Databricks at $38 billion shows AI spending starts with data

Databricks raised $1.6 billion at a $38 billion valuation, about 63 times recurring revenue. Investors are paying for the data layer under AI projects.
Databricks is now worth $38 billion after raising $1.6 billion in a Series H round on 31 August. The company sells a cloud platform that combines data storage, analytics and machine learning, and it puts annual recurring revenue at about $600 million, so the new price is roughly 63 times revenue.
A multiple like that is paid only where spending has to flow. In most enterprises machine learning starts with clean, governed, accessible data, and the platform that holds the data captures budget before any model is trained. Investors are buying the layer every AI project depends on rather than the most visible model, and the speed of the repricing shows how firmly that view has set: seven months earlier Databricks raised $1 billion at $28 billion.
The customer base explains the confidence. More than 5,000 organizations use the platform, including over 40% of the 500 largest US companies, and total funding now stands near $3.6 billion. The money goes into the lakehouse architecture, built to replace the separate data warehouse, and into new markets. Databricks calls itself the data and AI company, and the label is a claim about where AI budgets begin.
The February round carried a second signal. Two of the largest cloud providers came in among the new investors, backing a neutral data platform that runs on all three major clouds rather than on their own services alone. When the companies that rent out compute fund the platform that sits across them, the data layer has become infrastructure in its own right.
That sets the terms for builders. The data layer is consolidating around a few platforms, and the winning move is to build with them, not against them. Applications that sit on a lakehouse or warehouse and turn a company's own data into decisions, forecasts or automated work sell into budgets those platforms have already unlocked. Tools for data quality, governance, lineage and feature management answer pain every machine learning team shares, and buyers will expect a product to run where their data already lives.
A price of 63 times revenue still has to clear three tests. Databricks has to hold its recent growth pace as it moves from data engineering into the warehouse market. Public investors have to price data platforms in the coming months at something near this multiple, the real check on a private valuation of this size. And the cloud providers' own data and machine learning services must not slow the shift to neutral platforms.
Fanaura backs AI-first products that get better with every customer's data, and the founders we work with treat access to that data, and to the platforms that hold it, as strategy from the first customer. This round priced the foundation of the AI economy, and the foundation is data.
Building an AI-native company? Apply in five minutes. →
More from Fanaura
Building AI products on top of the data platforms enterprises already use? Send us the project.
Fanaura invests its own capital and the capital of its limited partners (LPs), usually through a separate SPV for each deal. If we see a fit, we get in touch.
Get the news by email
One short note each weekday. No marketing, and you can unsubscribe at any time.