Key Takeaways
- • CME Group and Silicon Data are launching the world's first computing power futures market, a new tradable asset class tied to GPU-based computing capacity central to AI infrastructure and model development.
- • The futures contracts will be benchmarked against Silicon Data's daily GPU pricing indices for on-demand rental rates, enabling investors, AI developers, cloud providers, and financial institutions to hedge and speculate on compute cost volatility.
- • CME Chief Executive Terry Duffy positioned computing power as an emerging core asset class comparable to energy's role in previous industrial cycles, reflecting industry recognition of compute as a foundational commodity in the digital economy.
- • The initiative addresses acute scarcity in high-performance computing resources driven by surging AI infrastructure demand, which has already led to usage caps and massive capital deployment in data centers and chip manufacturing.
- • Market launch is planned for later in 2026, subject to regulatory approval, with the framework extending CME's commodities futures model to GPU-based capacity.
Summary
CME Group, the world's leading derivatives exchange operator, is partnering with Silicon Data, a provider of real-time GPU benchmarking services, to create the first futures market for computing power. The initiative represents a significant expansion of the futures market framework into artificial intelligence infrastructure, a resource increasingly recognized as a critical commodity underpinning the global AI ecosystem.
The proposed futures contracts will enable market participants—including institutional investors, AI developers, cloud service providers, and financial institutions—to hedge exposure to and speculate on GPU computing capacity costs. Silicon Data has developed proprietary indices that track real-time on-demand GPU rental pricing, which will serve as the underlying benchmark for these contracts. By standardizing pricing mechanisms for compute resources, the initiative addresses a structural gap in markets where AI infrastructure demand has outpaced transparent price discovery mechanisms.
The timing reflects acute supply-demand imbalances in high-performance computing. Surging demand for AI model training and deployment has created compute shortages severe enough to trigger usage caps on some cloud services and prompted unprecedented capital investment in data center buildout and specialized semiconductor manufacturing. CME Chief Executive Terry Duffy framed computing power as an emerging core asset class, drawing a parallel to energy's centrality in prior industrial cycles. This positioning aligns with recent commentary from industry figures including BlackRock CEO Larry Fink, who has suggested that compute futures markets will emerge as a natural response to persistent scarcity and demand intensity across the AI infrastructure sector.
Futures contracts—the standardized instruments through which CME will offer compute exposure—have proven effective in commodities markets including crude oil and metals for enabling price discovery, volatility hedging, and capital-efficient risk transfer. Extending this framework to GPU-based computing capacity represents a natural application as compute transitions from a captive internal resource to a tradable commodity with transparent, volatile pricing. The launch is anticipated during 2026, pending regulatory approval.
The proposed futures contracts will enable market participants—including institutional investors, AI developers, cloud service providers, and financial institutions—to hedge exposure to and speculate on GPU computing capacity costs. Silicon Data has developed proprietary indices that track real-time on-demand GPU rental pricing, which will serve as the underlying benchmark for these contracts. By standardizing pricing mechanisms for compute resources, the initiative addresses a structural gap in markets where AI infrastructure demand has outpaced transparent price discovery mechanisms.
The timing reflects acute supply-demand imbalances in high-performance computing. Surging demand for AI model training and deployment has created compute shortages severe enough to trigger usage caps on some cloud services and prompted unprecedented capital investment in data center buildout and specialized semiconductor manufacturing. CME Chief Executive Terry Duffy framed computing power as an emerging core asset class, drawing a parallel to energy's centrality in prior industrial cycles. This positioning aligns with recent commentary from industry figures including BlackRock CEO Larry Fink, who has suggested that compute futures markets will emerge as a natural response to persistent scarcity and demand intensity across the AI infrastructure sector.
Futures contracts—the standardized instruments through which CME will offer compute exposure—have proven effective in commodities markets including crude oil and metals for enabling price discovery, volatility hedging, and capital-efficient risk transfer. Extending this framework to GPU-based computing capacity represents a natural application as compute transitions from a captive internal resource to a tradable commodity with transparent, volatile pricing. The launch is anticipated during 2026, pending regulatory approval.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.