# GPU cloud pricing: GPU Compute Price List

> The GPU Compute Price List carries 15 published figures read from 5 sources' own pages, from 0.48 to 12.42 USD, read 12 October 2026.

AI Implementation Cost · https://aiimplementationcost.com/datasets/gpu-compute-price-list/ · computed 2026-10-23

**Range of the published figures:** 0.48 to 12.42 USD (range 0.48 to 12.42 USD; 14 published figures; 5 of 24 sources read publish one).

5 of the 24 named sources read in full publish a figure; 19 publish none, read 12 October 2026.

## By accelerator

| Breakdown | Median | Range | Sources |
|---|---|---|---|
| accelerator H100 SXM | 2.69 USD | 2.69 to 2.69 USD | 1 |
| accelerator A100 | 1.09 USD | 1.09 to 1.09 USD | 1 |
| accelerator H100 PCIe | 2.01 USD | 2.01 to 2.01 USD | 1 |
| accelerator H100 SXM5 | 2.57 USD | 2.57 to 2.57 USD | 1 |
| accelerator HGX B200 (2 weeks-1 year commitment) | 8.87 USD | 8.87 to 8.87 USD | 1 |
| accelerator P100 | 1.28 USD | 1.28 to 1.28 USD | 1 |
| accelerator H100 | 3.2 USD | 3.2 to 3.2 USD | 1 |
| accelerator B200 SXM6 | 9.12 USD | 9.12 to 9.12 USD | 1 |
| accelerator B300 | 12.42 USD | 12.42 to 12.42 USD | 1 |
| accelerator A100 80GB | 1.48 USD | 1.48 to 1.48 USD | 1 |
| accelerator H100 SXM5 (spot) | 1.98 USD | 1.98 to 1.98 USD | 1 |
| accelerator A100 80GB (spot) | 1.24 USD | 1.24 to 1.24 USD | 1 |
| accelerator B200 SXM6 (spot) | 3.73 USD | 3.73 to 3.73 USD | 1 |
| accelerator RTX 4090 | 0.48 USD | 0.48 to 0.48 USD | 1 |

## Every figure, with its source

One row per source: the figure the source's own page publishes, the page, and the day it was read.

| Source | Figure | Breakdown | Read |
|---|---|---|---|
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 12.42 USD | B300 | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 9.12 USD | B200 SXM6 | 2026-10-12 |
| [GPU cloud pricing: rent NVIDIA H100, H200, and B200](https://lambda.ai/pricing) | 8.87 USD | HGX B200 (2 weeks-1 year commitment) | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 3.73 USD | B200 SXM6 (spot) | 2026-10-12 |
| [NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates](https://www.thundercompute.com/blog/nvidia-h100-pricing) | 3.2 USD | H100 | 2026-10-12 |
| [NVIDIA H100 Price Guide 2026: GPU Costs, Cloud Pricing & Buy vs Rent](https://jarvislabs.ai/blog/h100-price) | 2.69 USD | H100 SXM | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 2.64 USD | H100 SXM5 | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 2.5 USD | H100 SXM5 | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 2.01 USD | H100 PCIe | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 1.98 USD | H100 SXM5 (spot) | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 1.48 USD | A100 80GB | 2026-10-12 |
| [Cloud Price List](https://www.oracle.com/cloud/price-list/) | 1.28 USD | P100 | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 1.24 USD | A100 80GB (spot) | 2026-10-12 |
| [NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates](https://www.thundercompute.com/blog/nvidia-h100-pricing) | 1.09 USD | A100 | 2026-10-12 |
| [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) | 0.48 USD | RTX 4090 | 2026-10-12 |

## The sentence each figure was read from (15)

- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "B300 (Spheron) $12.42/hr $4.19/hr Varies, compare live rates"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Spheron (B200 SXM6) $9.12/hr $3.73/hr Live rates"
- [GPU cloud pricing: rent NVIDIA H100, H200, and B200](https://lambda.ai/pricing), read 2026-10-12: "NVIDIA HGX B200 2 weeks – 1 year 256+ $8.87 Talk to our team"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Cheapest B200 per hour: Spheron B200 is $3.73/hr on spot and $9.12/hr on-demand."
- [NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates](https://www.thundercompute.com/blog/nvidia-h100-pricing), read 2026-10-12: "Thunder Compute $3.20 $32.00"
- [NVIDIA H100 Price Guide 2026: GPU Costs, Cloud Pricing & Buy vs Rent](https://jarvislabs.ai/blog/h100-price), read 2026-10-12: "An NVIDIA H100 SXM with 80 GB of GPU memory costs $2.69 per GPU-hour to rent on Jarvislabs, billed by the minute."
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "H100 SXM5 (Spheron) $2.64/hr $1.98/hr Varies, compare live rates"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Spheron (H100 SXM5) $2.50 $1.03 8-way HGX; lowest spot rate"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Azure charges ~$12.29/hr per GPU on their ND H100 v5 instances. On Spheron, H100 PCIe is $2.01/hr on-demand."
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Cheapest H100 SXM5 per hour: Spheron H100 is $1.98/hr on spot and $2.64/hr on-demand as of 12 Oct 2026."
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "A100 80GB (Spheron) $1.48/hr $1.24/hr Varies, compare live rates"
- [Cloud Price List](https://www.oracle.com/cloud/price-list/), read 2026-10-12: "VM.GPU2.1 1x NVIDIA P100 Pascal N/A 16 GB 12 78 GB 25 Gb/sec $1.275"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Cheapest A100 80GB per hour: Spheron A100 is $1.48/hr on-demand and $1.24/hr on spot."
- [NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates](https://www.thundercompute.com/blog/nvidia-h100-pricing), read 2026-10-12: "Provider GPU Listed $/GPU-hr 10-Hour Cost Thunder Compute A100 $1.09 $10.90"
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/), read 2026-10-12: "Cheapest RTX 4090 per hour: Vast.ai marketplace hosts go as low as $0.31/hr; Spheron lists RTX 4090 on-demand at $0.48/hr, and Runpod Community Cloud at $0.34/hr ($0.69/hr on Secure Cloud)."

## Read in full, publishes no figure

- [VM instance pricing](https://cloud.google.com/compute/gpus-pricing), read 2026-10-12: every candidate on the page was refused
- [Pricing - Linux Virtual Machines](https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/), read 2026-10-12: every candidate on the page was refused
- [Amazon EC2 P5 Instances](https://aws.amazon.com/ec2/instance-types/p5/), read 2026-10-12: every candidate on the page was refused
- [Rent NVIDIA GPUs on demand: H100, H200, and B200](https://lambda.ai/service/gpu-cloud), read 2026-10-12: read in full and no accepted figure on the page
- [CoreWeave Cloud Pricing](https://www.coreweave.com/pricing), read 2026-10-12: every candidate on the page was refused
- [GPU Cloud Pricing](https://www.runpod.io/pricing), read 2026-10-12: every candidate on the page was refused
- [nebius.com](https://nebius.com/prices), read 2026-10-12: every candidate on the page was refused
- [GPU Pricing](https://vast.ai/pricing), read 2026-10-12: every candidate on the page was refused
- [GPU Droplets Pricing](https://www.digitalocean.com/pricing/gpu-droplets), read 2026-10-12: every candidate on the page was refused
- [Cloud GPU Rental Price Index](https://aimultiple.com/gpu-index), read 2026-10-12: every candidate on the page was refused
- [Cloud GPU Pricing - Compare 79 Providers and 4,900+ Prices (2026)](https://getdeploying.com/gpus), read 2026-10-12: every candidate on the page was refused
- [Pricing](https://www.paperspace.com/pricing), read 2026-10-12: read in full and no accepted figure on the page
- [Price list: A comparison of our Public Cloud offers](https://www.ovhcloud.com/en/public-cloud/prices/), read 2026-10-12: every candidate on the page was refused
- [GPU Instances Pricing](https://www.scaleway.com/en/pricing/gpu/), read 2026-10-12: read in full and no accepted figure on the page
- [Server with GPU: for your AI and machine learning projects.](https://www.hetzner.com/dedicated-rootserver/matrix-gpu/), read 2026-10-12: read in full and no accepted figure on the page
- [Plan Pricing](https://www.modal.com/pricing), read 2026-10-12: every candidate on the page was refused
- [GPU pricing for AI compute & inference](https://www.crusoe.ai/cloud/pricing), read 2026-10-12: every candidate on the page was refused
- [Civo Prices - Transparent Cloud Pricing - Civo.com](https://www.civo.com/pricing), read 2026-10-12: every candidate on the page was refused
- [Pricing](https://www.together.ai/pricing#gpu-clusters), read 2026-10-12: every candidate on the page was refused

## Methodology

Every row is a verbatim sentence read from the named source's own page through our fetch service on the date shown, hash-pinned to the stored read (the sha of the page's visible text; the raw page is mirrored to R2). 5 of the 24 named sources that could be read published a usable figure; 19 published none and are recorded as not published, never filled in; 1 could not be read (blocked, dead or unreachable) and count nowhere. The headline figures (median, minimum, maximum, quartiles) are over ONE figure per source for each breakdown it prices, in USD (a source's median where its page yielded several accepted sentences for one breakdown), so a page that gives figures for several breakdowns gives one for each and N counts 14 figures, never sentences; 382 candidate sentence(s) were refused by a reviewer with the reason recorded. The fuller derivation this dataset is chartered to grow into (not yet what the figures above are; they are each source's own published figure as read): one row per accelerator per provider at the provider's own published on-demand hourly rate, with the reserved and spot rate where published, normalised to a price per accelerator-hour (the hyperscalers quote whole 8-GPU instances and the specialists quote single cards, so the two are not comparable as published) and to a price per unit of memory; the hyperscaler-to-specialist premium, which is the finding; repriced-since

## Sources

- [VM instance pricing](https://cloud.google.com/compute/gpus-pricing) (the publisher's own terms; quoted for reporting and comment)
- [Pricing - Linux Virtual Machines](https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/) (the publisher's own terms; quoted for reporting and comment)
- [Amazon EC2 P5 Instances](https://aws.amazon.com/ec2/instance-types/p5/) (the publisher's own terms; quoted for reporting and comment)
- [Rent NVIDIA GPUs on demand: H100, H200, and B200](https://lambda.ai/service/gpu-cloud) (the publisher's own terms; quoted for reporting and comment)
- [CoreWeave Cloud Pricing](https://www.coreweave.com/pricing) (the publisher's own terms; quoted for reporting and comment)
- [GPU Cloud Pricing](https://www.runpod.io/pricing) (the publisher's own terms; quoted for reporting and comment)
- [nebius.com](https://nebius.com/prices) (the publisher's own terms; quoted for reporting and comment)
- [GPU Pricing](https://vast.ai/pricing) (the publisher's own terms; quoted for reporting and comment)
- [Cloud Price List](https://www.oracle.com/cloud/price-list/) (the publisher's own terms; quoted for reporting and comment)
- [GPU Droplets Pricing](https://www.digitalocean.com/pricing/gpu-droplets) (the publisher's own terms; quoted for reporting and comment)
- [NVIDIA H100 Price Guide 2026: GPU Costs, Cloud Pricing & Buy vs Rent](https://jarvislabs.ai/blog/h100-price) (the publisher's own terms; quoted for reporting and comment)
- [NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates](https://www.thundercompute.com/blog/nvidia-h100-pricing) (the publisher's own terms; quoted for reporting and comment)
- [GPU cloud pricing: rent NVIDIA H100, H200, and B200](https://lambda.ai/pricing) (the publisher's own terms; quoted for reporting and comment)
- [GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr](https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/) (the publisher's own terms; quoted for reporting and comment)
- [Cloud GPU Rental Price Index](https://aimultiple.com/gpu-index) (the publisher's own terms; quoted for reporting and comment)
- [Cloud GPU Pricing - Compare 79 Providers and 4,900+ Prices (2026)](https://getdeploying.com/gpus) (the publisher's own terms; quoted for reporting and comment)
- [Pricing](https://www.paperspace.com/pricing) (the publisher's own terms; quoted for reporting and comment)
- [Price list: A comparison of our Public Cloud offers](https://www.ovhcloud.com/en/public-cloud/prices/) (the publisher's own terms; quoted for reporting and comment)
- [GPU Instances Pricing](https://www.scaleway.com/en/pricing/gpu/) (the publisher's own terms; quoted for reporting and comment)
- [Server with GPU: for your AI and machine learning projects.](https://www.hetzner.com/dedicated-rootserver/matrix-gpu/) (the publisher's own terms; quoted for reporting and comment)
- [www.genesiscloud.com](https://www.genesiscloud.com/pricing) (the publisher's own terms; quoted for reporting and comment)
- [Plan Pricing](https://www.modal.com/pricing) (the publisher's own terms; quoted for reporting and comment)
- [GPU pricing for AI compute & inference](https://www.crusoe.ai/cloud/pricing) (the publisher's own terms; quoted for reporting and comment)
- [Civo Prices - Transparent Cloud Pricing - Civo.com](https://www.civo.com/pricing) (the publisher's own terms; quoted for reporting and comment)
- [Pricing](https://www.together.ai/pricing#gpu-clusters) (the publisher's own terms; quoted for reporting and comment)

## Cite or embed

- The GPU Compute Price List carries 15 published figures read from 5 sources' own pages, from 0.48 to 12.42 USD, read 12 October 2026.
- 5 of the 24 named sources read in full publish a figure; 19 publish none, read 12 October 2026.
- GPU Compute Price List, accelerator H100 SXM: median 2.69 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator A100: median 1.09 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator H100 PCIe: median 2.01 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator H100 SXM5: median 2.57 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator HGX B200 (2 weeks-1 year commitment): median 8.87 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator P100: median 1.27 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator H100: median 3.20 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator B200 SXM6: median 9.12 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator B300: median 12.42 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator A100 80GB: median 1.48 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator H100 SXM5 (spot): median 1.98 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator A100 80GB (spot): median 1.24 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator B200 SXM6 (spot): median 3.73 USD across 1 source, read 12 October 2026.
- GPU Compute Price List, accelerator RTX 4090: median 0.48 USD across 1 source, read 12 October 2026.

Cite as: "aiimplementationcost.com GPU Compute Price List", updated 2026-10-23, https://aiimplementationcost.com/datasets/gpu-compute-price-list/.

CSV: https://aiimplementationcost.com/datasets/gpu-compute-price-list.csv
JSON: https://aiimplementationcost.com/datasets/gpu-compute-price-list.json
Embed (plain HTML, static link back): see https://aiimplementationcost.com/datasets/gpu-compute-price-list/#median