Databricks acquires data optimization startup Tabular in fresh challenge to Snowflake

Ali Ghodsi, co-founder and CEO of Databricks, speaks at the company’s Spark and AI Summit in San Francisco in April 2019.

Databricks

Data analytics software maker Databricks said in a statement on Tuesday that it’s acquiring Tabular, a startup that helps optimize data stored in the cloud.

The gesture could help Databricks more quickly bring out products as it faces competition from Snowflake and other entities.

Databricks is paying over $1 billion to buy Tabular, Databricks CEO Ali Ghodsi said in an interview. The Wall Street Journal reported the size of the deal earlier on Tuesday. Snowflake was also bidding on Tabular, as was Confluent, a person familiar with the matter told CNBC. Snowflake declined to comment. Confluent did not respond to a request for comment.

Billions in venture capital have helped Databricks finance this and other recent deals, including the acquisitions of database replication startup Arcion and artificial intelligence efficiency startup MosaicML for $100 million and $1.3 billion, respectively. In September, Databricks said it had raised new funding at a $43 billion valuation, making it worth more than most startups and some publicly held enterprise software sellers.

The costs of running queries before exploring and making charts with data can add up. Developers built an open-source format called Apache Iceberg that stores data in tables that a variety of tools can then work with.

Tabular co-founders Ryan Blue and Dan Weeks came up with Iceberg while at Netflix. Tabular adds business-friendly features to Iceberg and keeps the tables in the Amazon or Google clouds. From there, organizations can connect tables to Snowflake and other systems, enabling less expensive queries.

Last week, software stocks tumbled as executives from Salesforce, MongoDB and Okta issued fresh warnings to investors about economic turbulence. Databricks is behaving differently. It’s growing faster and using its capital to take market share. Databricks told media outlets in March that it generated $1.6 billion in revenue in the year that ended on Jan. 31, up more than 50%.

Snowflake executives have said some large clients want to move data out of the company’s native storage layer and into Iceberg tables that live in a separate place, such as object storage in the Amazon Web Services cloud. That might result in lower storage revenue for Snowflake. But as organizations become comfortable with running queries on Iceberg tables, that could spur revenue growth from computing workloads running on a large quantity of data stored elsewhere.

“Net-net, we think it’s going to be a positive,” Mike Scarpelli, Snowflake’s finance chief, said at a Morgan Stanley event in March.

At its Summit conference in San Francisco on Monday, Snowflake announced that within 90 days it would release open-source catalog software for finding Iceberg tables. Tabular was among the exhibiting partners at the conference.

Databricks has promoted its own open-source initiative called Delta Lake. Focusing more on Iceberg tables might allow Databricks to take business from Snowflake clients that embrace the format. For now, the goal is to provide “full interoperability” between the Delta Lake and Iceberg projects, Ghodsi said.

Tabular was established in 2021 and has raised over $30 million in funding, including from Altimeter Capital, Andreessen Horowitz and Zetta Venture Partners.

Tabular has hundreds of customers, and it will work with Databricks to figure out what to do with the product, Ghodsi said.

WATCH: Part of Snowflake’s downfall is related to CEO Frank Slootman’s retirement: Jefferies’ Brent Thill