Databricks: From Open-Source Software to a $62 Billion Startup

Updated 5 November 2026 · Standard formula, shown below

Databricks grew from research at the University of California, Berkeley, into a cloud software company by commercializing tools for large-scale data processing and analytics. Its $62 billion valuation was reported in December 2024, when the company announced a $10 billion funding round; the rise reflects demand for a managed platform built around open-source projects, not a transformation of free software into a single paid product.

From Berkeley research to a company

The story began with Apache Spark, an open-source computing engine developed by researchers and students at UC Berkeley’s AMPLab. Spark was designed to process large datasets across clusters of computers and became known for supporting multiple kinds of data workloads, including batch processing, SQL analytics, streaming, and machine learning. The project was released as open-source software and later became an Apache Software Foundation project.

In 2013, several people involved in Spark’s development—including Ali Ghodsi, Matei Zaharia, and Ion Stoica—co-founded Databricks to make the technology easier for organizations to use. The company’s early proposition was practical: running distributed data systems required infrastructure, configuration, monitoring, and specialized expertise. A managed service could take on much of that operational work while offering a place for teams to collaborate.

Databricks was not the first company to sell software related to big data. Its advantage was its close connection to Spark’s creators and its ability to turn an influential research project into a supported cloud service. The company subsequently broadened its product beyond Spark itself.

What “free software” means here

Calling Databricks itself “free software” blurs two different things. Apache Spark is open-source software that can be downloaded and run without buying a Databricks subscription. Databricks is a commercial company that sells a hosted data and AI platform, support, and related services. Its platform incorporates open-source technologies, but the company’s managed environment and commercial features are not simply the same thing as the free Spark project.

This distinction is common in the technology industry. Open-source code can be freely available under a license, while a company earns revenue by making that code simpler to operate, integrating it with other tools, providing security and governance features, and supporting customers. Users can choose to self-manage open-source components, pay a provider to manage them, or combine the approaches.

Open-source project: Spark is available under the Apache License 2.0.
Commercial platform: Databricks sells a managed environment and enterprise capabilities.
Customer choice: A team can use Spark independently, or use a managed service to reduce infrastructure work.

How the platform expanded

Databricks built a broader platform around data engineering, analytics, and machine learning. Its lakehouse approach aims to bring together characteristics associated with data lakes—flexible, relatively low-cost storage for varied data—and data warehouses, which are designed for structured analytics and governance. The goal is to let organizations use shared data for workloads that might otherwise be split among separate systems.

Delta Lake, an open-source storage framework associated with Databricks, adds features such as transaction support to data stored in data lakes. The company also developed tools for collaborative notebooks and managed data workflows. Over time, its product messaging increasingly included artificial intelligence and generative AI, building on its existing data and machine-learning infrastructure.

These products address a business problem larger than “running Spark.” Companies often store data across many systems and need to prepare it, control access, track its quality, and make it available to analysts or machine-learning applications. A platform that brings more of those tasks into one environment can be valuable, although customers still need to assess integration, governance, performance, and cost for their particular workloads.

Why investors valued it at $62 billion

A startup valuation is not the same as cash in the company’s bank account, annual sales, or a guaranteed sale price. It is a negotiated estimate of a company’s equity value in a financing transaction, influenced by the terms of the deal and expectations about future growth.

In December 2024, Databricks announced that it had raised $10 billion in a Series J financing round at a $62 billion valuation. The company said the funding would support growth, acquisitions, and expansion of its AI products. The figure was a private-market valuation reported for that round—not a public-market share price, and not proof that the company was already worth that amount in a completed sale.

Investors were betting on several related trends: organizations’ growing data volumes, cloud adoption, demand for analytics and machine learning, and the possibility that generative AI would increase the need for well-organized business data. Databricks had also established itself as a prominent vendor in a competitive market. A high valuation reflects expectations as well as current performance, so it can change as growth, competition, and market conditions change.

What customers pay for—and what they should compare

Customers generally pay for cloud usage and the managed services they consume, rather than buying Spark as a proprietary replacement for its open-source release. Actual costs depend on the selected cloud, workload size, runtime, storage, usage patterns, and contract or service configuration. There is no single price that describes every Databricks deployment.

Before choosing a platform, organizations commonly compare:

Operational effort: How much infrastructure setup, maintenance, and troubleshooting can the team handle itself?
Workload fit: Does the service support the team’s data engineering, SQL, and machine-learning needs?
Portability: How easily can data and code move to other tools or environments?
Governance: Can the organization manage permissions, data lineage, and compliance requirements?
Total usage cost: What do compute, storage, data transfer, and ongoing administration add up to?

Alternatives include self-managed Spark and other cloud data warehouses, lakehouse platforms, and data-processing services. The best choice depends on existing infrastructure and skills; a high company valuation does not mean its product is automatically the right fit for every organization.

Frequently Asked Questions

Is Databricks open source?
Databricks is a commercial company and platform. It develops or supports open-source projects, including Delta Lake, while selling managed services and commercial capabilities. Apache Spark itself is open source and can be used separately.

What was Databricks’ $62 billion valuation based on?
It was the valuation announced for the company’s December 2024 Series J financing. A private funding-round valuation is a negotiated estimate tied to that transaction, not a public stock-market price or a guaranteed sale value.

Did Databricks invent Apache Spark?
Spark originated in research at UC Berkeley’s AMPLab. Researchers and students involved in its development later founded Databricks, which helped commercialize and manage Spark-based services.

Can I use Apache Spark without Databricks?
Yes. Spark is available as open-source software under the Apache License 2.0. Users can run it themselves or choose a managed service, depending on their technical resources and requirements.

Why do companies use Databricks instead of running Spark themselves?
A managed platform can reduce the work of setting up and maintaining infrastructure and can provide integrated tools for data workflows and collaboration. Whether that convenience justifies the cost depends on the workload and the organization.

Does a $62 billion valuation mean Databricks is profitable?
No. Valuation does not, by itself, establish profitability. It reflects what investors agreed to in a financing round and their expectations; profitability requires separate financial information.

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.