Key Insights

  • CME compute futures remained under CFTC review as their originally planned Oct. 5 trade date arrived.
  • The H100 and B200 contracts are designed around 730 GPU-hours and settle financially against Silicon Data rental indexes.
  • Silicon Data’s latest index page showed B200 rental costs at $5.86 per GPU-hour versus $2.78 for H100.

CME compute futures tied to Nvidia H100 and B200 rental costs did not reach their originally planned Oct. 5 launch with regulatory review still open. The Commodity Futures Trading Commission listed both NYMEX products as “Approval Pending (90)” after a Sept. 21 status update.

CME Group and Silicon Data announced the contracts on Aug. 11 with an Oct. 5 target, explicitly subject to regulatory review. CME’s own Globex notices also conditioned the listing on completion of all review periods.

The delay matters because the products would create an exchange-traded benchmark for a fast-growing AI infrastructure expense. Fusion Market News has also examined AI stocks tied to accelerating infrastructure spending. Until approval is complete, users cannot rely on the announced date as an active trading launch.

CME compute futures remain pending at CFTC

The CFTC database continued to show the Silicon Data H100 Rental Index Futures and B200 Rental Index Futures as pending. Both contracts are assigned to NYMEX and classified by the regulator under “Other Natural Resource.”

CME’s contract specifications set the H100 code at GPU1 and the B200 code at GPU2. Each contract represents 730 GPU-hours, roughly one GPU running for an average month. The contracts settle financially, so no physical GPU capacity changes hands.

The minimum move is $0.01 per GPU-hour, making one tick worth $7.30 per contract. CME said monthly contracts would extend 36 months once listed.

GPU rental prices show the risk the contracts target

Silicon Data’s latest public index page showed H100 neo-cloud rental pricing at $2.78 per GPU-hour. Its B200 benchmark stood at $5.86, more than twice the H100 level.

Latest Silicon Data GPU Rental Benchmarks

H100 — $2.78/GPU-hour

B200 — $5.86/GPU-hour

B200 rental costs remain more than double H100. Source: Silicon Data.

The gap illustrates why a standardized hedge could matter for AI builders and infrastructure operators. Rental costs vary by GPU generation, provider, region and contract structure, creating basis risk even when an index captures the broader market.

Fusion Market News has also tracked how AI data-center operators are expanding contracted compute capacity, increasing the financial importance of infrastructure pricing and project financing.

Regulatory clearance becomes the next catalyst

The next material event is not an Nvidia product announcement or a rental-price move. It is the completion of the CFTC review and a revised effective trading date from CME.

The regulatory question is unusually important for this market. The CFTC has separately raised questions about compute fungibility, standardization, liquidity and pricing power across providers. The agency is also shaping new federal rules for leveraged crypto products. Those issues go directly to whether GPU rental access can support a durable derivatives market.

CME’s original Oct. 5 target therefore should not be treated as an active launch date while the products remain pending. A new CME notice or CFTC status change would provide the clearest confirmation that GPU1 and GPU2 can begin live trading.

Sources: Commodity Futures Trading Commission designated contract market filings; CME Group Compute Futures specifications and Globex notices; Silicon Data GPU Rental Price Indices.

Sharron Kendi is a seasoned crypto market analyst and writer with over three years of experience decoding blockchain trends, price movements, and market dynamics. She holds a Bachelor’s Degree in Commerce (Finance) from Kenyatta University, blending a solid academic foundation with a sharp eye for technical analysis and a deep understanding of on-chain data. Her work delivers clear, data-driven insights that empower investors to navigate the fast-evolving digital asset space with confidence.