The table below compares current hourly H100 pricing for a single NVIDIA H100 80GB GPU across major U.S. cloud providers.
To help you compare Nvidia H100 prices across clouds, this guide normalizes per-GPU rates (even when only multi-GPU instances are listed) and uses each provider's currently published U.S. pricing references.
| Provider | SKU/Instance | $/GPU-hr* | Notes |
|---|---|---|---|
| Thunder Compute | H100 80GB | $1.38 | Lowest on-demand price with top-tier reliability |
| Vast.ai | H100 80GB (third-party hosts) | ~$1.53 | Marketplace current price per H100 GPU. Reliability varies. |
| RunPod | H100 80GB PCIe | $1.99 | Community cloud price; Secure Cloud lists $2.39. |
| TensorDock | H100 80GB PCIe | $1.99 | Public listed single-GPU equivalent. |
| Paperspace | H100 80GB (Dedicated instance) | $2.24 | Public listed price. |
| Sesterce | H100 80GB PCIe | $2.25 | Public listed single-GPU equivalent. |
| FluidStack | H100 80GB PCIe | $2.30 | Public estimate for on-demand access. |
| Lambda | NVIDIA H100 PCIe (1x) | $3.99 | Public listed 1-GPU rate on Lambda pricing page. |
| Verda | H100 80GB PCIe | $3.39 | Public listed single-GPU equivalent. |
| Hyperstack | H100 80GB PCIe | $3.50 | Public listed single-GPU equivalent. |
| Nebius | H100 80GB PCIe | $3.50 | Public listed single-GPU equivalent. |
| Lyceum | H100 80GB PCIe | $3.69 | Public listed single-GPU equivalent. |
| JarvisLabs | H100 80GB PCIe | $3.80 | Pay-as-you-go H100 pricing. |
| Hyperbolic | H100 80GB PCIe | $1.50 | Marketplace-style H100 pricing. |
| Crusoe Cloud | H100 80GB PCIe | $4.29 | Public listed hourly pricing. |
| CoreWeave | NVIDIA HGX H100 (8 x H100) | $6.16 | $49.24/hr node price, normalized per GPU. |
| AWS | p5.4xlarge | $6.88 | Capacity Blocks pricing page rate in U.S. regions. |
| Azure | NC40ads_H100_v5 VM | $8.59 | Single H100 GPU VM in East US region. |
| Oracle Cloud | BM.GPU.H100.8 | $10.00 | 8-GPU bare-metal node: $80.00/hr. |
| Google Cloud | a3-highgpu-1g | $11.06 | 1x H100 80GB in US-central (on-demand). |
| *Normalized May 2026 cost per single H100 GPU, even when only multi-GPU instances are offered by the provider.* | |||
Methodology (why you can trust these numbers)
<ul><li><strong>Published list rates:</strong> Provider pages only, with normalized per-GPU pricing.</li><li><strong>Same class of silicon:</strong> All prices refer to NVIDIA H100 80GB GPUs.</li><li><strong>Normalized across SKUs:</strong> Where providers list only 8-GPU nodes, we divide by eight to get a single-GPU equivalent.</li><li><strong>USD in U.S. regions:</strong> Rates elsewhere can differ by 5-20 %.</li></ul>
Why this matters for developers
This chart compares how much it costs to run an NVIDIA H100 for 10 hours across major cloud providers.
Price sources: provider pricing pages linked in References below. Vast.ai values are marketplace estimates and vary by host and region.
H100 workloads are expensive by default, so even small hourly price differences have an outsized impact. Whether you're training large models or running high-throughput inference, choosing a lower-cost provider can dramatically extend your budget.
AWS P5 Instances
AWS offers H100 GPUs in its EC2 P5 Instances (P5 with H100 and P5en with H200). All P5 instances have:
<ul><li>NVIDIA H100 or H200 GPUs</li><li>100-3200 GB/s network performance</li><li>1-8 x 4 TB NVMe SSD storage with 16 GB/sec read</li><li>16-192 vCPUs</li><li>2048 GB RAM</li></ul>
As usual for AWS instances, egress fees are $0.09/GB after the first free 100GB. Read a full analysis of AWS P5 instances.
| Provider | SKU | GPUs | Memory per GPU | Hourly price | Price per-GPU |
|---|---|---|---|---|---|
| AWS | p5.4xlarge | 1 × H100 | 80 GB | $6.88 | $6.88 |
| AWS | p5.48xlarge | 8 × H100 | 80 GB | $55.04-$68.80** | $6.88-$8.60 |
| AWS | p5en.48xlarge | 8 × H200 | 80 GB | $63.30-$79.12** | $7.91-$9.89 |
Google Cloud A3 Instances
Through A3 instances, GCP offers accelerator-optimized compute with NVIDIA H100 GPUs:
<ul><li>A3 High (a3-highgpu-1g) with 1-8xH100.</li><li>A3 Mega (a3-megagpu-8g) with 8xH100.</li><li>A3 Ultra (a3-ultragpu) with 8xH200.</li></ul>
| Instance name | CPUs | RAM | GPUs | SSD Storage | Price/hr |
|---|---|---|---|---|---|
| a3-highgpu-1g | 26 | 236GB | 1 | 2x375GB | $11.06 |
| a3-highgpu-2g | 52 | 468GB | 2 | 4x375GB | $22.12 |
| a3-highgpu-4g | 104 | 936GB | 4 | 8x375GB | $44.25 |
| a3-highgpu-8g | 208 | 1872GB | 8 | 16x375GB | $48.74 |
| a3-megagpu-8g | 96 | 1360GB | 8 | 8x375GB | $88.49 |
Azure NC Instances
Through NC instances, Azure offers GPU-optimized compute with NVIDIA H100 GPUs:
<ul><li>NC40ads_H100_v5 with 1xH100.</li><li>NC80adis_H100_v5 with 2xH100.</li></ul>
| Instance name | CPUs | RAM | GPUs | SSD Storage | Price/hr |
|---|---|---|---|---|---|
| NC40ads_H100_v5 | 40 | 320GB | 1 | 3576GB | $8.59 |
| NC80adis_H100_v5 | 80 | 640GB | 2 | 7152GB | $17.17 |
Takeaways
<ul><li>Thunder Compute's H100 rate is <strong>about 2.8x cheaper than AWS</strong>, <strong>about 5x cheaper than Azure</strong>, and <strong>about 10x cheaper than GCP</strong>.</li><li>Specialized providers like RunPod and Lambda are closer than hyperscalers, but still higher than Thunder Compute on listed H100 rates.</li><li>Bookmark this page, we refresh the numbers quarterly so you do not have to.</li><li>Building a startup? See our analysis of <strong>Startup-Friendly GPU Cloud Providers</strong> for credit offers, and spin up an H100 on <a href="https://www.thundercompute.com/">Thunder Compute</a></li></ul>
Read a full comparison of AWS P5 instances powered by NVIDIA H100 and H200 GPUs.
Final Thoughts on NVIDIA H100 Pricing
The NVIDIA H100 remains relevant for AI training and inference workloads. Cloud provider prices range from $1.38/GPU-hr on specialized providers to $11.06/GPU-hr on hyperscalers. A nearly 8x difference for the same silicon.
For teams running large training runs or inference pipelines, that gap translates directly into either runway or wasted spend. The major hyperscalers (AWS, Azure, and GCP) offer deep ecosystems, but at a premium.
Before committing to a provider, benchmark your specific workload, account for egress fees and reserved pricing options.