Index144.4
12 live

Compare providers

Sixteen providers, eighteen parts, one pricing type. On the NVIDIA H100 SXM the cheapest tracked provider sits 6.7× below the most expensive; on the NVIDIA L40 the entire market fits inside 1.1×.

Each cell is that provider’s median on-demand price for the SKU today, per single GPU per hour. The lowest price in each row is highlighted — that spread is why this site exists. Spot and community tiers live on each SKU page.

Scroll sideways to compare providers →

SKUAzureAzureCoreWeaveCoreWeaveVerda (formerly DataCrunch)VerdaDenvrDenvrHot AisleHot AisleHyperstackHyperstackLambdaLambdaMassed ComputeMassed ComputeOVHcloudOVHcloudVast.aiVast.aiVoltage ParkVoltage ParkVultrVultrSpread
NVIDIA B300——$13.05——$6.00——————2.2×
NVIDIA B200—$8.60$9.75———$7.89—————1.2×
NVIDIA H200 SXM—$6.30$7.04—————————1.1×
NVIDIA H200 NVL———————$3.72—$4.93——1.3×
NVIDIA H100 SXM$11.68$6.16$5.51$3.87—$3.99$4.74$3.14—$1.73$1.99—6.7×
NVIDIA H100 NVL$6.98——————$3.01—$3.46——2.3×
AMD Instinct MI300X————$2.99————————
NVIDIA H100 PCIe—————$2.60$3.79——$2.73——1.5×
NVIDIA A100 80GB SXM$4.10—$2.47$3.00—$1.85$3.19$1.38—$1.16——3.5×
NVIDIA RTX PRO 6000 Blackwell——$2.98——$2.50—$2.19—$1.73——1.7×
NVIDIA A100 40GB SXM$3.40—$1.73$2.00——$1.99——$0.82——4.1×
NVIDIA L40S—$2.25$2.08————$0.97$1.80$0.70—$1.673.2×
NVIDIA A100 40GB PCIe———$1.50——$1.99——$0.67——3.0×
NVIDIA A100 80GB PCIe$3.67————$1.40———$0.87——4.2×
NVIDIA L40—$1.25———$1.35——————1.1×
NVIDIA RTX 5090—————————$0.87———
NVIDIA RTX A6000—————$1.00$1.09$0.57—$0.48——2.3×
NVIDIA RTX 4090—————————$0.51———

Multi-config providers (1x/2x/4x/8x instances) show their median across configs. Prices are as listed at scrape time; availability isn’t guaranteed. Links marked ↗ may be affiliate links — they never affect prices shown or ordering.

Tracked providers

  • AzureAzureVisit
    Hyperscaler
    5 SKUs15,160 obs2026-11-22 11:16 UTC
  • CoreWeaveCoreWeaveVisit
    Dedicated GPU cloud
    7 SKUs9,003 obs2026-11-22 11:16 UTC
  • Verda (formerly DataCrunch)VerdaVisit
    Dedicated GPU cloud · formerly DataCrunch
    8 SKUs48,236 obs2026-11-22 11:16 UTC
  • DenvrDenvrVisit
    Dedicated GPU cloud
    4 SKUs3,028 obs2026-11-22 11:16 UTC
  • Hot AisleHot AisleVisit
    Dedicated GPU cloud
    1 SKUs757 obs2026-11-22 11:16 UTC
  • HyperstackHyperstackVisit
    Dedicated GPU cloud
    8 SKUs6,080 obs2026-11-22 11:16 UTC
  • LambdaLambdaVisit
    Dedicated GPU cloud
    7 SKUs13,511 obs2026-11-22 11:16 UTC
  • Massed ComputeMassed ComputeVisit
    Resells rented capacity
    9 SKUs8,291 obs2026-11-22 11:16 UTC
  • NebiusNebiusVisit
    Dedicated GPU cloud
    4 SKUs2,240 obs2026-10-12 02:22 UTC
  • OVHcloudOVHcloudVisit
    Dedicated GPU cloud
    1 SKUs2,274 obs2026-11-22 11:16 UTC
  • RunPodRunPodVisit
    Marketplace of independent hosts
    19 SKUs20,260 obs2026-10-05 18:24 UTC
  • SaladCloudSaladCloudVisit
    Marketplace of independent hosts
    2 SKUs836 obs2026-09-09 18:34 UTC
  • Thunder ComputeThunder ComputeVisit
    Resells rented capacity
    2 SKUs1,046 obs2026-09-29 01:55 UTC
  • Vast.aiVast.aiVisit
    Marketplace of independent hosts
    19 SKUs90,947 obs2026-11-22 03:07 UTC
  • Voltage ParkVoltage ParkVisit
    Dedicated GPU cloud
    1 SKUs740 obs2026-11-22 11:16 UTC
  • VultrVultrVisit
    Dedicated GPU cloud
    1 SKUs2,256 obs2026-11-22 11:16 UTC

Missing a provider you use? Request it. Each one is an independent adapter — and sources that can’t give us a number we can stand behind stay out until they can (see methodology).