Kalshi, the prediction markets exchange, has introduced a forward curve that tracks the future price of computing power, marking a significant step toward financializing GPU rental costs. This tool uses weekly and monthly event contracts related to compute prices, extending up to a year into the future. An algorithm stitches those contracts into a single curve that can serve as a reference for futures, options, and other derivatives. The move positions Kalshi alongside established exchanges like CME Group and Intercontinental Exchange (ICE) in the race to create a standardized benchmark for AI infrastructure costs.
The Rise of Compute Commoditization
The concept of treating computing power as a commodity has gained traction as artificial intelligence workloads have exploded. High-performance GPUs, such as Nvidia's H100 and A100 series, are in high demand, with rental rates fluctuating dramatically based on supply and demand dynamics. The market for GPU capacity remains fragmented, with cloud providers like AWS, Google Cloud, and Microsoft Azure offering different pricing models, alongside specialized GPU brokers and data center operators. This fragmentation leads to opacity and makes it difficult for buyers and sellers to hedge against price volatility. A forward curve provides a shared view of where prices are headed, which is the foundation for hedging and risk management.
Kalshi&x27;s Prediction Market Approach
Kalshi&x27;s approach differs from its larger rivals in one important respect. CME and ICE are building traditional futures contracts that require regulatory approval, while Kalshi is using its existing prediction market framework to construct the curve from event contracts that are already trading. Udesh Jha, Kalshi’s chief risk officer, told Bloomberg: “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.” This method leverages the wisdom of crowds, as traders place bets on future price levels, generating a continuous stream of price discovery. The algorithm then aggregates these bets into a smooth curve, akin to those used in oil and gas markets.
Prediction markets have a history of aggregating information effectively, often outperforming traditional surveys and expert opinions. Kalshi, founded in 2018, is a regulated exchange that allows trading on event outcomes such as elections, economic indicators, and now commodity prices. By using event contracts, Kalshi avoids the need to design a full-fledged futures contract from scratch, instead letting market participants define the price expectations. This flexibility allows for rapid iteration and customization, which is valuable in a nascent market like GPU compute.
The Competition: CME and ICE
Kalshi is not the only exchange moving on compute. CME Group announced compute futures in May, partnering with Silicon Data to build contracts linked to an index tracking the hourly cost of renting high-end GPUs. Days later, Intercontinental Exchange said it would team with Ornn to launch its own cash-settled compute futures, making at least three serious entrants in the race to establish the benchmark contract for AI computing power. CME, the world's largest derivatives exchange, brings deep liquidity and a vast network of institutional participants. Its compute futures are designed to be cash-settled against a benchmark index, similar to its successful energy and agricultural commodity contracts. ICE, known for its dominance in energy and soft commodities, offers a competing product that leverages its expertise in creating transparent pricing benchmarks.
The underlying dynamic driving all three efforts is the same. AI infrastructure spending is projected to reach trillions of dollars within the next decade, and the companies buying and selling GPU capacity have no standardised way to hedge against price swings. GPU rental rates have been volatile, influenced by new chip releases, data center construction, and shifts in AI model demand. For example, after Nvidia's quarterly earnings pointed to supply shortages, spot rental prices for H100s surged by 30% in some regions. Without a forward curve, participants are forced to rely on opaque bilateral deals or spot pricing, exposing them to significant risk.
Why a Forward Curve Matters
A forward curve does more than just provide prices; it enables a range of financial activities. Hedgers, such as cloud providers or AI startups, can lock in compute costs for future projects, making budgeting more predictable. Speculators can bet on price movements, adding liquidity to the market. Moreover, the curve can underpin structured products like swaps and options, further deepening the market. Historically, every commodity—from crude oil to corn to electricity—evolved from physical spot trading to futures and options. Compute power is following a similar trajectory, driven by the need for price transparency and risk management in a fast-growing sector.
The creation of a forward curve also attracts institutional investors who view compute as an asset class. Pension funds, endowments, and hedge funds are increasingly exploring alternative assets that offer uncorrelated returns. If compute futures gain traction, they could provide a way to gain exposure to the AI boom without directly buying GPUs or investing in volatile tech stocks. This could bring billions of dollars in new capital into the market, further stabilizing prices and encouraging investment in infrastructure.
Technical Implementation and Challenges
Building a reliable forward curve for compute is not trivial. Unlike oil, which has established delivery points and quality grades, GPU compute is highly heterogeneous. Different GPU models (e.g., H100 vs. A100), cloud regions, contract lengths, and service levels all affect price. Kalshi's algorithm must weigh multiple event contracts across different parameters to produce a coherent curve. The company has not disclosed full details of its methodology, but it likely uses techniques like spline interpolation and liquidity-weighted averaging. CME and ICE face similar challenges, but they benefit from decades of experience in constructing commodity indexes.
Another challenge is liquidity. Prediction markets can suffer from thin trading, especially for far-dated contracts. Kalshi will need to incentivize market makers to provide continuous quotes, possibly through fee rebates or revenue sharing. CME and ICE can rely on their existing market maker programs, but they also need to attract participants new to compute derivatives. Education and outreach will be critical to build understanding of how these instruments work.
Regulatory oversight also varies. Kalshi is regulated by the Commodity Futures Trading Commission (CFTC) as a designated contract market (DCM), allowing it to list event contracts. However, its prediction market model is relatively novel, and the CFTC has been cautious about expanding the range of allowed events. CME and ICE have established compliance frameworks for futures and options, making it easier to launch new products. Nonetheless, all three exchanges must ensure that their contracts cannot be manipulated and that pricing indices are transparent.
Broader Implications for AI Infrastructure
The financialization of compute has profound implications beyond trading desks. It could accelerate investment in data centers, as developers can use futures to lock in revenue streams and hedge construction costs. Utilities and energy companies might use compute derivatives to manage the power demands of AI workloads. Even chip manufacturers like Nvidia and AMD could benefit from more predictable demand signals. In the long run, a standardized benchmark could reduce the cost of capital for AI projects, making the technology more accessible to startups and researchers.
The race between Kalshi, CME, and ICE mirrors the historical competition in oil markets, where Brent and WTI emerged as the dominant benchmarks. Similarly, the compute forward curve that attracts the most liquidity will likely become the industry standard. Early movers have an advantage, but success will depend on user trust, low fees, and robust infrastructure. Kalshi's use of prediction markets gives it a nimble advantage, while CME and ICE bring scale and reputation. The next year will be crucial as these three players battle for market share.
For an asset class that did not exist two years ago, the financial infrastructure is assembling remarkably fast. The forward curve is not just a tool for traders; it is a sign that compute is maturing from a scarce resource into a fungible commodity. Whether you are a hedge fund manager, a cloud architect, or a policymaker, the development of this market will shape the future of AI. As more participants join, the curve will become more accurate, providing a vital compass for one of the most important inputs of the digital age.