Aquanode GPU Index

Figures on this page are computed live from Aquanode's marketplace aggregation feed, which polls 9 GPU cloud providers for on-demand list pricing. Each row groups all current listings for one GPU model and reports the lowest and median hourly price across those listings, the number of live offers, how many distinct providers carry that model, how many listings currently show stock, and how many distinct regions it's offered in.

All prices are normalized to a per-GPU hourly rate using each offer's authoritative GPU count, so that raw price is divided by the number of GPUs it actually covers to get a comparable $/hr-per-GPU figure; Akash reports its price as already per-GPU, so it is used as-listed.

On marketplace-style providers (e.g. Vast.ai, SimplePod), prices are host-set and legitimately disperse widely for the same GPU model, so we report lowest, median, and p90 per model rather than a single average or a raw maximum, so that one unusually priced listing can't distort the picture. A high-side outlier priced more than 3x its model's median is excluded from those figures instead of being published. The guard is high-side only: a genuinely cheap listing is never suppressed. An offer priced more than 20x its model's median is also logged internally as a likely data or unit-conversion issue for follow-up. No offers were excluded from this snapshot for a missing or invalid price. 7 were excluded as high-side price outliers.

These are on-demand list prices only. Spot/interruptible pricing and committed-use/reserved pricing are not included. Region labels are normalized for whitespace but otherwise reported as providers report them, so region granularity varies (some providers report a country, others a city or data-center code). This page regenerates at most once per hour.