Skip to content
Independent counter-quote desk for GPU clusters, bare-metal fleets & colocation · free on standard sourcing

Cloud GPU prices on October 31, 2026

On-demand list prices per GPU-hour from 11 GPU clouds as captured on October 31, 2026, with the change from the day and the week before.

No snapshot exists for October 30, 2026, so 1-day change shows n/a. No snapshot exists for October 24, 2026, so 7-day change shows n/a. A price marked "as of" could not be re-checked that day and shows its last good value.

Cheapest on-demand price per GPU

Lowest published on-demand list price per GPU-hour across the providers we track, USD, before tax and discounts.

GPUCheapest on-demandProvider1-day7-dayMedian across providersProviders
NVIDIA GB300 NVL72$18.00Oracle Cloud (OCI)n/an/a$18.001
NVIDIA GB200 NVL72$10.50CoreWeaven/an/a$16.003
NVIDIA B300$7.40Hyperstackn/an/a$9.756
NVIDIA B200$6.00Hyperstackn/an/a$8.507
NVIDIA H200$3.99Hyperstackn/an/a$6.0910
NVIDIA H100 SXM$3.20Hyperstackn/an/a$4.4610
NVIDIA H100 NVL$6.98Microsoft Azuren/an/a$6.981
NVIDIA H100 PCIe$2.50Hyperstackn/an/a$2.902
NVIDIA H100 (form factor not stated)$2.99OVHcloudn/an/a$2.991
NVIDIA GH200$2.29Lambdan/an/a$4.392
NVIDIA A100 80GB$1.35Hyperstackn/an/a$2.938
NVIDIA A100 40GB$1.99Lambdan/an/a$2.743
NVIDIA RTX PRO 6000 Blackwell$1.80Nebiusn/an/a$2.174
NVIDIA L40S$1.50Crusoen/an/a$1.807
NVIDIA L40$1.00Hyperstackn/an/a$1.122
NVIDIA L4$0.80AWSn/an/a$0.902
NVIDIA A10G$1.01AWSn/an/a$1.011
AMD Instinct MI355X$8.60Oracle Cloud (OCI)n/an/a$8.601
AMD Instinct MI325X$3.80DigitalOcean (incl. Paperspace)n/an/a$3.801
AMD Instinct MI300X$2.59DigitalOcean (incl. Paperspace)n/an/a$4.724

By provider

Each provider's lowest on-demand price per GPU-hour for that GPU, with the instance or plan it comes from. Multi-GPU instances are divided by their GPU count.

NVIDIA GB300 NVL72

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Oracle Cloud (OCI)$18.00n/an/aB112140 OCI - Compute - GPU - GB3001All OCI regions

NVIDIA GB200 NVL72

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
CoreWeave$10.50n/an/aNVIDIA GB200 NVL72 (4 GPU)4North America
Oracle Cloud (OCI)$16.00n/an/aB110979 OCI - Compute - GPU - GB2001All OCI regions
Microsoft Azure$27.04n/an/aStandard_ND128isr_NDR_GB200_v64Azure westus3

NVIDIA B300

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$7.40n/an/aNVIDIA B3001CA/EU/US
Crusoe$8.96n/an/aNVIDIA B300 288GB HGX1US
Nebius$9.50n/an/aNVIDIA HGX B3001EU/US
Together AI$9.99n/an/aGPU Clusters HGX B3001US
Oracle Cloud (OCI)$15.00n/an/aB112237 Oracle Cloud Infrastructure - Compute - GPU - B3001All OCI regions
AWS$17.80n/an/ap6-b300.48xlarge8AWS us-east-1

NVIDIA B200

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$6.00n/an/aNVIDIA B2001CA/EU/US
Lambda$6.69n/an/aB200 SXM6 instance1US
Together AI$8.19n/an/aGPU Clusters HGX B2001US
Nebius$8.50n/an/aNVIDIA HGX B2001EU/US
CoreWeave$8.60n/an/aNVIDIA HGX B200 (8 GPU)8North America
Oracle Cloud (OCI)$14.00n/an/aB110978 OCI - Compute - GPU - B2001All OCI regions
AWS$14.24n/an/ap6-b200.48xlarge8AWS us-east-1

NVIDIA H200

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$3.99n/an/aNVIDIA H200 SXM1CA/EU/US
Crusoe$4.29n/an/aNVIDIA H200 141GB HGX1US
DigitalOcean (incl. Paperspace)$4.47n/an/aGPU Droplet NVIDIA HGX H2001US/Canada
Nebius$5.40n/an/aNVIDIA HGX H2001EU/US
Together AI$5.99n/an/aGPU Clusters HGX H2001US
OVHcloud$6.20n/an/ah200-19208France (Gravelines)
CoreWeave$6.30n/an/aNVIDIA HGX H200 (8 GPU)8North America
AWS$7.91n/an/ap5en.48xlarge8AWS us-east-1
Oracle Cloud (OCI)$10.00n/an/aB110519 OCI - Compute - GPU - H2001All OCI regions
Microsoft Azure$10.60n/an/aStandard_ND96isr_H200_v58Azure westus2

NVIDIA H100 SXM

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$3.20n/an/aNVIDIA H100 SXM1CA/EU/US
Crusoe$3.90n/an/aNVIDIA H100 80GB HGX1US
Lambda$3.99n/an/aH100 SXM instance1US
Together AI$3.99n/an/aGPU Clusters HGX H1001US
DigitalOcean (incl. Paperspace)$4.41n/an/aGPU Droplet NVIDIA HGX H1001US/Canada
Nebius$4.50n/an/aNVIDIA HGX H1001EU/US
CoreWeave$6.16n/an/aNVIDIA HGX H100 (8 GPU)8North America
AWS$6.88n/an/ap5.48xlarge8AWS us-east-1
Oracle Cloud (OCI)$10.00n/an/aB98415 Oracle Cloud Infrastructure - Compute - GPU - H1001All OCI regions
Microsoft Azure$12.29n/an/aStandard_ND96isr_H100_v58Azure westus3

NVIDIA H100 NVL

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Microsoft Azure$6.98n/an/aStandard_NC40ads_H100_v51Azure eastus

NVIDIA H100 PCIe

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$2.50n/an/aNVIDIA H1001CA/EU/US
Lambda$3.29n/an/aH100 PCIe instance1US

NVIDIA H100 (form factor not stated)

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
OVHcloud$2.99n/an/ah100-3801France (Gravelines)

NVIDIA GH200

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Lambda$2.29n/an/aGH200 instance1US
CoreWeave$6.50n/an/aNVIDIA GH200 (1 GPU)1North America

NVIDIA A100 80GB

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$1.35n/an/aNVIDIA A1001CA/EU/US
Crusoe$2.00n/an/aNVIDIA A100 80GB PCIe1US
CoreWeave$2.70n/an/aNVIDIA A100 (8 GPU)8North America
Lambda$2.79n/an/aA100 SXM instance1US
OVHcloud$3.07n/an/aa100-1801France (Gravelines)
AWS$3.43n/an/ap4de.24xlarge8AWS us-east-1
Microsoft Azure$3.67n/an/aStandard_NC24ads_A100_v41Azure westus3
Oracle Cloud (OCI)$4.00n/an/aB95907 Compute - GPU - A100 - v21All OCI regions

NVIDIA A100 40GB

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Lambda$1.99n/an/aA100 SXM instance1US
AWS$2.74n/an/ap4d.24xlarge8AWS us-east-1
Microsoft Azure$3.40n/an/aStandard_ND96asr_v48Azure westus2

NVIDIA RTX PRO 6000 Blackwell

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Nebius$1.80n/an/aNVIDIA RTX PRO 60001EU/US
Hyperstack$1.85n/an/aNVIDIA RTX Pro 6000 SE1CA/EU/US
CoreWeave$2.50n/an/aNVIDIA RTX PRO 6000 Blackwell Server Edition (High Memory) (8 GPU)8North America
Oracle Cloud (OCI)$4.50n/an/aB112613 OCI - Compute - GPU - RTX PRO 60001All OCI regions

NVIDIA L40S

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Crusoe$1.50n/an/aNVIDIA L40S 48GB1US
Nebius$1.55n/an/aNVIDIA L40S with Intel CPU1EU/US
DigitalOcean (incl. Paperspace)$1.57n/an/aGPU Droplet NVIDIA L40S1US/Canada
OVHcloud$1.80n/an/al40s-901France (Gravelines)
AWS$1.86n/an/ag6e.xlarge1AWS us-east-1
CoreWeave$2.25n/an/aNVIDIA L40S (8 GPU)8North America
Oracle Cloud (OCI)$3.50n/an/aB109479 Compute - GPU - L40S1All OCI regions

NVIDIA L40

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Hyperstack$1.00n/an/aNVIDIA L401CA/EU/US
CoreWeave$1.25n/an/aNVIDIA L40 (8 GPU)8North America

NVIDIA L4

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
AWS$0.80n/an/ag6.xlarge1AWS us-east-1
OVHcloud$1.00n/an/al4-901France (Gravelines)

NVIDIA A10G

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
AWS$1.01n/an/ag5.xlarge1AWS us-east-1

AMD Instinct MI355X

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
Oracle Cloud (OCI)$8.60n/an/aB111758 OCI- Compute - GPU - MI355X1All OCI regions

AMD Instinct MI325X

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
DigitalOcean (incl. Paperspace)$3.80n/an/aGPU Droplet AMD Instinct™ MI325X1US/Canada

AMD Instinct MI300X

ProviderPer GPU-hour1-day7-dayInstance or planGPUsRegion
DigitalOcean (incl. Paperspace)$2.59n/an/aGPU Droplet AMD Instinct™ MI300X1US/Canada
Crusoe$3.45n/an/aAMD MI300X 192GB1US
Microsoft Azure$6.00n/an/aStandard_ND96isr_MI300X_v58Azure eastus2
Oracle Cloud (OCI)$6.00n/an/aB109485 OCI - Compute - GPU - MI300X1All OCI regions

List price is the ceiling, not the deal

Reserved terms, volume and timing usually come in well under these on-demand rates. Send us your GPU, count and term and we'll bring back firm quotes.

Get a quote

Sources, each checked once a day: AWS (AWS Price List Bulk API (EC2 offer file, us-east-1)); Microsoft Azure (Azure Retail Prices API); CoreWeave (Public pricing page); Crusoe (Public pricing page); DigitalOcean (incl. Paperspace) (Public pricing page); Hyperstack (Public pricing page); Lambda (Public pricing page); Nebius (Public pricing page); Oracle Cloud (OCI) (Oracle Cloud price list API); OVHcloud (Public pricing page (Public Cloud GPU instances)); Together AI (Public pricing page (GPU Clusters, on-demand)). On-demand list prices only (no spot, preemptible, reserved or committed-use rates). This page is a fixed record of the prices we captured on October 31, 2026 and is not updated afterwards. Download as CSV.

Make 3+ providers bid for your servers

Send your spec, invoice or renewal notice. Quotes normalized into one 36-month cost within one business day.

Get counter-quotes