# H100 reference price

> The GPUQuant H100 reference price: the on-demand list price of one H100 for one hour on the eight-GPU node, averaged across AWS, Azure, Google Cloud and Oracle, monthly since 2023.

Updated 2026-09-06. Canonical: https://gpuquant.com/docs/market-data/h100-reference-price

## What it measures

One NVIDIA H100 80GB SXM, rented for one hour on the eight-GPU node, at the on-demand list price for Linux, in US dollars. It is the price a provider publishes on its own price list before any negotiation, which makes it the ceiling of the market and the one figure for an H100-hour that can be checked against a URL.

The node is held constant on every provider: AWS p5, Azure ND H100 v5, Google Cloud a3-highgpu-8g and Oracle’s BM.GPU.H100.8. The same silicon in the same NVLink-connected form factor is what makes the cross-provider figure meaningful, and the smaller and PCIe variants are excluded rather than merged in.

The series runs monthly from July 2023 and is the longest hyperscaler history on the site after the A100’s. Its largest move is AWS’s cut of June 2025, 44 percent across eleven regions, which the change log under the chart records to the day.

Oracle joined as the fourth provider from August 2026, which moved the figure from $12.21 to $11.66 without any provider changing a price; that is a coverage step, annotated on the chart, and not a price move.

## How it is built

In two steps, every month. Per region, GPUQuant takes the lowest eligible published price for the eight-GPU node and divides it by eight. Per provider, it averages that provider's regional prices, so a provider that publishes the H100 in forty regions counts no more than one that publishes it in six. The reference is the average of the provider figures, and the median is computed and stored beside it on every row.

A provider with no eligible offering is absent, never zero. A provider that has not republished holds its last price forward and is marked held. Google Cloud prices the accelerator separately from the machine, so its complete machine price is rebuilt from the GPU, core and memory SKUs before dividing; Oracle prices per GPU-hour for a whole shape and is used as published. The arithmetic is on [The reference price](https://gpuquant.com/docs/reference-price) and [Normalization](https://gpuquant.com/docs/normalization), and the calculation version is **v4**.

## Publication

**Source**

GPUQuant

**Unit**

USD / GPU-hour

**Frequency**

Monthly: the price lists are read daily, and a month’s figure is the last reading taken in it

**Ticker**

None; a GPUQuant figure

## What GPUQuant shows

The whole history, from the first month GPUQuant priced the part, with every provider's own regional average as a row under the reference, the provider comparison and every regional offering behind it under the chart, and the change log of every list-price move dated to the provider's own effective date. Nothing is licensed and nothing is windowed: every figure has a URL behind it.

On [Market Data](https://gpuquant.com/market-data) as GPUQuant Reference under H100.

## Reading it beside the others

-   [H100 neocloud index](https://gpuquant.com/docs/market-data/h100-neocloud-index): The list price of the same hour on the specialist clouds, roughly a third of the hyperscaler figure.
-   [Silicon Data H100 index](https://gpuquant.com/docs/market-data/silicon-data-h100-index): Silicon Data's modelled figure, which counts negotiated and committed-use deals and sits below any list price.
-   [Ornn H100 index](https://gpuquant.com/docs/market-data/ornn-h100-index): Ornn's transacted figure, the prices H100 rentals actually cleared at.
-   [H100 forward curve](https://gpuquant.com/docs/market-data/h100-forward-curve): What the market expects the H100 rental index to be out to three years.
-   [H100 residual value](https://gpuquant.com/docs/market-data/h100-residual-value): What one H100 is still worth, from that curve.

## Sources

-   [Data sources](https://gpuquant.com/docs/data-sources), the provider price lists every figure is read from.
-   [Coverage](https://gpuquant.com/docs/coverage), which provider publishes the part where, and every exclusion with its reason.
-   [Time and coverage](https://gpuquant.com/docs/history), how a held price and a missing provider are shown.