Databricks

Databricks is an American data and artificial intelligence company founded in 2013 by the creators of Apache Spark at the University of California, Berkeley. Its Data Intelligence Platform combines data engineering, analytics and machine learning in a single lakehouse architecture that unifies data lakes and data warehouses, and the company is among the most highly valued private technology firms in the world, raising funding at a US$188 billion valuation in July 2026.^[1][3]

History

Databricks grew out of the AMPLab research group at UC Berkeley, where students and researchers including Matei Zaharia developed Apache Spark, an open-source engine for large-scale data processing first released in 2010. The company was founded in 2013 by the core team behind Spark, with Ali Ghodsi as chief executive, and initially sold a managed version of Spark in the cloud.^[1]

Over the following decade the company expanded from a Spark vendor into a full data and AI platform. It introduced Delta Lake, an open table format that brought transactional reliability to data lakes, open-sourced the machine learning lifecycle tool MLflow in 2018, and popularised the lakehouse concept that merges lake and warehouse workloads. In June 2023 it acquired MosaicML, a generative AI startup, for approximately US$1.3 billion, followed by further acquisitions including the data management company Tabular and the serverless database firm Neon.^[1][5]

Databricks has raised capital at a rapid pace: a Series J round announced in December 2024 raised US$10 billion at a US$62 billion valuation, led by Thrive Capital, alongside a revenue run-rate of about US$3 billion. A Series K round in 2025 valued the company above US$100 billion, and a Series L round in December 2025 valued it at US$134 billion. In July 2026 the company announced a strategic round at a US$188 billion valuation, and in August 2026 it reported year-on-year revenue growth above 80 per cent, surpassing a US$7 billion annual run-rate with more than 20,000 customers.^[2][1][3][4]

Key Concepts and Technology

The lakehouse is Databricks' core architectural idea: instead of maintaining separate systems for structured warehouse data and unstructured lake data, a lakehouse stores data once in open formats such as Delta Lake and Apache Iceberg, and supports SQL analytics, machine learning, streaming and generative AI workloads over the same governed copy. The Unity Catalog provides fine-grained governance, lineage and access control across data, models and applications.^[7]

The platform's engine lineage traces back to Apache Spark, and its tooling spans the machine learning lifecycle through MLflow, model serving and feature engineering. More recently Databricks has extended into operational and agentic workloads: Lakebase, a serverless PostgreSQL database, stores application state and agent memory; Genie answers natural-language questions over enterprise data; Agent Bricks helps teams build and evaluate AI agents; and the Unity AI Gateway manages access to language models across providers.^[4][7]

Databricks is available on the major public clouds through its own control plane and through partnerships with Amazon Web Services, Microsoft Azure and Google Cloud. In July 2026 Databricks and Microsoft expanded their long-standing partnership to help enterprises connect business context to AI systems built on both companies' platforms.^[4]

Applications and Impact

Databricks is used by enterprises to consolidate fragmented data estates and to move artificial intelligence from pilot projects into production. Its customers include major firms across banking, telecommunications, retail, energy, transport and healthcare, and discussions at the company's user events in 2026 centred on the difficulty organisations face in managing data spread across transactional systems, analytics platforms and streaming pipelines.^[4][7]

The company's rapid valuation growth has made it a bellwether for enterprise AI spending, and analysts frequently describe its platform as the connective infrastructure layer for data-driven AI applications. Its model of combining open formats with commercial governance and management tools has also influenced competitors such as cloud data warehouses and open-source lakehouse projects.^[1][4]

>See Also

• MLflow
• Azure AI
• Amazon SageMaker
• Google Vertex AI
• Data Pipeline
• MLOps