Have you noticed how the conversation around artificial intelligence has shifted from pure software breakthroughs to something far more grounded? Suddenly everyone is talking about the physical power needed to train and run those models. Computing capacity has become the new scarce resource, and regulators are starting to treat it like a commodity that can be hedged, traded, and potentially manipulated. On August 19 the Commodity Futures Trading Commission took a measured but significant step by requesting public comment on how U.S. exchanges should list and oversee derivatives tied to that very capacity. I’ve been following commodity markets long enough to recognize when an agency is testing the water rather than diving in, and this feels exactly like that moment.
Why Computing Power Suddenly Matters to Futures Markets
Compute is not an abstract idea. It is the processing muscle that trains large language models and keeps inference running for millions of users. Most companies do not own the hardware. They rent access to high-end graphics processors through cloud providers or specialized data-center operators. That rental market has grown fast and remains largely private. Prices move according to bilateral deals that outsiders rarely see. When a market looks like that, the natural next question for any derivatives regulator is whether it can support transparent futures contracts.
The CFTC’s 19-page request, labeled RIN 3038-AF77, does not propose a rule and does not approve any specific contract. It simply asks market participants to weigh in on liquidity, benchmark reliability, manipulation risks, customer protection, and the possibility of perpetual compute futures. Comments will stay open for 60 days once the document appears in the Federal Register. As of mid-August that publication had not yet occurred, so the final deadline remains unknown. In my view this measured approach is wise. Rushing a new product class into the market before the underlying cash market is well understood has caused headaches before.
Treating Compute as an Emerging Commodity
The agency is clear that the underlying commodity would typically be access to rented computing capacity. Think of the hourly rental price of an advanced Nvidia processor, or even access to a defined quantity of AI inference tokens. Chairman Michael S. Selig put it bluntly: the United States cannot win the AI race without a robust derivatives market for compute. He framed the request as an early step toward clear rules of the road rather than a finished regulatory position. That language matters. It signals openness while keeping expectations in check.
Futures contracts built on these prices could let AI developers, cloud companies, and data-center operators lock in costs or protect against sudden spikes. Speculators would also find a new venue to express views on future demand for AI infrastructure. In theory the product solves a real hedging need. In practice the path from theory to liquid trading is rarely smooth.
The Liquidity and Transparency Puzzle
Here is where the questions get harder. Much of the economic value in compute markets still flows through private agreements. Prices differ by provider, region, processor model, and contract length. A cash-settled futures contract needs a reliable reference price that can be observed, verified, and monitored. If that price is drawn from opaque bilateral deals, settlement becomes vulnerable to distortion.
The CFTC wants concrete data on transaction volumes, market concentration, and the share of trades that occur at publicly disclosed prices. It also asks whether compute providers themselves could influence both available capacity and the inputs that feed any benchmark. Surveillance arrangements, information-sharing protocols, and safeguards against manipulated settlement prices all sit high on the agency’s list of concerns. I’ve seen similar debates around other emerging energy and agricultural products. Transparency is not optional; it is the foundation.
Clear rules of the road are essential if derivatives are to support rather than complicate the AI infrastructure build-out.
That sentiment captures the careful tone of the request. The agency is not declaring the market ready. It is asking whether the market can become ready under the right conditions.
Perpetual Contracts Enter the Conversation
One of the more interesting threads in the request concerns perpetual compute futures. These contracts have no fixed expiration date. Instead they rely on recurring funding payments or similar mechanisms to keep the futures price aligned with the underlying market. The CFTC asks whether such a structure would offer risk-management benefits that standard dated contracts cannot match, and whether it would introduce fresh risks that need additional protections.
Perpetual products already exist in certain cryptocurrency markets, though that policy remains under legal challenge. The compute discussion is separate, yet the broader classification of perpetual derivatives could influence how new commodity markets are treated. Market participants will need to think carefully about funding mechanisms, margin requirements, and the behavior of these contracts during periods of extreme stress in the rental market for high-end processors.
CME’s October Timeline Adds Urgency
While the comment period unfolds, CME Group and Silicon Data have announced plans to launch two cash-settled compute futures on October 5, subject to regulatory review. One contract would track one month of rental costs for Nvidia H100 processors. The other would follow rental costs for the newer Blackwell B200 processors. Silicon Data would supply the pricing benchmarks. The target date does not guarantee approval. Regulatory review can take longer than expected, especially when the underlying market is still being examined.
The coincidence of timing is worth noting. An exchange is ready to list products at the same moment the primary regulator is soliciting views on the broader product class. That creates a practical test of how the comment process interacts with actual listing decisions. Market participants watching both tracks will want to understand how feedback submitted to the agency might shape the review of specific contract filings.
Customer Protection and Broader Safeguards
Beyond the core market-structure questions, the CFTC is seeking input on retail disclosures, anti-money-laundering controls, position limits, and the geopolitical sensitivity of advanced computing capacity. Advanced chips sit at the center of national-security discussions. Any futures market built on those chips will inevitably attract attention from policymakers concerned about technology transfer and strategic supply. Empirical, data-driven comments are especially encouraged. Anecdote has its place, but the agency needs measurable evidence.
Position limits will be particularly important. Concentrated holdings in a new market can amplify volatility and raise the risk of squeezes. Disclosures for retail participants need to explain both the hedging potential and the unique risks of a product whose underlying market remains relatively young and opaque. Anti-money-laundering frameworks must adapt to a commodity that is intangible yet highly valuable.
What Successful Compute Futures Could Deliver
If the market develops carefully, the benefits could be real. AI developers facing unpredictable rental costs would gain a tool for budgeting and risk management. Cloud providers and data-center operators could hedge capacity utilization. Speculators would bring liquidity and price discovery. Over time a transparent futures curve might even help capital allocation decisions for new data-center construction. That is the optimistic case.
The less optimistic case is also easy to imagine. Thin liquidity, unreliable benchmarks, and occasional settlement disputes could leave the product underused. Worse, poorly designed contracts could create new avenues for manipulation that spill back into the physical rental market. The CFTC’s request is designed to surface both possibilities before any large-scale trading begins.
Key Questions the Agency Wants Answered
Reading through the request, several practical questions stand out. How liquid is the cash market for the specific processors that would serve as benchmarks? What proportion of transactions occurs at prices that can be independently verified? Can exchanges design surveillance systems capable of detecting attempts to influence those prices? Would perpetual structures improve or complicate risk management for end users? These are not theoretical concerns. They determine whether the product can function as intended.
- Cash-market depth and concentration across major providers
- Reliability and independence of proposed pricing benchmarks
- Surveillance tools needed to protect settlement integrity
- Customer-protection features suitable for both commercial and retail users
- Additional risks unique to perpetual compute futures
Answers to those points will shape whatever guidance or rulemaking eventually follows. They will also influence how exchanges design their own contract specifications and risk controls.
Geopolitical Overtones Cannot Be Ignored
Advanced computing capacity is not just another commodity. Access to the latest processors is subject to export controls and national-security reviews. A liquid futures market could, in theory, create new channels for price discovery and risk transfer. It could also raise questions about who holds economic exposure to critical technology. The CFTC has flagged this sensitivity and invited comment. Any durable framework will need to account for those broader policy considerations without turning the derivatives market into an extension of trade policy.
In my experience, regulators who acknowledge these intersections early tend to produce more resilient rules. Ignoring them invites later friction when other agencies or congressional committees take notice.
How Market Participants Should Approach the Comment Process
Once the Federal Register publication occurs, comments can be submitted through the usual channels. The most useful submissions will be specific. Volume data, concentration statistics, descriptions of existing bilateral pricing practices, and concrete proposals for surveillance or disclosure standards will carry more weight than general statements of support or opposition. Commercial users who actually rent capacity have a particular interest in explaining how futures might improve their risk management. Market makers and potential liquidity providers can describe the conditions under which they would commit capital.
Exchanges preparing contract filings will also benefit from the public record that develops. Feedback submitted during the comment period can highlight issues that need to be addressed in product design before formal review begins. That kind of iterative process is healthier than discovering problems after trading has started.
Looking Ahead to Possible Outcomes
Several paths are possible. The CFTC could conclude that existing principles for commodity futures are already sufficient and simply apply them to compute products with modest guidance. It could issue more detailed recommendations on benchmarks, surveillance, or customer protections. In a more formal direction, it could eventually propose specific rules tailored to this asset class. Or it could determine that the cash market is not yet mature enough and encourage further development before derivatives listing proceeds at scale.
None of those outcomes is predetermined. The quality of the public record will matter. So will the practical experience that emerges if and when the first contracts begin trading under existing review procedures. The October target dates set by one major exchange create a natural experiment. Regulators and market participants will both be watching how those products perform if they receive clearance.
Why This Matters Beyond the Trading Floor
AI infrastructure spending is measured in tens of billions of dollars and is still rising. Any tool that helps participants manage the associated cost volatility has potential systemic importance. At the same time, poorly designed markets can amplify rather than dampen risk. The CFTC’s decision to solicit input before the product class becomes entrenched is a recognition of that dual reality. It is also a reminder that derivatives regulation is rarely static. New underlying assets force regulators to revisit old assumptions about liquidity, transparency, and manipulation.
I’ve watched several commodity markets evolve from niche ideas into mainstream risk-management tools. The ones that succeeded usually had three ingredients: a genuine commercial need, sufficient cash-market transparency, and thoughtful contract design. Compute futures currently satisfy the first criterion. The second and third remain open questions. The public comment process is the agency’s way of gathering evidence on both.
Practical Implications for AI Companies and Investors
For companies that rent large volumes of processing capacity, the arrival of liquid futures would change budgeting conversations. Instead of absorbing every price swing, they could lock in a portion of expected costs. That certainty has value when training runs last weeks or months and model economics remain sensitive to infrastructure expense. Cloud providers and specialized operators could use the same contracts to manage utilization risk on the other side of the market.
Investors watching the AI sector should pay attention as well. A functioning derivatives market can improve price discovery and, over time, influence capital allocation toward data-center capacity. It can also introduce new sources of volatility if speculative flows dominate thin liquidity. Understanding the contract specifications, the quality of the benchmarks, and the depth of commercial participation will become part of fundamental analysis for anyone exposed to AI infrastructure themes.
The Role of Benchmark Providers
Reliable benchmarks sit at the center of any cash-settled product. The partnership between CME Group and Silicon Data illustrates one possible model. Independent pricing sources that aggregate transaction data across multiple providers can reduce the risk that any single participant distorts the settlement price. Yet independence alone is not enough. Coverage, methodology transparency, and resistance to gaming all matter. The CFTC’s questions about observable and verifiable prices speak directly to these design choices.
Market participants evaluating future contracts will want to examine how the benchmark is constructed, how many actual transactions feed it, and what contingency procedures exist when data are sparse. Those details often determine whether a contract attracts commercial hedging or remains primarily a speculative vehicle.
Balancing Innovation and Caution
Regulators face a familiar tension. Move too slowly and legitimate commercial needs go unmet. Move too quickly and poorly designed products create new problems. The current request for comment is an attempt to thread that needle. By gathering evidence before committing to a detailed regulatory framework, the agency preserves flexibility. At the same time, the existence of concrete listing plans keeps the process from becoming purely theoretical.
Perhaps the most interesting aspect is how this discussion sits alongside broader debates about perpetual derivatives and the classification of novel commodities. Compute is not cryptocurrency, yet some of the structural questions overlap. Lessons learned in one area can inform the other, provided the differences are respected. That cross-pollination of experience is one reason public comment periods remain valuable even when the product class feels entirely new.
What Success Would Look Like
A successful outcome would feature contracts that commercial users actually employ for hedging, supported by benchmarks that withstand scrutiny, and overseen by surveillance systems capable of detecting and deterring manipulation. Liquidity would deepen gradually as confidence builds. Position limits and disclosure rules would protect the market without stifling participation. Geopolitical sensitivities would be acknowledged without turning every trade into a policy statement.
That is a high bar. Meeting it will require ongoing dialogue among exchanges, commercial users, benchmark providers, and regulators. The comment process is only the first chapter. The real test will come when the first contracts begin trading and market participants discover how the product behaves under real conditions.
Until then, the best contribution anyone can make is clear, data-supported input. The agency has asked for it. The market’s ability to deliver useful risk-management tools for the AI era may depend on how thoughtfully that request is answered.
In the end, computing capacity has become too important to remain outside the derivatives conversation. Treating it as a commodity that can be hedged is a logical evolution. Doing so carefully is the only responsible path forward. The CFTC’s request for comment is the formal beginning of that careful process. How the market and the public respond will shape what comes next.
