CME Group to Launch AI Compute Power Futures Contracts

A new commodity market for AI infrastructure

CME Group is preparing to launch the first futures contracts tied to the cost of artificial intelligence computing power, treating GPU capacity as a tradable commodity alongside oil, electricity, and agricultural products. The exchange plans to introduce two compute futures contracts on October 5, pending regulatory approval, through a partnership with Silicon Data.

The contracts will be based on the rental cost of Nvidia's H100 and newer Blackwell B200 graphics processing units. Each contract will represent one month's rent for an Nvidia H100 GPU, with pricing derived from Silicon Data indexes that track hourly GPU rental rates across the market.

Why it matters

The absence of transparent pricing in the GPU rental market has meant companies purchasing identical computing capacity often pay vastly different amounts with no reliable way to compare deals. Futures contracts create a public benchmark that brings price discovery to an opaque market, while giving AI developers and data center operators tools to hedge against volatile infrastructure costs. This marks computing power's evolution from a procurement expense into a financialized asset class with standardized pricing mechanisms.

Addressing market opacity

"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," Carmen Li, CEO of Silicon Data, said in a statement. "They will now have a benchmark to check that against."

Li added that compute futures provide "a public, tradable reference price for the resource every AI system runs on" — something the market has lacked until now.

Part of a broader financial ecosystem

The futures launch arrives as Wall Street develops multiple channels for financing AI infrastructure expansion. Nvidia has been collaborating with major asset managers on an initiative that could direct up to $500 billion toward AI infrastructure investments.

Compute futures add a distinct layer to this emerging financial architecture. Rather than investing directly in physical data centers, semiconductor manufacturing, or technology companies, market participants can gain exposure to the price of computing capacity itself. AI developers could use the contracts to lock in future costs, while data center operators might hedge against revenue fluctuations tied to GPU rental rates.

The standardization of compute pricing through exchange-traded contracts could accelerate the maturation of AI infrastructure markets, providing the price transparency and risk management tools that characterize established commodity sectors.

These details were first reported by CNBC.