Databricks at $190B Sets the Price of an Agent Platform, and Names the Three Products You Will Be Sold

Databricks said in an Aug. 13 press release that it raised $5 billion in a strategic round at a $190 billion valuation, led by Coatue ^[1][9]. In the same release the company said it had surpassed a $7 billion revenue run-rate and grew 80% year over year in the second quarter ^[6], which means the number every enterprise AI platform now gets measured against is roughly 27 times revenue ^[1].

That is the useful part of this announcement. Valuation marks in this category have been asserted more than derived, and here both halves are on the record from the same issuer on the same day ^[1][6]. Anyone raising into the agent-infrastructure story will be asked why they deserve a multiple near 27x, and anyone selling into it will be quoted prices that assume the buyer accepts that arithmetic. At 80% growth held for another year, the same price is about 15 times forward revenue ^[6], which is the number the round's supporters are presumably underwriting rather than the trailing one. The $5 billion raise itself is about 2.6% of the stated valuation ^[7].

The step-ups are steep. The company said the mark is up from $134 billion in February and $62 billion in January 2025 ^[2][3]: about 1.4x and $56 billion of added value in the more recent interval ^[2], and roughly 3.1x since the January 2025 round ^[3]. It is also $2 billion above the $188 billion Databricks said on July 16 it was raising at, in a round it then expected to close by the end of summer ^[11][4].

The product list is the operator-relevant disclosure. Databricks said the money supports its serverless Postgres database for AI agents, Lakebase; its AI coworker, Genie; and its gateway for multi-AI governance and cost controls, Unity AI Gateway ^[4]. The release framed those three as the foundation, context, smart routing and cost controls companies need to deploy agents ^[5]. Read that as the shape of the coming sales motion: a transactional database positioned under agents, an assistant layer above the warehouse, and a control plane that meters and governs model calls. The third one is the tell. Gateways exist because model spend is hard to attribute and hard to cap, and a vendor pitching cost controls is pitching against your current bill.

Chief Executive Ali Ghodsi said enterprises want AI agents working across their business and that Databricks offers the foundation they need ^[7], and that investor demand for the round shows the company's AI strategy is winning the market ^[8]. Coatue Co-Founder Thomas Laffont said Databricks is now the infrastructure the industry builds and scales on ^[10]. Both are seller-side characterizations of the same round.

Two adjacent facts frame the deployment. Databricks said in July it would also use the capital for future AI acquisitions and deeper AI research ^[12], and on July 23 it said it had expanded its decade-long Microsoft partnership through the 2030s, running core business operations on Azure Databricks and using Azure Cobalt for performance and efficiency, with Microsoft continuing to integrate the Databricks platform across its products ^[13].