Databricks, the analytics and AI giant, has acquired data management company Tabular for an undisclosed sum. (CNBC reports that Databricks paid over $1 billion.)
According to Tabular co-founder Ryan Blue, he and Tabular’s other two co-founders, Daniel Weeks and Jason Reid, will be joining Databricks in some capacity. There, they’ll work to unify Tabular’s and Databrick’s customer bases and communities.
“Joining Databricks means that there will be more contributions from our new colleagues,” Blue writes in a blog post. “While doing this, we assure that our approach to [our community] is not changing.”
Tabular, which was founded by Blue, Weeks and Reid in 2021, offers data management products built on Apache Iceberg, a project Blue and Weeks developed while at Netflix and later donated to the Apache Software Foundation. Iceberg is an open source, high-performance format for databases that optimizes tables in databases for big data while at the same time allowing data engines to work with the tables.
Iceberg competed with Databricks’ Delta Lake in the format wars for data lakehouses — data architectures built to store large amounts of raw data while providing structure and management functions. While both Iceberg and Delta Lake use the Apache Parquet data storage format, they’re incompatible in key aspects.
Soon, however, Delta Lake and Iceberg will converge into one. Databricks and Tabular are pledging to work toward a common standard in light of the acquisition news.
“[We will be] working to improve Iceberg support throughout the Databricks platform,” Blue said. “Our goal is to improve interoperability so that you can take advantage of the amazing work of both communities and don’t need to worry about the underlying format.”
The market for data lakehouses is enormous — according to MIT Tech Review, about 74% of organizations have one — and so, from Databricks’ perspective, bringing Tabular into its corporate family was probably the clear choice. Fewer competing data lakehouse formats — or, alternatively, platforms with strong support for multiple formats — makes Databricks’ platform more attractive to corporate clients, after all, even if those formats aren’t vendor-proprietary.
In a blog post co-authored by Databricks CEO Ali Ghodsi and chief architect Reynold Xin, Databricks says that it intends to “work closely” with the Iceberg and Delta Lake communities to “bring interoperability to the formats themselves.”
“This acquisition highlights our commitment to open formats and open source data in the cloud,” the blog post reads. “This is a long journey, one that will likely take several years to achieve in [the data lakehouse] communities.”
Prior to the acquisition, San Jose-based Tabular had raised $37 million in venture capital from investors, including Andreessen Horowitz, Zetta Venture Partners and Altimeter Capital. Databricks says that it expects the purchase to close sometime in Q2 2024, subject to customary closing conditions.
Comment