Datasets:
provider stringlengths 3 18 | gpu stringclasses 8
values | offer stringclasses 3
values | price_usd_per_gpu_hour float64 0.16 14.2 | listing stringlengths 2 68 | source_url stringlengths 25 50 | retrieved stringdate 2026-10-26 00:00:00 2026-10-26 00:00:00 |
|---|---|---|---|---|---|---|
Runpod | H100 SXM | on-demand | 3.99 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | H100 PCIe | on-demand | 2.89 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | H200 | on-demand | 5.29 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | B200 | on-demand | 7.99 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | A100 80GB | on-demand | 1.79 | Secure Cloud, PCIe and SXM | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | L40S | on-demand | 1.09 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | RTX 4090 | on-demand | 0.89 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | RTX 5090 | on-demand | 1.19 | Secure Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | H100 SXM | spot-community | 2.69 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | H100 PCIe | spot-community | 1.99 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | H200 | spot-community | 3.59 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | B200 | spot-community | 5.98 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | A100 80GB | spot-community | 1.19 | Community Cloud, PCIe | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | L40S | spot-community | 0.79 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | RTX 4090 | spot-community | 0.34 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Runpod | RTX 5090 | spot-community | 0.69 | Community Cloud | https://www.runpod.io/pricing | 2026-10-26 |
Lambda | H100 SXM | on-demand | 3.99 | 8x instance, per GPU (1x: $4.29) | https://lambda.ai/pricing | 2026-10-26 |
Lambda | H100 PCIe | on-demand | 3.29 | 1x instance | https://lambda.ai/pricing | 2026-10-26 |
Lambda | B200 | on-demand | 6.69 | 8x instance, per GPU (1x: $6.99) | https://lambda.ai/pricing | 2026-10-26 |
Lambda | A100 80GB | on-demand | 2.79 | 8x SXM instance, per GPU | https://lambda.ai/pricing | 2026-10-26 |
CoreWeave | H100 SXM | on-demand | 6.155 | HGX H100, 8x instance $49.24 / 8, North America | https://www.coreweave.com/pricing | 2026-10-26 |
CoreWeave | H200 | on-demand | 6.305 | HGX H200, 8x instance $50.44 / 8, North America | https://www.coreweave.com/pricing | 2026-10-26 |
CoreWeave | B200 | on-demand | 8.6 | HGX B200, 8x instance $68.80 / 8, North America | https://www.coreweave.com/pricing | 2026-10-26 |
CoreWeave | A100 80GB | on-demand | 2.7 | 8x instance $21.60 / 8, North America | https://www.coreweave.com/pricing | 2026-10-26 |
CoreWeave | L40S | on-demand | 2.25 | 8x instance $18.00 / 8, North America | https://www.coreweave.com/pricing | 2026-10-26 |
Together AI | H100 SXM | on-demand | 3.99 | HGX H100 GPU cluster, pay as you go | https://www.together.ai/pricing | 2026-10-26 |
Together AI | H200 | on-demand | 5.99 | HGX H200 GPU cluster, pay as you go | https://www.together.ai/pricing | 2026-10-26 |
Together AI | B200 | on-demand | 8.19 | HGX B200 GPU cluster, pay as you go | https://www.together.ai/pricing | 2026-10-26 |
Together AI | H100 SXM | spot-community | 1.99 | preemptible HGX H100 | https://www.together.ai/pricing | 2026-10-26 |
Together AI | H200 | spot-community | 2.99 | preemptible HGX H200 | https://www.together.ai/pricing | 2026-10-26 |
Together AI | B200 | spot-community | 4.09 | preemptible HGX B200 | https://www.together.ai/pricing | 2026-10-26 |
Nebius | H200 | on-demand | 5.4 | on-demand | https://nebius.com/prices | 2026-10-26 |
Nebius | B200 | on-demand | 8.5 | on-demand | https://nebius.com/prices | 2026-10-26 |
Nebius | L40S | on-demand | 1.55 | from, Intel CPU configuration | https://nebius.com/prices | 2026-10-26 |
DigitalOcean | H100 SXM | on-demand | 4.41 | HGX H100 GPU Droplet | https://www.digitalocean.com/pricing/gpu-droplets | 2026-10-26 |
DigitalOcean | H200 | on-demand | 4.47 | HGX H200 GPU Droplet | https://www.digitalocean.com/pricing/gpu-droplets | 2026-10-26 |
DigitalOcean | L40S | on-demand | 1.57 | GPU Droplet | https://www.digitalocean.com/pricing/gpu-droplets | 2026-10-26 |
Hyperstack | H100 SXM | on-demand | 3.2 | on-demand | https://www.hyperstack.cloud/gpu-pricing | 2026-10-26 |
Hyperstack | H200 | on-demand | 3.99 | H200 SXM, on-demand | https://www.hyperstack.cloud/gpu-pricing | 2026-10-26 |
Hyperstack | B200 | on-demand | 6 | on-demand | https://www.hyperstack.cloud/gpu-pricing | 2026-10-26 |
Hyperstack | A100 80GB | on-demand | 1.35 | A100 80GB PCIe, on-demand (SXM: $1.60) | https://www.hyperstack.cloud/gpu-pricing | 2026-10-26 |
Hyperstack | A100 80GB | spot-community | 1.08 | spot | https://www.hyperstack.cloud/gpu-pricing | 2026-10-26 |
Thunder Compute | H100 SXM | on-demand | 3.2 | GPU instance | https://www.thundercompute.com/pricing | 2026-10-26 |
Thunder Compute | A100 80GB | on-demand | 1.09 | GPU instance | https://www.thundercompute.com/pricing | 2026-10-26 |
JarvisLabs | H100 SXM | on-demand | 3.49 | on-demand | https://jarvislabs.ai/pricing | 2026-10-26 |
JarvisLabs | H200 | on-demand | 4.59 | H200 SXM, on-demand | https://jarvislabs.ai/pricing | 2026-10-26 |
JarvisLabs | A100 80GB | on-demand | 1.49 | on-demand | https://jarvislabs.ai/pricing | 2026-10-26 |
Verda (DataCrunch) | H100 SXM | on-demand | 4.606 | H100 SXM5 instance | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | H200 | on-demand | 6.006 | H200 SXM5 instance | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | B200 | on-demand | 8.523 | B200 SXM6 instance | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | A100 80GB | on-demand | 2.16 | A100 SXM4 80GB instance | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | L40S | on-demand | 1.827 | instance | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | H100 SXM | spot-community | 2.303 | spot | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | H200 | spot-community | 3.003 | spot | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | B200 | spot-community | 4.262 | spot | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | A100 80GB | spot-community | 1.08 | spot | https://verda.com/pricing | 2026-10-26 |
Verda (DataCrunch) | L40S | spot-community | 0.9135 | spot | https://verda.com/pricing | 2026-10-26 |
Massed Compute | H100 SXM | on-demand | 3.14 | H100 SXM5, x1 | https://vm.massedcompute.com/pricing | 2026-10-26 |
Massed Compute | H200 | on-demand | 3.62 | H200 NVL, x1 | https://vm.massedcompute.com/pricing | 2026-10-26 |
Massed Compute | A100 80GB | on-demand | 1.35 | A100 80GB, x1 | https://vm.massedcompute.com/pricing | 2026-10-26 |
Massed Compute | L40S | on-demand | 0.97 | x1 | https://vm.massedcompute.com/pricing | 2026-10-26 |
Oblivus | H100 SXM | on-demand | 4.31 | H100 80GB SXM5 | https://www.oblivus.com/pricing/ | 2026-10-26 |
Oblivus | H100 PCIe | on-demand | 3.5 | H100 80GB PCIe | https://www.oblivus.com/pricing/ | 2026-10-26 |
Oblivus | H200 | on-demand | 3.99 | H200 141GB SXM5 | https://www.oblivus.com/pricing/ | 2026-10-26 |
Oblivus | A100 80GB | on-demand | 1.47 | A100 80GB PCIe | https://www.oblivus.com/pricing/ | 2026-10-26 |
Oblivus | RTX 4090 | on-demand | 0.64 | on-demand | https://www.oblivus.com/pricing/ | 2026-10-26 |
CloudRift | RTX 4090 | on-demand | 0.39 | on-demand | https://www.cloudrift.ai/gpu-rentals | 2026-10-26 |
CloudRift | RTX 5090 | on-demand | 0.65 | on-demand | https://www.cloudrift.ai/gpu-rentals | 2026-10-26 |
Salad | RTX 4090 | spot-community | 0.16 | distributed, lowest (batch) priority; high priority $0.330 | https://salad.com/pricing | 2026-10-26 |
Salad | RTX 5090 | spot-community | 0.25 | distributed, lowest (batch) priority; high priority $0.500 | https://salad.com/pricing | 2026-10-26 |
Lium | RTX 4090 | spot-community | 0.65 | marketplace, lowest live ask | https://lium.io/gpu/rtx-4090 | 2026-10-26 |
AWS | H100 SXM | reserved-capacity-block | 5.97 | p5.48xlarge Capacity Block, US East (N. Virginia), $47.757 / 8 | https://aws.amazon.com/ec2/capacityblocks/pricing/ | 2026-10-26 |
AWS | H200 | reserved-capacity-block | 6.866 | p5e.48xlarge Capacity Block, US East (Ohio), $54.924 / 8 | https://aws.amazon.com/ec2/capacityblocks/pricing/ | 2026-10-26 |
AWS | B200 | reserved-capacity-block | 14.208 | p6-b200.48xlarge Capacity Block, US East (N. Virginia), $113.666 / 8 | https://aws.amazon.com/ec2/capacityblocks/pricing/ | 2026-10-26 |
AWS | A100 80GB | reserved-capacity-block | 2.546 | p4de.24xlarge Capacity Block, US East (N. Virginia), $20.369 / 8 | https://aws.amazon.com/ec2/capacityblocks/pricing/ | 2026-10-26 |
EverXCloud GPU Rental Price Report, November 2026
Public list prices for renting NVIDIA data center and consumer GPUs in the cloud, normalized to USD per GPU-hour. The dataset contains 75 prices for eight GPUs (H100 SXM, H100 PCIe, H200, B200, A100 80GB, L40S, RTX 4090 and RTX 5090) across 16 cloud providers, read on 2026-10-26 from each provider's official pricing page. Every row records the provider, the GPU, the offer type, the price, the exact listing used and the source URL, so each figure can be traced and checked. A per-GPU summary gives the number of providers, minimum, median, maximum and spread for each offer type.
The dataset is published by EverXCloud Inc., a US-based GPU rental marketplace. It contains third-party public prices only; no EverXCloud prices are included.
- Report page: https://everxcloud.ai/en/reports/gpu-rental-price-report-2026-11
- Source CSV: https://everxcloud.ai/media/reports/everxcloud-gpu-rental-price-report-2026-11.csv
- Website: https://everxcloud.ai
Key findings
- The median on-demand H100 SXM price across 11 providers is $3.99 per GPU-hour, from $3.14 to $6.16: the most expensive listing costs 2.0× the cheapest.
- Spot, preemptible and community H100 SXM capacity has a median of $2.30 per GPU-hour across 3 providers, 42% below the on-demand median.
- The H200 carries a 24% premium over the H100 SXM at the median ($4.94 vs $3.99 on demand, 10 providers).
- On-demand B200 prices range from $6.00 to $8.60 per GPU-hour across 7 providers, with a median of $8.19.
- The A100 80GB shows the widest relative spread among data center GPUs: 2.6× between the cheapest ($1.09) and the most expensive ($2.79) on-demand listing, across 9 providers.
- AWS Capacity Blocks, a short-term reservation, list the H100 at $5.97 and the B200 at $14.21 per GPU-hour, 1.5× and 1.7× the on-demand medians of the other providers.
On-demand summary (USD per GPU-hour)
| GPU | Providers | Min | Median | Max | Spread (max ÷ min) |
|---|---|---|---|---|---|
| H100 SXM | 11 | 3.14 | 3.99 | 6.155 | 1.96× |
| H100 PCIe | 3 | 2.89 | 3.29 | 3.50 | 1.21× |
| H200 | 10 | 3.62 | 4.94 | 6.305 | 1.74× |
| B200 | 7 | 6.00 | 8.19 | 8.60 | 1.43× |
| A100 80GB | 9 | 1.09 | 1.49 | 2.79 | 2.56× |
| L40S | 6 | 0.97 | 1.56 | 2.25 | 2.32× |
| RTX 4090 | 3 | 0.39 | 0.64 | 0.89 | 2.28× |
| RTX 5090 | 2 | 0.65 | 0.92 | 1.19 | 1.83× |
Spot, preemptible and community statistics, and the AWS Capacity Block prices, are in data/summary_2026_11.csv.
Files
| File | Description |
|---|---|
data/gpu_rental_prices_2026_11.csv |
75 price rows, one per provider, GPU and offer type. Byte-identical to the CSV published with the report. |
data/gpu_rental_prices_2026_11.json |
The same 75 rows as a JSON array, with prices as numbers. |
data/summary_2026_11.csv |
Statistics per GPU and offer type, computed from the price rows. |
scripts/check.py |
Integrity check (Python standard library only): row count, columns, values and summary consistency. |
CITATION.cff |
Citation metadata. |
LICENSE |
Creative Commons Attribution 4.0 International legal code. |
Column dictionary
gpu_rental_prices_2026_11.csv / .json
| Column | Type | Description |
|---|---|---|
provider |
string | Cloud provider name. |
gpu |
string | GPU model: H100 SXM, H100 PCIe, H200, B200, A100 80GB, L40S, RTX 4090 or RTX 5090. |
offer |
string | Offer type: on-demand; spot-community (spot, preemptible or community capacity); reserved-capacity-block (short-term reservation, AWS Capacity Blocks only). |
price_usd_per_gpu_hour |
number | Public list price in USD per GPU-hour, excluding tax. |
listing |
string | The listing the price was taken from, including the per-GPU conversion when the provider prices a multi-GPU instance (for example 8x instance $49.24 / 8). |
source_url |
string | Official pricing page the price was read from. |
retrieved |
date | Date the price was read (2026-10-26). |
summary_2026_11.csv
| Column | Type | Description |
|---|---|---|
gpu |
string | GPU model. |
offer |
string | Offer type (same values as above). |
providers |
integer | Number of providers with a price for this GPU and offer type (one price per provider). |
min_usd_per_gpu_hour |
number | Lowest price. |
median_usd_per_gpu_hour |
number | Median price (mean of the two middle values when the count is even). |
max_usd_per_gpu_hour |
number | Highest price. |
spread_max_over_min |
number | Highest price divided by lowest price, rounded to three decimals. |
Methodology
- Prices were read on 2026-10-26 on each provider's official public pricing page, in USD per GPU-hour, excluding tax. No price was estimated, negotiated or taken from a third-party aggregator.
- One price per provider, GPU and offer type. When a provider lists several configurations for the same GPU, the lowest listed price was kept, and the listing is recorded in the data.
- Multi-GPU instances are converted to a per-GPU price by dividing the instance price by the number of GPUs (for example CoreWeave and AWS 8-GPU instances).
- On-demand and spot, preemptible or community prices are reported separately. AWS prices are Capacity Block rates, a short-term reservation, and are not included in the on-demand statistics.
- Providers whose pricing page could not be read, showed prices only after sign-in or on request, or did not map prices to GPUs unambiguously were left out: Vast.ai, Crusoe, Vultr, TensorDock, Fluidstack, Genesis Cloud, Google Cloud, Microsoft Azure, Voltage Park, Cudo Compute, Scaleway, Paperspace and SF Compute.
- List prices change often. The data reflects the retrieval date only and does not capture discounts, committed-use contracts or regional variation beyond what each listing states.
Providers and sources
| Provider | Pricing page |
|---|---|
| AWS | https://aws.amazon.com/ec2/capacityblocks/pricing/ |
| CloudRift | https://www.cloudrift.ai/gpu-rentals |
| CoreWeave | https://www.coreweave.com/pricing |
| DigitalOcean | https://www.digitalocean.com/pricing/gpu-droplets |
| Hyperstack | https://www.hyperstack.cloud/gpu-pricing |
| JarvisLabs | https://jarvislabs.ai/pricing |
| Lambda | https://lambda.ai/pricing |
| Lium | https://lium.io/gpu/rtx-4090 |
| Massed Compute | https://vm.massedcompute.com/pricing |
| Nebius | https://nebius.com/prices |
| Oblivus | https://www.oblivus.com/pricing/ |
| Runpod | https://www.runpod.io/pricing |
| Salad | https://salad.com/pricing |
| Thunder Compute | https://www.thundercompute.com/pricing |
| Together AI | https://www.together.ai/pricing |
| Verda (DataCrunch) | https://verda.com/pricing |
Usage
from datasets import load_dataset
prices = load_dataset("everxcloud/gpu-rental-prices", "prices", split="train")
summary = load_dataset("everxcloud/gpu-rental-prices", "summary", split="train")
import pandas as pd
prices = pd.read_csv("data/gpu_rental_prices_2026_11.csv")
prices[prices.offer == "on-demand"].groupby("gpu").price_usd_per_gpu_hour.median()
Run python3 scripts/check.py to verify the files.
Update cadence
Monthly. Each edition is published as new dated files (*_YYYY_MM.csv); earlier editions are kept unchanged.
How to cite
Suggested citation: "Source: EverXCloud GPU Rental Price Report, November 2026", with a link to https://everxcloud.ai/en/reports/gpu-rental-price-report-2026-11.
@misc{everxcloud_gpu_rental_price_report_2026_11,
author = {{EverXCloud Inc.}},
title = {GPU Rental Price Report, November 2026},
year = {2026},
month = nov,
version = {2026.11},
howpublished = {\url{https://everxcloud.ai/en/reports/gpu-rental-price-report-2026-11}},
note = {Prices retrieved 2026-10-26. Licensed under CC BY 4.0}
}
License
Creative Commons Attribution 4.0 International (CC BY 4.0). You may share and adapt the data for any purpose, including commercial use, provided you give appropriate credit. See LICENSE and https://creativecommons.org/licenses/by/4.0/.
Provider names and trademarks belong to their respective owners. Prices are public list prices as displayed on the retrieval date and may have changed since.
- Downloads last month
- 43