CME Group to launch 2 compute futures as AI power becomes tradable

CME Group's Oct. 5 launch of two compute futures contracts turns AI processing power into a hedgeable commodity for the first time.

AI computing capacity is becoming a tradable commodity, with CME Group set to launch two futures contracts on Oct. 5 that let companies hedge GPU rental costs.

"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, chief executive officer of Silicon Data, said. "Compute futures give the market something it's never had: a public, tradable reference price for the resource every AI system runs on."

The contracts, listed on NYMEX and pending regulatory review, track hourly rental prices for Nvidia's H100 and newer Blackwell B200 chips based on indexes published by Silicon Data, the GPU market intelligence firm backed by trading house DRW. Each contract represents one month's rent for the respective chip.

The launch comes as Wall Street builds new ways to finance the AI infrastructure buildout. Nvidia is working with some of the world's largest asset managers on an effort that could channel as much as $500 billion into AI infrastructure, and compute futures add a hedging layer to that emerging financial structure.

Compute Joins Oil as a Hedged Commodity

"Compute has become the currency of the AI age, and this 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 at CME Group, said. He drew a parallel to oil, which evolved from spot trading into a global derivatives market over the 20th century.

Explosive demand has driven sharp swings in compute prices, exposing a gap in the risk-management toolkit for companies building AI infrastructure. The futures close that gap, letting AI developers and hyperscalers lock in costs and giving investors a window into future AI spending.

The H100 has become the workhorse of the AI buildout, powering the training runs behind large language models from OpenAI, Anthropic and Google. The B200, Nvidia's successor chip, is expected to take over as demand for inference — running trained models — accelerates. Pricing for both has swung sharply as supply shortages and hyperscaler buying sprees collide.

The contracts will be listed and subject to the rules of NYMEX, extending CME Group's benchmark franchise — which spans interest rates, equity indexes, foreign exchange, energy, agricultural products and metals — into a new asset class. Silicon Data's indexes, which measure hourly rental costs, become the reference price that buyers and sellers trade against.

Who Wins From a Tradable Compute Market

Rather than investing directly in data centers, chips or the companies building them, investors could gain exposure to the price of the underlying computing capacity itself. AI developers and data-center operators could use the contracts to hedge their costs or revenues.

For Nvidia, whose H100 is central to today's AI buildout and whose Blackwell B200 is the successor, a transparent rental market could smooth the boom-bust cycles that have characterized GPU pricing. The company is simultaneously working with some of the world's largest asset managers on a separate effort that could channel as much as $500 billion into AI infrastructure, according to CNBC.

The launch also gives Silicon Data, backed by global trading firm DRW, a central role in pricing the AI economy. Its benchmarks make the price real; CME makes it tradable, in the words of CEO Carmen Li — turning compute from something enterprises negotiate blindly into a market they can plan around.

The futures curve could also become a leading indicator for AI infrastructure investment, much as oil futures signal expectations for energy demand. As hyperscalers lock in compute costs months ahead, the contracts give the market a way to price future AI spending — a barometer for the AI economy as much as a hedging tool.

This article is for informational purposes only and does not constitute investment advice.