loading...
RISING$33.0M · 5 Jul 2026 · 4 MIN
Trillion-Dollar Market. No Price Discovery. Ornn Is the Fix.
Two MIT graduates put GPU spot prices on Bloomberg Terminal, partnered with ICE for compute futures, and raised $33M to build financial infrastructure for the compute economy.
ORNNornn.com ↗
KEY METRICOCPI on Bloomberg Terminal; 400+ platform users; ICE GPU futures partnership
LATEST ROUND$33M Seed, June 2026
TOTAL RAISED$33.0M · 1 rounds
OPPORTUNITY SNAPSHOTenter market
ENTRY ANGLES
Build compute financing using OCPI as pricing oracle · Create GPU-backed lending for signed capacity agreements · Develop structured products on top of GPU compute futures · Build compute escrow services for AI training runs
VERTICALS
AI infrastructureFintechGPU cloud operatorsNeoclouds
CAPABILITIES
Capital markets expertise, Commodity derivatives knowledge, Relationships with neocloud operators, Risk modeling for compute capacity default
KUSH BAVARIA, CEO
“Compute has grown into a trillion-dollar market, yet it still lacks the pricing and risk-transfer infrastructure that every other major commodity relies on.”
01 /The Concept
The company selling you GPU time does not know what its GPUs are worth. Not in the sense that it has guessed wrong – in the sense that there is no reference price from which to start. GPU compute transactions happen privately, prices are set in bilateral negotiations, and there is no benchmark an operator can reference to know whether what it is charging or paying reflects the market. Every other major resource economy solved this problem before it reached the scale GPU compute has already passed.
Oil has NYMEX. Natural gas has Henry Hub. Corn has the Chicago Board of Trade. GPU compute – now priced at the trillion-dollar scale – has none of that. No price index, no futures market, no risk-transfer mechanism.
Kush Bavaria and Wayne Nelms, two MIT graduates, founded Ornn in 2025 to build it. Bavaria brings capital markets fluency; Nelms spent years as an equity options trader at Susquehanna International Group. Their first product was the Ornn Compute Price Index – OCPI – built exclusively from live, printed GPU spot transactions across H100, H200, B200, and RTX 5090 hardware. OCPI is now distributed on the Bloomberg Terminal. In May 2026, the Intercontinental Exchange announced a partnership with Ornn to launch cash-settled, dollar-denominated GPU compute futures contracts referenced against OCPI, pending regulatory approval. Ornn also operates a compute marketplace that aggregates GPU capacity across public clouds and neoclouds, with a secondary market for capacity transfers and on-demand subleasing. A $33 million seed round led by a16z Crypto closed in June 2026, with participation from Galaxy Ventures, Nordstar, SV Angel, and others; 400 data center operators, investors, and AI companies are already on the platform.
02 /Why It Matters
Getting OCPI into Bloomberg's data infrastructure signals to the institutional players who build commodity derivatives that GPU compute pricing is ready to be treated as a structured financial product – not because of what OCPI is technically, but because of who Bloomberg's data feeds reach. It is a category declaration in the language commodity traders understand.
Futures contracts solve a specific problem: operators who need certainty about future costs can hedge it forward rather than carry price risk on a volatile spot market. An AI company that needs 10,000 H100 hours over the next six months cannot currently lock in that price. It buys spot, absorbs volatility, and models the uncertainty into planning. Cash-settled futures change that calculus – for buyers, sellers, and anyone who wants to take a financial position on compute prices without touching a single GPU.
The 400 users already on OCPI – before the futures market exists – suggest the benchmark is valuable in its own right, not only as a derivative reference price. That is exactly how commodity benchmarks develop: price discovery comes first, derivatives follow.
03 /Opportunities
Commodity financialization follows a consistent sequence: spot transparency, then forwards, then futures, then options, then structured products. Ornn is at step three. The instruments that follow – GPU-backed lending, compute escrow for training runs, capacity insurance against delivery failure – do not exist.
The most immediate and concrete opportunity is compute financing. An AI startup that has signed a six-month capacity agreement with a neocloud holds, today, a commitment that serves as collateral for nothing. If that capacity is priced by OCPI and the counterparty is creditworthy, the contract becomes financeable. A lender that uses OCPI as its pricing oracle, underwrites loans against signed capacity agreements, and converts GPU commitments into working capital would be the first institution of its kind – and it would target precisely the companies that are compute-constrained not because they can't access GPUs but because they can't pay for them upfront.
Building that requires relationships with neocloud operators (to validate counterparty creditworthiness), a risk model for compute capacity default, and access to Ornn's OCPI data. None of those pieces require a16z to be your lead investor.
Funding$33M seed
Platform users400+ data centers, investors, AI companies
Price indexOCPI on Bloomberg Terminal
Hardware trackedH100, H200, B200, RTX 5090
Lead investora16z Crypto
Exchange partnershipIntercontinental Exchange (ICE), pending regulatory approval
Founded2025
RELATED DRILLS · 6