Wall Street Is Starting to Trade AI Computing Power Like Stock

Major financial exchanges are building markets where people can buy and sell contracts based on the cost of renting computers to run artificial intelligence systems. This is new — and it suggests that AI computing has become important enough to need the same financial tools that traders use for oil, wheat, and other commodities.

CME Group, one of the world's largest futures exchanges, has partnered with a company called Silicon Data to create contracts tied to the price of GPU computing power. (GPUs are the specialized computer chips that do most of the heavy lifting in AI systems.) Think of it like a farmer locking in the price of corn before harvest — except here, tech companies are locking in the cost of renting computing power for their AI projects.

Silicon Data's index tracks what it costs to rent these powerful chips. Separately, both CME Group and Intercontinental Exchange — another major futures exchange — are preparing to launch their own contracts based on computing rental prices.

Why This Is Happening Now

When you train an AI system or use one to make predictions, you need to rent computing power from cloud companies like Amazon or Google. The prices for that power swing wildly — sometimes doubling in a few weeks if lots of companies are trying to use AI at the same time. For a business trying to plan its budget, that unpredictability is a headache.

Futures contracts solve this problem. They let a company lock in a price today for computing power they will use months from now, the same way an airline might lock in jet fuel prices to protect itself from price shocks. Both the company that needs computing power and the company that sells it benefit from knowing what the price will be.

We saw a similar pattern back in the late 1990s, when the internet was being built out and companies suddenly needed to trade bandwidth the same way they traded other resources. As a technology becomes essential and expensive, financial markets emerge to help people manage the risk.

The Global Race to Control AI Markets

China's Shanghai Futures Exchange is building its own version of this market. Instead of tracking the cost of renting chips, it is focusing on "AI tokens" — a smaller unit of measurement for the computational work that happens inside an AI model.

The Shanghai exchange's research is still early, but the effort signals how seriously both the United States and China see AI computing as a strategic asset. The way you price and trade these resources has become part of the broader competition between the two countries in artificial intelligence.

Who Makes the Rules

The U.S. Commodity Futures Trading Commission, the government agency that oversees futures markets, has set up a task force to figure out how to regulate AI-related financial products. The task force coordinates with other agencies, including the Securities and Exchange Commission, to make sure these new markets work safely.

The commission has also published guidance on how AI should be treated in financial markets, signaling that the government is paying attention as these products get developed.

Meanwhile, the appetite for trading digital assets has been growing. In May 2025, CME Group launched futures contracts for XRP, a cryptocurrency, and crypto-focused exchanges have been building similar products. These AI computing markets fit into the same broader trend.

What This Means Going Forward

Right now, if a company needs computing power for an AI project, it either negotiates a long-term deal with a cloud provider or buys power on the spot market — paying whatever the price is that day. Futures contracts would change that. Companies could lock in prices months in advance.

This matters most for smaller AI companies and research institutions that do not have the bargaining power of a giant tech firm. With futures contracts, they could plan budgets more reliably and protect themselves from sudden price jumps.

There are real challenges ahead. Unlike oil or wheat, which are standardized products, computing power varies depending on what kind of AI work you are doing. One company's computing needs might not be the same as another's. Getting the measurements right so that contracts actually work is going to be complicated.

Still, the basic pattern here is familiar. As technologies move from experimental to mainstream, they develop financial tools to help manage risk and cost. AI infrastructure is moving through that transition faster than most technologies have in the past, which creates both opportunity and uncertainty as these markets get built.

For now, traders and technology companies will be watching how these first contracts work and whether they actually help solve the real problem — making it easier and cheaper to run AI systems reliably.