Price tracker
GPU rental prices by provider
Listed rental prices for NVIDIA H100, H200, B200, A100 and L40S and AMD MI300X from 10 providers, in dollars per GPU-hour, each linked to the page it came from. As of Oct 15, 2026.
Lowest listed on-demand price per GPU-hour
Lowest across every variant and instance size that is not stale. Prices were last checked Oct 15, 2026. A listing not re-checked within 7 days is marked stale and left out of every lowest-price figure. Providers are not like for like; see how prices are collected.
H100 · lowest on-demand$2.50 /GPU-hrHyperstack · checked Oct 15, 2026H200 · lowest on-demand$3.99 /GPU-hrHyperstack (SXM) · checked Oct 15, 2026B200 · lowest on-demand$6.00 /GPU-hrHyperstack · checked Oct 15, 2026A100 80GB · lowest on-demand$1.35 /GPU-hrHyperstack · checked Oct 15, 2026L40S · lowest on-demand$1.50 /GPU-hrCrusoe · checked Oct 15, 2026MI300X · lowest on-demand$2.59 /GPU-hrDigitalOcean · checked Oct 15, 2026
All listed prices
| Provider | GPU | Configuration | Pricing | $ per GPU-hour | Region | Checked | Source |
|---|---|---|---|---|---|---|---|
| Nebius | NVIDIA H100 | Configuration not stated HGX H100 |
Spot | $0.79 listed: from $0.79 (preemptible) |
not stated | nebius.com | |
| Together AI | NVIDIA H100 | Configuration not stated HGX H100 |
Spot | $1.99 listed: $1.99 (listed as preemptible) |
not stated | together.ai | |
| Microsoft Azure | NVIDIA H100 | 8 GPUs per instance · 80 GB each Standard_ND96isr_H100_v5 |
Spot | $2.2712 from retailPrice 18.169536 USD per 1 Hour (Spot) |
eastus | prices.azure.com | |
| CoreWeave | NVIDIA H100 | 8 GPUs per instance · 80 GB each HGX H100 |
Spot | $2.4638 from Spot Price: $19.71 / Hour |
North America | coreweave.com | |
| Hyperstack | NVIDIA H100 | 80 GB each H100 |
On-demand | $2.50 | not stated | hyperstack.cloud | |
| Hyperstack | NVIDIA H100 | NVLink · 80 GB each · NVLink (in SKU name) H100 NVLink |
On-demand | $2.60 | not stated | hyperstack.cloud | |
| Together AI | NVIDIA H100 | Configuration not stated HGX H100 |
Reserved, 91-180 days | $3.19 listed: $3.19 (reserved, 91-180 days) |
not stated | together.ai | |
| Hyperstack | NVIDIA H100 | SXM · 80 GB each H100 SXM |
On-demand | $3.20 | not stated | hyperstack.cloud | |
| DigitalOcean | NVIDIA H100 | 80 GB each gpu-h100x1-80gb |
Reserved, 12 months | $3.26 listed: $3.26/GPU/hour (12-month commitment) |
not stated | digitalocean.com | |
| Lambda | NVIDIA H100 | PCIe · 1 GPU per instance · 80 GB each 1x H100 |
On-demand | $3.29 | not stated | lambda.ai | |
| Together AI | NVIDIA H100 | Configuration not stated HGX H100 |
Reserved, 31-90 days | $3.45 listed: $3.45 (reserved, 31-90 days) |
not stated | together.ai | |
| Together AI | NVIDIA H100 | Configuration not stated HGX H100 |
Reserved, 3-30 days | $3.69 listed: $3.69 (reserved, 3-30 days) |
not stated | together.ai | |
| Crusoe | NVIDIA H100 | HGX · 80 GB each H100 HGX |
On-demand | $3.90 listed: $3.90/GPU-hr |
not stated | crusoe.ai | |
| Lambda | NVIDIA H100 | SXM · 8 GPUs per instance · 80 GB each 8x H100 |
On-demand | $3.99 | not stated | lambda.ai | |
| Together AI | NVIDIA H100 | Configuration not stated HGX H100 |
On-demand | $3.99 | not stated | together.ai | |
| Lambda | NVIDIA H100 | SXM · 1 GPU per instance · 80 GB each 1x H100 |
On-demand | $4.29 | not stated | lambda.ai | |
| DigitalOcean | NVIDIA H100 | 80 GB each gpu-h100x1-80gb |
On-demand | $4.41 listed: $4.41/GPU/hour |
not stated | digitalocean.com | |
| Nebius | NVIDIA H100 | Configuration not stated HGX H100 |
On-demand | $4.50 | not stated | nebius.com | |
| CoreWeave | NVIDIA H100 | 8 GPUs per instance · 80 GB each HGX H100 |
On-demand | $6.155 from On-Demand Price: $49.24 / Hour |
North America and Europe | coreweave.com | |
| AWS | NVIDIA H100 | 8 GPUs per instance · 80 GB each p5.48xlarge |
On-demand | $6.88 from $55.04 per instance-hour |
us-east-1 | aws.amazon.com | |
| AWS | NVIDIA H100 | 1 GPU per instance · 80 GB each p5.4xlarge |
On-demand | $6.88 from $6.88 per instance-hour |
us-east-1 | aws.amazon.com | |
| Microsoft Azure | NVIDIA H100 | NVL · 1 GPU per instance Standard_NC40ads_H100_v5 |
On-demand | $6.98 from retailPrice 6.98 USD per 1 Hour |
eastus | prices.azure.com | |
| Oracle Cloud | NVIDIA H100 | 8 GPUs per instance · 80 GB each · 8x2x200 Gb/s RDMA BM.GPU.H100.8 |
On-demand | $10.00 listed: $10 per GPU per hour (Compute - GPU - H100) |
not stated | apexapps.oracle.com | |
| Microsoft Azure | NVIDIA H100 | 8 GPUs per instance · 80 GB each Standard_ND96isr_H100_v5 |
On-demand | $12.29 from retailPrice 98.32 USD per 1 Hour |
eastus | prices.azure.com | |
| Nebius | NVIDIA H200 | Configuration not stated HGX H200 |
Spot | $0.79 listed: from $0.79 (preemptible) |
not stated | nebius.com | |
| CoreWeave | NVIDIA H200 | 8 GPUs per instance · 141 GB each HGX H200 |
Spot | $2.6163 from Spot Price: $20.93 / Hour |
North America | coreweave.com | |
| Together AI | NVIDIA H200 | Configuration not stated HGX H200 |
Spot | $2.99 listed: $2.99 (listed as preemptible) |
not stated | together.ai | |
| DigitalOcean | NVIDIA H200 | 141 GB each gpu-h200x1-141gb |
Reserved, 12 months | $3.40 listed: $3.40/GPU/hour (12-month commitment) |
not stated | digitalocean.com | |
| Hyperstack | NVIDIA H200 | SXM · 141 GB each H200 SXM |
On-demand | $3.99 | not stated | hyperstack.cloud | |
| Together AI | NVIDIA H200 | Configuration not stated HGX H200 |
Reserved, 91-180 days | $3.99 listed: $3.99 (reserved, 91-180 days) |
not stated | together.ai | |
| Together AI | NVIDIA H200 | Configuration not stated HGX H200 |
Reserved, 31-90 days | $4.15 listed: $4.15 (reserved, 31-90 days) |
not stated | together.ai | |
| Crusoe | NVIDIA H200 | HGX · 141 GB each H200 HGX |
On-demand | $4.29 listed: $4.29/GPU-hr |
not stated | crusoe.ai | |
| DigitalOcean | NVIDIA H200 | 141 GB each gpu-h200x1-141gb |
On-demand | $4.47 listed: $4.47/GPU/hour |
not stated | digitalocean.com | |
| Together AI | NVIDIA H200 | Configuration not stated HGX H200 |
Reserved, 3-30 days | $4.99 listed: $4.99 (reserved, 3-30 days) |
not stated | together.ai | |
| Nebius | NVIDIA H200 | Configuration not stated HGX H200 |
On-demand | $5.40 | not stated | nebius.com | |
| Together AI | NVIDIA H200 | Configuration not stated HGX H200 |
On-demand | $5.99 | not stated | together.ai | |
| CoreWeave | NVIDIA H200 | 8 GPUs per instance · 141 GB each HGX H200 |
On-demand | $6.305 from On-Demand Price: $50.44 / Hour |
North America and Europe | coreweave.com | |
| AWS | NVIDIA H200 | 8 GPUs per instance · 141 GB each p5en.48xlarge |
On-demand | $7.912 from $63.296 per instance-hour |
us-east-1 | aws.amazon.com | |
| Oracle Cloud | NVIDIA H200 | 8 GPUs per instance · 141 GB each · 8x1x400 Gb/s RDMA BM.GPU.H200.8 |
On-demand | $10.00 listed: $10 per GPU per hour (Compute - GPU - H200) |
not stated | apexapps.oracle.com | |
| Microsoft Azure | NVIDIA H200 | 8 GPUs per instance · 141 GB each Standard_ND96isr_H200_v5 |
On-demand | $10.60 from retailPrice 84.8 USD per 1 Hour |
eastus2 | prices.azure.com | |
| Nebius | NVIDIA B200 | Configuration not stated HGX B200 |
Spot | $0.99 listed: from $0.99 (preemptible) |
not stated | nebius.com | |
| Together AI | NVIDIA B200 | Configuration not stated HGX B200 |
Spot | $4.09 listed: $4.09 (listed as preemptible) |
not stated | together.ai | |
| CoreWeave | NVIDIA B200 | 8 GPUs per instance · 180 GB each HGX B200 |
Spot | $4.2638 from Spot Price: $34.11 / Hour |
North America | coreweave.com | |
| Hyperstack | NVIDIA B200 | 192 GB each B200 |
On-demand | $6.00 | not stated | hyperstack.cloud | |
| Lambda | NVIDIA B200 | SXM6 · 8 GPUs per instance · 180 GB each 8x B200 |
On-demand | $6.69 | not stated | lambda.ai | |
| Together AI | NVIDIA B200 | Configuration not stated HGX B200 |
Reserved, 91-180 days | $6.79 listed: $6.79 (reserved, 91-180 days) |
not stated | together.ai | |
| Lambda | NVIDIA B200 | SXM6 · 1 GPU per instance · 180 GB each 1x B200 |
On-demand | $6.99 | not stated | lambda.ai | |
| Together AI | NVIDIA B200 | Configuration not stated HGX B200 |
Reserved, 31-90 days | $7.79 listed: $7.79 (reserved, 31-90 days) |
not stated | together.ai | |
| Together AI | NVIDIA B200 | Configuration not stated HGX B200 |
Reserved, 3-30 days | $7.99 listed: $7.99 (reserved, 3-30 days) |
not stated | together.ai | |
| Together AI | NVIDIA B200 | Configuration not stated HGX B200 |
On-demand | $8.19 | not stated | together.ai | |
| Nebius | NVIDIA B200 | Configuration not stated HGX B200 |
On-demand | $8.50 | not stated | nebius.com | |
| CoreWeave | NVIDIA B200 | 8 GPUs per instance · 180 GB each HGX B200 |
On-demand | $8.60 from On-Demand Price: $68.80 / Hour |
North America and Europe | coreweave.com | |
| Oracle Cloud | NVIDIA B200 | 8 GPUs per instance · 180 GB each BM.GPU.B200.8 |
On-demand | $14.00 listed: $14 per GPU per hour (Compute - GPU - B200) |
not stated | apexapps.oracle.com | |
| AWS | NVIDIA B200 | 8 GPUs per instance p6-b200.48xlarge |
On-demand | $14.2416 from $113.9328 per instance-hour |
us-east-1 | aws.amazon.com | |
| Microsoft Azure | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each Standard_ND96amsr_A100_v4 |
Spot | $1.0552 from retailPrice 8.441552 USD per 1 Hour (Spot) |
eastus | prices.azure.com | |
| CoreWeave | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each A100 |
Spot | $1.2063 from Spot Price: $9.65 / Hour |
North America | coreweave.com | |
| Hyperstack | NVIDIA A100 80GB | 80 GB each A100 |
On-demand | $1.35 | not stated | hyperstack.cloud | |
| Hyperstack | NVIDIA A100 80GB | NVLink · 80 GB each · NVLink (in SKU name) A100 NVLink |
On-demand | $1.40 | not stated | hyperstack.cloud | |
| Hyperstack | NVIDIA A100 80GB | SXM · 80 GB each A100 SXM |
On-demand | $1.60 | not stated | hyperstack.cloud | |
| Crusoe | NVIDIA A100 80GB | PCIe · 80 GB each A100 PCIe |
On-demand | $2.00 listed: $2.00/GPU-hr |
not stated | crusoe.ai | |
| Crusoe | NVIDIA A100 80GB | SXM · 80 GB each A100 SXM |
On-demand | $2.30 listed: $2.30/GPU-hr |
not stated | crusoe.ai | |
| CoreWeave | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each A100 |
On-demand | $2.70 from On-Demand Price: $21.60 / Hour |
North America and Europe | coreweave.com | |
| Lambda | NVIDIA A100 80GB | SXM · 8 GPUs per instance · 80 GB each 8x A100 80 GB |
On-demand | $2.79 | not stated | lambda.ai | |
| AWS | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each p4de.24xlarge |
On-demand | $3.4309 from $27.4471 per instance-hour |
us-east-1 | aws.amazon.com | |
| Oracle Cloud | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each BM.GPU.A100-v2.8 |
On-demand | $4.00 listed: $4 per GPU per hour (Compute - GPU - A100 - v2) |
not stated | apexapps.oracle.com | |
| Microsoft Azure | NVIDIA A100 80GB | 8 GPUs per instance · 80 GB each Standard_ND96amsr_A100_v4 |
On-demand | $4.0963 from retailPrice 32.77 USD per 1 Hour |
eastus | prices.azure.com | |
| CoreWeave | NVIDIA L40S | 8 GPUs per instance · 48 GB each L40S |
Spot | $0.985 from Spot Price: $7.88 / Hour |
North America | coreweave.com | |
| Crusoe | NVIDIA L40S | 48 GB each L40S |
On-demand | $1.50 listed: $1.50/GPU-hr |
not stated | crusoe.ai | |
| Nebius | NVIDIA L40S | Intel CPU · 48 GB each L40S with Intel CPU |
On-demand | $1.55 listed: from $1.55 |
not stated | nebius.com | |
| DigitalOcean | NVIDIA L40S | 1 GPU per instance · 48 GB each L40S |
On-demand | $1.57 listed: $1.57/GPU/hour |
not stated | digitalocean.com | |
| AWS | NVIDIA L40S | 1 GPU per instance · 48 GB each g6e.xlarge |
On-demand | $1.861 from $1.861 per instance-hour |
us-east-1 | aws.amazon.com | |
| CoreWeave | NVIDIA L40S | 8 GPUs per instance · 48 GB each L40S |
On-demand | $2.25 from On-Demand Price: $18.00 / Hour |
North America and Europe | coreweave.com | |
| AWS | NVIDIA L40S | 4 GPUs per instance · 48 GB each g6e.12xlarge |
On-demand | $2.6232 from $10.4926 per instance-hour |
us-east-1 | aws.amazon.com | |
| Oracle Cloud | NVIDIA L40S | 4 GPUs per instance · 48 GB each BM.GPU.L40S.4 |
On-demand | $3.50 listed: $3.5 per GPU per hour (Compute - GPU - L40S) |
not stated | apexapps.oracle.com | |
| AWS | NVIDIA L40S | 8 GPUs per instance · 48 GB each g6e.48xlarge |
On-demand | $3.7664 from $30.1312 per instance-hour |
us-east-1 | aws.amazon.com | |
| Microsoft Azure | AMD Instinct MI300X | 8 GPUs per instance · 192 GB each Standard_ND96is_MI300X_v5 |
Spot | $1.1088 from retailPrice 8.8704 USD per 1 Hour (Spot) |
eastus2 | prices.azure.com | |
| DigitalOcean | AMD Instinct MI300X | 192 GB each gpu-mi300x1-192gb |
Reserved, 12 months | $1.91 listed: $1.91/GPU/hour (12-month commitment) |
not stated | digitalocean.com | |
| DigitalOcean | AMD Instinct MI300X | 192 GB each gpu-mi300x1-192gb |
On-demand | $2.59 listed: $2.59/GPU/hour |
not stated | digitalocean.com | |
| Crusoe | AMD Instinct MI300X | 192 GB each MI300X |
On-demand | $3.45 listed: $3.45/GPU-hr |
not stated | crusoe.ai | |
| Microsoft Azure | AMD Instinct MI300X | 8 GPUs per instance · 192 GB each Standard_ND96is_MI300X_v5 |
On-demand | $6.00 from retailPrice 48.0 USD per 1 Hour |
eastus2 | prices.azure.com | |
| Oracle Cloud | AMD Instinct MI300X | 8 GPUs per instance · 192 GB each BM.GPU.MI300X.8 |
On-demand | $6.00 listed: $6 per GPU per hour (OCI - Compute - GPU - MI300X) |
not stated | apexapps.oracle.com |
Click a column heading to sort. Greyed rows marked Stale have not been re-checked in 7 days.
Download: current prices (CSV) · current prices (JSON) · daily history (CSV) · price-change feed (RSS). Free to reuse with attribution under CC BY 4.0.
Tools
- GPU rental cost calculator: GPUs × hours × rate across providers, plus a buy-versus-rent breakeven from a price you enter.
- GPU price changes: what moved this week, from the daily history, with an RSS feed.
- Embeddable price charts for reporters, and open data downloads.
- Compute futures explained: the GPU rental index futures filed with the CFTC.
Embed this chart
<iframe src="https://datacentercensus.com/embed/gpu-prices" title="GPU rental prices" width="100%" height="520" style="border:0;max-width:640px" loading="lazy"></iframe>
<p style="font:13px/1.4 system-ui,sans-serif;margin:6px 0 0">Source: <a href="https://datacentercensus.com/gpu-prices">Data Center Census</a></p>
<script>addEventListener("message",function(e){if(e.origin==="https://datacentercensus.com"&&e.data&&e.data.dcc){document.querySelectorAll('iframe[src^="https://datacentercensus.com/embed/"]').forEach(function(f){if(f.contentWindow===e.source)f.style.height=e.data.h+"px"})}})</script>Free with the credit line. Preview the embed.
How these prices are collected
- What counts
- Only prices a provider publishes on a public pricing page or in its public price list, opened by the Census on the day shown. Marketplace offers, quotes and “contact sales” rows are not listed. Purchase prices are not listed at all.
- $ per GPU-hour
- The listed price per GPU-hour. Where a provider lists a price per instance, the Census divides it by the GPUs in that instance and shows the original figure under it. Nothing else is converted or estimated.
- Pricing types
- On-demand, spot (including tiers a provider calls preemptible or low-priority) and reserved with the term the provider states. Reserved rows without a stated term are not listed.
- Variants
- SXM, PCIe, NVL and similar versions are kept separate when the provider names them. A provider that does not name the variant is shown without one.
- Freshness
- Every listing carries a check date. 7 days after that date it is marked stale: still shown, but left out of “lowest” figures and the calculator’s default results.
- History
- Each daily run saves a snapshot, so the model pages can chart the lowest listed on-demand price over time and the change log can show what moved. History began Oct 15, 2026.
- Differences between providers
- Listed prices are not like for like: regions, minimum instance sizes, networking, storage, commitments and availability differ and are not priced here. Check the provider before relying on a figure.
Providers checked but not listed
- Google Cloud: The compute pricing pages (cloud.google.com/compute/gpus-pricing and vm-instance-pricing) showed no GPU prices when opened.
- RunPod: The pricing page shows two different sets of on-demand prices for the same GPUs (the visible cards and the page’s structured data), so no single listed price could be confirmed.
- Vast.ai: Its pricing page shows only relative claims ("50%+ cheaper"), no listed price. Marketplace prices change by offer.
- Voltage Park: Its pricing pages say "Contact for pricing" for every GPU; one homepage line says "Starting at $1.99/hr" without a GPU model or unit.
GPU rental prices: common questions
How much does it cost to rent an H100?
Among listings checked in the last 7 days, the lowest listed on-demand price for an H100 is $2.50 per GPU-hour at Hyperstack, checked Oct 15, 2026. Other providers list higher prices, and spot and reserved prices differ; the H100 page shows every listing.
Which providers are included?
AWS, Microsoft Azure, Oracle Cloud, CoreWeave, Lambda, Together AI, Nebius, Crusoe, DigitalOcean, Hyperstack. A provider is included only when its public pricing page or public price list shows a price for at least one of these GPUs. Checked but not listed: Google Cloud, RunPod, Vast.ai, Voltage Park, because no single listed price could be confirmed.
How often are the prices updated?
The Census re-checks every provider page in its daily update run and records the date next to each price. A price not re-checked within 7 days is marked stale and excluded from the lowest-price figures.
Does the Census list GPU purchase prices?
No. Street prices for buying a GPU vary by seller and are not published on a stable public price page. The calculator’s buy-versus-rent section uses a purchase price you type in.
Related: NVIDIA H100 prices · NVIDIA H200 prices · NVIDIA B200 prices · NVIDIA A100 80GB prices · NVIDIA L40S prices · AMD Instinct MI300X prices
Cite this page
Last updated Oct 15, 2026. Figures may change as records are updated — include the date you retrieved them.
| APA | Data Center Census. (2026). GPU rental prices by provider. Retrieved from https://datacentercensus.com/gpu-prices |
|---|---|
| MLA | “GPU rental prices by provider.” Data Center Census, Oct 15, 2026, https://datacentercensus.com/gpu-prices. |
| Link | https://datacentercensus.com/gpu-prices |
Republishing a figure or table? Attribute it to Data Center Census with a link to this page. The underlying data is downloadable as JSON and CSV; the press kit has standing figures and contact details.