Kalshi builds forward curve for computing power as exchanges race to financialize AI infrastructure

Kalshi Constructs Forward Curve Using Prediction Markets

Kalshi has built a forward curve for GPU compute costs, marking a significant step in the commoditization of AI infrastructure. The prediction market platform uses weekly and monthly event contracts related to compute prices, extending up to a year into the future1

1

. An algorithm stitches those contracts into a single curve that can serve as a reference for derivatives and other financial instruments. "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," Udesh Jha, Kalshi's chief risk officer, told Bloomberg1

1

.

Racing Against CME and ICE to Define the Benchmark Contract

Kalshi is not alone in attempting to turn computing power into a tradable commodity. CME Group announced compute futures contracts in May, partnering with Silicon Data to build contracts linked to an index tracking the hourly cost of renting high-end GPUs1

1

. Days later, Intercontinental Exchange (ICE) revealed plans to team with Ornn to launch its own cash-settled compute futures1

1

. The race among these exchanges will determine which curve becomes the industry benchmark contract, much as competing oil contracts settled into the Brent and WTI duopoly that still defines energy markets1

1

.

Strategic Advantage Through Existing Event Contracts

Kalshi's approach differs from its larger rivals in one important respect. While CME and ICE are building traditional futures contracts that require regulatory approval, Kalshi is using its existing prediction market framework to construct the curve from event contracts that are already trading1

1

. Jha called it a key enabler for hedging, risk management, and speculative activity alike. This allows companies to anticipate where the cost of compute is headed and potentially lock in prices with providers before spikes occur2

2

.

Unprecedented Demand Drives Infrastructure Financialization

The underlying dynamic driving all three efforts stems from explosive demand for GPU capacity. AI infrastructure spending is projected to reach trillions of dollars within the next decade, yet companies buying and selling GPU capacity have no standardised way to hedge against price swings1

1

. Former Intel CEO Pat Gelsinger recently told CNBC that demand is "almost unlimited," while a recent report from Apollo Global Management described current compute capacity as "effectively sold out"2

2

. This bottleneck has caused GPU rental rates to climb faster than new capacity becomes available, creating volatility across cloud providers, data centre operators, and GPU brokers who price capacity through bilateral deals with little transparency1

1

.

Energy Constraints Compound Pricing Pressures

The pressure on computing power extends beyond simple availability. Recent research suggests that agentic AI tools consume up to 136.5 times more energy per query than most generative AI models2

2

. This energy intensity compounds the challenge facing AI labs and companies relying on compute-intensive workloads. A functioning forward curve for GPU compute costs gives buyers and sellers a shared view of where prices are headed, providing the foundation on which hedging and risk management are built1

1

. For an asset class that did not exist two years ago, the financial infrastructure is assembling remarkably fast, suggesting that the future price of computing power will become as trackable and tradable as traditional commodities1

1

.