CIComputeRadar
Live multi-GPU configuration index

2× H100 GPU Server Rental Prices

Compare 2× NVIDIA H100 GPU rental offers with live pricing and capacity.

Available offers143
Providers13
Lowest listed price now$0.217 / hour
Last catalog refresh2026-09-07 21:50 UTC

At the latest catalog refresh, the lowest listed matching offer is $0.217 / hour from Baseten. Availability and final checkout pricing can change after this timestamp.

Vultr4181Hivelocity2695UpCloud2610OVHcloud2081Gcore1816Scaleway864Sesterce719Alibaba Cloud549Cherry Servers381Servers.com291Exoscale234Spheron200Vast.ai150Ori142Verda128Timeweb Cloud100Hyperstack79Northflank57Akash Network43RunPod42bunny.net Storage30Prime Intellect21IDrive e219Replicate19Baseten18SaladCloud17Wasabi16Paperspace15Render15Thunder Compute13Modal11Hostwinds10Melbicom10Fly.io8fal.ai7Worldstream7Impossible Cloud6Limestone Networks6Serverspace6GMI Cloud5Together AI5InterServer4Railway4TensorDock4Hyperbolic3Filebase1Storj1
⌕
143 matching offers17,643 indexed · 45/47 sources with data · 2 syncing · catalog 12:50 AM · showing 1–50
BasetenGPU

Baseten · H100:2 - 2 H100 GPUs, 160 GiB VRAM

Location not specified · Available

$0.217/ hour≈ $158.16/mo
CPU32 vCPUvirtual CPU allocation
RAM236 GBsystem memory
GPU2× H100160 GB VRAM / GPU
StorageManagedRuntime storage is managed by Baseten; /v1/instance_type_prices does not publish a fixed disk capacity for this instance type
NetworkManaged Baseten networking · /v1/instance_type_prices does not publish fixed per-instance bandwidthprovider-reported network terms; port speed not reported
Available · updated 12:44 AM
Full specs
BasetenGPU

Baseten · H100:4 - 4 H100 GPUs, 320 GiB VRAM

Location not specified · Available

$0.433/ hour≈ $316.32/mo
CPU64 vCPUvirtual CPU allocation
RAM472 GBsystem memory
GPU4× H100320 GB VRAM / GPU
StorageManagedRuntime storage is managed by Baseten; /v1/instance_type_prices does not publish a fixed disk capacity for this instance type
NetworkManaged Baseten networking · /v1/instance_type_prices does not publish fixed per-instance bandwidthprovider-reported network terms; port speed not reported
Available · updated 12:44 AM
Full specs
HyperstackGPUCANADA-1

n3-H100x4

CANADA-1 · Available · 1 config(s)

$0.459/ hour≈ $335.07/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100-80G-PCIe80 GB VRAM / GPU
Storage100 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 1
Pricing noteFull specs
VerdaGPUSpot

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$3.25/ hour≈ $2,372.50/mo
CPU80 vCPU80 physical cores
RAM340 GBsystem memory
GPU2× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · data-crunch

2× H100 SXM5 · data-crunch

Location not specified · Marketplace offer listed

$4.13/ hour≈ $3,015.85/mo
CPU80 vCPUvirtual CPU allocation
RAM340 GBsystem memory
GPU2× H100 SXM5VRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

2× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$4.41/ hour≈ $3,217.40/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

2× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$4.41/ hour≈ $3,217.40/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

2× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$5.29/ hour≈ $3,859.58/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

2× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$5.29/ hour≈ $3,859.58/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
GcoreGPU

Gcore · g3-ai-48-464-2500-h100-80-2 · Sines-3

Sines-3 · 2 capacity reported

$5.50/ hour$3,957.10/moprovider bills in EUR · FX 2026-09-07
CPU48 vCPUvirtual CPU allocation
RAM475,136 GBsystem memory
GPU2× h100VRAM not reported
Storage2,500 GBstorage not reported
NetworkInfiniBand 2x400provider-reported network terms; port speed not reported
Available · updated 12:49 AM · stock 2
Full specs
SesterceGPUvm · AZ_05

2× H100 · vm

CAN-Montreal-2 · Available · CAN-Montreal-2

$5.50/ hour≈ $4,015.00/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H10080 GB VRAM / GPU
Storage1,600 GB1600 GB instance storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
OriGPU

Ori · ogc-h100sxm-160-if

dallas-2 · Template available · dallas-2

$5.80/ hour≈ $4,234.00/mo
CPU48 vCPUvirtual CPU allocation
RAM480 GBsystem memory
GPU2× H100SXM 80GB80 GB VRAM / GPU
Storage3,000 GBNVMe 3000 GB · SSD 0 GB
NetworkPublic/private networking · no ingress/egress feesprovider-reported network terms; port speed not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · massed-compute

2× H100 PCIe · massed-compute

Location not specified · Marketplace offer listed

$5.95/ hour≈ $4,344.96/mo
CPU40 vCPUvirtual CPU allocation
RAM256 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SesterceGPUvm · AZ_13

2× H100 · vm

USA-Desmoines-1 · Available · USA-Desmoines-1

$6.01/ hour≈ $4,384.38/mo
CPU40 vCPUvirtual CPU allocation
RAM256 GBsystem memory
GPU2× H10080 GB VRAM / GPU
Storage2,500 GB2500 GB instance storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
Thunder ComputeGPUOn-demand

NVIDIA H100

Thunder Compute cloud · available

$6.40/ hour≈ $4,672.00/mo
CPU8 vCPUvirtual CPU allocation
RAM64 GBsystem memory
GPU2× H10080 GB VRAM / GPU
Storage100 GB100–1000 GB configurable storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
VerdaGPUOn-demand

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$6.50/ hour≈ $4,745.00/mo
CPU80 vCPU80 physical cores
RAM340 GBsystem memory
GPU2× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
VerdaGPUSpot

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$6.50/ hour≈ $4,745.00/mo
CPU176 vCPU176 physical cores
RAM680 GBsystem memory
GPU4× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · sesterce

2× H100 PCIe · sesterce

Location not specified · Marketplace offer listed

$6.60/ hour≈ $4,818.00/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · sesterce

2× H100 PCIe · sesterce

Location not specified · Marketplace offer listed

$7.21/ hour≈ $5,261.26/mo
CPU40 vCPUvirtual CPU allocation
RAM256 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · data-crunch

2× H100 SXM5 · data-crunch

Location not specified · Marketplace offer listed

$7.96/ hour≈ $5,814.01/mo
CPU80 vCPUvirtual CPU allocation
RAM340 GBsystem memory
GPU2× H100 SXM5VRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · data-crunch

2× H100 SXM5 · data-crunch

Location not specified · Marketplace offer listed

$8.13/ hour≈ $5,934.02/mo
CPU80 vCPUvirtual CPU allocation
RAM340 GBsystem memory
GPU2× H100 SXM5VRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
Prime IntellectGPUon-demand

Prime Intellect · H100_80GB × 2

us-southeast-1 · Available

$8.38/ hour≈ $6,117.40/mo
CPU52 vCPUvirtual CPU allocation
RAM450 GBsystem memory
GPU2× H100 80GB160 GB VRAM / GPU
Storage5,632 GBprovider default local disk
NetworkProvider/network-specific · current availability row does not publish port speedprovider-reported network terms; port speed not reported
Available · updated 12:50 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

2× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$8.40/ hour≈ $6,132.00/mo
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

4× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$8.81/ hour≈ $6,429.40/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

4× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$8.81/ hour≈ $6,429.40/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SesterceGPUvm · AZ_07

2× H100 · vm

us-southeast-1 · Available · us-southeast-1

$9.22/ hour≈ $6,729.14/mo
CPU52 vCPUvirtual CPU allocation
RAM450 GBsystem memory
GPU2× H10080 GB VRAM / GPU
Storage5,632 GB5632 GB instance storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ms

2× H100 SXM5 · spheron-ms

Location not specified · Marketplace offer listed

$10.06/ hour≈ $7,340.88/mo
CPU52 vCPUvirtual CPU allocation
RAM450 GBsystem memory
GPU2× H100 SXM5VRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

4× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$10.57/ hour≈ $7,713.98/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

4× H100 PCIe · spheron-ai

Location not specified · Marketplace offer listed

$10.57/ hour≈ $7,713.98/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
ReplicateGPUcommitted-spend capacity

Replicate · gpu-h100-2x

Location not specified · Available

$10.98/ hour≈ $8,015.40/mo
CPU—CPU not reported
RAM—system memory not reported
GPU2× H10080 GB VRAM / GPU
StorageManagedReplicate publishes hardware CPU/RAM/GPU specs but not a fixed disk capacity per hardware SKU
NetworkManaged Replicate networking · fixed per-hardware port speed/traffic allowance is not publishedprovider-reported network terms; port speed not reported
Available · updated 12:43 AM
Pricing noteFull specs
SesterceGPUvm · AZ_05

4× H100 · vm

CAN-Montreal-2 · Available · CAN-Montreal-2

$11.00/ hour≈ $8,030.00/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H10080 GB VRAM / GPU
Storage3,300 GB3300 GB instance storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · sesterce

2× H100 SXM5 · sesterce

Location not specified · Marketplace offer listed

$11.06/ hour≈ $8,074.97/mo
CPU52 vCPUvirtual CPU allocation
RAM450 GBsystem memory
GPU2× H100 SXM5VRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
OriGPU

Ori · ogc-h100sxm-320-if

dallas-2 · Template available · dallas-2

$11.60/ hour≈ $8,468.00/mo
CPU96 vCPUvirtual CPU allocation
RAM960 GBsystem memory
GPU4× H100SXM 80GB80 GB VRAM / GPU
Storage4,000 GBNVMe 4000 GB · SSD 0 GB
NetworkPublic/private networking · no ingress/egress feesprovider-reported network terms; port speed not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · massed-compute

4× H100 PCIe · massed-compute

Location not specified · Marketplace offer listed

$11.90/ hour≈ $8,689.92/mo
CPU64 vCPUvirtual CPU allocation
RAM512 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
Thunder ComputeGPUOn-demand

NVIDIA H100

Thunder Compute cloud · available

$12.80/ hour≈ $9,344.00/mo
CPU60 vCPUvirtual CPU allocation
RAM480 GBsystem memory
GPU4× H10080 GB VRAM / GPU
Storage100 GB100–2000 GB configurable storage
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
VerdaGPUOn-demand

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$13.00/ hour≈ $9,490.00/mo
CPU176 vCPU176 physical cores
RAM680 GBsystem memory
GPU4× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
VerdaGPUSpot

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$13.00/ hour≈ $9,490.00/mo
CPU176 vCPU176 physical cores
RAM1,360 GBsystem memory
GPU8× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
SpheronGPUOn-demand · sesterce

4× H100 PCIe · sesterce

Location not specified · Marketplace offer listed

$13.20/ hour≈ $9,636.00/mo
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100 PCIeVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
SpheronGPUOn-demand · spheron-ai

8× H100 NVL · spheron-ai

Location not specified · Marketplace offer listed

$20.81/ hour≈ $15,189.40/mo
CPU252 vCPUvirtual CPU allocation
RAM1,440 GBsystem memory
GPU8× H100 NVLVRAM not reported
Storage—storage not reported
Network—not reported
Available · updated 12:48 AM
Pricing noteFull specs
ReplicateGPUcommitted-spend capacity

Replicate · gpu-h100-4x

Location not specified · Available

$21.96/ hour≈ $16,030.80/mo
CPU—CPU not reported
RAM—system memory not reported
GPU4× H10080 GB VRAM / GPU
StorageManagedReplicate publishes hardware CPU/RAM/GPU specs but not a fixed disk capacity per hardware SKU
NetworkManaged Replicate networking · fixed per-hardware port speed/traffic allowance is not publishedprovider-reported network terms; port speed not reported
Available · updated 12:43 AM
Pricing noteFull specs
OriGPU

Ori · ogc-h100sxm-640-if

dallas-2 · Template available · dallas-2

$23.20/ hour≈ $16,936.00/mo
CPU168 vCPUvirtual CPU allocation
RAM1,500 GBsystem memory
GPU8× H100SXM 80GB80 GB VRAM / GPU
Storage6,000 GBNVMe 6000 GB · SSD 0 GB
NetworkPublic/private networking · no ingress/egress feesprovider-reported network terms; port speed not reported
Available · updated 12:49 AM
Pricing noteFull specs
Limestone NetworksGPU8-GPU dedicated node

Limestone AI · 8× H100 SXM5 80GB

Location not specified · Published / orderability not live-verified

$23.60/ hour$17,228.00/mo
CPUConfigurableDedicated single-tenant node CPU is configuration-specific; Limestone does not publish CPU count/model in the standard AI tariff
RAMConfigurableSystem RAM is configuration-specific; Limestone does not publish RAM capacity in the standard AI tariff
GPU8× H100 SXM580 GB VRAM / GPU
StorageConfigurableLocal/storage configuration is built to spec and is not published in the standard AI tariff
Network400 Gb/s InfiniBand · RoCE availableprovider-reported network terms; port speed not reported
Available · updated 10:43 PM
Full specs
VerdaGPUOn-demand

H100 SXM5 80GB

Location not specified · Catalog listed · live capacity not verified anonymously

$26.00/ hour≈ $18,980.00/mo
CPU176 vCPU176 physical cores
RAM1,360 GBsystem memory
GPU8× H10080 GB VRAM / GPU
StorageSee detailsdynamic
Network—not reported
Available · updated 12:49 AM
Pricing noteFull specs
ReplicateGPUcommitted-spend capacity

Replicate · gpu-h100-8x

Location not specified · Available

$43.92/ hour≈ $32,061.60/mo
CPU—CPU not reported
RAM—system memory not reported
GPU8× H10080 GB VRAM / GPU
StorageManagedReplicate publishes hardware CPU/RAM/GPU specs but not a fixed disk capacity per hardware SKU
NetworkManaged Replicate networking · fixed per-hardware port speed/traffic allowance is not publishedprovider-reported network terms; port speed not reported
Available · updated 12:43 AM
Pricing noteFull specs
fal.aiGPUDedicated Compute

fal Compute · 8xH100-SXM

Location not specified · Published / orderability not live-verified

—price unavailable
CPU128 vCPUvirtual CPU allocation
RAM1,600 GBsystem memory
GPU8× H100 SXM80 GB VRAM / GPU
Storage8,000 GB8 TB SSD included
NetworkInfiniBand available between same-sector 8-GPU instances for multi-node workloads · exact link rate not publishedprovider-reported network terms; port speed not reported
Available · updated 10:43 PM
Pricing noteFull specs
HyperstackGPUCANADA-1

n3-H100x2-spot

CANADA-1 · Available · 25 config(s)

—price unavailable
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100-80G-PCIe-spot80 GB VRAM / GPU
Storage100 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 25
Pricing noteFull specs
HyperstackGPUCANADA-1

n3-H100x4-spot

CANADA-1 · Available · 8 config(s)

—price unavailable
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100-80G-PCIe-spot80 GB VRAM / GPU
Storage100 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 8
Pricing noteFull specs
HyperstackGPUCANADA-1

n3-H100x2-bigroot

CANADA-1 · Available · 2 config(s)

—price unavailable
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100-80G-PCIe80 GB VRAM / GPU
Storage1,600 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 2
Pricing noteFull specs
HyperstackGPUCANADA-1

n3-H100x2-bigroot-spot

CANADA-1 · Available · 25 config(s)

—price unavailable
CPU60 vCPUvirtual CPU allocation
RAM360 GBsystem memory
GPU2× H100-80G-PCIe-spot80 GB VRAM / GPU
Storage1,600 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 25
Pricing noteFull specs
HyperstackGPUCANADA-1

n3-H100x4-bigroot

CANADA-1 · Available · 1 config(s)

—price unavailable
CPU124 vCPUvirtual CPU allocation
RAM720 GBsystem memory
GPU4× H100-80G-PCIe80 GB VRAM / GPU
Storage3,300 GBstorage not reported
Network—not reported
Available · updated 12:49 AM · cached (Syncing) · stock 1
Pricing noteFull specs
How to read this page

Compare the full configuration, not only the headline price

This page requires at least 2 NVIDIA H100 GPUs in the normalized offer. It is designed for searches where the complete multi-GPU configuration matters, not only a per-GPU rate.

GPU count does not prove NVLink, NVSwitch, InfiniBand or equivalent topology. Verify interconnect, host RAM, local storage and networking before ordering.

Selection notes

What to check before renting

01

Check whether the displayed price is per GPU or for the full host.

02

Confirm NVLink/NVSwitch or other interconnect requirements with the provider.

03

For training clusters, compare CPU, RAM, storage throughput and network fabric as well as GPU count.

Related searches

Compare nearby options

GPU configuration8× H100 GPU servers→GPU configuration4× H100 GPU servers→GPU configuration8× H200 GPU servers→GPU configuration8× B200 GPU servers→GPU configuration8× A100 GPU servers→
FAQ

Common questions

Where can I rent 2× H100 GPU servers?

The live matching offers above come from currently connected provider sources and are refreshed with the catalog.

Does an 8-GPU result guarantee all GPUs are on one host?

It represents an offer reporting at least the requested GPU count, but provider topology can vary. Confirm host and interconnect details at checkout.