Databricks completes $10B funding round, raises $5.25B in debt – News

Databricks Inc. has closed the mammoth $10 billion funding round it announced last month.

The company disclosed the milestone today alongside the completion of a new $5.25 billion debt raise. JPMorgan Chase led the latter deal with participation from Barclays, Citi and other financial institutions. 

Rumors of Databricks’ funding round first emerged in November. At the time, the company was expected to raise “at least” $5 billion. Reuters reported in mid-December that the number could top $9.5 billion and, a few days later, Databricks confirmed that it’s raising a $10 billion Series J round.

Lead investor Thrive Capital was joined in the deal by more than a half-dozen other institutional backers. Databricks revealed today that the round also included the participation of the Qatar Investment Authority, Temasek, Macquarie Capital and Meta Platforms Inc. Databricks co-founder and Chief Executive Officer Ali Ghodsi (pictured) told Reuters that the investment from Meta will deepen the companies’ artificial intelligence collaboration, which focuses on the Facebook parent’s Llama series of large language models.

Databricks provides a popular cloud-based platform for storing and analyzing data. It implements a data lakehouse architecture that can hold structured, unstructured and semi-structured records. Companies can run SQL queries on the information they keep in Databricks, use AI models to find useful patterns and perform other analytics tasks.

The platform supports a data reliability standard called ACID. It reverses data modifications that are not completed successfully, which means the erroneous information such modifications often produce is deleted. Additionally, ACID prevents data modifications that are carried out simultaneously from interfering with one another.

In 2023, Databricks acquired a venture-backed AI startup called MosaicML Inc. for $1.3 billion. It has since expanded its feature set with a slew of AI capabilities. On occasion of today’s funding milestone, Ghodsi told Reuters that thousands of customers are running Llama models on the company’s platform.

Databricks provides features that enable enterprises to fine-tune, or customize, AI models using the data they keep in its platform. It also uses an open-source tool called DSPy to automate prompt tuning. This is a machine learning technique that boosts AI models’ output by providing them with instructions on how to process user prompts. 

Databricks has integrated one of Meta’s Llama models directly into its platform. The company provides prepackaged SQL functions, or programs, powered by Llama 3.1 70B. Customers can use those functions to summarize information stored in Databricks, translate it and perform other tasks without having to manually deploy an LLM.

“Organizations are modernizing their data and AI infrastructure because they recognize the immense potential of generative AI,” said Ghodsi. “Data intelligence is critical to both unlocking this potential and to helping enterprises reach their business goals.”

The company will use the proceeds from its $10 billion funding round to build new AI products. Additionally, Databricks plans to make acquisitions and grow its international go-to-market operations. Some of the capital will be used to provide liquidity for current and former employees.

The new funding makes a public offering less urgent for Databricks. Shortly before rumors of its funding round emerged in November, Ghodsi stated that the company plans to list its shares in the second half of 2025 at the earliest. 

Photo: Databricks