Databricks tops a $100B valuation with a $1B raise—and says its AI revenue run‑rate just crossed $1B

International Desk — September 9, 2025

Databricks has pulled off the rare feat of getting bigger and clearer at the same time. The data‑and‑AI platform confirmed it has closed a $1 billion Series K at a valuation above $100 billion, while also disclosing a $4 billion annual revenue run‑rate, more than 50% higher than a year ago. In the same breath, the company said revenue tied specifically to its AI products has itself reached a $1 billion run‑rate, and that the business generated positive free cash flow over the past twelve months. For a still‑private software firm, that’s a striking combination of growth and discipline—and it explains why crossover investors keep crowding the cap table. ReutersDatabricks

The investor roster reads like a who’s who of late‑stage tech backers: Andreessen Horowitz, Insight Partners, MGX, Thrive Capital, and WCM Investment Management co‑led the round, according to the company. That mirrors independent reporting from major outlets and follows weeks of pre‑close whispers that pegged the valuation in 11‑figure territory. The Wall Street Journal, which closely tracks Databricks’ finances, has framed the new run‑rate as a clean step‑up from the company’s internal targets earlier this year—another signal that enterprise spending on AI infrastructure isn’t a passing sugar high. DatabricksThe Wall Street Journal

What will the money actually buy? If you’ve watched Databricks this summer, the plan is hiding in plain sight. The company is putting fresh capital into Agent Bricks—a unified workspace for building production‑grade AI agents that run on a customer’s own data—and into Lakebase, a new Postgres‑based operational database (OLTP) the company says is engineered for the kinds of agents businesses now want to deploy. In other words, it’s not a vanity raise; it’s fuel for a product pivot from generic “gen‑AI” to agents that remember, reason, and plug into workflows customers already live in. The press release ties those bets directly to the raise. Databricks

There’s also a telling detail in the customer math. Databricks says it now has net revenue retention above 140%, with 650+ customers each spending over $1 million per year on the platform. Those are the sorts of numbers that make CFOs more comfortable funding “AI” as a line‑item, not a pilot—and they help explain how the company could grow faster than 50% at its scale without lighting cash on fire. The firm also claims 20,000+ organizations use its platform globally, a reach that makes every new feature—agents, databases, governance—instantly relevant to a very large base. Databricks

Context helps here. Back in December 2024, Databricks announced a $10 billion Series J at a $62 billion valuation and talked openly about crossing a $3 billion run‑rate. The new numbers suggest the company not only hit those marks, but kept accelerating. Moving from a $62B valuation to north of $100B in less than a year is aggressive—but the jump is less surprising when you consider the broader backdrop: enterprises are rebuilding data stacks to make AI useful in day‑to‑day operations, and they want a single pane of glass for warehousing, governance, and model‑driven apps. Databricks

Rivalry also sharpens the picture. Over the past ten days, Snowflake shares have surged on signs that customers are spending more to simplify AI adoption on top of their data warehouses. That enthusiasm spills over to Databricks because the two companies increasingly compete for the same budgets. When one proves that the “AI + data” thesis is converting to dollars, it raises confidence that the other can keep comping big quarters, too. The real competition is less about marketing slogans and more about which vendor lands deeper inside the daily work of data teams, app developers, and business users. Reuters

Databricks’ message to those teams is that the era of flashy demos is winding down; the next phase is agents that operate on secure, governed data and survive the rough edges of production. “AI that can pass a lab test but falls apart in a workflow doesn’t help anyone” is the subtext. That’s why the company keeps pairing AI features with workmanlike plumbing: Unity Catalog to govern data and models in one place, Delta Lake and lakehouse patterns so analytics and apps sit on the same substrate, and now a transactional database (Lakebase) tuned for agent read‑and‑write behaviors. Agree with the vision or not, it’s coherent—and, crucially, it’s built to be measured in ROI terms.

On the capital markets side, staying private while printing public‑grade metrics gives Databricks leverage. Reuters notes outside investors already see it as an IPO‑ready story, especially in a market newly receptive to profitable growth after the Figma listing reset expectations. But the company doesn’t need to rush; with positive free cash flow and a cap table full of patient backers, it can time a listing for market structure, not for survival. That alone distinguishes it from the last generation of unicorns that hit the public tape still searching for a business model. Reuters

None of this makes the road easy. AI tailwinds are real, but they also invite scrutiny: how much of that $1B AI run‑rate is new money versus spend recaptured from point tools? Do agents reduce help‑desk tickets and cycle time, or do they just add one more service fee to the stack? Can Lakebase win workloads from incumbent operational databases that have decades of muscle memory in enterprise IT? Those are fair questions, and the answers will arrive in the only place that matters—customer renewals and expansions.

What’s clear is that Databricks has moved the AI conversation back to where most buyers live: data first, outcomes measured. It isn’t trying to out‑meme the model labs or chase every headline feature. The pitch is plainer—and, to many CIOs, more attractive: give teams a governed foundation, let them ship agents that actually do a job, and keep everything under one roof so costs and risks are visible. If the company can keep net retention above 140% while expanding the number of million‑dollar accounts, the $100B badge will feel less like a trophy and more like table stakes for the category leader. Databricks

The bottom line: Databricks’ new round is less about valuation theater than operating proof. Crossing a $4B run‑rate with positive FCF while investing in a specific agent‑centric roadmap is exactly the profile growth investors want to see in late 2025. The company still has to win the long game—integration is hard, and incumbents never stand still—but for now the signals point in the same direction: customers are not just experimenting with AI; they’re betting their data on it. And Databricks is positioning itself as the place where that bet becomes a system, not a slide. ReutersDatabricks

Sources used in this report: Databricks’ official press releases (Series K and run‑rate metrics), plus corroborating coverage by Reuters and The Wall Street Journal. DatabricksReutersThe Wall Street Journal