Databricks Closes $5B Round at $190B Valuation as Agent Demand Accelerates

On August 13, 2026, Databricks announced two numbers that frame the current state of enterprise AI better than almost anything else this quarter: a $7 billion revenue run-rate growing more than 80% year over year, and a freshly closed $5 billion strategic funding round at a $190 billion valuation. The round was led by Coatue, with Blackstone, MGX, T. Rowe Price, and new investor Sixth Street Growth participating.

The valuation step-up is steep. Just six months ago, Databricks closed a $4 billion Series L at $134 billion. In February it completed the round with $2 billion in debt financing led by JPMorgan, bringing its early-2026 war chest to roughly $7 billion with JPMorganChase, Goldman Sachs, Microsoft, and the Qatar Investment Authority on the cap table. Now, at $190 billion, Databricks sits comfortably among the most valuable private technology companies in the world — and it got there without an IPO.

What the money is for

Unlike many AI-era mega-rounds that quietly fund compute bills, Databricks has named exactly where the capital goes: three products that together form what CEO Ali Ghodsi calls the operating layer for enterprise AI agents.

• Lakebase — a serverless Postgres database purpose-built for AI agent workloads, giving agents real-time read/write access to operational data rather than stale snapshots. The product has already crossed a $100 million run-rate and traces its lineage to Databricks’ roughly $1 billion acquisition of Neon in 2025.
• Genie — an AI coworker that turns business data into answers and actions. Genie One, launched in June as a self-improving agentic coworker, is the consumer-facing tip of this effort.
• Unity AI Gateway — a multi-AI governance layer with model routing and cost controls, which reached general availability in August. As enterprises juggle models from multiple providers, the gateway has become the toll booth for AI spending.

“Enterprises want agents working across their business that remember context, deliver accurate results and execute work without blowing through budgets,” Ghodsi said in the announcement. “That requires real-time operational data with Lakebase, context from across the business with Genie, and multi-AI cost controls with Unity AI Gateway.”

The numbers behind the round

Databricks’ growth story is increasingly built on large contract concentration. The company now counts more than 1,000 customers spending above $1 million annually, and over 100 customers spending above $10 million — a cohort structure that looks more like an enterprise infrastructure incumbent than a startup.

The lakehouse core remains the engine. Databricks’ data warehousing product hit a $1.5 billion run-rate, growing more than 100% year over year, meaning the classic analytics business is still compounding even as the narrative shifts to agents.

The acquisition strategy has filled gaps deliberately: Neon for serverless Postgres, Tecton for feature engineering, Mooncake Labs for storage infrastructure, Quotient AI for data quality, and the recently completed Panther acquisition to establish a security lakehouse. Capital Bricks-style investments aside, the through-line is a single governed environment spanning data, features, models, and now agents.

The platform convergence thesis

The round lands at a moment when agentic AI deployment is converging toward unified platforms that embed governance, data lineage, and semantic capabilities natively — rather than bolting them on afterward. Databricks is positioning itself at the center of that shift with a stack spanning lakehouse architecture, serverless Postgres, and agentic tooling.

A second pattern is governance-by-design. Analysts at VKTR note that managing AI agents through a single interface is rapidly becoming a baseline enterprise expectation, not a premium feature. Unity AI Gateway is Databricks’ answer to exactly that expectation.

There is also an unspoken pressure behind the raise. Enterprise AI ROI remains stubbornly hard to demonstrate: MIT research found that despite nearly $40 billion spent on generative AI over two years, only 5% of enterprises could show real business returns. Vendors that can credibly claim to convert agent enthusiasm into measurable outcomes — with cost controls attached — stand to capture a disproportionate share of the next budget cycle.

Beyond infrastructure: CustomerLake and martech

Perhaps the most strategically interesting move is Databricks’ expansion into marketing technology. In June the company launched CustomerLake, an agentic customer data platform built natively inside the lakehouse, bringing Customer 360 profiles, identity resolution, audience segmentation, campaign automation, and channel activation into the same governed environment where enterprises already manage data and models.

The pitch is architectural: no more copying customer data into a standalone CDP. Profile Agents turn raw data into business-ready customer profiles, while Campaign Agents build audiences and drive what Databricks calls “infinity campaigns” — continuous, goal-driven engagement loops that respond to signals instead of waiting for manually configured one-time pushes.

Gartner’s assessment is blunt: it expects 80% of net-new enterprise CDP deployments by 2030 to follow the Databricks-style pattern. If that prediction holds, the $190 billion valuation may look conservative against the martech TAM it opens up.

What it means

Three takeaways worth watching:

1. Private markets are still deep for AI infrastructure with real revenue. An 80%-growth, $7B-run-rate business commanding a 27x revenue multiple signals that late-stage investors see durable cash flows, not just GPU-fueled speculation.
2. The agent stack is the new platform war. Databricks is not selling models; it is selling the substrate agents run on — operational data, business context, and cost governance. That is a defensible position exactly where enterprise budgets are consolidating.
3. The IPO question is deferred, not dead. Databricks chose another private round over listing, and the secondary dynamics suggest it can keep compounding on private capital. Whenever the filing does arrive, it will arrive as one of the largest software IPOs in history.

For now, Databricks has the balance sheet, the customer concentration, and the product surface to make its case that the next decade of enterprise software belongs to whoever runs the agents. The $190 billion question is whether the agents, in turn, deliver returns the dashboards can finally prove.