Lakebase Postgres Doubles Databricks SQL

This growth rate shows Databricks has found a faster path into the application database layer than it found in analytics warehousing. Databricks SQL had to win analyst workloads from Snowflake and legacy BI setups, which meant long enterprise sales and careful migration. Lakebase Postgres starts closer to developers and AI apps, where teams can spin up a serverless Postgres database in minutes and usage ramps with every new agent, app, and test environment.1 2 3

• Databricks SQL hit $100M ARR in April 2023, roughly one year after launch, and $400M ARR a year later. By late 2025, Databricks SQL was already a $1B run rate product, but it still represented only one expansion line inside a much larger data platform business.1 2 4
• Lakebase comes from the May 2025 Neon acquisition, which gave Databricks a serverless Postgres engine built for fast database creation and branching. That fits AI agent workflows especially well, because agents create many short lived databases for experiments, previews, and production backends, which directly drives usage based revenue.2 3
• The bigger implication is product mix. Databricks was built around Spark jobs, notebooks, and ML pipelines, but newer lines like warehousing, AI, and now Postgres let it charge not just for analyzing data, but for the live application database underneath software built on the platform.1 2 4

From here, the center of gravity keeps moving up the stack. If Lakebase continues compounding faster than SQL did, Databricks becomes less of a place to process data after the fact, and more of a full runtime for building and operating AI software, with database, model, and app revenue bundled together.1 2 3