CME Group bets on computing power futures, BlackRock CEO: This is the next trillion-dollar asset

Original Author: Chamath Palihapitiya

Original Compilation: Deep Tide TechFlow

Deep Tide Guide: AI capital expenditure has first surpassed oil and gas, but the price of computing power is highly volatile, and the market has almost no hedging tools. The CME Group plans to launch computing power futures, which is a key experiment to determine whether computing power can become the next trillion-dollar asset class. This article points out that for computing power futures to be successful, two issues must first be resolved: concentration and interchangeability, which are risk variables that all investors laying out AI infrastructure must understand.

"I actually believe a new asset class will emerge, which is purchasing computing power futures. Right now, we just don't have enough computing power." ------ BlackRock CEO Larry Fink

This week, he was proven right.

CME Group, the world's leading derivatives market, announced in conjunction with Silicon Data, an industry leader in GPU market intelligence and benchmarking, plans to launch computing power futures contracts on October 5, 2026, pending regulatory review.

Why does computing power need a financial market?

In 2026, AI capital expenditure is expected to reach $765 billion, surpassing oil and gas for the first time at $681 billion. By 2031, it is expected to nearly double. Morgan Stanley predicts that the diffusion of AI in the global economy will create $40 trillion in opportunities. And this opportunity relies on a key resource: computing power.

Silicon Data's index shows that since the beginning of this year, even older generations of GPUs have seen a sharp increase in demand for computing power:

When so much capital flows into an industry, those spending money need a way to protect themselves from adverse price fluctuations.

Today, oil producers can buy futures contracts to lock in prices before delivery. If spot prices fall, the contracts can stabilize income. Buyers on the other side use the same market to limit fuel costs. Both parties strip price volatility from their businesses.

Computing power does not yet have such tools, which exposes anyone building or purchasing AI infrastructure to three types of risk exposure:

GPU rental prices are highly volatile, soaring during spikes in demand or plummeting when supply is ample or new chips are released, making it difficult for AI companies to accurately budget for their largest cost item.

Whenever Nvidia releases faster chips, the rental value of the previous generation chips declines, and the collateral behind hardware loans shrinks accordingly.

A data center takes two to three years to build, but developers have very limited means to lock in their computing power costs or revenues. Each such decision is a bet of billions of dollars.

These risk exposures create demand for computing power futures. But before this market can scale, it must confront the same two issues that have historically limited other futures markets: concentration and interchangeability.

Attempts to establish futures markets around onions, uranium, DRAM memory chips, and bandwidth have encountered one or both of these issues.

For computing power, the concentration issue is more complex. Buyers are becoming dispersed as inference demand spreads across thousands of companies running production workloads. The seller camp is broad and still growing, with new cloud vendors' revenues expected to exceed $25 billion by 2025, covering over 60 providers. However, the underlying supply remains highly concentrated, with Nvidia supplying most AI chips.

The second issue is interchangeability.

Today, computing power prices are quoted in GPU hourly rates, which is the cost of renting a single GPU for one hour. However, two GPUs of the same model may provide different computing power in one hour.

Silicon Data, in collaboration with academic partners, ran the same workload on 3,500 GPUs from 11 cloud providers. Even within the same chip model, they found significant differences. In one test, the performance difference of the H100 reached as high as 34.5%, with the largest gap in the entire study reaching 38%.

The first batch of sustainable contracts may need to define several tiers, just as energy markets use different fuels, locations, and delivery periods.

But the bigger question is, what happens if computing power futures work? Can computing power become the next asset class with a nominal trading volume of trillions of dollars and accelerate the AI economy?