Databricks to Acquire Neon, Targeting Serverless Postgres for AI Agent Workloads

Databricks has announced plans to acquire Neon, a serverless Postgres provider purpose-built for developer workflows increasingly driven by automation and AI agents. The acquisition reflects how the role of databases is changing as more workloads are initiated by AI, not people.

Neon was designed around that shift. Internal telemetry shows that over 80% of the databases created on its platform are provisioned automatically by AI agents. These workloads operate differently than human-driven use cases. They move faster, scale unpredictably, and require infrastructure that doesn’t get in the way. Neon’s ability to spin up fully isolated Postgres instances in under 500 milliseconds, fork schema and data instantly, and separate compute from storage makes it a natural fit for machine-generated operations.

Cost efficiency is another factor. With traditional databases, costs often scale linearly with time or infrastructure, regardless of actual usage. For AI agents generating thousands of short-lived queries and test environments, that model breaks down quickly. Neon’s architecture is designed to be usage-based, letting organizations pay for the queries that run—not the idle time in between.

Neon is also fully Postgres-compatible and supports common extensions out of the box. That means developers don’t have to rework existing tools or retrain teams to take advantage of it. This openness aligns well with Databricks’ own approach of supporting open standards and giving teams flexibility across data and AI pipelines.

Integrating Neon into the Databricks Data Intelligence Platform will remove friction for developers building with AI agents at scale. Instead of dealing with infrastructure that slows them down or doesn’t scale to automated demand, teams will be able to provision, test, and deploy using a native serverless database built to keep up with real-time agent workflows.

Once the acquisition closes, Neon’s team is expected to join Databricks, where they’ll continue supporting Neon’s user base while accelerating the development of tools tailored to AI-native systems.