Designed as a fully managed operational database, Lakebase is built on Postgres and tightly integrated with Databricks’ lakehouse architecture. It represents a move to eliminate the long-standing divide between transactional systems and analytical environments—an increasingly important shift as intelligent agents and real-time data applications become more prevalent.
Traditional operational databases were not designed with AI in mind. They are often rigid, difficult to scale dynamically, and disconnected from the analytical systems where machine learning models are trained and evaluated. Lakebase addresses these limitations by introducing a serverless, cloud-native architecture that separates compute and storage. This separation enables high concurrency and low latency, while autoscaling compute resources ensure responsiveness to changing workload demands. The result is an operational layer capable of supporting modern, AI-enhanced applications on the same platform where data is analysed and models are developed.
Postgres foundation with AI-oriented enhancements
Lakebase is based on the widely used Postgres engine, providing developers with a familiar SQL environment and access to a broad ecosystem of extensions and tools. The use of Postgres supports flexibility and ease of adoption, particularly for teams already working with open source technologies. Beyond the core database functionality, Lakebase includes features designed specifically for AI development workflows. One of the most notable is branching, a copy-on-write system that enables users to instantly create database clones for testing, staging, or recovery. These branches do not duplicate data and can be created at specific points in time, making them well suited for isolated development or audit scenarios.
The platform also supports near-instant startup and usage-based billing through serverless infrastructure, allowing developers to create and discard database environments as needed. This supports a more agile approach to development, especially in projects where testing with production-like data is required. Lakebase environments can be spun up in seconds, scaled automatically, and decommissioned without manual intervention.
Integration with the lakehouse plays a central role in the utility of Lakebase. Data can be synchronised automatically between Lakebase and Unity Catalog-managed tables, making it possible to feed operational data into analytical processes or serve machine learning features and model predictions directly from Lakebase. This capability reduces the need for custom pipelines or data duplication and ensures that operational and analytical layers remain aligned in real time.
Operational reliability and streamlined innovation
Lakebase offers high availability features such as multi-zone replication and point-in-time recovery. All data is encrypted and stored with regional durability, protecting against data loss or service disruption. Readable secondaries can be configured to distribute workloads across zones, improving performance and resilience. These capabilities are managed entirely by Databricks, reducing the operational burden on development teams and enabling them to focus on application logic rather than infrastructure.
Developers also benefit from integrated monitoring tools that provide visibility into key database metrics, including throughput, open connections, and resource utilisation. Security is managed through the same enterprise-grade frameworks used across the Databricks platform, including support for PrivateLink, IP access controls, and consistent identity management via OAuth and Unity Catalog integration.
Lakebase offers a multi-cloud solution without requiring replatforming or architecture changes. By providing a unified, developer-friendly stack that reduces complexity and accelerates innovation, Databricks is helping organisations maximise the value they derive across their entire data estate. The combination of real-time operational capabilities and seamless analytical integration positions Lakebase as a key enabler of intelligent, scalable applications in the AI era.
