insights.
Today, we are dismantling this barrier with the General Availability (GA) of Azure Databricks Lakebase, a significant milestone announced by Microsoft. Lakebase is a managed, serverless PostgreSQL solution optimized for the Databricks Platform on Azure. This innovative architecture separates compute from storage, enabling direct writing of operational data to lakehouse storage. By bridging the gap between transactional systems and analytics, Azure Databricks Lakebase completes the puzzle for a cohesive data architecture. As a first-party service within the Microsoft ecosystem, it enhances existing Azure investments while significantly boosting developer productivity. Features such as instant branching and zero-copy clones empower teams to iterate on production-grade data without the infrastructure delays that have traditionally impeded innovation.
“Azure Databricks Lakebase gave us one governed foundation for apps, analytics, and AI, so we stopped duplicating data and shipped real‑time features faster.”
— Simon Gilles Fassot, Head of Global Data and Analytics, Hafnia
Why Lakebase? The Database for Modern Apps
Unlike traditional cloud databases that function as isolated entities, Lakebase is seamlessly integrated into the Azure ecosystem. With Lakebase and the lakehouse sharing the same storage layer, the complexities of building and maintaining intricate data pipelines are eliminated, ensuring that data jobs remain in sync. Insights can now be derived from operational database systems without compromising the performance of operational workloads.
Serverless Efficiency with Autoscaling and Scale-to-Zero
Lakebase offers an enterprise-ready PostgreSQL experience, enhanced by the efficiency of a serverless model. The platform automatically scales to accommodate high application traffic and can scale down to zero during idle periods, ensuring that compute resources align with actual demand. This usage-based pricing model guarantees the lowest total cost of ownership (TCO), as users only pay for the compute they utilize while Azure manages the underlying infrastructure and availability.
Developer Agility with Branching and Recovery
In the fast-paced world of modern development, speed and safety are paramount. Lakebase facilitates instant clones and data branching, allowing teams to create zero-copy branches of production data in mere seconds. This capability enables safe, isolated environments for testing schema migrations or debugging queries without affecting live users. Additionally, Lakebase features instant Point-in-Time Recovery (PITR), enabling immediate restoration of the database to a specific moment, thereby facilitating recovery from errors or incidents.
Standard Postgres and Open Ecosystem
Built on standard PostgreSQL, Lakebase ensures full compatibility with existing tools and libraries. It supports numerous popular extensions, including pgvector for AI-driven search and PostGIS for advanced geospatial analysis. By embracing the standard PostgreSQL ecosystem, Lakebase allows developers to harness the latest open-source innovations while Azure manages security, identity, and networking requirements.
Unified Governance via Unity Catalog
Security should be cohesive rather than fragmented across various database engines. With Lakebase, operational data is governed under the same Unity Catalog as analytical and AI workloads. This unified governance model provides consistent access control, automated lineage, and enterprise-grade auditing across the entire Azure Databricks data estate.
“Azure Databricks Lakebase gives enterprise teams a clear path from Lakehouse to relational, governed data without a costly migration. As AI agents start operating directly on investment data, that foundation matters. We’ve already seen what it does to the speed and quality of traditional analysis at Quantum.”
— Ian Brown, Head of Digital Engineering, Quantum Capital Group
Powering AI Agents
The unification of the database and the lakehouse through Lakebase opens up new possibilities for developers crafting the next generation of intelligent software:
- AI agent memory and state: Store agent conversation history and tool logs in a high-performance, governed environment, allowing agents to access real-time operational context with the reliability of a production database.
- Vector-driven context with pgvector: Full support for pgvector enables developers to create RAG (Retrieval-Augmented Generation) workflows that utilize the most current data directly from the operational source.
- Low-latency feature serving: Leverage Lakebase as a high-performance online store for your feature store, ensuring machine learning models have immediate access to fresh features for real-time inference without the complexity of managing separate serving infrastructure.
- Operational analytics with synced tables: Utilizing synced tables ensures that models can be trained and BI dashboards updated using the same data generated by the application in real-time, eliminating manual processes, reducing data duplication, and maintaining synchronization between operational and historical context.
Built on the Enterprise Trust of Azure
Azure Databricks Lakebase allows developers to continue using familiar tools and libraries such as pgAdmin, DBeaver, and the PostgREST API, while Azure manages security, identity, networking, and compliance. By integrating with Microsoft Entra ID and Azure networking protections, Lakebase accelerates application delivery while simplifying the underlying DevOps burden.
“The data platform we’ve built with Azure Databricks Lakebase gives us a treasure trove of usable, enriched data that sets us apart from anyone else in the industry. Lakebase is the intelligent foundation powering our ability to solve problems no one else can.”
— Grant Veazey, CTO, Ensemble
Get Started Today
The General Availability of Azure Databricks Lakebase presents a new foundation tailored for the pace and complexity of modern data systems. It serves as the most straightforward path for Azure customers to develop intelligent, real-time applications directly on their lakehouse infrastructure.
Ready to build?
Azure Databricks Lakebase is integrated into the Azure Databricks experience and can be provisioned directly within your workspaces. Begin your first project today and discover how the dissolution of barriers between applications and analytics can propel your innovation forward.
