On August 19, 2026 the Commodity Futures Trading Commission issued a Request for Comment on the Listing of Compute Derivatives Contracts, release number 9286-26. It was published in the Federal Register on August 21, 2026 under document number 2026-17163, categorized as a Proposed Rule with the action listed as “Request for comment,” and comments are due on or before October 20, 2026. The legal status is worth stating exactly: this is not a rule, not a proposed rule text, and not an approval of any contract. It is the stage at which an agency asks the market what a future framework should address, which is also the stage at which comments carry the most weight.
The Commission says it is seeking public responses “to better inform its understanding and oversight of derivatives markets in compute.” Four areas are named: the size, liquidity, and other considerations of compute cash markets; market oversight and manipulation concerns; customer protection; and perpetual compute futures. The request also invites comment on all aspects of compute markets generally. The inclusion of perpetual futures is notable, since perpetuals are a crypto-native contract structure without a fixed expiry, and asking about them in the same breath as a physical-input commodity signals that the CFTC expects compute contracts to be designed by people from both worlds.
The framing comes from CFTC Chairman Michael S. Selig, who is quoted saying, “America cannot win the AI race without a robust derivatives market for compute,” and, “Just as American markets helped establish the gold standard for trading the commodities that powered the industrial economy, we will do the same for the commodity that will power the intelligence economy.” He calls the request “the first step toward establishing clear rules of the road for American compute markets.” Nothing in the release names an exchange, a contract specification, a settlement index, or a unit of account, and it does not mention data centers, GPUs, or specific chip vendors.
For a business or technical leader, this is a signal about what compute is becoming rather than about any immediate compliance burden. Treating GPU hours as a hedgeable commodity would give buyers a way to lock in cost and sellers a way to monetize idle capacity, and it would eventually produce something the industry currently lacks entirely: a public forward price curve for compute. That would change how AI capital expenditure is justified and how compute contracts are negotiated. The unproven parts are large. The cash market is opaque, heavily bilateral, and dominated by a handful of sellers, which is exactly the profile regulators associate with manipulation risk. Compute is also not fungible in the way gold or oil is, since a contract must specify chip generation, interconnect, region, and availability window before it can settle honestly. A request for comment is the beginning of that argument, not the resolution of it, and the practical test will be whether a designated contract market files a contract that survives self-certification.