GPU rental prices in United States
26 American providers pricing 32 GPUs. 274 listings on machines physically in United States, across 25 GPUs. The cheapest local H100 is $1.99/hr at io.net.
Prices collected 2026-10-11 · every figure is a provider's own published rate
Cheapest listing per GPU from American providers
| GPU | Cheapest at | $ / GPU / hr | $ / month | Local median | Global | vs global |
|---|---|---|---|---|---|---|
 AMD MI355X |  Oracle Cloud | $8.60 | $6,192 | $8.601 provider | $8.60 | +0% |
 NVIDIA B300 |  io.net | $4.50 | $3,240 | $8.0410 providers | $8.00 | −44% |
| NVIDIA B200 |  Beam | $4.11 | $2,959 | $6.8417 providers | $6.99 | −41% |
| NVIDIA H200 | io.net | $2.00 | $1,440 | $4.5417 providers | $4.50 | −56% |
| AMD MI325X |  DigitalOcean GPU Droplets | $3.80 | $2,736 | $3.801 provider | $3.80 | +0% |
| NVIDIA H200 NVL |  RunPod Community Cloud | $0.50 | $360 | $3.764 providers | $3.79 | −87% |
| NVIDIA H100 SXM | io.net | $1.99 | $1,433 | $4.0724 providers | $3.50 | −43% |
| NVIDIA GH200 |  Lambda | $2.29 | $1,649 | $4.402 providers | $3.45 | −34% |
| AMD MI300X | RunPod Community Cloud | $0.50 | $360 | $2.595 providers | $3.02 | −83% |
| NVIDIA V100 32GB | io.net | $1.57 | $1,130 | $2.753 providers | $2.90 | −46% |
| NVIDIA H100 NVL |  Vast.ai | $2.33 | $1,681 | $3.015 providers | $2.80 | −17% |
| NVIDIA H100 PCIe | io.net | $1.59 | $1,145 | $2.146 providers | $2.29 | −30% |
| NVIDIA RTX PRO 6000 Blackwell | Vast.ai | $0.94 | $679 | $2.5011 providers | $2.23 | −58% |
| NVIDIA A100 40GB | io.net | $0.79 | $569 | $2.109 providers | $1.82 | −57% |
| NVIDIA A100 80GB | Vast.ai | $0.37 | $270 | $2.1220 providers | $1.79 | −79% |
| NVIDIA A10G |  Amazon Web Services | $1.01 | $724 | $1.522 providers | $1.52 | −34% |
| NVIDIA A10 |  Modal | $1.10 | $793 | $1.295 providers | $1.29 | −15% |
| NVIDIA L40S | Vast.ai | $0.54 | $386 | $1.5715 providers | $1.29 | −58% |
| NVIDIA RTX 6000 Ada | Vast.ai | $0.49 | $355 | $0.896 providers | $0.89 | −45% |
| NVIDIA L40 | io.net | $0.66 | $475 | $0.846 providers | $0.88 | −25% |
| NVIDIA RTX 5090 | Vast.ai | $0.44 | $319 | $0.697 providers | $0.81 | −45% |
| NVIDIA L4 | Vast.ai | $0.27 | $194 | $0.809 providers | $0.79 | −66% |
| NVIDIA V100 16GB | Vast.ai | $0.12 | $87 | $0.799 providers | $0.78 | −84% |
| NVIDIA T4 | $0.16 | $115 | $0.598 providers | $0.59 | −73% | |
| NVIDIA A40 | RunPod Community Cloud | $0.35 | $252 | $0.504 providers | $0.59 | −41% |
| NVIDIA RTX A6000 | RunPod Community Cloud | $0.33 | $238 | $0.549 providers | $0.58 | −43% |
| NVIDIA RTX 4090 | io.net | $0.30 | $216 | $0.378 providers | $0.48 | −38% |
| NVIDIA RTX 5080 | Vast.ai | $0.22 | $157 | $0.384 providers | $0.38 | −42% |
| NVIDIA RTX 4080 | $0.15 | $108 | $0.275 providers | $0.27 | −44% | |
| NVIDIA RTX A5000 | RunPod Community Cloud | $0.16 | $115 | $0.256 providers | $0.26 | −37% |
| NVIDIA RTX 3090 | Vast.ai | $0.14 | $99 | $0.236 providers | $0.22 | −37% |
| NVIDIA RTX A4000 | Vast.ai | $0.08 | $56 | $0.175 providers | $0.16 | −51% |
Cheapest = lowest in-stock on-demand listing from a provider headquartered in United States. Local median = median of those providers' own medians. Global = the index median across every provider we track.
RunPod Community Cloud25 GPUs · 28 configs · -28.3% vs global marketB300 $6.94 · B200 $5.98 · H200 $3.59
RunPod Secure Cloud23 GPUs · 25 configs · +22.7% vs global marketB300 $8.99 · B200 $7.99 · H200 $5.29
Vast.ai22 GPUs · 460 configs · -13.6% vs global marketB300 $9.94 · B200 $7.50 · H100 SXM $6.15
io.net22 GPUs · 48 configs · -8.4% vs global marketB200 $4.89 · B300 $4.50 · H100 SXM $3.87
B300 $6.60 · B200 $5.43 · H200 NVL $3.72
Oracle Cloud12 GPUs · 12 configs · +116.2% vs global marketB300 $15.00 · B200 $14.00 · H200 $10.00
Amazon Web Services11 GPUs · 43 configs · +99.9% vs global marketB200 $14.24 · H200 $7.91 · H100 SXM $6.88
Beam10 GPUs · 13 configs · -18.2% vs global marketB200 $4.11 · H100 SXM $3.50 · H200 $2.09
Modal10 GPUs · 10 configs · +10.7% vs global marketB300 $7.10 · B200 $6.25 · H200 $4.54
Cerebrium10 GPUs · 10 configs · +5.3% vs global marketB200 $6.01 · H200 $4.20 · H100 SXM $3.40
B300 $6.00 · B200 $5.00 · H100 SXM $2.73
Lambda9 GPUs · 21 configs · +20.2% vs global marketB200 $6.84 · H100 SXM $4.14 · H100 PCIe $3.29
Microsoft Azure8 GPUs · 52 configs · +131.9% vs global marketH100 SXM $11.06 · H100 NVL $6.98 · A100 80GB $3.67
CoreWeave8 GPUs · 8 configs · +51% vs global marketB200 $8.60 · GH200 $6.50 · H200 $6.31
Lightning AI7 GPUs · 7 configs · +23.4% vs global marketH100 SXM $4.68 · H200 $4.50 · A100 80GB $2.71
DigitalOcean GPU Droplets6 GPUs · 6 configs · +18.2% vs global marketH200 $4.47 · H100 SXM $4.41 · MI325X $3.80
Replicate5 GPUs · 16 configs · +93.9% vs global marketH200 $5.49 · H100 SXM $5.49 · A100 80GB $5.04
Paperspace (DigitalOcean)5 GPUs · 7 configs · +242.6% vs global marketH100 SXM $4.52 · V100 16GB $2.30 · RTX A6000 $1.89
fal.ai5 GPUs · 5 configs · +43.7% vs global marketB300 $12.99 · B200 $7.99 · H200 $6.00
Crusoe5 GPUs · 6 configs · +11.5% vs global marketH200 $4.29 · H100 SXM $3.90 · MI300X $3.45
Baseten5 GPUs · 5 configs · +47.6% vs global marketB200 $9.98 · H100 SXM $6.50 · A100 80GB $4.00
SaladCloud5 GPUs · 10 configs · -27.5% vs global marketRTX 5090 $0.50 · RTX 4090 $0.33 · RTX 5080 $0.26
Together AI4 GPUs · 11 configs · +22.3% vs global marketB300 $9.99 · B200 $8.19 · H200 $5.99
Thunder Compute4 GPUs · 8 configs · -16.7% vs global marketH100 SXM $3.08 · A100 80GB $1.49 · L40 $0.89
B200 $5.00 · H200 $3.20 · H100 SXM $2.60
TensorDock3 GPUs · 3 configs · -20.7% vs global marketH100 SXM $2.25 · A100 80GB $1.80 · RTX 4090 $0.35
VultrNo readable price in the latest run
QubridNo readable price in the latest runMachines physically in United States
Listings whose host reports a location inside United States — marketplace machines and hyperscaler regions. This is the section that matters for latency and data residency; the provider may be headquartered anywhere.
| GPU | Cheapest at | Where | $ / GPU / hr | Local median | Listings | Global |
|---|---|---|---|---|---|---|
| AMD MI355X | Oracle Cloud | us-ashburn-1 | $8.60 | $8.60 | 1 | $8.60 |
| NVIDIA B300 | Oracle Cloud | us-ashburn-1 | $15.00 | $15.00 | 1 | $8.00 |
| NVIDIA B200 | Vast.ai | Virginia, US | $6.20 | $7.50 | 25 | $6.99 |
| NVIDIA H200 | Amazon Web Services | us-east-1 | $7.91 | $8.96 | 2 | $4.50 |
| NVIDIA H200 NVL | Vast.ai | Mississippi, US | $3.85 | $4.00 | 17 | $3.79 |
| NVIDIA H100 SXM | Vast.ai | US | $6.15 | $10.53 | 10 | $3.50 |
| AMD MI300X | Oracle Cloud | us-ashburn-1 | $6.00 | $6.00 | 1 | $3.02 |
| NVIDIA V100 32GB | Microsoft Azure | eastus | $2.75 | $3.33 | 2 | $2.90 |
| NVIDIA H100 NVL | Vast.ai | US | $2.33 | $3.01 | 6 | $2.80 |
| NVIDIA RTX PRO 6000 Blackwell | Vast.ai | North Dakota, US | $1.17 | $1.46 | 36 | $2.23 |
| NVIDIA A100 40GB | Amazon Web Services | us-east-1 | $2.74 | $3.22 | 4 | $1.82 |
| NVIDIA A100 80GB | Vast.ai | California, US | $0.37 | $1.07 | 21 | $1.79 |
| NVIDIA A10G | Amazon Web Services | us-east-1 | $1.01 | $1.83 | 8 | $1.52 |
| NVIDIA A10 | Oracle Cloud | us-ashburn-1 | $2.00 | $3.23 | 4 | $1.29 |
| NVIDIA L40S | Amazon Web Services | us-east-1 | $1.86 | $3.50 | 9 | $1.29 |
| NVIDIA RTX 6000 Ada | Vast.ai | California, US | $0.54 | $0.54 | 2 | $0.89 |
| NVIDIA RTX 5090 | Vast.ai | California, US | $0.44 | $0.53 | 19 | $0.81 |
| NVIDIA L4 | Vast.ai | Utah, US | $0.33 | $1.24 | 14 | $0.79 |
| NVIDIA V100 16GB | Vast.ai | United States, US | $0.12 | $0.27 | 21 | $0.78 |
| NVIDIA T4 | Microsoft Azure | eastus | $0.53 | $0.98 | 11 | $0.59 |
| NVIDIA RTX A6000 | Vast.ai | Arizona, US | $0.39 | $0.40 | 3 | $0.58 |
| NVIDIA RTX 4090 | Vast.ai | Texas, US | $0.40 | $0.64 | 24 | $0.48 |
| NVIDIA RTX 5080 | Vast.ai | United States, US | $0.22 | $0.33 | 12 | $0.38 |
| NVIDIA RTX 3090 | Vast.ai | California, US | $0.14 | $0.25 | 14 | $0.22 |
| NVIDIA RTX A4000 | Vast.ai | Michigan, US | $0.09 | $0.11 | 7 | $0.16 |
Every GPU, priced in USD
The whole index — 32 GPUs from every provider we track, each of which will rent to a customer in United States — with the global median and the single cheapest listing converted to USD. Tax is not included.
Latitude.sh
Packet.ai
HyperAI
OVHcloud
GPU.aiNearest providers by headquarters — North America
Denvr DataworksQuestions
How much does it cost to rent an H100 in United States?
From a provider headquartered in United States: $1.99 per GPU-hour at io.net on 2026-10-11, about $1,433 a month at 720 hours. On a machine physically located in United States: $6.15 per GPU-hour at Vast.ai (US), 10 such listings today. Across every provider we track, the global median is $3.50 and the cheapest listing anywhere is $1.68 at Latitude.sh — any of them will sell to a customer in United States.
Are there GPU cloud providers in United States?
26 of the 45 providers we price today are headquartered in United States: RunPod Community Cloud, RunPod Secure Cloud, Vast.ai, io.net, Massed Compute, Oracle Cloud, Amazon Web Services, Beam, Modal, Cerebrium, Akash, Lambda, Microsoft Azure, CoreWeave, Lightning AI, DigitalOcean GPU Droplets, Replicate, Paperspace (DigitalOcean), fal.ai, Crusoe, Baseten, SaladCloud, Together AI, Thunder Compute, GMI Cloud, TensorDock. On 30 of the 32 GPUs they price, the cheapest local listing is below the global median.
What does the USD price mean?
The provider's published dollar price, before tax. Providers that publish in euros are converted at the rate captured with the snapshot (2026-10-11).
Which GPU is the best value from United States?
For holding a model, memory per dollar: the NVIDIA RTX 3090 gives 109.1 GB per $1/hr. The cheapest GPU-hour of any kind is the NVIDIA RTX A4000 at $0.16. The full table above prices every GPU we track in USD; the cheapest-listing column links to who sells it.
Is the price the same everywhere in the world?
Nearly. Most providers publish one price list regardless of where you are, so the market table on this page is the same market a buyer in any country sees — the differences are tax, currency and, for the hyperscalers, region. What does change by country is who has hardware nearby: the hosted-hardware section lists machines actually located in United States, where any exist, which is what matters for latency and data residency.