Databricks Acquires Electric for AI Agents
Databricks has acquired Electric, a startup developing database technology for distributed applications and AI agents. The deal brings Electric’s PGlite database and real-time synchronization technology into Databricks’ platform.
Electric built PGlite as a lightweight Postgres database that runs inside applications, browser tabs, devices and AI agent sandboxes. As a result, agents can access local data with low latency instead of reaching a remote database. Databricks says the technology extends its Postgres capabilities from the lakehouse to the edge.
Why Electric matters for AI agents
Traditional applications usually follow predictable data access patterns. AI agents, however, decide what information they need while they work. Therefore, they can generate rapidly changing context and require frequent data access.
Moreover, agents often operate inside isolated sandboxes. A network connection to a central database can add latency. Electric’s approach gives each agent a lightweight local Postgres environment while maintaining centralized infrastructure.
Electric also developed a real-time synchronization engine. Consequently, distributed agents can share changing information with a central Lakebase Postgres system. This helps agent teams maintain current context and avoid duplicated work or stale decisions.
Databricks expands its agent infrastructure
The acquisition strengthens Databricks’ database strategy around Lakebase. Lakebase provides managed Postgres infrastructure. Meanwhile, PGlite adds an embedded option for agents that need local state.
The companies also share roots in the Postgres ecosystem. PGlite grew from foundational WebAssembly Postgres work associated with Stas Kelvich, who co-founded Neon. Electric later developed that concept into an embeddable Postgres product.
Notably, Electric’s developer adoption has grown quickly. Databricks says PGlite weekly downloads increased from 1 million to 13 million over the past year. Therefore, the acquisition gives Databricks technology with established developer traction.
The move reflects a wider shift in AI infrastructure. As agentic applications become more capable, databases must handle local context, distributed state and synchronization. Databricks is positioning Postgres across those layers, from Lakebase to lightweight databases running beside AI agents.
