Get GPU inventory with real-time availability

​

Get GPU series with real-time inventory aggregated by location and GPU count. This is the primary endpoint for the GPU order page.

For each series/location/GPU count combination, returns:

  • specs: Specs of the recommended SKU (cheapest available, auto-selected on order if no skuCode provided)
  • availableSkus: All orderable SKUs sorted cheapest first, each with their own specs

Presence in the response indicates availability — only locations and GPU counts with available capacity are included.

When multiple SKUs exist at a location, the frontend can display a SKU picker and send skuCode when creating a VM.

Query Parameters
  • code
    Type: string

    Filter by series code (e.g., "h100_series")

  • datacenterId
    Type: stringFormat: uuid

    Filter by datacenter ID

  • count
    Type: integerenum

    Filter by GPU count

    values
    • 1
    • 2
    • 4
    • 8
Responses
  • application/json
  • application/json
  • 403

    GPU feature not enabled for this entity

  • 404

    Datacenter is not available to the requesting entity (only returned when datacenterId is supplied as a query parameter).

  • application/json
Request Example for get/resource/gpu/inventory
curl 'https://api.gpcn.com/v1/resource/gpu/inventory?code=&datacenterId=&count=1' \
  --header 'X-API-Key: YOUR_SECRET_TOKEN'

{
  "success": true,
  "message": "GPU series inventory retrieved successfully",
  "data": {
    "series": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "name": "H100 Series",
        "code": "h100_series",
        "description": "Latest generation data center GPUs",
        "availability": [
          {
            "datacenterId": "660e8400-e29b-41d4-a716-446655440001",
            "datacenterName": "Kansas City",
            "datacenterCode": "KAN-1",
            "gpuCounts": [
              {
                "count": 1,
                "specs": {
                  "gpuDescription": "1x NVIDIA H100 80GB HBM3",
                  "vcpu": 28,
                  "memoryGiB": 200,
                  "storageGB": 1400
                },
                "availableSkus": [
                  {
                    "skuCode": "h100-sxm-80gb-1x",
                    "description": "1x NVIDIA H100 SXM 80GB HBM3",
                    "specs": {
                      "gpuDescription": "1x NVIDIA H100 SXM 80GB HBM3",
                      "vcpu": 28,
                      "memoryGiB": 200,
                      "storageGB": 1400
                    }
                  }
                ]
              },
              {
                "count": 8,
                "specs": {
                  "gpuDescription": "8x NVIDIA H100 80GB HBM3",
                  "vcpu": 224,
                  "memoryGiB": 1800,
                  "storageGB": 25000
                },
                "availableSkus": [
                  {
                    "skuCode": "h100-sxm-80gb-8x",
                    "description": "8x NVIDIA H100 SXM 80GB HBM3",
                    "specs": {
                      "gpuDescription": "8x NVIDIA H100 SXM 80GB HBM3",
                      "vcpu": 224,
                      "memoryGiB": 1800,
                      "storageGB": 25000
                    }
                  }
                ]
              }
            ]
          }
        ]
      }
    ]
  }
}

Successfully retrieved GPU series inventory. On live-provider outage the response is still 200 but data.series[].availability is empty and data.error is populated (Decision Log GPU-1).