CME Group is aiming to launch compute futures in October this year and Architect Financial Technologies, a multinational derivatives exchange group that focuses on the AI economy and perpetual futures, has applied for regulatory approval to launch a U.S. exchange for trading futures and options on compute costs.

On 11 August CME said in a statement that it plans to launch two compute futures on 5 Octobe 2026, pending regulatory review, in partnership with Silicon Data, which provides GPU market intelligence and benchmarking and is backed by global trading firm DRW.

Don Wilson, founder and chief executive of DRW, has predicted that compute will become the largest commodity in the world. He said in a statement: “The exponential growth in spending on data centers as we move towards that reality has been hampered by the lack of a hedging vehicle.”

My thanks to LS for alerting me to this eye-popping chart from Goldman Sachs.
The Good News: There is no mechanism as powerful as the US capital markets for mobilizing investment, drawing both domestic and foreign capital.
The Challenge: All else being equal, financing needs of… pic.twitter.com/EFhQOdElTJ

— Mohamed A. El-Erian (@elerianm) August 11, 2026

Pete Keavey, global head of energy and environmental products at CME Group, said in a statement: “Compute has become the currency of the AI age, and this innovative market will bring transparency to the current and future costs that AI builders and hyperscalers need to hedge as they grow.’

Keavey continued that oil fueled the 20th century economy and evolved from spot trading into a global derivatives market. In a similar fashion, he expects futures will turn compute into a standardized, tradable commodity that will provide global businesses with a reliable, regulated venue to manage price risk.

Carmen Li, chief executive of Silicon Data, said in a statement that companies buying the same graphics processing unit (GPU) capacity can pay wildly different prices due to the lack of a benchmark. She added: “Compute futures give the market something it’s never had: a public, tradable reference price for the resource every AI system runs on. That turns compute from something enterprises negotiate blindly into a market they can actually plan around.”

I've been asked "who is the buyer of compute?" at least 50 times in the last week. Here's a simple diagram:

Futures = lock in a price before you actually buy/sell compute

Options = protect against the downside/upside risk in prices (= insurance) pic.twitter.com/ZWmz54hE2c

— Ish (@Ishanee) July 2, 2026

Intercontinental Exchange also said in May this year that it plans to launch a suite of GPU compute futures contracts based on Ornn’s Compute Price Index (OCPI), subject to regulatory approval. OCPI tracks live-traded spot prices for GPU compute across major hardware types and the firm said it is the first compute index to be built only from printed transactions.

Kush Bavaria, co-founder and chief executive of Ornn, said in a statement that compute has grown into a trillion-dollar market, yet it lacks the pricing and risk-transfer infrastructure that every other major commodity relies on. Bavaria added: “Listing futures on ICE puts the risk-transfer layer in front of the institutional buyers and operators who need it most.”

American Innovation Exchange

CME and ICE are incumbents in derivatives markets but on 28 May 2026 Architect said it had acquired IMX Health, a U.S. designated contract market authorized by the Commodity Futures Trading Commission. Following the acquisition, Architect intends to launch the American Innovation (AI) Exchange, pending regulatory review.

Brett Harrison, founder and chief executive of Architect, told Markets Media that in traditional commodities markets, the annual notional turnover in the futures market is typically a multiple of the annual production of the underlying spot commodity. For example, electricity derivatives trade approximately 2.5 times the amount of electricity that is produced each year, with the equivalent of between 15 and 18 times in crude oil markets. Harrison said: “My guess is that compute will be somewhere between electricity and oil.”

This year capital expenditure on compute is expected to be between $600bn and $800bn, and projected to reach $2 trillion by 2030, according to Harrison. Therefore, he estimated that compute futures could reach a notional value of $10 trillion per year by the end of the decade.

“This is enormous, so it could easily be one of the largest commodities in futures markets,” he added.

How do you size the US compute futures market? Looking at ratios of annual derivatives volume to annual production for a variety of commodities, the answer is somewhere between power and oil:

US power: 2.5
Crude oil: 26
Aluminum: 29
Soybeans: 34
Copper: 56
Gold: 86,014 pic.twitter.com/SsP6fzfBJ7

— Brett Harrison (@BrettHarrison) July 17, 2026

 

American Innovation Exchange aims to be the first CFTC-regulated exchange for futures and options trading on compute costs tied to multiple GPU vendors and models, as well as other inputs in the artificial intelligence supply chain. In addition to enabling capital-efficient hedging through portfolio margining compute derivatives, the AI Exchange will also offer cross-margining on related metals, energy, and power derivative contracts.

A Guide to the US Financialization of Compute

We frequently receive questions about compute trading from neoclouds, GPU-as-a-service providers, data center operators, training and inference companies, energy companies, HFTs, brokerages, investment banks, FCMs, CTAs, RIAs, ETF… pic.twitter.com/TLTRnqPAOS

— Brett Harrison (@BrettHarrison) June 29, 2026

Subsequently in July Architect partnered with Compute Desk to launch ComputeConnect, which it said was the U.S. financial industry’s first compute exchange-for-physical network.

The partnership builds on Architect’s collaboration with Compute Desk on the AI Exchange. ComputeConnect will enable American Innovation Exchange customers to convert compute futures positions into real GPU capacity, providing physical settlement for the exchange’s centrally cleared futures markets and decreasing overall fragmentation of capacity delivery.

Architect said its CFTC-regulated compute futures will provide the market’s first standardized instrument to all participants’ financial exposure, while ComputeConnect augments these futures with physical delivery capabilities, establishing compute as a bankable, financeable asset.

There have been concerns that compute markets are not heterogeneous, like most commodity markets. On LinkedIn Harrison gave the example that an “H100 hour” could represent many different goods: SXM or PCIe, spot or reserved, hyperscaler or neocloud, US or Asia. However, he said a hedge does not need to be perfect but only needs to remove enough risk at scale to be worth its initial cost.

“The success of US compute futures/options markets primarily depends on CFTC-regulated exchanges’ agility and competency at collaborating with index providers native to chip configurations, neocloud procurement, and timeseries interpolation on a continuous basis,” he added. “Designing a futures contract that minimizes basis risk and maximizes liquidity formation is an antecedent requirement.”

Harrison also gave Markets Media the example  hat there are 130 different places to draw natural gas around the U.S, but natural gas futures use Henry Hub as the standard. Similarly, there are thousands of different electricity markets but there are also standardized futures contracts.

“There is an open question around whether compute should follow the oil or electricity model in derivatives, and our view is that it will be somewhere in the middle,” he added.

Once compute futures gain sufficient open interest, Architect aims to  launch options. Harrison also highlighted that a number of ETF issuers have already filed preliminary prospectuses for futures on compute, so he expects the market to expand further.

Although Architect faces competition from CME and ICE in compute futures, he argued that the firm is small and nimble. Harrison said: “We can build products and services that are directly addressing the needs of actual compute buyers, sellers and hedgers and not just financial speculators, hedge funds and retail.”

In addition, because chip technology advances so quickly Harrison claimed that it is important for a company to be fully dedicated to this sector as its flagship product in order to keep pace. He added: “That is where we think we will have a competitive advantage.”

He also believes that the launch of ComputeConnect, which provides an exchange-for-physical network, is an important differentiator.

“We think having both a standardized cash market and a bilateral compute physical compute market will be key in building out the full derivatives complex,” Harrison added.

AI infrastructure financing

At CME the Silicon Data H100 Rental Index Futures and Silicon Data B200 Rental Index Futures will track indexes measuring hourly rental GPU costs published by Silicon Data. Each contract will represent a month’s worth of rent for the Nvidia H100 and the next-generation Nvidia Blackwell B200, respectively.

Jensen Huang, founder and chief executive of NVIDIA, said on X on 10 August 2026 that NVIDIA AI factory compute is becoming an investable asset class.

https://t.co/pppEbPVtfE

— Jensen Huang (@JensenHuang) August 10, 2026

“Every industrial revolution has been built on infrastructure: electricity, transportation, communications and computing, with every buildout enabled by external financing,” added Huang. “AI factories are the infrastructure of the intelligence era.”

On 10 August 2026 NVIDIA announced partnerships with financial institutions – Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR – to establish independent financing platforms designed to mobilize over $500bn of third-party capital to support the buildout of AI infrastructure over time. Huang described this as the “beginning of an open capital market for AI infrastructure.”

He argued that NVIDIA compute is not just a chip but is a complete AI factory platform including accelerated computing, networking, systems software, AI frameworks and a global developer ecosystem. According to Huang the value of NVIDIA AI factories is not fixed at installation as they continuously improve their output; the installed base remains productive well beyond its initial depreciation period; and the same standard architecture serves a deep, growing global market of AI workloads.

“These are the characteristics of an investable infrastructure asset: it produces revenue, serves a broad market, improves in performance over time and can be redeployed,” he added.

Compute Exchange

Silicon Data’s Li is also chief executive of Compute Exchange. On 8 July 2026 Compute Exchange said it launched a dedicated hardware marketplace for used and refurbished GPUs to help organizations source physical AI hardware through a single, market-neutral platform and provide pricing transparency.

“For many organizations, used and refurbished hardware offers the fastest and most economical path to expanding AI infrastructure,” she added. “We see the secondary GPU market as a natural step toward making AI infrastructure more efficient, transparent, and accessible.”

The new marketplace comes as Compute Exchange is seeing growing demand for older generation NVIDIA GPUs, particularly H100s and A100s, as enterprises, cloud providers, AI startups, and infrastructure operators look to expand capacity without purchasing the latest hardware.

Compute Exchange performs machine-level testing and provides the required specifications for the chips to match buyers and sellers, who then negotiate the price and handle the physical delivery.

Li told Markets media: “As the  market is becoming more financialized, people want hedges against rising prices because of the investment they need to make.”