Injective Launches Nvidia GPU Rental-Fee Perpetual Futures With Up to 10x Leverage
Summary
- Injective said it has launched a perpetual futures market that allows investors to trade swings in Nvidia GPU rental fees.
- The product is based on the Super Intelligence Compute Index (SICI), and investors can use up to 10x leverage.
- The launch was described as a product that offers a way to invest in and manage risk tied to AI computing costs and GPU rental-fee volatility.
Forecast Trend Report by Period



Layer-1 blockchain Injective has launched a perpetual futures market that lets investors trade swings in rental fees for Nvidia's graphics processing units, or GPUs.
Cointelegraph reported on October 27 that Injective now offers perpetual futures based on the Super Intelligence Compute Index, or SICI, which tracks rental prices for five Nvidia GPUs. Investors can use up to 10x leverage to bet on both rises and declines in GPU rental fees.
The SICI reflects rental rates for five Nvidia GPU models: the H100, B200, H200, A100 and RTX 4090. The index uses pricing data collected by London-based market data firm LO:TECH from computing-resource trading platforms.
The H100 carries the largest weight in the index at 40%, followed by the B200 at 25%. The index is updated hourly based on the average rental fee over the previous seven days, and USDC is used as trading collateral.
The launch expands on a product Injective introduced in August 2025 that was linked to H100 rental fees. By broadening trading from a single GPU model to five, the product enables investment and risk management across AI computing costs more broadly.
Injective said the product could help AI developers and GPU suppliers reduce exposure to cost swings. AI developers can hedge against rising rental fees, while GPU suppliers can hedge the risk of declining rental income.
At the same time, concerns are growing about profitability as investment in AI infrastructure accelerates. Gartner warned this month that more than 80% of neocloud providers specializing in AI cloud services could stop operating independently by 2030 as they struggle to generate returns on investment. Against that backdrop, efforts to trade GPU rental-fee volatility through financial products also appear to be expanding.
Suehyeon Lee
shlee@bloomingbit.ioI'm reporter Suehyeon Lee, your Web3 Moderator.