# H100 neocloud index > The GPUQuant H100 neocloud index: the on-demand list price of one H100 for one hour on the specialist GPU clouds, read from each provider’s own price list and averaged, monthly since April 2023. Updated 2026-09-06. Canonical: https://gpuquant.com/docs/market-data/h100-neocloud-index ## What it measures One H100-hour at the on-demand list price on the specialist GPU clouds, read from each provider’s own price list, normalized to the same eight-GPU node the reference price holds constant, and averaged with every provider counting once. It is the second tier beside the hyperscaler reference, and the same GPU-hour sells for roughly a third of the hyperscaler list price on it. Only an on-demand rate against a stated configuration counts. A published “from” floor, a reserved or committed rate, a spot or preemptible rate, and a rate in a currency other than dollars are stored and labelled and never enter the index; each provider’s other rates sit under the index as rows of their own. The history runs from April 2023, read from Internet Archive captures of the same pages the live reader parses, because no neocloud publishes a price archive. CoreWeave’s 2023 page priced the GPU as a component with host rates beside it, so the adapter adds the node’s share of host at those printed rates to keep the figure comparable with the complete machines elsewhere. A month no capture recorded is gone, which is why the live pages are read daily. This is the like-for-like comparison for the Silicon Data H100 index and the CME H100 contract, both of which are neo-cloud readings; a list price against a modelled transacted figure, for the same segment. ## How it is built From each provider's own price list, read every day. Only an on-demand rate against a stated configuration counts; a “from” floor, a reserved or committed rate, a spot or preemptible rate and a rate in a currency other than dollars are stored, labelled and kept out. Every quote is normalized to one GPU-hour on the eight-GPU node, a provider that has not republished holds its last reading forward, and the index is the average of the providers with every provider counting once, the same estimator as the reference price. The median is stored beside it. 16 providers are read; those whose published figure cannot be used are kept with the reason stated on their row. The rules and every provider's reading are on [The neocloud index](https://gpuquant.com/docs/neocloud-index), and the calculation version is **v2**. ## 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, with every provider's on-demand rate as a row under the index and its other rates, spot, reserved and committed, as rows of their own, read straight from its price list and never averaged. A rate that stopped being published ends where it ended. On [Market Data](https://gpuquant.com/market-data) as GPUQuant Reference under H100. ## Reading it beside the others - [H100 reference price](https://gpuquant.com/docs/market-data/h100-reference-price): The hyperscaler list price of one H100-hour, the ceiling of the market. - [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 - [The neocloud index](https://gpuquant.com/docs/neocloud-index), every provider read and the source page of each.