Databricks is in the process of closing a fresh round at a $100 billion valuation, sources confirmed to TechCrunch. The round was originally reported by the Wall Street Journal.

A source familiar with the deal tells TechCrunch exclusively the new round is about $1 billion, and was wildly oversubscribed. Databricks, best known for its data analytics products, refrained from selling even more equity because it didn’t need cash for operations after its once record-breaking $10 billion raise at a $62B valuation in January, according to the source. (OpenAI has since squashed the record with a $40 billion raise in March.)

The round was co-led by both Thrive and one of Databrick’s early investors, Insight Partners, TechCrunch has learned. These two firms led the last round, as well. The company has now raised about $20 billion since it was founded in 2013.

This was a primary round, meaning it didn’t include employees selling their shares. However, sources close to the company say Databricks has already had two secondary rounds for employees in 2025. Those offers allowed employees to sell up to 40%, 50%, or up to 60% of their shares, depending on the size of their holdings. 

In both cases, the source said, the full funds available for the secondary round were not maxed out, meaning employees held onto more shares than they could have sold. While Databricks clearly isn’t in a hurry to IPO, employees have had two recent chances to cash out shares. 

This new round, however, was raised to pursue two specific projects — a database for AI agents and its AI agent platform — Databricks co-founder and CEO Ali Ghodsi told TechCrunch in an interview. 

The company will invest heavily in its database for AI agents, making it generally available to all customers. It launched the product, known as Lakebase, in June at its annual tech conference. Lakebase, which is based on the open source database Postres, is enterprise-grade and supports corporate developers’ vibe coding projects. This makes it a competitor to Supabase. 

“The database market is $105 billion of TAM, of revenue, sitting there, kind of unaffected in the last 40 years,” Ghodsi told TechCrunch, giving a subtle nod to how database giant Oracle has had a lock on the market for decades. TAM refers to the total addressable market.

“Here’s the interesting statistic nobody’s paying attention to: a year ago, we saw in the data that 30% of the databases were not created by humans. For the first time, they were created by AI agents. And this year, the statistic is 80%,” he said, adding that he predicts this stat to increase to 99% of new databases within a year. 

“There’s a new user. The user is not human. It’s an AI agent, and if we just double down on making that user persona successful, that’s the wedge to disrupt that TAM,” he said.

As for how Lakebase will differentiate from Supabase and others already building Postgres-based databases for agents, Ghodsi said the key is “separated compute and storage.” 

By untying the pricey compute from the lower-cost storage, Databricks can affordably let users create many databases. “Because these agents are super fast. They just spin up lots of databases, much faster than humans can, but you don’t want to go bankrupt because you’re doing that,” he explained.

The second project Databricks will be investing heavily in is AI agent platform Agent Bricks, also launched in June. “Everybody’s super focused on super intelligence,” Ghodsi said. “But that’s not what we need in organizations.”

Rather than artificial general math geniuses or cancer-curing scientists, what companies need are agents that can reliably handle, unaided, mundane tasks like onboarding employees or answering personalized questions about HR benefits. 

“I think that’s a much bigger opportunity, actually, for the worldwide GDP and for organizations,” he said. He believes that such focus will give Agent Bricks a competitive advantage. 

He also raised the extra cash so Databricks can get into the AI poaching wars. “As you know, it’s pretty expensive to hire AI talent right now,” he smiled.