CME and ICE plan to launch computing power futures, but the review process has been postponed to January 2027.
2026-11-20 20:58:39

The Market Basis for the Transformation of Computing Power into Commodity Attributes
Computing power is currently primarily achieved through the leasing of graphics processing units (GPUs) within data centers. NVIDIA's H100 and the newer B200 chips serve as key benchmarks, with leasing priced hourly. Buyers are often model developers or labs, while sellers are cloud service providers or dedicated data center operators. Most large transactions are completed through non-public agreements, and public indices only capture a portion of observable prices. Antti Belt, head of technology and AI at the Boston Consulting Group, points out that large agreements can significantly impact the overall cost structure but are not necessarily reflected in indices. Steve Hou, head of research at Silicon Data, states that the natural gas and corporate loan markets also have a large number of bilateral transactions, but widely cited index systems have been established, with their company indices incorporating transaction data beyond listed prices. Compared to mature commodity markets, computing power has a "use-and-go" characteristic, similar to electricity trading, but the standardization connecting physical delivery and financial flows remains low. Existing indices do not adequately cover private agreements, resulting in an incomplete price discovery mechanism. Peter Kivy, global head of energy and environmental products at the CME Group, said at the Dow Jones Energy Forum in October that the company received inquiries from several exchange-traded fund (ETF) issuers immediately after the contract was announced, focusing on the launch timing. This development indicates that market participants are viewing computing power costs as a manageable risk exposure, rather than simply an operating expense.Progress of Exchange Product Design and Regulatory Review
The Chicago Mercantile Exchange Group (CME Group) and Silicon Data plan to launch cash-settled futures contracts based on the H100 and B200 leasing indices, with each contract representing approximately one month's worth of continuous leasing hours for a single chip. The Intercontinental Exchange (ICE) is partnering with AORN to launch contracts based on its calculated price index. Both exchanges emphasize that futures can help buyers lock in future costs, sellers lock in cash flow, and improve market efficiency in the long term. However, the U.S. Commodity Futures Trading Commission (CFTC) extended its review period by two months in November, focusing on liquidity, manipulation risks, and the index's representativeness for private trading. The review is now extended to January 8, 2027, meaning the original October 5 listing date for some contracts will no longer be achievable. Regulators previously released a request for comments in August, focusing on the maturity of the underlying spot market, price monitoring capabilities, and whether the index adequately reflects actual costs. Industry feedback indicates that the computing power market is highly concentrated, a few suppliers could influence the index, and there is insufficient historical data to validate contract rollover and term structure. Despite these uncertainties, exchanges believe that computing power has become a core investment in AI systems, requiring standardized tools to transfer risk.First-mover competition among issuers of exchange-traded open-ended funds
Langhill Financial, Grayscale Investments, Granite, and Pro-Tech have submitted prospectuses for products linked to computing power futures, which have yet to be approved. Steve Vanurni, head of indices at Grayscale Investments, stated that the team is refining parameters such as the maturity date, but product design flexibility is limited until the specific structure and listing date of the futures are clarified. Will Lind, founder and CEO of Granite, said the company has applied for three products, tracking the H100, B200, and a hybrid of both. The US exchange-traded fund (ETF) market has exceeded $16 trillion, but funds are highly concentrated in large issuers. Market analysis indicates that while the absolute value of the market share contested by other participants is considerable, competition is fierce. The average duration of ETFs liquidated this year was only about one year and nine months, shorter than before. Early adopters can lock in market share through liquidity and brand advantages, while laggards need to continuously invest to maintain differentiation, otherwise they face pressure to exit due to insufficient scale.Volatility Characteristics and Rollover Mechanism
The 30-day average volatility of Silicon Data's B200 Rental Price Index was close to 25% in early November, about half that of Brent crude oil contracts and close to the S&P 500 index during the same period. Exchange-traded funds (ETFs) holding futures-based contracts may face negative rolling returns when contracts are rolled over: if the price of a forward contract is higher than that of the near-month contract, the fund must roll over its exposure at a higher cost, thus eroding its net asset value. Furthermore, the opacity of the spot market may cause the index to deviate from actual transaction prices, and the economic impact of large bilateral agreements may not be reflected in settlement references in a timely manner.- Risk Warning and Disclaimer
- The market involves risk, and trading may not be suitable for all investors. This article is for reference only and does not constitute personal investment advice, nor does it take into account certain users’ specific investment objectives, financial situation, or other needs. Any investment decisions made based on this information are at your own risk.