Hyperliquid Policy Center Proposes CFTC-Regulated AI Compute Futures

The proposal suggests that rented computing capacity qualifies as a commodity under the Commodity Exchange Act. By introducing perpetual contracts, which lack fixed expiration dates, the organization aims to match the continuous, hourly nature of GPU consumption and inference model usage. CEO Jake Chervinsky and senior counsel Brad Bourque argue that these tools are vital for lenders and operators managing price exposure in an industry where hardware costs fluctuate based on chip generation and availability.

To mitigate risks in thin markets, the policy center recommends strict safeguards, including leverage limits, funding-rate caps, and transparent, independent pricing benchmarks. The letter highlights that while private, negotiated agreements currently dominate the market, standardized derivatives would provide the price discovery necessary for scaling infrastructure. The submission follows a broader CFTC consultation on compute derivatives, an area Chairman Michael Selig has identified as essential to maintaining American competitiveness in the artificial intelligence sector.