What better way to ensure you have a growing pipeline of customers than to invest them? Databricks, a San Francisco decacorn, recently announced a new investment division, Databricks Ventures, to support other startups building on its ecosystem.
The 8-year-old San Francisco company, last valued at $38 billion, hasn’t raised a specific amount for the fund but will draw from its own balance sheet when it decides to invest. It won’t lead any funding rounds, either, but will participate when it makes strategic sense to do so, Andrew Ferguson, who oversees the new venture arm, told me.
"It's about finding founding teams and companies that share our vision of an open data AI, machine learning ecosystem,” he said. “And if that vision is aligned with Databricks and our customer base, that's really where we can partner with them to help accelerate their business, and then it's also good for Databricks as well as being a good financial investment. That's really the win-win."
Databricks hasn’t publicly announced any of its investments yet but says it has participated in several rounds. The focus is startups building on Databricks' products — specifically, data lakehouses.
Lakehouses are a form of data management technology that cut down on redundant and inefficient storage. Essentially, a data lakehouse takes organized and unorganized information and creates a centralized repository that can be accessed and processed more efficiently across multiple teams within an organization. And a core tenet of the lakehouse is also enabling open-source technology.
This differs from traditional data warehouses that store organized, or structured, information in many different places which can lead to multiple versions and slow down processing speeds. In contrast, large repositories of unstructured data have traditionally been stored in so-called data lakes.
Google's cloud division explains data lakehouses as a new form of data storage "which combines key benefits of data lakes and data warehouses" that "offers low-cost storage in an open format accessible by a variety of processing engines."
The demand for this type of data management has grown with the acceleration of cloud storage and machine learning.
"It's about moving towards simplicity, getting all of that architectural complexity out of our environment, moving towards open so that the data is always in your control and you're always able to take advantage of what the latest innovations are out there in the data and AI space, and it's about being able to become a much more collaborative organization," Joel Minnick, Databricks' vice president of product marketing, told me.
Databrick's competition in this space includes Snowflake, which went public in 2020 and moved from San Mateo to Montana earlier this year. All the major cloud storage providers in the U.S. — including Google Cloud, AWS and Microsoft — offer lakehouse storage and partner with Databricks to implement it.
Gartner released two reports earlier this month assessing the market leaders in cloud data management systems as well as data science and machine learning platforms. Databricks was categorized as a "leader" (the top level) in each category alongside larger competitors including AWS, Google, Microsoft, Oracle, IBM and Snowflake.
The company’s investors include Andreessen Horowitz, Tiger Global Management, Salesforce Ventures, AWS, Blackrock, Microsoft, Coatue and DCVC, according to Crunchbase. It has raised $2.6 billion this year over two rounds through its Series H.
As to the question of if and when Databricks might go public itself, the company is staying tight-lipped about its plans for now.
