When Kush Bavaria and Wayne Nelms set out to solve what seemed like a straightforward pricing problem, they probably didn't expect to end up courting Goldman Sachs.
But that's where Ornn, the New York-based startup the two MIT alumni founded, finds itself after closing a $33 million seed round on June 24. Led by a16z crypto, the funding attracted Galaxy Ventures, Nordstar, and SV Angel, along with returning backers Vine Ventures, Crucible Capital, Link Ventures, and Box Group. The company had previously raised $5.7 million in a seed round in October 2025—a respectable raise that set the stage for what Bavaria and Nelms now describe as an effort to financialize GPU compute itself.
The pitch is audacious in its simplicity: treat computing power the way markets treat oil, gold, or natural gas. Create reliable benchmarks. Build futures contracts around them. Let companies hedge their exposure and investors bet on capacity the way they might on wheat harvests or interest rates.
Whether that vision materializes remains an open question. But Ornn has moved with unusual speed—and attracted unusual attention.
The Index That Started It All
The company's origin story hinges on a problem that anyone shopping for high-performance GPUs knows intimately: pricing is a mess. There was no transparent, transaction-based way to benchmark what an Nvidia H100 actually cost to rent at any given moment. So Bavaria and Nelms built one.
In April 2026, Ornn's Compute Price Index—OCPI for short—went live on the Bloomberg Terminal. The index tracks real transaction prices for H100 GPUs, offering what the company bills as the first reliable pricing signal in a market that has historically operated more like informal backchannels than an organized exchange.
That benchmark quickly became more than a data product. On May 19, Intercontinental Exchange announced plans to launch cash-settled GPU compute futures tied to OCPI, pending regulatory approval. Eight days later, FalconX executed the first over-the-counter compute forward trade referencing OCPI's H100 pricing.
Suddenly, Ornn wasn't just publishing numbers. It was providing the infrastructure for derivative markets.
From Data to Derivatives
The seed funding will help Ornn shift from indexing to operating—building out a marketplace where companies can buy, sell, and hedge compute capacity using the benchmarks it has established. As a16z crypto framed it in its investment announcement, the firm sees Ornn constructing "the market stack" around GPU capacity: trusted pricing, hedging tools, and the capital-markets scaffolding that could turn compute into a tradeable asset class.
Ornn has layered on additional products with a pace that suggests urgency, perhaps even impatience. On June 16, it launched OTPI—Ornn Token Price Indices—benchmarking the realized cost of inference tokens from Anthropic and OpenAI. The company also runs Ornn Compute, a platform aggregating dedicated GPU capacity, and has partnered with ProCap Financial to distribute its pricing data to independent investors.
According to a report in The Information, both Goldman Sachs and JPMorgan are exploring ways to participate in what remains a nascent, uncertain market. The publication was careful to note that widespread adoption is far from guaranteed—a caveat worth taking seriously given how many financial products have launched with fanfare only to fizzle.
Not the Only Player
Ornn faces competition, and from a formidable source. CME Group announced on May 12 that it plans to launch its own compute futures later this year, partnering with Silicon Data and using that firm's GPU rental-rate benchmarks. The parallel efforts underscore a broader industry shift: as AI workloads scale and computing infrastructure becomes mission-critical, the market is starting to resemble energy or metals trading more than it does traditional cloud services.
The company lists approximately 19 employees on LinkedIn, though such figures are self-reported and can fluctuate—a lean team for the ambitions it's articulated. In mid-June, Ornn made an unusual move, reserving the ticker "ORNN" on the New York Stock Exchange. It's a signaling exercise, not an imminent IPO filing. But it reflects Bavaria and Nelms' stated intention to build Ornn as a public-market company from the outset, a strategy that carries its own risks and rewards.
For now, the company is betting that the chaotic, opaque world of GPU procurement is ready to be tamed by the tools of finance. Whether Wall Street—and the hyperscalers, AI labs, and cloud providers who actually buy this capacity—will agree is the question that $33 million is meant to help answer.
