Cloud data platform provider Databricks Inc. today announced that it’s raising a very late-stage $10 billion funding round at a $62 billion valuation.

The Series J investment doesn’t come as a surprise. Reports that Databricks is in the process of raising capital first emerged last month, when sources told CNBC that the company was seeking at least $5 billion. Last week, Reuters reported that that the sum could exceed $9.5 billion thanks to higher-than-expected investor interest. 

Thrive Capital, which recently backed a multibillion-dollar funding round for OpenAI, is co-leading Databricks’ raise with Andreessen Horowitz, DST Global, GIC, Insight Partners and WCM Investment Management. The Ontario Teachers’ Pension Plan, ICONIQ Growth, MGX, Sands Capital and Wellington Management are participating as well. The deal ranks as one of the largest startup funding rounds on record.

Enterprises use Databricks’ platform to store and analyze large volumes of data, as well as build artificial intelligence models. At its core, the software is based on a so-called data lakehouse. This is a type of data management system that combines features from two earlier product categories: data lakes and data warehouses. 

Data lakes are optimized to store unstructured and semi-structured information in a cost-efficient manner. A data warehouse, in turn, is a system geared towards storing structured records. Such platforms also support ACID, a reliability standard that helps ensure information isn’t lost in the event of an outage. 

Databricks disclosed today that its annualized revenue run rate is set to top $3 billion at the end of its current quarter. The company’s top-line growth is driven partly by Databricks SQL, the subset of its feature set geared toward creating data warehouse environments. Databricks says the offering’s revenue run rate is up more than 150% from a year ago, to $600 million.

Databricks SQL is one of several products the company has built atop its core data lakehouse over the past few years. AI is another major focus of the software maker’s product roadmap. Databricks’ engineering efforts in this market draw on technology and talent from Mosaic AI Inc., a machine learning startup it bought last year for $1.3 billion.

The company has used the acquisition to launch an AI research unit. In March, the unit released DBRX, an open-source large language model based on a mixture-of-experts architecture. When it receives a query, the model activates only some of its components to generate an answer rather than all of them as is the usual practice, which reduces hardware usage. 

Databricks has also integrated technology from Mosaic into its core platform. In June, the company introduced two AI tools for comparing the performance of LLMs and building RAG applications. More recently, it added an application programming interface for generating synthetic training datasets.

“More and more, the center of gravity is going to shift towards the data and AI you have built rather than the app you’re building around it. The app becomes less important,” Databricks co-founder and Chief Executive Officer Ali Ghodsi (pictured) said in a recent interview on SiliconANGLE Media’s theCUBE. “The particular UI, how things look like won’t matter that much. Even the shape of the data doesn’t matter so much because the gen AI is able to understand it.”

Databricks said today that part of the proceeds from its new funding round will go toward building more AI products. Additionally, the company plans to make acquisitions and significantly grow its international go-to-market operations. Another portion of the capital will be used to provide liquidity for current and former employees.

With another $10 billion on its books, Databricks could take more time to enhance its balance sheet before going public. The company disclosed today that it expects to achieve a positive free cash flow for the first time this quarter. Last month, Ghodsi stated that Databricks plans to go public in the second half of 2025 at the earliest.

Photo: Databricks