Exchanges race to establish financial benchmarks for AI computing power
Prediction market operator Kalshi has joined traditional rivals in constructing forward curves for graphics processing unit rental costs, marking a rapid shift toward treating artificial intelligence infrastructure as a standardised financial asset.
Kalshi has constructed a forward curve tracking the future price of computing power, joining a growing cohort of exchanges attempting to financialise artificial intelligence infrastructure. The prediction market platform uses weekly and monthly event contracts tied to compute prices, projecting up to a year ahead. An algorithm then weaves these individual contracts into a unified curve that can act as a reference point for derivatives.
“We are using prediction markets to build the forward curve, which will provide the market a view of what compute costs will be in the future for different grades and time-frames of GPUs,” stated Udesh Jha, Kalshi’s chief risk officer. Forward curves are traditionally foundational to commodity markets, plotting expected prices for assets like natural gas or crude oil. The emergence of such a tool for graphics processing unit rentals highlights the rapid commoditisation of AI infrastructure.
Kalshi is not acting alone in this emerging sector. In May, CME Group announced plans for compute futures in partnership with Silicon Data, creating contracts tied to an index of hourly high-end GPU rental costs. Shortly thereafter, Intercontinental Exchange revealed a collaboration with Ornn to introduce its own cash-settled compute futures.
The methodology separates Kalshi from its larger institutional competitors. While CME and ICE are developing traditional futures contracts that must navigate regulatory approval processes, Kalshi leverages its existing prediction market framework. This allows the exchange to construct the curve from event contracts that are already actively trading.
This financial engineering responds to a massive underlying economic shift. Artificial intelligence infrastructure expenditure is projected to reach trillions of dollars over the next decade. Currently, entities purchasing and selling GPU capacity lack a standardised mechanism to hedge against severe price volatility.
The existing compute market remains highly fragmented. Cloud providers, data centre operators, and specialised brokers currently price capacity through opaque bilateral agreements. A reliable forward curve would offer buyers and sellers a transparent, shared expectation of future pricing, establishing the necessary foundation for effective risk management.
The outcome of this competition will shape the future of technology finance. Whichever exchange captures the most liquidity will likely establish the definitive industry benchmark, mirroring the Brent and WTI duopoly that governs global energy markets. For an asset class that barely existed two years ago, this financial infrastructure is materialising at an unprecedented pace.
