Ornn Raises $33M Led by a16z to Build GPU Compute Futures Market

GPU compute spot prices don't behave like software subscriptions. Bloomberg reports that Andreessen Horowitz led a $33 million round in Ornn, a startup building financial infrastructure to let AI companies, data center operators, and investors trade and hedge against GPU price swings. The round follows an earlier $5.7 million seed and brings in co-investors including Galaxy Ventures, Nordstar, and SV Angel.

Ornn's core product is the Ornn Compute Price Index (OCPI), a benchmark built exclusively from live GPU spot transactions covering H100, H200, B200, and RTX 5090 hardware, already distributed on Bloomberg Terminal. In May 2026, the Intercontinental Exchange announced a partnership with Ornn to launch cash-settled, U.S. dollar-denominated GPU compute futures contracts referenced against the OCPI, pending regulatory approval. According to the ICE announcement, Trabue Bland, SVP of Futures Markets at ICE, described the compute market as "in desperate need of a globally accepted pricing mechanism and risk management tool." The contracts would use Asian-style settlement based on the arithmetic average of daily index values.

The case for hedging tools is legible in the price data. The Next Web reported that Nvidia Blackwell spot rental surged 48% from mid-February to mid-April 2026, moving from $2.75 to $4.08 per GPU-hour. Any company planning a large training run months out faces real financial exposure from swings like that. Futures contracts let buyers and sellers lock in prices in advance, the same mechanism energy traders have used in oil and gas markets for decades.

The honest caveat is that "subject to regulatory approval" is doing substantial work here. The ICE compute futures are not yet live, and the reporting does not give a timeline for when regulatory clearance might come. What the $33 million buys in the meantime is index infrastructure, broader participant reach (Ornn reportedly has 400-plus data center operators, investors, and AI companies on its platform), and the credibility that comes with a16z's lead and exchange-grade Bloomberg Terminal distribution.

If approval comes through and liquidity builds, the concrete beneficiaries are AI labs that need to budget training runs in advance, data center operators who can presell capacity at locked-in prices, and financial institutions seeking a regulated vehicle for exposure to AI infrastructure. Whether four separate GPU-type contracts can each sustain enough volume to be useful is the question the reporting leaves open.