The round was confirmed quietly, the way Databricks has done these things since it stopped needing to shout. The company, which sells software that helps enterprises organize data for AI, said it had closed a new round of funding at a valuation of $188 billion, nearly triple the $62 billion it commanded in its previous round. The investors are strategic rather than purely financial, according to people familiar with the matter, a choice that reflects the company’s ambition to sit inside every large company’s AI plans.
The jump is striking even by the standards of a market that has repriced software companies repeatedly. Databricks was valued at $62 billion in late 2024, when it raised $10 billion in what was then one of the largest private rounds in the industry. Since then the business has grown quickly. The company’s annualized revenue run rate has climbed to roughly $4 billion and keeps growing, executives have said, and its customer base now spans most of the world’s largest enterprises.
The thesis behind the valuation is simple: models are interchangeable, data is not. Enterprises experimenting with AI have discovered that the value sits in their own information, from sales records and support logs to internal documents, and Databricks sells the platform for storing, cleaning, and querying that data, plus the tools to train or fine-tune models on top of it. That positioning has made the company the default infrastructure choice for a growing number of large customers, according to analysts.
It has also put Databricks in direct competition with Snowflake, its longtime rival in the market for cloud data platforms. Both companies have spent the past two years repositioning themselves as AI companies, and both claim to be the layer where enterprise AI actually happens. The rivalry has produced a steady stream of product launches, price skirmishes, and public sniping between the two chief executives. The competition has already claimed casualties in the open-source world: Databricks bought Tabular, the company behind the Iceberg table format, and gave the technology to the Apache Foundation, a move that forced Snowflake to change course.
Analysts said the new valuation reflects investors’ belief that Databricks is winning that argument. The company’s platform has become the place where many large companies store the data their AI applications depend on, and its Mosaic AI product line lets customers fine-tune models without moving data to another vendor. Snowflake has countered with its own AI features and has argued that its approach is simpler for customers already standardized on its platform, but the funding gap between the two companies has widened sharply.
The round also funds a war chest. Databricks has been acquisitive, and its largest purchase, of MosaicML in 2023, gave it the AI training stack it now sells. Analysts said the company is likely to keep buying, with the candidates being tools that strengthen its grip on enterprise data and AI workflows. The capital also covers compute commitments; running AI workloads for customers means reserving GPU capacity at scale, and those contracts are expensive.
The valuation places Databricks among the most valuable private technology companies in the world, though below the paper valuations of AI labs such as Anthropic, which has been reported to be worth close to $1 trillion in private markets. The comparison is imperfect: Databricks sells software with real revenue and real customers, while the labs trade largely on potential. That distinction has become central to how investors think about the AI sector, with platforms and model makers competing for the same dollars.
For investors, the round raises a familiar question: how much of the AI boom is real revenue and how much is expectation. Databricks has genuine customers and genuine growth, analysts said, but the multiple embedded in $188 billion assumes that enterprise AI spending will compound for years and that Databricks will capture a large share of it. The company also faces competition from the cloud giants, which bundle their own data and AI tools, and from startups attacking pieces of its stack.
The timing matters. Databricks has long been expected to go public, and each new round resets the benchmark for that event. A person familiar with the company’s plans said an IPO remains on the table but no date has been set, and the strategic nature of the new investors suggests the company is lining up relationships that will matter once it is public.
For now, the message from the round is about the market’s center of gravity. The most valuable private AI companies are no longer just the model labs; they are the platforms that make models useful inside companies. Databricks has bet its valuation on that idea, and the market has agreed to pay for it. Whether Snowflake’s investors are convinced is another question.
