Silicon Data Raises $30 Million for CME AI Compute Futures
Silicon Data raises $30 million to price GPU rentals and plans CME compute futures on October 5, giving Wall Street a hedge for changing AI compute costs.
Summary
Silicon Data closed a $30 million Series A to establish a reference price for GPU rentals and an index against which a Wall Street futures contract could settle. TechCrunch reported on August 19, 2026, that annual data center and GPU investment totals hundreds of billions of dollars, making compute the largest cost for AI product builders, yet firms lack a simple pricing benchmark or a way to hedge changing prices.
Pending regulatory approval, Silicon Data plans to launch compute futures trading on the CME on October 5. In a 31 minute TechCrunch Equity podcast, host Rebecca Bellan and Silicon Data research chief Steve Hou examine the AI infrastructure expansion and why available data conflicts with headlines warning of depreciating chips and stalled data centers.
Positives
- Silicon Data secured $30 million in Series A funding to develop GPU rental pricing and compute derivatives.
- October 5 is the planned CME launch date for compute futures, subject to regulatory approval.
- A GPU rental reference price could give AI companies and investors a consistent compute benchmark.
- Steve Hou says available data challenges reports of depreciating chips and stalled data centers.
Risks & concerns
- Hundreds of billions of dollars in annual data center and GPU spending have made compute the largest cost for AI product builders.
- GPU rental markets still lack a straightforward reference price for valuing compute.
- Firms currently lack a futures instrument for hedging exposure when compute prices change.
- CME trading cannot begin on October 5 unless regulators approve Silicon Data’s planned compute futures.