Ornn's $33M seed puts two 20-something founders in the middle of Wall Street's compute trade

Why it matters

Ornn is attacking a real financing gap in the AI buildout: GPU capacity is being bought through opaque contracts while trillions of dollars of infrastructure spend needs benchmarks, hedges and collateral values.

Kush Bavaria (@bavaria_kush) and Wayne Nelms (@wayne_nelmz) have raised a $33 million seed round for Ornn, a compute-finance startup aiming to make GPU capacity priceable and hedgeable through market machinery more like oil, power, and metals.

Axios reported on July 6 that Ornn is building a marketplace for the computing power behind AI, with lenders using Ornn for benchmarks and buyers and sellers using it to hedge exposure to GPU prices. Axios describes Ornn as Andreessen Horowitz-backed.

Bavaria and Nelms are 20-something founders, according to Axios. The bet is that compute is becoming an asset class faster than the financial plumbing around it.

The round is a bet on price discovery before liquidity

Per Axios, Ornn has integrated its pricing data with the Bloomberg Terminal and other data providers so traders can check GPU prices in the tools they already use. Lenders can use that data for benchmarking, and counterparties can use it to hedge.

List prices for cloud GPUs tell a buyer what is advertised, not necessarily what cleared between counterparties across discount, duration, and geography. Ornn is trying to center cleared transactions.

So far, AI companies have tried to lock up supply and prices through long-term pre-purchasing agreements, Axios notes.

ICE gives Ornn a path into regulated derivatives

Axios reports that, pending regulatory approval, Intercontinental Exchange plans to launch GPU compute futures tied to Ornn's pricing index. Because GPU hours cannot be stored like barrels of oil, physical delivery would be complicated; standardized contracts in such markets often rely on cash settlement.

Axios also reports that Ornn is able to operate under a de minimis exemption while larger firms work through regulatory approval.

Ornn is early, and the incumbents have arrived

Ornn is not alone in trying to financialize compute. Axios notes that CME Group plans to launch compute futures tied to Silicon Data's benchmark, pending regulatory review.

That puts Ornn in an unusual early-stage position: a seed-stage startup selling into a market where the biggest exchanges already see a product line. If GPU capacity becomes a recognized hedging instrument, the company that supplies trusted pricing and transaction rails can sit close to the flow of capital.

The technical challenge is real. Compute is not a uniform barrel: Nvidia GPU generations change economics quickly, and unused capacity disappears. Any benchmark has to chase a depreciating, non-storable resource.

Goldman Sachs estimates roughly $7.6 trillion of global investment between 2026 and 2031 across compute, data centers, and power. Those stakes explain why Wall Street wants instruments around compute before the market has fully standardized.

Axios also attributes a geopolitical stance to Bavaria: he sees compute finance as part of America's potential advantage over China and says Ornn does not work with Chinese AI labs.

For founders building AI companies, the pitch is straightforward: GPU prices are volatile and opaque, and commitments can make or break a runway. For data-center operators, the question is whether a benchmark and hedging layer can make future revenue easier to finance. Ornn still has to prove adoption, but a $33 million seed buys time to try.