Databricks’ $175B Valuation Echoes Crypto’s 2021 Circular Logic Problem

Published by James Harris on June 9, 2026June 9, 2026

Databricks is quietly shopping a new funding round that would value the company at up to $175 billion, a 30% jump from the $134 billion valuation it secured just four months ago. No fundamental business event triggered the reset. Investors are simply willing to pay more.

The company’s underlying numbers are real enough: $5.4 billion in annualized revenue growing at 65% year-over-year is a legitimate business, not a fiction. But the valuation math is getting harder to ignore. At $175 billion, Databricks would trade at roughly 32 times forward revenue, a multiple that assumes the growth rate holds, competition from Snowflake stays manageable, and the broader enterprise AI spending cycle does not turn. That last assumption is where the circularity problem enters. A significant portion of the cloud commitments underpinning AI revenue across Microsoft, Amazon, Google, and Oracle traces back to OpenAI and Anthropic, companies that spend nearly as much on cloud infrastructure as they earn in revenue, funded by the same hyperscalers recording that spending as growth. The loop is not hidden. Analysts have flagged it. The capital keeps flowing anyway.

The gap between OpenAI’s $60 billion cloud bill and its $25 billion in revenue is not a rounding error. It is the business model.

This matters beyond the private markets. Crypto has its own history with reflexive valuation structures, most visibly in the 2021 DeFi and NFT cycles where protocol treasuries held their own tokens as primary assets, inflating book value until liquidity dried up. The AI private market is running a slower, more institutionally dressed version of the same dynamic: valuations justified by revenue, revenue partially justified by the same capital that set the valuation. When Databricks eventually files for an IPO, public market investors will have to price that dependency honestly in a way that a Qatar Investment Authority-led private round does not require. CEO Ali Ghodsi’s reported comment that 2026 is the “worst year” to go public, given IPO competition from SpaceX and others, may also be a reasonable read on how much scrutiny a public filing would invite right now.

The confirmed IPO target is 2027, which gives Databricks another funding cycle to grow into its multiple before facing quarterly earnings calls. Whether the revenue composition looks cleaner by then depends largely on whether enterprise AI adoption outside the hyperscaler ecosystem accelerates enough to replace the circular flows with genuinely independent demand.