Databricks closed a $5 billion strategic round on Thursday at a $190 billion valuation, the company confirmed alongside second-quarter numbers showing its revenue run-rate crossing $7 billion, up more than 80% year on year.
The headline number is the valuation: $190 billion is 42% above the $134 billion Databricks carried after its Series L closed in December, and it lands the company among the most valuable private firms on earth. Coatue led. Blackstone, MGX, accounts advised by T. Rowe Price and new investor Sixth Street Growth joined, along with BOND, Clearlake Capital, Point72, Premji Invest and TPG. Existing backers Andreessen Horowitz, GIC, Temasek and Thrive Capital came back in.
But the valuation is not the interesting number. The ratio is.
Do the division
At $134 billion against a $5.4 billion run-rate in February, Databricks was priced at roughly 25 times revenue. At $190 billion against a run-rate above $7 billion, it is priced at roughly 27 times. The valuation jumped 42%; the multiple moved less than three turns.
That is a rarer outcome in 2026 than it sounds. The pattern across the AI cohort this year has mostly been the reverse — valuations sprinting ahead of the revenue underneath them, or revenue sprinting ahead of a valuation investors are no longer willing to re-mark. Databricks did the boring version: it grew into the price and then asked for a slightly higher one.
Our take: A 42% valuation increase backed by ~30% run-rate growth is a repricing, not a re-rating. That is what a durable enterprise software business looks like when the AI narrative happens to be pointed at it. The multiple is still 27x — expensive by any pre-2023 standard — but it is the kind of expensive that revenue can catch, and the composition of this round says the buyers know it.
What the money is actually for
Databricks named three products as the destination for the capital, and each one is a land-grab against a different incumbent:
- Lakebase, its serverless Postgres for AI workloads, which the company says has passed a $100 million run-rate. That is small against $7 billion — but it is an operational database sitting next to the analytics estate, which is a different budget line entirely.
- Lakehouse, the data warehousing product, now past a $1.5 billion run-rate. This is the direct shot at the warehouse incumbents, and it is the single largest disclosed product line.
- Unity AI Gateway, governance and controls for running multiple models. Unglamorous, and precisely the thing enterprises stall on before they deploy agents at scale.
Read together, those three are one argument: that the company holding the data will end up holding the AI budget, because governance and proximity beat model choice. Every model vendor is arguing the opposite.
What to watch
- Whether the run-rate compounds or the comp gets harder. Growth above 80% off a $7 billion base is the claim that justifies 27x. The next print is the test.
- Lakebase's slope, not its size. $100 million is a rounding error on the total; how fast it triples is the signal on whether the operational-database push is real.
- The IPO clock. A $5 billion private round at $190 billion buys a lot of patience — and gives late investors a reference price the public market will be measured against.
The AI market spent 2026 arguing about which valuations are defensible. This one at least shows its work.