In 2013, seven researchers from the University of California, Berkeley did something that sounded like a bad business plan. They took Apache Spark, the fast distributed computing engine they had built, and kept it free and open. Then they started a company.
Thirteen years later, that company - Databricks - is valued at $134 billion, employs roughly 10,000 people, and runs at an annualized revenue run rate of about $6.9 billion. It is one of the most valuable private technology companies in the world, and it got there by selling something adjacent to the code it gave away: a managed, governed, cloud-native platform that makes big data and AI usable for ordinary enterprises.
Databricks describes what it sells as the Data Intelligence Platform, built on an idea the company itself named: the “lakehouse.” The pitch is straightforward. For years, organizations kept two separate systems - cheap, flexible data lakes for raw storage, and expensive, reliable data warehouses for analytics. Databricks argued you could have both in one place. Store everything in open formats, add a governance and transaction layer on top, and run analytics, business intelligence, and machine learning against the same data.
That single architectural bet is the through-line of the company's history. It explains the acquisitions, the open-source strategy, the rivalry with Snowflake, and the pivot into generative AI. And it explains why so many enterprises - across banking, healthcare, retail, and manufacturing - now treat Databricks as default infrastructure rather than an experiment.
What makes the story unusual is who is telling it. Databricks is still run by its academic founders. CEO Ali Ghodsi is an adjunct professor at Berkeley. CTO Matei Zaharia created Spark as part of his PhD. The research DNA is not marketing - it shows up in the products.
