Injective Debuts GPU Rental Index with 10x Leverage

The index, developed alongside London-based firm LO:TECH, calculates a benchmark hourly rate derived from a seven-day average of weighted rental-price data. Nvidia’s H100 SXM 80GB dominates the basket with a 40% weighting, followed by the B200 at 25% and the H200 at 15%. The remaining 20% is split equally between the A100 SXM 80GB and the consumer-grade RTX 4090, which the company includes to account for smaller AI workloads and open-source model training.

LO:TECH sources raw data from various compute marketplaces, applying outlier filters and validation checks before normalizing prices against a July 22, 2026, baseline. To ensure transparency, Injective maintains that the methodology allows for full reproducibility from archived observations. If the index methodology is adjusted, the platform will launch a separate version rather than overwriting existing historical data.

This launch expands Injective’s existing suite of AI-focused financial products, following a previous Nvidia-linked derivative introduced in August 2025. The move coincides with broader industry interest in computing-based derivatives; the Hyperliquid Policy Center recently petitioned U.S. regulators to establish standardized contracts for GPU rental and inference costs. Meanwhile, the physical hardware market continues to see collateralized lending activity, highlighted by a recent $100 million facility from Bullish to finance GPU-backed loans.