Databricks

Databricks, Inc. is a global data, analytics and artificial intelligence company founded by the original creators of Apache Spark.

The company provides a cloud-based platform to help enterprises build, scale, and govern data and AI, including generative AI and other machine learning models.

Databricks pioneered the data lakehouse, a data and AI platform that combines the capabilities of a data warehouse with a data lake, allowing organizations to manage and use both structured and unstructured data for traditional business analytics and AI workloads.

Databricks acquired MosaicML for $1.4 billion in June 2023, its largest acquisition.

In November 2023, Databricks unveiled the Databricks Data Intelligence Platform, a new offering that combines the unification benefits of the lakehouse with MosaicML’s Generative AI technology to enable customers to better understand and use their own proprietary data.

The company develops Delta Lake, an open-source project to bring reliability to data lakes for machine learning and other data science use cases.

History
Databricks grew out of the AMPLab project at University of California, Berkeley that was involved in making Apache Spark, an open-source distributed computing framework built atop Scala. The company was founded by Ali Ghodsi, Andy Konwinski, Arsalan Tavakoli-Shiraji, Ion Stoica, Matei Zaharia, Patrick Wendell, and Reynold Xin.

In November 2017, the company was announced as a first-party service on Microsoft Azure via integration Azure Databricks.

In June 2020, Databricks acquired Redash, an open-source tool designed to help data scientists and analysts visualize and build interactive dashboards of their data.

In February 2021 together with Google Cloud, Databricks provided integration with the Google Kubernetes Engine and Google's BigQuery platform. Fortune ranked Databricks as one of the best large "Workplaces for Millennials" in 2021. At the time, the company said more than 5,000 organizations used its products.

In August 2021, Databricks finished its eighth round of funding by raising $1.6 billion and valuing the company at $38 billion.

In October 2021, Databricks made its second acquisition of German no-code company 8080 Labs. 8080 Labs makes bamboolib, a data exploration tool that does not require coding to use.

In response to the popularity of OpenAI's ChatGPT, in March 2023, the company introduced an open-source language model, named Dolly after Dolly the sheep, that developers could use to create their own chatbots. Their model uses fewer parameters to produce similar results as ChatGPT, but Databricks had not released formal benchmark tests to show whether its bot actually matched the performance of ChatGPT.

Databricks acquired data security startup Okera in May 2023 to extend its data governance capabilities. The next month, it acquired an open-source generative AI startup MosaicML for $1.4billion.

In October 2023, Databricks acquired data replication startup Arcion for $100 million.

Databricks reported $1.6 billion in revenue for the 2023 fiscal year, more than doubling its previous level.

Funding
In September 2013, Databricks announced it raised $13.9 million from Andreessen Horowitz and said it aimed to offer an alternative to Google's MapReduce system. Microsoft was a noted investor of Databricks in 2019, participating in the company's Series E at an unspecified amount. The company has raised $1.9 billion in funding, including a $1 billion Series G led by Franklin Templeton at a $28 billion post-money valuation in February 2021. Other investors include Amazon Web Services, CapitalG (a growth equity firm under Alphabet Inc.) and Salesforce Ventures.

Products
Databricks develops and sells a cloud data platform using the marketing term "lakehouse", a portmanteau based on the terms "data warehouse" and "data lake". Databricks' lakehouse is based on the open source Apache Spark framework that allows analytical queries against semi-structured data without a traditional database schema. In October 2022, Lakehouse received FedRAMP authorized status for use with the U.S. federal government and contractors.

Databricks' Delta Engine launched in June 2020 as a new query engine that layers on top of Delta Lake to boost query performance. It is compatible with Apache Spark and MLflow, which are also open source projects Databricks employees helped create.

In November 2020, Databricks introduced Databricks SQL (previously known as SQL Analytics) for running business intelligence and analytics reporting on top of data lakes. Analysts can query data sets directly with standard SQL or use product connectors to integrate directly with business intelligence tools like Holistics, Tableau, Qlik, SigmaComputing, Looker, and ThoughtSpot.

Databricks offers a platform for other workloads, including machine learning, data storage and processing, streaming analytics, and business intelligence.

The company has also created Delta Lake, MLflow and Koalas, open source projects that span data engineering, data science and machine learning. In addition to building the Databricks platform, the company has co-organized massive open online courses about Spark and a conference for the Spark community called the Data + AI Summit, formerly known as Spark Summit.

In early 2024, Databricks released a portfolio of new tools to help customers customize, fine-tune or build their own AI systems, including: Mosaic AI Vector Search, which enables companies to build RAG models, Mosaic AI Model Serving, a unified service for deploying, governing, querying and monitoring models fine-tuned or pre-deployed by Databricks, and Mosaic AI Pretraining, a platform for enterprises to create their own LLMs.

In March 2024, Databricks released DBRX, an open source foundation model. It relies on a mixture-of-experts architecture and is built on the MegaBlocks open source project.

DBRX cost $10 million to create. At the time of launch, it was the fastest open source LLM, based on commonly-used industry benchmarks. It beat other models like LlaMA2 at solving logic puzzles and answering general knowledge questions, among other tasks. And while it’s a 136 billion parameters model, it uses only an average of 36 billion to generate outputs.

DBRX also serves as a foundation for companies to build or customize their own AI models. Companies can also use proprietary data to generate higher-quality outputs for specific use cases.

Operations
Databricks is headquartered in San Francisco. It also has operations in Canada, the United Kingdom, Netherlands, Singapore, Australia, Germany, France, Japan, China, South Korea, India, Brazil, Switzerland, Costa Rica and Serbia.