Databricks wanted to raise $1 billion. Investors offered $15 billion. It took $5 billion.

Key points

• Databricks closed a $5 billion funding round in 2025 at a $190 billion valuation, led by investment firm Coatue.
• The company originally planned to raise only $1 billion before a news leak triggered $15 billion in investor interest.
• Databricks reported $7 billion in annualised run-rate revenue, meaning the pace at which it earns money projected across a full year, growing at 80 percent.
• Its database tool for AI agents, called Lakebase, hit $100 million in annualised revenue shortly after launching in June 2025.
• CEO Ali Ghodsi says the company still plans to go public one day, but has not set a date.

Databricks, a company that helps large businesses store and analyse data using AI, raised $5 billion from roughly two dozen investors. The round closed at a valuation of $190 billion, making it one of the most valuable private technology companies in the world.

That figure crept up from $188 billion. The company had quietly announced that earlier valuation in July without disclosing how much it had actually raised. Thursday brought the full picture.

How did a $1 billion raise turn into $5 billion?

Databricks never planned to raise this much. CEO Ali Ghodsi told TechCrunch he originally wanted $1 billion. Then The Information ran a story in June saying the company was planning a large fundraise, right in the middle of a Databricks conference.

"My phone blew up," Ghodsi said. "It was like the worst timing for us because we were busy with our conference."

The story became a self-fulfilling prophecy. From the pool of investors Databricks actually considered, demand hit $15 billion. Turning away that many long-standing backers risks bad feelings, so the company issued more shares and accepted more money than it wanted.

The round was led by Coatue, a major technology-focused investment firm, with participation from Blackstone, MGX, several T. Rowe Price accounts and new backer Sixth Street Growth, the firm founded by former Goldman Sachs chief investment officer Alan Waxman.

Why does a company earning billions still need more cash?

Databricks is profitable on a cash-flow basis, meaning it brings in more money than it spends on daily operations. Its core product, a cloud data warehouse (software that stores and processes huge amounts of business data online), generates $1.5 billion in annualised revenue and is still growing at 100 percent year over year.

AI is expensive in ways that go beyond normal operations. The company holds multi-billion dollar commitments with all three major cloud providers to run its computing workloads, and it employs 100 AI researchers in a field where talent costs are sky-high.

Acquisitions add further pressure. This week Databricks bought Electric, the maker of a lightweight database tool called PGlite that lets AI agents, software programs that carry out multi-step tasks on their own, spin up small databases quickly. In June it bought AI cybersecurity firm Panther, and it made two more purchases in March.

What does this mean for ordinary people?

If your employer uses Databricks software, more investment likely means faster product updates and new AI tools built into platforms your company already runs.

For anyone watching the broader AI economy, the numbers are worth sitting with. Databricks has now raised $20 billion over the past 20 months. As we reported on 13 August, Anthropic investors expect a $2 trillion valuation when that company goes public, a sign that private AI valuations have entered territory that would've seemed fictional a few years ago. A round that would've been considered enormous five years ago is now a rounding error.

Ghodsi has said publicly he wants to take the company through an initial public offering, a stock market listing, eventually. With two dozen investors waiting to cash out, he hasn't got much choice. For now he's keeping the company private and spending heavily on AI.

Given that he can summon $15 billion in demand with no effort, the wait may continue a while longer.

Common questions

What is Databricks and why should I care about it?

Databricks makes the software that many large companies use to store and make sense of their data, and it's now building AI tools on top of that. If you use a bank or a retailer that runs on modern data systems, there's a reasonable chance Databricks is somewhere in the background.

What is a valuation and does $190 billion mean the company is worth that?

A valuation is the price investors agreed to pay per share, multiplied by all the shares that exist. It reflects what investors believe the company could be worth, not a confirmed sale price. Private companies like Databricks have no public share price to check, so the number can shift with each new funding round.

Will Databricks ever go public?

Ghodsi has said yes, one day. No date has been set. The company has enough investor demand and cash to stay private for now, but a growing roster of shareholders who want a return makes a public listing increasingly likely over the next few years.