Databricks

Databricks sells a platform for data engineering, analytics and machine learning. Seven researchers from UC Berkeley founded it in 2013, including Ali Ghodsi, Matei Zaharia and Ion Stoica. They had created Apache Spark, an open source engine for processing large amounts of data.

The company promotes the lakehouse approach, which combines the cheap storage of a data lake with the management features of a data warehouse. Delta Lake and MLflow, both open source projects started by Databricks, support this work. Unity Catalog manages access and governance across data and models.

In 2023 Databricks bought MosaicML to train language models for customers and released its own open model, DBRX. The platform runs on AWS, Azure and Google Cloud. Databricks is privately held and is based in San Francisco, with large operations in other countries. Ghodsi is chief executive. Databricks has raised money at valuations well into the tens of billions of dollars, and it has bought companies such as Tabular and Neon. The firm has said it is preparing for an eventual public listing.