As investment in artificial intelligence infrastructure continues to expand, GPU computing power is transitioning from a basic infrastructure cost in the AI supply chain to a financial asset that can be priced, traded, and hedged.

CME Group Inc and GPU market data firm Silicon Data announced on Tuesday plans to launch two computing futures contracts on October 5, pending regulatory approval.

The two contracts are the Silicon Data H100 Rental Index Futures and the Silicon Data B200 Rental Index Futures, which will be listed on CME Group Inc and subject to the rules of the New York Mercantile Exchange (NYMEX).

The contracts will track the H100 and B200 GPU rental price indices compiled by Silicon Data, using the hourly GPU rental cost as the core pricing metric.

This means market participants may soon be able to hedge against AI computing price fluctuations using futures instruments. CME Group Inc stated that the new products aim to help traders, financial institutions, AI developers, and cloud service providers manage price risks in the rapidly growing computing market.

CME Group Inc has previously referred to computing power as "the new oil of the 21st century" and noted that computing resources are becoming an independent emerging asset class.

GPU rental prices are becoming a key cost indicator for the AI industry. For companies involved in AI model training and inference, beyond capital expenditures like purchasing GPUs and building data centers, leasing GPUs directly from cloud providers or specialized computing service firms is a vital way to access computational power.

Therefore, changes in the hourly GPU rental price can reflect computing supply and demand dynamics to some extent.

Silicon Data currently publishes daily GPU rental benchmarks covering major AI accelerators like the H100, A100, B200, and AMD MI300X, standardizing prices by collecting data from cloud service providers, hyperscale cloud firms, colocation data centers, and the private rental market.

The B200 is a core product of Nvidia's Blackwell architecture, offering higher computational performance than the previous-generation H100. As Blackwell capacity and deployment remain in an expansion phase, B200 rental prices are currently influenced by both supply constraints and AI training demand.

Silicon Data stated that the B200 rental price index reflects a standardized hourly price from various computing supply channels.

The significance of the futures market lies in establishing a public forward price discovery mechanism. Currently, the GPU rental market is highly fragmented, with significant price differences among cloud providers, computing service firms, and across different lease terms. Companies also lack mature tools, similar to crude oil futures, to lock in future computing costs.

When CME Group Inc and Silicon Data first announced their collaboration in May, they stated that the new contracts are designed to help AI companies and cloud service providers manage the risk of computing price volatility. Silicon Data subsequently launched GPU forward price curves covering models like the H100, B200, and A100, with terms of up to 36 months.

For computing suppliers, if they expect GPU rental prices to decline in the future, they can use futures for hedging. For AI companies that need to lease large amounts of GPU power, futures may provide a tool to lock in future computing costs.

Meanwhile, financial institutions and traders can also participate in price trading based on their assessments of AI demand, GPU supply, and data center construction cycles.

This also signifies that the financialization of the AI industry is extending from chips, data centers, and related stocks down to the most fundamental computing resource itself.

If CME Group Inc's products are successfully launched and generate sufficient liquidity, the H100 and B200 futures prices could become important market indicators for observing AI computing supply and demand, data center utilization rates, and returns on AI infrastructure investment.

However, computing power differs from traditional commodities in a key way: GPU computing cannot be stored like oil or gold. Once computing capacity is idle, the value of its hourly service disappears, so the pricing mechanism, liquidity, and hedging effectiveness of GPU futures will differ significantly from traditional commodity futures.

A recent study on AI computing asset pricing also noted that because computing power is non-storable, the no-arbitrage pricing relationships in traditional commodity futures cannot be directly applied to the computing market.

Therefore, if these two contracts are listed on October 5 as planned, their significance may extend beyond just adding two new futures products, potentially establishing a public financial benchmark for AI computing, which previously lacked a unified market price.