Injective Launches AI Compute Trading Index
UpGatePositiveTechnology innovation

Injective Launches AI Compute Trading Index

Reading time: 3 min

Injective Launches New Benchmark for AI Compute Rental Rates

Injective has unveiled the Super Intelligence Compute Index, a novel benchmark designed to track the rental rates of five different NVIDIA GPUs. This index allows traders to take leveraged positions on-chain through perpetual futures contracts.

The new index consolidates the rental rates of five NVIDIA GPUs into a single, tradable figure. Functionally, it mirrors a stock market index, but instead of company shares, its components are the hourly prices for renting computing chips.

Exposure to this market is facilitated through perpetual futures, commonly known as “perps.” These are derivative contracts that do not have an expiration date, enabling traders to maintain long or short positions indefinitely as long as they remain adequately funded.

The underlying proposition is straightforward: traders who anticipate an increase in the cost of AI compute can go long on the index. Conversely, those expecting a decrease can go short. Companies that purchase compute resources might also utilize these positions to hedge against fluctuations in their expenditure.

This latest index is not an isolated development; Injective has been exploring the concept of trading compute prices for some time.

On August 18, 2025, the network introduced a perpetual futures market linked to tokenized NVIDIA H100 GPU rental rates. This was described in research as the first on-chain perps market built around this specific data input. The market operates on Helix, Injective’s decentralized exchange, and reportedly offers leverage of up to 5x. By late 2025, the H100 market had recorded over $20.3 million in cumulative trading volume.

Pricing data for the H100 market is sourced from Squaretower, a firm that provides a custom decentralized oracle to refresh rental rates hourly. Oracles act as crucial bridges between the external world and blockchain networks, as smart contracts cannot independently access real-world data like GPU rental prices. The reliability of the entire market hinges on the accuracy and consistency of the data fed by these oracles.

Earlier, on January 30, 2025, Injective launched the AIX AI Index perpetual. This product combined AI-focused cryptocurrency tokens with traditional equities, including NVIDIA. While the AIX index allowed traders to speculate on the broader AI theme, the H100 market focused specifically on the rental price of a single chip. The Super Intelligence Compute Index now expands this scope by encompassing five GPUs, offering a more comprehensive view.

The H100 product was conceptualized to serve both compute buyers and cloud service providers, who are positioned on opposite ends of the price spectrum. Startups developing AI models may be concerned about rising rental costs, while GPU providers leasing out their hardware might worry about falling prices. In theory, a liquid derivatives market can offer both parties a degree of price certainty.

The effectiveness of this theory, however, depends on market liquidity. While over $20.3 million in cumulative H100 volume represents a significant start, substantial corporate hedging activities would likely require considerably greater depth. The expansion of the index to include five GPUs could bolster liquidity by appealing to a wider range of participants than a single-chip market.

The design of the oracle is a critical factor. Unlike widely traded assets like Bitcoin or Apple stock, GPU rental prices are not consistently quoted on a single public exchange. The methodology Injective employs for sourcing, weighting, and updating the five rental rates will significantly influence trader confidence in the index.

Furthermore, the question remains whether compute effectively behaves as a commodity. GPU rental prices can vary based on the provider, contract duration, and geographical region. Condensing these diverse factors into a single tradable number involves inherent complexities and subjective judgments.

Tags:UpGatePositiveTechnology innovation
Copied