A Fundraise Enlarged by Demand

Databricks originally planned to raise $1 billion, but overwhelming investor demand pushed the AI data company to close a $5 billion round at a $190 billion valuation.

CEO and co-founder Ali Ghodsi told TechCrunch that the process was accelerated after The Information reported on a potential large fundraise during Databricks’ June conference. The company, he said, was focused on the event rather than fundraising, but investor calls quickly surged. From just a selected group of investors Databricks considered, Ghodsi said there was $15 billion of interest.

That created a familiar late-stage startup dilemma: decline too many long-time backers and risk damaging relationships, or sell more shares than initially intended. Databricks chose the latter. In July, it announced a new round at a $188 billion valuation without disclosing the amount; it later confirmed the raise totaled $5 billion and the valuation had moved to $190 billion.

Why Investors Wanted In

The round was led by Coatue, with participation from Blackstone, MGX, accounts associated with T. Rowe Price, new investor Sixth Street Growth, and roughly two dozen named investors. Sixth Street was founded by former Goldman Sachs chief investment officer Alan Waxman.

The investor logic is straightforward: Databricks sits at the intersection of enterprise data infrastructure and AI adoption. Ghodsi said the company has reached $7 billion in annualized run-rate revenue, is growing at 80%, and is cash-flow positive. Annualized run-rate revenue is a common software metric that estimates yearly revenue based on current revenue levels.

Key figures disclosed include:

• $7 billion annualized run-rate revenue, growing 80%;
• $1.5 billion run-rate from its core cloud data warehouse, growing 100% year over year;
• $100 million run-rate for Lakebase, its database for agents;
• Strong adoption for Genie, its AI chatbot tool for business analysis.

A cloud data warehouse is a cloud-based system for storing and analyzing large volumes of enterprise data. For companies trying to deploy AI, the data layer matters because models and AI agents need clean, accessible, and governed data to produce useful results.

Why Raise More When the Business Is Working?

Ghodsi’s answer was blunt: AI is expensive.

Databricks has multi-billion-dollar cloud commitments with all three major hyperscalers. A hyperscaler is a large cloud provider that operates global computing infrastructure at massive scale. AI workloads depend heavily on that infrastructure for compute, storage, and data processing.

Research is another cost center. Ghodsi said Databricks has a 100-person AI research team, an expensive and highly competitive area. The company is also active in M&A. This week it announced the acquisition of Electric, maker of PGlite, a lightweight Postgres database designed to help agents spin up databases; terms were not disclosed. In June, Databricks bought AI cybersecurity company Panther, and in March it acquired two startups.

Those moves suggest Databricks is using capital not simply as a cushion, but to broaden its AI data stack: cloud infrastructure commitments at the base, database and security capabilities in the middle, and AI tools for business workflows at the top.

The IPO Question

Databricks has raised $20 billion over the past 20 months. In an earlier software era, a $1 billion round would have looked extraordinary; in today’s AI market, where infrastructure costs and investor appetite have both expanded, even that amount can appear modest.

The company’s repeated private-market fundraising has become a running joke in Silicon Valley, with some observers quipping that it is running out of alphabet letters for new rounds. Ghodsi has said he still wants to take Databricks public one day, and its expanding investor base will eventually want liquidity.

For now, however, remaining private gives Databricks more room to fund AI research, cloud commitments, and acquisitions without the quarterly pressure of public markets. The broader signal is that capital is flowing not only to model companies, but also to the data infrastructure layer that makes enterprise AI usable. As long as AI spending remains high, leading infrastructure companies can continue to command large private rounds and premium valuations. The next challenge will be proving that Databricks’ $190 billion valuation can translate into durable revenue growth, operating discipline, and eventually a credible public-market story.