CME Launching Futures Market For AI Compute

US derivatives exchange CME Group and index provider Silicon Data have teamed up to create a futures market for computing power, a key source driving the AI boom. Bloomberg reports that the futures will help traders, financial firms, AI builders and cloud providers manage volatility and price swings. Indexes from market-intelligence firm Silicon Data will help underpin the products. The project is still pending regulatory review.

“As the backbone of the digital economy, compute is the new oil of the 21st century,” CME CEO Terry Duffy said in the statement. “Every AI model trained, every transaction cleared and every byte of data processed runs on compute, which is becoming a fast-emerging asset class in its own right.”

CME’s addition of compute futures signals a broader shift to make the asset tradable like other commodities. Creating a futures market can help make the costs more transparent and raise overall market efficiency. 

Computing power, better known as “compute”, has seen soaring demand as AI companies use it to power their systems. BlackRock CEO Larry Fink said last week that a new asset class will likely be buying futures of compute given the shortage and high demand.

Futures contracts, which give investors the ability to bet on the future value of a commodity such as oil or metals at a certain date, are traded on exchanges such as Chicago-based CME’s, and require a brokerage account that’s approved to trade futures.

The need for compute power has been growing as AI technologies have scaled. But until now, it’s been difficult to hedge against price swings and other costs.

The index from Silicon Data, which is backed by trading firm DRW Holdings, provides a benchmark for traders and others to follow, giving insight into the cost of goods for those building AI products or in need of graphics processing unit computing power.

Silicon Data, founded by former DRW trader Carmen Li, created daily GPU benchmarks for on-demand rental rates, giving customers insight into the cost of goods for those building AI products or in need of GPU computing power.

The company’s Silicon Data H100 Rental Index tracks the hourly cost of renting a GPU, which is the workhorse for training AI models.

A data center typically uses hundreds or even thousands of processors, which can cost thousands of dollars.