TL;DR
ICE, the owner of the New York Stock Exchange, is partnering with index provider Ornn to launch cash-settled futures contracts tied to GPU computing costs. The move comes days after rival CME Group announced its own compute futures, signalling that Wall Street is racing to turn AI computing power into a standardised, tradable commodity.
Intercontinental Exchange, the parent company of the New York Stock Exchange, is preparing to launch futures contracts tied to the cost of computing power, marking the latest sign that Wall Street sees AI infrastructure as the next great commodity market.
ICE announced on Monday that it will team with Ornn, a financial-infrastructure firm whose index products track GPU computing costs in real time, to develop the new contracts. The futures will be US dollar-denominated, cash-settled, and referenced against Ornn’s indexes covering a variety of major GPU types. The plans remain subject to regulatory approval.
What ICE and Ornn are building
The partnership pairs one of the world’s largest exchange operators with a startup that has quietly built the plumbing for compute price discovery. Ornn, formally Ornn AI Inc, publishes the Ornn Compute Price Index, which tracks live traded spot prices for GPU compute across hardware types including Nvidia’s H100, H200, and B200 chips. The index, now available on the Bloomberg Terminal, draws on real transaction data from live GPU markets and has attracted more than 400 data centre operators, investors, and AI companies to its platform.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
Trabue Bland, senior vice president of futures markets at ICE, framed the move as a response to a market that has outgrown its informal pricing mechanisms. The compute market, he said, is “in desperate need of a globally accepted pricing mechanism and risk management tool” as AI shifts from research labs to becoming a central driver of the global economy.
The contracts will settle in cash rather than through physical delivery, a structure familiar from energy and financial futures. For AI companies planning large model training runs or cloud providers locking in capacity, the instruments would offer a way to hedge against the kind of volatile compute costs that have accompanied Big Tech’s $650 billion capex surge in 2026.
A two-horse race with CME
ICE is not alone in spotting this opportunity. CME Group, the world’s largest derivatives exchange, announced its own compute futures contracts on 12 May, partnering with Silicon Data to build products based on daily GPU benchmark rental rates. CME’s contracts will reference the Silicon Data H100 Rental Index, which tracks the cost of renting high-end GPUs used for AI training workloads.
The fact that two of the world’s most established futures exchanges have moved on compute within days of each other signals that institutional conviction in compute-as-commodity has reached a tipping point. It mirrors the early days of energy futures in the 1980s, when competing exchanges raced to establish benchmark contracts for crude oil and natural gas. The exchange that captures the most liquidity early on will likely set the reference price for the industry, just as ICE Brent and CME WTI did for oil.
The competitive dynamic also extends beyond the big two. Architect Financial Technologies partnered with Ornn in January to launch exchange-traded perpetual futures on GPU and RAM prices through its AX platform, and prediction market Kalshi has offered contracts allowing users to wager on Nvidia GPU compute prices. But ICE and CME bring something the newer entrants lack: deep institutional liquidity, regulatory credibility, and the clearing infrastructure that large-scale GPU-as-a-service providers and their customers will demand.
Why compute needs a futures market
Kush Bavaria, co-founder and CEO of Ornn, put the scale of the problem bluntly. Compute, he said, “has grown into a trillion-dollar market, yet it still lacks the pricing and risk-transfer infrastructure that every other major commodity relies on.”
That gap has real consequences. GPU rental prices have been wildly volatile, with Ornn’s own index showing the Nvidia Blackwell spot rental price surging 48% between mid-February and mid-April 2026, from $2.75 to $4.08 per GPU-hour. For AI companies whose training runs can cost tens of millions of dollars, that kind of price swing can blow through budgets with little warning. Cloud providers, data centre operators, and the lenders financing billions of dollars in AI infrastructure buildouts face similar exposure.
A functioning futures market would allow these participants to lock in forward prices, transfer risk to willing counterparties, and plan capital expenditure with greater certainty. It would also generate transparent price signals that the broader market currently lacks, giving investors, analysts, and policymakers a clearer view of where compute costs are heading.
Broader implications for the AI economy
The emergence of compute futures reflects a deeper structural shift. As AI moves from an experimental technology to core economic infrastructure, the inputs that power it are being financialised in much the same way that energy, metals, and agricultural products were in previous decades. The surging demand for advanced semiconductors has already reshaped chip supply chains and driven record capital investment across the technology sector.
Futures contracts add a new layer to this ecosystem. They create standardised benchmarks that can underpin lending decisions, insurance products, and investment strategies tied to AI infrastructure. A bank financing a new data centre, for instance, could use compute futures to assess the facility’s projected revenue against forward GPU prices, much as energy lenders use oil futures to evaluate drilling projects.
There are complications, of course. Unlike oil sitting in a tank, compute is what traders call a flow commodity, one that is consumed in real time and cannot be stored. Ornn has addressed this by designing its futures with Asian-style settlement, meaning contracts settle on the arithmetic average of daily index values across the contract’s tenor rather than on a single expiry-day price. This structure aligns the financial instrument with the way compute is actually purchased and consumed.
Whether ICE or CME ultimately captures the lion’s share of this market will depend on liquidity, the breadth of GPU types covered, and which index providers gain the most institutional trust. But the direction of travel is clear. Computing power, the resource that underpins everything from energy-hungry AI data centres to autonomous vehicle development, is being transformed from a bespoke procurement headache into a standardised, tradable financial asset. For an industry accustomed to negotiating GPU access through opaque, bilateral deals, that is a significant change.
Stephan is the sports journalist for the Maple Grove Report.
