The Commodity Futures Trading Commission (CFTC) published a request for comment on the listing of compute derivatives contracts on August 21, 2026, and its questions point at the product rather than at the path to market: whether a futures contract may settle to an index administered by the same firms that control supply of the underlying commodity, and whether the Commission could surveil that index at all.

The document is RIN 3038-AF77, published at 91 FR 54259, running six pages under 17 CFR parts 1 and 38, with comments due October 20, 2026. Its action line reads “Request for comment,” not a proposed rule text. The Commission’s own press release frames it as a first step toward “clear rules of the road for American compute markets.” The text does something narrower and harder. It puts 23 questions to the market, and the load-bearing ones ask whether a cash-settled compute contract can satisfy Designated Contract Market (DCM) Core Principle 3 at all.

Methodology and sources

This analysis rests on the full text of the request for comment as published in the Federal Register on August 21, 2026, read in its entirety including the 46 footnotes, rather than on summaries. Supporting primary documents are CFTC press release 9286-26 of August 19, 2026; the Commission’s ISDAFIX settlement orders of May 20, 2015 and May 25, 2016; Regulation (EU) 2016/1011 as published on EUR-Lex, including Article 19, Article 42 and Annex II; the Financial Conduct Authority’s benchmarks pages, last updated August 6, 2026; and Parts 6AA and 12 of Singapore’s Securities and Futures Act 2001, current version as at August 29, 2026. Jurisdictional scope is the United States, the European Union, the United Kingdom and Singapore. Two caveats. Comment letters filed against RIN 3038-AF77 could not be retrieved because the CFTC’s public comment portal returned an access block, so no commenter positions are characterised here. The institutional affiliation of the academic quoted at footnote 5 could not be independently confirmed and is therefore not stated.

What the request for comment actually asks

DCM Core Principle 3 is a listing gate, not a disclosure duty. Under CEA section 5(d)(3), a designated contract market may list only contracts that are “not readily susceptible to manipulation.” The Commission has never issued a rule defining that phrase for cash-settled products; instead it published the Appendix C Guidance to part 38, “Demonstration of Compliance That a Contract is Not Readily Susceptible to Manipulation,” alongside the DCM core principles final rule at 77 FR 36612 (June 19, 2012). Appendix C paragraph (c)(1) requires rules that fully describe the essential economic characteristics of the underlying commodity and how the final settlement price is calculated. Paragraph (c)(2) states that a cash-settled contract’s utility for risk management and price discovery “would be significantly impaired if the cash settlement price is not a reliable or robust indicator of the value of the underlying commodity.” The burden sits on the exchange, and it is discharged before listing, not after.

That is the frame in which the questions bite. Question 1(c) asks: “Would it be appropriate to permit trading in a derivative contract settling to a price computed from data that the Commission may not be able to observe, verify, or surveil, in whole or in part?” Question 1(b) records the Commission’s preliminary understanding that “non-public, bilateral agreements carry the majority of economic value but tend to be undisclosed and negotiated privately.” Question 2(b) asks whether anything would prevent a compute capacity provider “from manipulating a cash settlement index by adjusting a posted rate, directing capacity onto or away from a venue whose transactions the index calculation methodology treats as input data, or by executing or declining to execute transactions during the observation window.”

Question 1(f) is the structural one. It asks how the Commission should “consider whether the parties best positioned to influence the reference price are the same parties that supply capacity or contribute transactions or posted rates from which price is computed.” In a conventional cash-settled contract, the price reporting agency, the physical suppliers and the futures participants are at least partly distinct populations. In compute, on the Commission’s preliminary reading, they collapse into one. That is not a data-quality problem a better index solves; it is a conflict written into the supply structure.

Question 2(g) closes the loop. It restates the Appendix C test in full — a settlement price that is “reliable, acceptable, publicly available, and timely,” computed from a cash market “that is sufficiently liquid and not itself readily susceptible to manipulation” — asks whether any compute price series meets it, then adds the sentence carrying the most weight in the document: “What steps should the Commission take, if any, if no such cash price series is available?” A regulator expecting an answer would not need to ask what happens when there is none.

Question 1(g) sets a second, quieter trap. It asks, “Regarding DCM Core Principle 5, what would be an appropriate deliverable supply estimate methodology to evaluate the necessity and appropriateness of position limits or position accountability levels?” Deliverable supply is the classic input to position limits, and it presumes a commodity that can be accumulated and held. Yet question 1(e)(i) asks respondents whether the commodity is storable at all, and footnote 45 reaches for the Commission’s 2012 not-for-profit electric utilities proposal, which observed that electric energy “is not capable of being purchased in large commercial quantities ahead of time, delivered, and stored for later consumption or use.” Rented accelerator hours expire unused; a machine idle at 3am produced nothing deliverable in December. Answering 1(g), and question 2(h) on estimated deliverable supply, means building a deliverable-supply concept for a commodity the same document treats as non-storable.

How four regimes treat an index its own suppliers administer

The comparison that matters is not how regulators view AI. It is which regime places a legal duty on the administrator of the settlement index itself, rather than on the exchange listing a contract that references it.

Jurisdiction / regulator	Instrument and date	Scope	Key requirement	Penalty / sanction
US (CFTC)	DCM Core Principle 3 and the Appendix C Guidance to 17 CFR part 38, adopted at 77 FR 36612 (June 19, 2012)	The listing DCM only; no authorisation regime for the index administrator	CEA section 5(d)(3) and 17 CFR 38.200–38.201: list only contracts not readily susceptible to manipulation; Appendix C paragraph (c)(2) settlement-price reliability	Manipulation and false reporting under CEA section 6(c)(1) and 17 CFR 180.1; largest benchmark-submission penalty $250 million (Citibank, May 25, 2016)
EU (ESMA and national competent authorities)	Regulation (EU) 2016/1011 of June 8, 2016; applies from January 1, 2018	Administrators of, and supervised contributors to, commodity benchmarks used in EU financial instruments	Article 19 and Annex II: published methodology, minimum transaction-data thresholds, and under Annex II paragraph 1(g) express criteria for periods when contributors supply “a significant proportion of the total input data”	Article 42(2)(h)(i): at least €1 million or 10% of total annual turnover for legal persons; Article 42(2)(f): at least three times profits gained or losses avoided
UK (FCA)	UK Benchmarks Regulation, onshored from the EU regime that took full effect January 1, 2018	UK benchmark administrators, users and contributors; third-country administrators need recognition or endorsement by December 31, 2030	Authorisation or registration of UK administrators, application deadline January 1, 2020, and supervision under the FCA’s BENCH sourcebook	Refusal, suspension or removal from the UK Benchmarks Register; administering a benchmark without authorisation or registration is not permitted
Singapore (MAS)	Securities and Futures Act 2001, Part 6AA, current version as at August 29, 2026	Benchmarks the Authority designates under section 123B, and their administrators and submitters	Authorisation of administrators under sections 123D to 123F, a Code on designated benchmark under section 123O, and record, reporting and notification duties under sections 123P to 123V	Criminal prohibition on manipulation of financial benchmarks under section 207, with penalties prescribed by section 210

Sources: 91 FR 54259; Regulation (EU) 2016/1011; FCA, Benchmarks (last updated August 6, 2026); Securities and Futures Act 2001. Last updated: August 29, 2026.

Read across the rows and the asymmetry is plain. Three of the four regimes regulate the administrator. The United States regulates the exchange that references the administrator’s output. The European Union went furthest: Annex II paragraph 1(g) obliges a commodity benchmark administrator to publish criteria for assessment periods in which one or more contributors supply a significant proportion of total input data, and to define what “significant” means for each calculation. That is a codified answer to the Commission’s question 2(c), which asks what contributor-concentration thresholds “would be appropriate for a compute settlement reference price.” The EU wrote that rule in 2016 and applied it from 2018. The CFTC is asking the market to propose one in 2026, because the Commodity Exchange Act gives it no equivalent lever over an index provider that is not a registered entity.

This is the same perimeter gap running through several recent Commission workstreams: the conflicts proposal aimed at exchange-owned futures commission merchants, where the remedy again attaches to the registered entity rather than the affiliate creating the conflict, and the SEC-CFTC memorandum of understanding on dual registrants, which added process without an enforceable obligation. Perimeter, not intent, usually limits what a US market regulator can reach — the argument we set out in Registered is not regulated.

“America cannot win the AI race without a robust derivatives market for compute. 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. This request for comment is the first step toward establishing clear rules of the road for American compute markets.”

— Michael S. Selig, Chairman, Commodity Futures Trading Commission (CFTC release 9286-26, August 19, 2026)

The distance between that statement and the document it announces is the story. The press release is about winning a race. The Federal Register text is about whether a contract can be surveilled, and says so in the Commission’s own preliminary findings: compute markets “are fragmented and price formation primarily occurs in opaque bilateral transactions”; “dominant market participants may wield significant pricing power that may lead to manipulability, preferential pricing arrangements, and, in turn, unfair market dynamics”; and compute “may not yet exhibit certain of the characteristics of commodities that typically underlie a commodity derivatives market, including fungibility, standardization, and sufficient liquidity.”

Enforcement context: what happens when submitters trade the fix

The Commission does not have to speculate about what goes wrong when the population setting a benchmark overlaps with the population profiting from its level. It has litigated it. On May 25, 2016 the CFTC issued an order against Citibank, N.A. finding that from January 2007 through January 2012 the bank repeatedly attempted to manipulate, and made false reports concerning, the US Dollar International Swaps and Derivatives Association Fix (USD ISDAFIX), a global benchmark for interest rate products. The order imposed a $250 million civil monetary penalty and a cease-and-desist. The mechanism is the one question 2(b) describes: Citibank skewed its submissions as a panel bank in the USD ISDAFIX setting process to benefit its own positions “at the expense of its derivatives counterparties,” and traded targeted interest rate products around the setting window. It was the second such case, following a $115 million settlement with Barclays PLC, Barclays Bank PLC and Barclays Capital Inc. on May 20, 2015. By the Commission’s own tally in that release, benchmark actions against banks and brokers had by then exceeded $5.08 billion across 17 cases addressing ISDAFIX, foreign exchange and LIBOR abuses.

“The CFTC’s order demonstrates that we will vigorously continue to investigate any efforts to manipulate financial benchmarks, and we will take action where possible to protect the integrity of these benchmarks,” said Aitan Goelman, then the CFTC’s Director of Enforcement, announcing the Citibank order.

The compute case is harder in three ways. ISDAFIX submitters were CFTC- or prudentially regulated entities; compute capacity providers are not registrants and owe the Commission nothing. ISDAFIX submissions went into a defined window on an identified panel; compute posted rates are published unilaterally on commercial pricing pages and can change without notice. And the Commission’s premise at question 2(b) is that compute price series draw on “posted or listed rates that the compute capacity providers themselves administer,” with the remainder from “venues that a small number of participants operate or dominate.” Hence question 2(d), which asks whether a DCM “should be expected or required to maintain an information-sharing arrangement with each compute venue and each compute capacity provider” feeding the settlement price, and whether such surveillance is “presently feasible from a technological, operational, and legal perspective.”

What this means for exchanges, intermediaries and compliance teams

For DCMs and prospective listing venues. Self-certification under 17 CFR 40.2 remains available, but the Commission has published a 23-question record of what it considers unresolved. A certification filed after October 20, 2026 that does not address contributor concentration, index governance, observation-window controls and information-sharing arrangements is filed against a documented set of concerns. Expect to have to show a written index methodology, defined minimum transaction thresholds, an alternative procedure for windows with no transaction data, and a contractual feed from every venue or provider whose prices enter the calculation. Core Principle 4 duties under 17 CFR 38.250–38.258 attach on day one of listing, not once liquidity arrives.

For futures commission merchants and introducing brokers. Section 3 of the request is aimed at intermediaries. Question 3(a) asks what “heightened anti-money laundering and know your customers concerns” arise in compute markets and what challenges firms face “in implementing a BSA/AML program for compute futures.” Question 3(b) raises disclosure obligations for a contract “settling against a geopolitically sensitive commodity, including to retail participants,” and invites comparison with oil. Firms already running enhanced due diligence for sanctioned-jurisdiction exposure should assume the export-control status of accelerator hardware becomes an onboarding question, not only a supply-chain one.

For fund managers and treasury hedgers. The hedging case is real and the Commission does not dispute it: a compute futures market “may therefore provide a means for managing and assuming price risks, discovering prices, or disseminating pricing information as to general trends in AI adoption.” But a hedge is only as good as the correlation between the settlement index and the price a firm actually pays. Where real cost sits in a multi-year reserved-capacity agreement negotiated privately and the index is built from on-demand posted rates, basis risk is not a residual, it is the position. Question 1(a) asks respondents to distinguish “on-demand, spot, reserved, committed purchase modes” precisely because they do not move together.

For legal and compliance teams. The deliverable has a date. Comments must be received on or before October 20, 2026, and the Commission “particularly encourages commenters to provide empirical and data-driven input.” A firm holding transaction-level compute pricing data that does not file leaves the record to the parties with the strongest interest in a permissive index standard. Related perimeter analysis sits in our coverage of where the regulatory perimeter actually bites.

“The economic scale of AI compute is already macroeconomically material. Based on our calculations, the 2025-Q4 installed compute stock already implies a gross compute service flow of around $430 billion to $1.3 trillion per year, or approximately 1.4% to 4.0% of U.S. GDP.”

— Federico M. Bandi, author of (Early) AI Compute Asset Pricing (July 1, 2026), quoted by the Commission at footnote 5 of 91 FR 54259

That range is the steelman for moving quickly, and it deserves full strength: a service flow between $430 billion and $1.3 trillion a year is larger than several commodity complexes that have had listed futures for decades, and the factor-of-three spread between the low and high estimate is itself an argument for price discovery. Read the other way, it measures how little is observable. An estimate that wide is what a market looks like before it has a price.

What is next

The comment file closes on October 20, 2026. Nothing commits the Commission to a rulemaking afterwards, and a request for comment carries no compliance date. The plausible next steps are a proposed rule amending part 38, guidance updating the Appendix C standard for indices of this kind, or nothing at all while exchanges test the self-certification route. The absence of a go-live date is the point: this is the stage at which the record is built.

Two related files bear watching. The Commission published a companion request on June 25, 2026 — RIN 3038-AF75, at 91 FR 38334 — on extending standard futures to 24/7 trading and on perpetual contracts referencing physically delivered or storable energy commodities, closed to comment on July 27, 2026. Its title is instructive: that consultation confined perpetuals to commodities that are physically delivered or storable. Section 4 of the compute request now asks whether perpetual compute futures should exist at all, for an underlier that is neither. The definitional question remains live on the securities side too, as we covered when US crypto perpetuals went live before the swap definition was settled.

The second is the Commission’s own composition. The voting summary records that “Chairman Selig voted in the affirmative” and that “No Commissioner voted in the negative,” wording consistent with a Commission acting without a full complement. A single-vote agency can move quickly on a request for comment; a contested final rule on a novel product class is a different exercise, and that is where the thinness of the pricing record gets tested. Firms tracking the wider direction should also follow the move to end the swap execution facility order-book requirement for permitted transactions and the twin event-contract rules that ended the no-action era.

FAQ

Does the CFTC’s request for comment approve compute futures?

No. The document is a request for comment under RIN 3038-AF77, published at 91 FR 54259 on August 21, 2026. Its action line is “Request for comment,” it proposes no regulatory text and sets no effective or go-live date. It gathers a record on 23 questions. A designated contract market’s ability to list a contract by self-certification under 17 CFR 40.2, or by seeking approval under 17 CFR 40.3, is unchanged by it.

What is DCM Core Principle 3 and why does it matter for compute?

Core Principle 3, at CEA section 5(d)(3) and 17 CFR 38.200 to 38.201, allows a designated contract market to list only contracts “not readily susceptible to manipulation.” For cash-settled contracts the Appendix C Guidance to part 38 asks whether the settlement price is reliable, acceptable, publicly available and timely, and drawn from a cash market that is itself sufficiently liquid and not manipulable. Compute strains that test because the firms contributing posted rates to an index are also the firms controlling supply.

Why is deliverable supply a problem for a compute contract?

Question 1(g) asks for a deliverable supply estimate methodology to size position limits under Core Principle 5, and question 2(h) asks for estimated deliverable supply “at the relevant pricing point or points.” Deliverable supply assumes a commodity that can be accumulated. Question 1(e)(i) asks respondents whether compute is storable at all, and footnote 45 draws the parallel with electric energy, which the Commission has described as not capable of being purchased ahead of time, delivered and stored.

How does the EU approach differ from the US approach here?

The European Union regulates the benchmark administrator directly. Regulation (EU) 2016/1011 has applied since January 1, 2018, and Annex II paragraph 1(g) requires a commodity benchmark administrator to publish criteria for periods when one or more contributors supply a significant proportion of total input data. The United States has no equivalent authorisation regime for a commodity index provider; the obligation attaches to the exchange listing the contract, which is why the CFTC asks the market to propose a standard rather than applying one.

What did the ISDAFIX cases establish?

That a benchmark set by submissions from firms trading against it can be, and was, distorted. The CFTC’s order of May 25, 2016 required Citibank, N.A. to pay a $250 million civil monetary penalty for attempted manipulation and false reporting of USD ISDAFIX between January 2007 and January 2012, following a $115 million settlement with Barclays entities on May 20, 2015. By the Commission’s own count, benchmark actions against banks and brokers had by then exceeded $5.08 billion across 17 cases.

Are perpetual compute futures part of this consultation?

Yes, as the fourth and shortest section. Two questions ask whether perpetual compute futures would offer commercial risk management features fixed-date contracts cannot, and whether they pose unique risks requiring extra safeguards. The framing matters because the Commission’s June 2026 companion request, at 91 FR 38334, limited its perpetual-contract questions to energy commodities that are physically delivered or storable. Compute, on the Commission’s own preliminary analysis, is neither.

This article is informational analysis only and does not constitute legal, regulatory, tax, or investment advice. Regulatory frameworks change frequently and interpretation depends on facts and circumstances; primary documents and official regulator guidance always supersede summaries. Firms should consult qualified legal counsel and their relevant supervisory authority before taking any action based on the analysis above.