Databricks Ceo Says Fresh $1b Will Help Him Attack A New Ai Database Market

Databricks is successful nan process of closing a caller information astatine a $100 cardinal valuation, sources confirmed to TechCrunch. The information was primitively reported by nan Wall Street Journal.

A root acquainted pinch nan woody tells TechCrunch exclusively nan caller information is astir $1 billion, and was wildly oversubscribed. Databricks, champion known for its information analytics products, refrained from trading moreover much equity because it didn’t request rate for operations aft its erstwhile record-breaking $10 cardinal raise astatine a $62B valuation successful January, according to nan source. (OpenAI has since squashed nan grounds with a $40 cardinal raise successful March.)

The information was co-led by some Thrive and 1 of Databrick’s early investors, Insight Partners, TechCrunch has learned. These 2 firms led nan past round, arsenic well. The institution has now raised astir $20 cardinal since it was founded successful 2013.

This was a superior round, meaning it didn’t see labor trading their shares. However, sources adjacent to nan institution opportunity Databricks has already had 2 secondary rounds for labor successful 2025. Those offers allowed labor to waste up to 40%, 50%, aliases up to 60% of their shares, depending connected nan size of their holdings. 

In some cases, nan root said, nan afloat costs disposable for nan secondary information were not maxed out, meaning labor held onto much shares than they could person sold. While Databricks intelligibly isn’t successful a hurry to IPO, labor person had 2 caller chances to rate retired shares. 

This caller round, however, was raised to prosecute 2 circumstantial projects — a database for AI agents and its AI supplier level — Databricks co-founder and CEO Ali Ghodsi told TechCrunch successful an interview. 

The institution will put heavy successful its database for AI agents, making it mostly disposable to each customers. It launched nan product, known arsenic Lakebase, in June astatine its yearly tech conference. Lakebase, which is based connected nan unfastened root database Postres, is enterprise-grade and supports firm developers’ vibe coding projects. This makes it a competitor to Supabase. 

“The database marketplace is $105 cardinal of TAM, of revenue, sitting there, benignant of unaffected successful nan past 40 years,” Ghodsi told TechCrunch, giving a subtle motion to really database elephantine Oracle has had a fastener connected nan marketplace for decades. TAM refers to nan full addressable market.

“Here’s nan absorbing statistic nobody’s paying attraction to: a twelvemonth ago, we saw successful nan information that 30% of nan databases were not created by humans. For nan first time, they were created by AI agents. And this year, nan statistic is 80%,” he said, adding that he predicts this stat to summation to 99% of caller databases wrong a year. 

“There’s a caller user. The personification is not human. It’s an AI agent, and if we conscionable double down connected making that personification persona successful, that’s nan wedge to disrupt that TAM,” he said.

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

By untying nan pricey compute from nan lower-cost storage, Databricks tin affordably fto users create galore databases. “Because these agents are ace fast. They conscionable rotation up tons of databases, overmuch faster than humans can, but you don’t want to spell bankrupt because you’re doing that,” he explained.

The 2nd task Databricks will beryllium investing heavy successful is AI supplier level Agent Bricks, besides launched successful June. “Everybody’s ace focused connected ace intelligence,” Ghodsi said. “But that’s not what we request successful organizations.”

Rather than artificial wide mathematics geniuses aliases cancer-curing scientists, what companies request are agents that tin reliably handle, unaided, mundane tasks for illustration onboarding labor aliases answering personalized questions astir HR benefits. 

“I deliberation that’s a overmuch bigger opportunity, actually, for nan worldwide GDP and for organizations,” he said. He believes that specified attraction will springiness Agent Bricks a competitory advantage. 

He besides raised nan other rate truthful Databricks tin get into the AI poaching wars. “As you know, it’s beautiful costly to prosecute AI talent correct now,” he smiled.