Ahead of an IPO, the maths behind Databrick’s $190 billion valuation doesn’t add up
The data and AI company is one of the world’s strongest private companies, but a PitchBook report says it’s unattractive at the current price.
Databricks has built one of the strongest private software companies in the world, but its $190 billion price tag is a bit steep, argues a new PitchBook report.
“Databricks is a high-quality business but an unattractive investment at the current price,” said Harrison Rolfes, the report’s author and senior analyst covering late-stage companies.
A Databricks IPO remains far off. The company’s CEO Ali Ghodsi has publicly ruled out a public listing for 2026 telling Bloomberg Television that this was “a terrible year to go public,” due to a crowded calendar of SpaceX and Anthropic. 2027 is the earliest the company could IPO, Ghodsi has said, but the company hasn’t even filed an S-1 confidentially.
In August, Databricks raised $5 billion at a $190 billion valuation, led by private equity firm Coatue, up 41.8% from Databricks’ last valuation of $134 billion priced in February. The company said in August that its annualized run rate rose over the same stretch from $5.4 billion to $7.0 billion, a 29.6% increase.
Rolfes’ report estimates the business is closer to a $68.7 billion operating value (what the business itself is worth before accounting for debt, cash, and preferred shares that get paid ahead of common shareholders), a 64% discount from the $190 billion valuation.
Looking ahead, the most optimistic scenario—which requires Databricks to achieve 80% gross margins, continue hitting its growth targets, and see market conditions cooperate—still values the company at only $182.1 billion.
The gap comes down to how the report estimates Databricks’ future profits.
Rolfes projected Databricks’ revenue and cash flow over the next 10 years, then discounted them back to today’s dollars to account for changes in the value of money. The forecast also makes assumptions about how Databricks will behave and expand in the future, estimating that recognized revenue growth next year settles at 48% and then into single digits by the mid-2030s.

Source: PitchBook. Data as of Sept. 10, 2026.
Putting all that together, the model has Databricks generating $9.3 billion of cash a year by 2035. Real money, but not enough to justify $190 billion today.

Source: Morningstar. Data as of Sept. 14, 2026.
Snowflake SNOW, Databricks’ chief competitor, offers the closest thing to a reality check. The public data-warehousing company was worth about $124 billion in early September, trading at about 20 times the revenue it expects this fiscal year.
Databricks’ $190 billion is premised on its $7 billion run rate, which works out to a valuation multiple of 27 times. Databricks is growing faster than Snowflake, so Rolfes argues some of the premium is warranted. But Snowflake’s numbers are all audited and reported, unlike Databricks’ self-reported figures, which leave an incomplete picture.
“The central question is no longer whether Databricks is a great business,” Rolfes said. “It is whether the price leaves enough return for the next investor.”