DEVELOPERS
RESOURCES
AI is scaling faster than any infrastructure category in modern history. But the most critical input powering that growth—GPU compute—remains opaque, fragmented, and difficult to price.
Today, we’re excited to announce that Squaretower is now a data publisher on Stork, bringing high-fidelity compute indices for GPU rental pricing on-chain.
This integration unlocks a new class of markets, financial products, and risk management tools built around one of the most essential commodities of the next century.
Squaretower is building the financial infrastructure for the compute economy. Their platform aggregates GPU rental pricing across multiple sources, including GPU cloud providers, normalizing non-standard configurations into standardized, comparable pricing metrics.
The result: trusted, granular indices for GPU compute, delivered with the transparency and consistency required to power financial markets.
With Squaretower publishing directly to Stork, this data becomes natively available to on-chain applications with the same guarantees developers expect from mission-critical price feeds: low latency, redundancy, and cryptographic integrity.
Squaretower’s NVIDIA H100 pricing index is now live on Stork.
The index captures on-demand H100 GPU rental pricing in near-real time, aggregating across a diverse set of providers and updating hourly to reflect real market behavior.
Example use-cases include:
Squaretower data is now available through Stork's push and pull based oracles, with support for EVM, SVM, Move, and CosmWasm chains, as well as via API for off-chain use cases.
To get started:
