Building a Compute Futures Market Bigger Than Oil

🔑 Enhanced Key Takeaways

•Silicon Data has partnered with CME Group, the world's leading derivatives marketplace, to launch a first-in-class compute futures market later in 2026, pending regulatory review.
•Carmen Li, CEO of Silicon Data and Compute Exchange, holds a Ph.D. in Mathematical Physics and previously managed global data partnerships at Bloomberg, where she identified the critical need for risk management tools in the compute economy, drawing parallels to established commodity markets like oil.
•Silicon Data, backed by global trading firms DRW and Jump Trading Group, is the industry leader in GPU market intelligence and benchmarking, providing the foundational indices for these futures contracts, including the SDH100RT (H100 Rental Index) which tracks daily GPU rental rates.
•Compute Exchange, co-founded by Carmen Li and DRW founder Don Wilson, operates as an open, decentralized marketplace for trading GPU compute resources, utilizing an auction-based system and blockchain for transparent and secure settlement of spot and futures contracts.
•The broader GPU compute market remains highly fragmented, characterized by significant price volatility, a lack of standardized tradable units, and the coexistence of multiple, sometimes conflicting, price indices from providers like Silicon Data, Ornn AI, Compute Desk, and SemiAnalysis.

📊 Competitor Analysis▸ Show

Aspect	Compute Exchange / Silicon Data (CME Group Partnership)	Intercontinental Exchange (ICE) / Ornn
Role	Marketplace (Compute Exchange), Index Provider (Silicon Data), Derivatives Exchange (CME Group)	Derivatives Exchange (ICE), Index Provider (Ornn)
Index Basis	Silicon Data's indices (e.g., SDH100RT for H100 rental rates)	Ornn's OCPI index (covering H100, H200, RTX 5090)
Contract Type	Futures and options on futures (CME)	GPU computing power futures (ICE)
Settlement	Cash-settled (CME contract)	Not explicitly stated for ICE, but typically cash-settled for such commodities.
Market Focus	Managing volatility and price risk for traders, financial institutions, AI builders, cloud providers	Managing price volatility and providing risk management for AI labs and compute operators
Underlying Asset	On-demand GPU rental rates	Actual transaction records for computing power
Launch Date	Later this year (2026), pending regulatory review (CME)	Plans to launch (ICE)

🛠️ Technical Deep Dive

Silicon Data GPU indices track pricing for major AI GPU categories, including NVIDIA A100, H100, and B200 (Bloomberg tickers: SDA100RT, SDH100RT, SDB200RT).
The indices are built on a foundation of hundreds of thousands of verified pricing records, spanning 50 countries/regions and covering 50-100 platforms, including hyperscalers, neoclouds, and marketplaces.
Data captured for index calculation includes pricing structures (lease type and durations), providers and locations, and granular hardware specifications.
The flagship SDH100RT (H100 Rental Index) specifically monitors the hourly cost of renting a top-tier GPU model.
Silicon Data offers programmatic access to its GPU index data via an API, enabling integration into dashboards, trading algorithms, and analytics workflows.
Compute Exchange operates an auction-based system to facilitate real-time price discovery for compute resources.
The platform utilizes blockchain technology for transparency and security in settling spot and futures contracts.
Compute Exchange supports flexible pricing models designed to dynamically match supply and demand.
The platform lists available inventories and dynamically pings providers to obtain pricing and comprehensive specifications, including details on performance and thermal latency degradation.
While the market is evolving, a truly standardized tradable unit for compute, such as a "chip instance hour," is still emerging, with current offerings varying by region, configuration, and contract terms.

🔮 Future ImplicationsAI analysis grounded in cited sources

The emergence of compute futures markets will significantly reduce price volatility and improve long-term planning for AI builders and cloud providers.

Futures markets provide essential hedging tools, allowing participants to lock in prices and manage risk, thereby stabilizing compute costs for long-term projects, similar to how oil futures benefit industries like aviation.

Increased standardization and transparency in compute pricing will accelerate innovation in AI by democratizing access to resources.

By making compute resources more accessible and predictable through standardized indices and transparent marketplaces, smaller startups and researchers can better plan and afford necessary GPU capacity, fostering a more competitive and innovative AI ecosystem.

The compute futures market could lead to a fundamental shift in resource allocation logic, moving from reliance on spot purchases to being more influenced by financial market price signals.

As compute becomes a financialized commodity, its allocation will increasingly be driven by forward pricing and hedging strategies, optimizing capital deployment for data center buildouts and long-term capacity planning.

⏳ Timeline

2022

Carmen Li leaves Citi, identifying a gap in compute risk management.

2024

Silicon Data is founded by Carmen Li.

2025-01

Compute Exchange launches as the world's first open exchange for GPUs.

2025-03

Silicon Data raises $4.7M in seed funding.

2025-09

Silicon Data publishes the industry's first GPU compute index on Bloomberg Terminal.

2026-04

Silicon Data unveils its first GPU Forward Curve.

2026-05-12

CME Group and Silicon Data announce partnership to launch a compute futures market later in 2026.

2026-05-19

Intercontinental Exchange (ICE) announces plans to launch GPU computing power futures based on Ornn's OCPI index.

Building a Compute Futures Market Bigger Than Oil

⚡ 30-Second TL;DR

🧠 Deep Insight

🔑 Enhanced Key Takeaways

🛠️ Technical Deep Dive

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline

📎 Sources (29)

👉Related Updates

Meta to sunset Llama API public preview

South Korea Plans Semiconductor Growth Fund from Tax Windfall