just now
CME Group Takes On ICE in the Race to Set the World's Compute Benchmark
CME Group and GPU market intelligence and benchmarking firm Silicon Data have announced plans to launch two Compute futures contracts on 5 October 2026, subject to regulatory review. The contracts will give AI developers, cloud operators and investors a regulated way to hedge the cost of processing power, an input that has moved from a line item in a technology budget to one of the largest variable costs in the global economy.
Silicon Data H100 Rental Index Futures and Silicon Data B200 Rental Index Futures will track indices measuring hourly GPU rental costs published by Silicon Data, which is backed by global trading firm DRW. Each contract represents a month's worth of rent for the Nvidia H100, the chip at the centre of the current AI build out, and the next generation Nvidia Blackwell B200 respectively. The contracts are cash settled against the indices rather than involving any physical delivery of hardware, and will be listed under the rules of NYMEX.
An input that has never had a public price
Renting AI compute has until now been a matter of private negotiation, with prices varying widely by provider, contract length and demand conditions.
"For years, two companies buying the exact same GPU capacity could pay wildly different prices with no way to know who got the better deal. They will now have a benchmark to check that against." - Carmen Li, Chief Executive Officer, Silicon Data
Li added that the contracts give the market a public, tradable reference price for the resource every AI system runs on, turning compute from something enterprises negotiate blindly into something they can plan around.
CME Group is pitching the launch in commodity terms, drawing the comparison with the development of oil from an opaque spot business into a global derivatives complex.
"Compute has become the currency of the AI age, and this innovative market will bring transparency to the current and future costs that AI builders and hyperscalers need to hedge as they grow." - Pete Keavey, Global Head of Energy and Environmental Products, CME Group
The CME vs ICE race to set the reference price
CME Group is not the only venue chasing this market. Intercontinental Exchange announced in May that it would launch a suite of GPU compute futures with Ornn, based on the Ornn Compute Price Index, which tracks live traded spot prices for GPU compute across major hardware types. ICE followed in early July with plans alongside NATIVX for cash settled GPU compute futures based on the COIL index, which measures tokenised, energy normalised compute and is expected to list beside ICE's existing power and natural gas contracts. Elsewhere, Architect Financial Technologies has listed perpetual futures on GPU and RAM prices, and prediction market Kalshi offers contracts on the direction of Nvidia compute prices.
The venue that captures early liquidity is likely to set the global reference price for compute, much as ICE Brent and CME WTI did in oil, while fragmented rival contracts risk splitting activity across platforms and leaving none of them deep enough to be useful.
The five largest hyperscalers are set to have combined capital expenditure of roughly $775bn to $800bn in 2026, around three times the level deployed in 2024, with approximately 75% of it AI specific. Investors have been less comfortable than the operators, with three of the four hyperscalers losing market value after their most recent earnings calls on concerns over the size of those commitments.
Until now, a firm wanting exposure to, or protection from, the cost of AI infrastructure has had to express it through proxies such as chipmaker equity or cloud provider earnings. A listed contract on rental cost gives a direct hedge instead, and once the contracts are live, trading platforms will be able to add them to their product suites.
Will compute futures attract investors from the retail market?
Found this interesting? Become a member of LiquidityFinder, join the community and join the discussion - join here.
