Author: Zen, PANews
In an era where AI is sprinting ahead, GPUs have literally become “digital crude oil.” Yet the underlying resource that underpins a trillion-dollar AI industry is still traded in a primitive “private negotiation” era, and some companies are starting to try to make this business open and transparent.
Recently, Ornn announced it had closed a $33 million seed funding round led by a16z Crypto, with participation from Galaxy Ventures, Nordstar and SV Angel, and follow-on investments from Vine Ventures, Crucible Capital, Link Ventures and Box Group.
Founded in 2025 by two MIT graduates, Kush Bavaria and Wayne Nelms, the company was able to raise a large early-stage investment from multiple VCs not only because of its founders’ prestigious university pedigree, but more importantly because they seized the wave of frenzied AI infrastructure expansion and trained their sights on an even more foundational layer — launching a trading market around compute power.
Why the compute market needs a new pricing system
Today, GPUs, data centers, electricity and network resources have shifted from being internal engineering issues at tech companies to one of the core cost items across the entire AI supply chain.
But unlike mature commodity markets such as oil, natural gas and electricity, the compute market remains highly opaque. Prices are usually determined through private bilateral negotiations, long-term contracts are hard to exit, and there is no unified price benchmark across different GPU models, regions, data center conditions and contract durations. Ornn is the emerging project trying to tackle exactly this problem.
Ordinary cloud compute rental platforms mainly operate on a business model of directly selling GPU compute hours to AI companies. Ornn’s core positioning is to transform the right to use a specific GPU model for a certain period from a privately negotiated, long-term locked contract resource into a commodity that can be priced, traded, financed and hedged.
With the rapid advance of the AI era, compute power has long since become one of the world’s most important commodities. However, the market mechanisms surrounding compute remain rather “primitive.” Traditional capital-intensive commodity markets typically evolve price benchmarks, risk transfer tools and investable assets over time, yet the AI compute market still lacks this infrastructure. a16z Crypto believes that what Ornn is doing is pushing compute power from “private negotiation and individual contracts” toward a truly functioning market.
From price index to trading platform: building liquidity for GPU compute resources
Solving the “pricing transparency” problem is Ornn’s starting point, and the key lies in its approach: Ornn’s idea is to directly change the liquidity status quo of compute power by building a trading platform.
One of Ornn’s core products is a price index. The company previously launched the Ornn Compute Price Index (OCPI) to track market prices for GPU compute resources. According to Ornn’s website and a16z Crypto’s official blog, OCPI is not simply scraping public listing prices; it is a price benchmark based on executed or settled transactions, covering major GPU types like H100, H200, B200 and RTX 5090, and is standardized by dimensions such as hardware, region and contract duration.
One of the biggest problems in the current compute market is the potentially large gap between “listed prices” and actual transaction prices. Large AI labs, cloud service providers, data center operators and small-to-medium enterprises get very different prices, and contract terms are highly customized. In this environment, without a unified reference price, buyers can hardly tell whether they are overpaying, sellers find it hard to gauge how to price future compute resources, and financial institutions cannot properly assess the collateral value or future cash flows of GPU assets.
This April, Ornn announced that OCPI had been integrated into the Bloomberg Terminal, further opening it to institutional users. According to the company’s announcement, OCPI tracks GPU hourly rental prices in cloud and on-premises markets, covering Nvidia H100, A100, H200, B200 and RTX series GPUs. Ornn said more than 400 data center operators, investors and AI companies already use its platform to track GPU prices.
Built on top of the price index, Ornn is rolling out a second layer of products: Ornn Compute, a trading platform for the right to use GPU compute capacity. A huge amount of GPU compute resources is currently locked up in private deals and long-term contracts, and Ornn Compute aims to act as a bridge between buyers and sellers. Buyers can lock in dedicated GPU compute resources of a specific model, region and duration under a single contract; when workloads change, they can also transfer or sublet the remaining portion of their usage rights to other users, turning what would otherwise be idle compute resources back into revenue.
This is also where Ornn’s model differs from traditional cloud providers. Traditional cloud services mostly sell compute power on an on-demand basis or through long-term cloud resource contracts; Ornn is trying to give GPU usage rights their own secondary market liquidity. For small-to-medium AI startups, this means they can secure short-term GPU compute resources more flexibly without having to commit to unduly long resource obligations from day one. For data centers and small-to-medium cloud service providers, it means they can sell idle or future available GPU compute resources in a more standardized way and receive more predictable income.
ICE enters the fray, compute power may head toward derivatives markets
From the price yardstick provided by OCPI to the liquidity support of a spot trading platform, the entry of another derivatives giant turns this into financial infrastructure, laying financial tracks for the AI economy.
This May, Intercontinental Exchange (ICE), the parent company of the New York Stock Exchange, announced it plans to launch GPU compute futures contracts based on OCPI with Ornn. The ICE announcement showed these contracts will be U.S. dollar-denominated, cash-settled, and reference the OCPI family of indices covering mainstream GPU types, with actual listing subject to regulatory approval.
If a price index marks the first step in forming a market, then futures and hedging instruments are important hallmarks of a mature commodity market. ICE’s participation lifts Ornn’s narrative from a compute trading platform to compute derivatives infrastructure.
For data center operators, a decline in GPU rental prices can hurt future revenue and financing capacity; for AI companies, rising GPU prices can drive up training and inference costs. Through futures contracts based on OCPI, buyers and sellers could theoretically lock in future prices in advance and reduce business volatility.
Viewed against the broader industry backdrop, AI infrastructure competition has entered a more sophisticated phase, the core of which is how to give highly capital-intensive assets financial measurability. Data centers need financing, lenders need valuations, AI companies need budget certainty, and investors need more direct exposure to AI infrastructure. In this process, GPUs can become more than just hardware — they can become an asset that generates cash flow, serves as collateral, forms a price curve and can be traded.
Beyond this, Ornn recently extended its index business to AI token costs, launching the Ornn Token Price Indices (OTPI) this June. The OCPI mentioned above measures the input side of the AI economy — the GPU-time cost required to train and run models — while OTPI measures the output side, capturing the actual cost of tokens produced by major model developers such as Anthropic and OpenAI. Together, the two families of indices provide the market with a cost curve from compute input to AI consumer demand.
Overall, Ornn’s rise reflects the AI compute market’s shift from a scramble for resources toward financialized pricing. As GPU, data center and electricity investments become one of the biggest cost items in the AI industry, market demand for transparent prices, flexible trading and risk management tools is also growing. What Ornn is betting on is precisely the financialization opportunity behind this change.
Ornn’s rise is not merely a fundraising story; it is an epitome of the underlying logic shift in the AI industry: compute power is undergoing the metamorphosis from “heavy-asset hardware” to a “financializable asset.”
