During the Data + AI Summit, Databricks launched Lakebase, a fully managed Postgres database designed for AI.
Lakebase is designed for AI workloads and adds an operational database layer to Databricks Data Intelligence. This allows developers to build AI applications and agents on an affordable, flexible, and cloud-native platform.
Database for AI Applications
“We think this will be a new architecture for databases,” says Databricks CEO Ali Ghodsi during the keynote. Lakebase, based on Neon technology, splits storage and computing power into affordable and open-source lakes in users’ own accounts.
Lakebase runs on Postgres and supports high availability, low latency of less than 10 ms, and more than 10,000 queries per second. Data is automatically synchronized with lakehouse tables. The database is fully managed, supports encryption, point-in-time recovery, and advanced security features.
Practical Applications
According to Databricks, hundreds of companies are already using Lakebase. For example, Heineken uses the technology to gain faster insights into operational processes, such as personalized recommendations and supply chain optimizations.
The public preview is available from today for all Databricks users. Further improvements will be rolled out in the coming months.
read also