Databricks Rolls Out General Availability of Zerobus Ingest to Simplify Real-Time Data Streaming

Zerobus Ingest removes the need for brokers, partitions, consumer group monitoring, and maintenance, enabling teams to focus on deriving insights faster

Databricks Rolls Out General Availability of Zerobus Ingest to Simplify Real-Time Data Streaming

Databricks has officially launched Zerobus Ingest, a fully managed, serverless data ingestion service now generally available as part of its Lakeflow Connect portfolio.

The offering is designed to help organisations stream high-volume event data directly into Delta Lakehouse tables, bypassing traditional message brokers and complex infrastructure.

Zerobus Ingest enables near real-time ingestion with sub-five-second latency and supports more than 10 GB per second of aggregate throughput to a single table, with up to 100 MB per second per connection — all without users needing to manage brokers like Apache Kafka or other intermediary layers.

"With Zerobus you can replace this whole pipeline. Just hit an API with individual data items you want to add and they'll appear in Parquet, Iceberg, Delta Lake on your Lakehouse in seconds! It can handle the 3 Vs (Volume, Variety, Velocity) of the data just like any bus, without the hassle of hosting, managing, operating a message bus." Ali Ghodsi, CEO at Databricks, said.

By shifting to a “single-sink” architecture, the service aims to cut infrastructure overhead, reduce operational complexity and eliminate the expertise and cost traditionally required for multi-component streaming stacks. Developers can integrate via native gRPC APIs, REST (in beta), or language-specific SDKs for Python, Java, Rust, Go and TypeScript, allowing applications to write directly into governed Delta tables managed by Unity Catalog.

Databricks says Zerobus Ingest removes the need for brokers, partitions, consumer group monitoring, and maintenance, enabling teams to focus on deriving insights faster. Use cases cited include telemetry streaming, IoT data, clickstream analytics, cybersecurity event feeds and other workloads where rapid data availability matters.

The GA release also introduces enterprise-ready support for production use, and Zerobus Ingest is available now on AWS and Azure, with Google Cloud support expected soon.

Earlier this month, Databricks said it has secured more than $7 billion in new investments as it reported strong growth and outlined plans to expand its AI-focused products. The San Francisco-based firm said it has crossed a $5.4 billion revenue run-rate, posting more than 65% year-over-year growth in the fourth quarter.