Silicon Data Tracks AI Chip Rental Market
The firm secured $30.5 million to expand its compute indices as demand for second-hand hardware grows.
Updated on Oct. 21, 2026 in Semiconductors
Live Poll
Is now a good time for businesses to invest in older technology instead of the newest?
Silicon Data has secured $30.5 million in Series A funding to expand its market intelligence services for AI compute. The startup provides indices tracking rental rates and residual values for hardware like Nvidia H100 and A100 chips.
Why it matters
Rising costs for newer AI hardware are driving increased demand for previous-generation compute, forcing companies to track chip depreciation and rental pricing as commodities. This shift highlights a maturing market where AI silicon is increasingly managed as a liquid financial asset rather than just capital expenditure.
As of mid-October 2026, Nvidia H100 units hold 61% of their purchase value, while six-year-old A100 chips retain 22.5%. Current market rates show H100 rentals at $2 or more per hour, trailing the $5 per hour benchmark for newer B200 GPUs.
The players
Silicon Data
A financial technology startup that tracks GPU rental rates and residual hardware values for the AI compute market.
Nvidia
The primary designer of the H100, A100, and B200 GPUs that currently dominate global AI training and inference infrastructure.
CME Group
A global derivatives exchange operator partnering with Silicon Data to launch GPU-linked financial products.
The details
Silicon Data tracks the fluctuating costs of AI compute by aggregating rental rates and hardware depreciation across the global market. The company converts these proprietary price data points into indices, allowing users to benchmark the hourly cost of training or inference workloads against current hardware valuations. These indices help buyers assess whether leasing older, depreciated hardware remains more cost-effective than investing in current-generation units.
Timeline
April 2024: Silicon Data is founded.
May 27, 2025: The H100 Rental Index is launched.
August 2026: Silicon Data secures $30.5 million in Series A funding.
mid-October 2026: H100 residual value is recorded at 61%.
fall 2026: Planned launch of cash-settled GPU futures and options.
The Tech Race
Silicon Data is attempting to build the foundational infrastructure for an AI compute derivatives market, which could eventually reach $6 to $30 trillion annually. This effort mirrors the historical transition of volatile industrial inputs into standardized, liquid financial commodities.
Enterprises and cloud providers can now use these indices to benchmark their hourly cloud GPU spending against market averages. The data serves as a guide for procurement teams deciding between purchasing new hardware or leveraging the secondary rental market.
The takeaway
As AI compute becomes a critical operating cost, hardware is shifting from static asset to tradable financial instrument. Monitor the fall 2026 launch of derivatives for the first real-world test of market demand for compute-based risk management.
What happens next
Watch for the planned launch of cash-settled GPU futures and options with CME Group in fall 2026.
Further reading
Explore the evolving hardware landscape in our Semiconductors section.
Live Poll
Is now a good time for businesses to invest in older technology instead of the newest?
