Understand any stock in simple English
Enter a ticker to learn what the company does and how it makes money.
Databricks sells cloud software that helps companies store, clean, and analyze huge amounts of data, and increasingly build and run AI on top of it.
Databricks is a software company that helps businesses make sense of enormous piles of data. Its customers are mostly big organizations — banks, retailers, manufacturers, media companies — that collect data from everywhere and need somewhere to put it, a way to clean it up, and tools to ask questions of it. The company grew out of a research project at the University of California, Berkeley, and its founders helped create an open-source technology called Apache Spark, which became a standard way to process giant datasets. That heritage matters: Databricks built its business on top of open tools that customers can adopt without feeling locked in, then wrapped them in a polished, managed service.
The core product is a "data lakehouse," a mouthful of a term that really means one place to hold both messy raw data and neatly organized tables, so a company doesn't have to run two separate systems and copy data between them. On top of that sits a workspace where data engineers, analysts, and data scientists collaborate, plus tools for machine learning. More recently, Databricks has pushed hard into AI: it sells services for building and running large language models, and in 2023 it bought MosaicML, a company that helps businesses train and deploy their own AI models, for about $1.3 billion. It also acquired Tabular, a data-management startup, in 2024. The pitch is that a company's data already lives in Databricks, so that's the natural place to build AI on it — rather than shipping sensitive data out to someone else. Who pays? Businesses subscribe, typically on multi-year contracts, and the bill scales with how much computing and storage they use. The company has grown fast — annualized revenue crossed roughly $3 billion in early 2025, up sharply from the year before — but it is not yet profitable on a net basis, because it spends heavily on sales and engineering to chase a huge market. Its main rival is Snowflake, which sells a competing data platform, alongside the cloud giants Amazon, Microsoft, and Alphabet, whose own data tools are bundled with their cloud services. Databricks is private, having raised billions from investors including Microsoft, and it has been widely expected to go public eventually.
In a university lab at Berkeley, a researcher named Matei Zaharia got fed up with how slow the existing tools for crunching huge amounts of data were, so he built a faster one called Apache Spark. It worked so well that people everywhere started using it — but it was free and community-run, which meant no company to call when things broke and a painful setup. So Zaharia and a group of colleagues, including Ali Ghodsi and Ion Stoica, started Databricks to give Spark the business backbone it was missing.
Databricks raised money and launched its first product, a service that made it far easier to process big data in the cloud instead of on messy machines you had to run yourself.
Microsoft built Databricks right into its Azure cloud as a first-party service, which meant a giant tech company was now selling Databricks as part of its own lineup. That put the startup in front of many more businesses.
Investors piled in — a billion dollars at a $28 billion valuation, then more soon after at $38 billion — and some of the biggest cloud companies on earth bought in despite being rivals. By then Databricks had thousands of customers and worked across multiple clouds, no longer tied to one.
When ChatGPT made the world care about AI that writes and talks, Databricks bought a startup called MosaicML for $1.4 billion to help companies build their own ChatGPT-style tools on their private data. It also released a free open model called Dolly, showing it could play the AI game too.
Databricks raised another $10 billion at a $62 billion valuation, a stunning number for a company not traded on any stock exchange. Behind it was real business: over a billion and a half in revenue with growth still above 80%.
Databricks launched Agent Bricks for building AI helpers and Lakebase for storing the data those helpers need. Then it raised money three times in under a year, each round bigger than the last, ending at a $134 billion valuation in December. It was becoming one of the most valuable private companies anywhere.
Another massive round pushed Databricks to a $190 billion valuation as annualized revenue jumped from $5.4 billion to $7.0 billion. CEO Ali Ghodsi said going public in 2026 would be a disaster and that 2027 was the earliest chance. Not everyone believed the price: analysts estimated the company was really worth closer to $69 billion. The hype and the doubt now sat side by side.
What makes Databricks strong is a mix of switching costs and mindshare. Once a company's data and workflows live inside its platform, ripping it out is painful, and its open-source roots give customers comfort that they aren't trapped. The risk is that the cloud providers it depends on — it runs mostly on their infrastructure — are also its competitors, and that AI spending could shift toward rivals. For now, Databricks is one of the most valuable private software companies in the world, valued around $62 billion in a 2024 funding round, and it's betting that the next decade of enterprise AI gets built on top of the data companies already trust it to hold.
Saved answer · last researched 2 Nov 2026