Databricks reaches $188B valuation on AI pivot momentum

Databricks announced on Thursday a new funding round that values the company at $188 billion（約30兆円）. The round was led by Coatue and involves roughly $3 billion（約4800億円）, though Databricks has not disclosed the exact amount; the company noted the money will not arrive until later in the summer. Unusually, Databricks announced its valuation before the funding closed, but a venture capitalist confirmed to TechCrunch that the deal is solid, with such strong demand from investors that the company felt no reason to keep its new valuation private.

This latest round caps a remarkable fundraising sprint. In February, Databricks closed a Series L raise of $5 billion（約8000億円） at a $134 billion（約21兆円） valuation. Five months before that, in September 2025, it raised $1 billion（約1600億円） at $100 billion（約16兆円）. Roughly nine months before September, in December 2024, it raised $10 billion（約1.6兆円） at $62 billion（約9.9兆円）. The rapid sequence of rounds became notable enough to spawn internet humor about the company running out of letters of the alphabet to name its funding rounds. "Turning on alerts for when we get a Series AA," one observer posted.

The fundraising surge reflects a genuine business transformation. Databricks, founded in 2013, originally succeeded during the big-data era by providing software that let enterprises store massive data volumes in the cloud and produce fast analytics. When enterprises began adopting AI, Databricks was well-positioned: it already had deep customer relationships and access to enormous enterprise datasets. The company began rolling out AI products—Lakebase (a database for AI agents), Unity (an AI gateway), and Omnigent (a multi-agent management system)—to serve this demand. Critically, Databricks also became known as a champion of affordable, open-weight models (models whose code is published and can be modified by anyone), positioning itself as a guide to cost control, a major enterprise trend in 2026. It particularly champions Z.ai's GLM 5.2 for coding tasks.

Last week, CEO Ali Ghodsi published the results of internal benchmarking he conducted to manage AI costs for his 3,000 software engineers. Databricks compared AI models on the actual coding tasks its programmers perform. The company found that open models, especially GLM 5.2, can now handle even the highest-difficulty coding tasks, and at a lower total cost than proprietary models from Anthropic and OpenAI. Databricks also found that the choice of agentic harness—the tool that wraps around a model and manages context and instructions—equally impacted costs. Open-source harness Pi emerged as one of the best at managing context and therefore one of the lowest-cost choices without sacrificing quality. "The lesson here isn't that one harness is always cheaper or that native harnesses are worse," Databricks wrote. "Instead, model choice is only one piece of the puzzle." This grounded, public benchmarking has enhanced Databricks' credibility as a practical guide to enterprise AI adoption, fueling investor enthusiasm and its ability to command a $188 billion（約30兆円） valuation.

Databricks' $188 billion（約30兆円） valuation represents a decisive victory for the company's bet on repositioning itself in the AI era. Founded in 2013 during the big-data boom, Databricks initially succeeded by helping enterprises manage and analyze vast data stores in the cloud. That legacy gave it a natural advantage when enterprises began demanding AI: the company already sat on deep customer relationships and access to troves of enterprise data—precisely what companies need to implement AI securely and governably. The company's pivot accelerated sharply in 2025 and 2026, producing a string of AI-focused products (Lakebase, Unity, Omnigent) and positioning itself as an authority on affordable, open-weight models.

What distinguishes Databricks' fundraising momentum is its concrete answer to a real enterprise concern: cost control. CEO Ali Ghodsi's public benchmarking—testing models on actual coding tasks his engineers perform—supplied evidence that open models paired with intelligent agentic harnesses can outperform expensive proprietary alternatives on both quality and price. This is not merely marketing halo; it is a grounded recommendation backed by internal testing. As a result, Databricks has earned credibility not as an AI lab, but as a trusted guide to practical AI deployment for cost-conscious enterprises. The rapid succession of rounds (December 2024, September 2025, February 2026, and now) reflects investor appetite for companies that can translate enterprise pain points into real business value.