apiCommonsPlans: '0.1'
provider:
  name: NVIDIA NIM
  id: nvidia-nim
  url: https://www.nvidia.com/en-us/ai-data-science/products/nim-microservices/
  sources:
  - https://build.nvidia.com
  - https://www.nvidia.com/en-us/data-center/products/ai-enterprise/
  - https://www.nvidia.com/en-us/ai-data-science/products/nim-microservices/
plans:
- id: developer-free
  name: Developer (Free)
  description: Free hosted inference on DGX Cloud through build.nvidia.com. Designed for prototyping and learning.
  entries:
  - geo: Global
    unit: 1
    label: Developer
    limit: 1
    price: 0
    metric: user
    timeFrame: month
    description: Free tier with shared rate-limited endpoints.
  elements:
  - name: Free NVIDIA developer account
  - name: 1,000 inference credits on signup
  - name: 40 requests per minute soft rate limit
  - name: Access to 100+ models through integrate.api.nvidia.com
  - name: Sandbox try-it-now playground at build.nvidia.com
  - name: OpenAI-compatible REST and gRPC endpoints
  url: https://build.nvidia.com/explore/discover
- id: nvidia-ai-enterprise
  name: NVIDIA AI Enterprise
  description: Production licensing for self-hosted NIM microservices with enterprise support and security updates.
  entries:
  - geo: Global
    unit: 1
    label: GPU
    price: 4500
    metric: gpu
    timeFrame: year
    description: Per-GPU per-year subscription licensing for production NIM deployment.
  - geo: Global
    unit: 1
    label: GPU (Perpetual)
    price: 13500
    metric: gpu
    timeFrame: perpetual
    description: Perpetual per-GPU license with bundled NVIDIA support.
  elements:
  - name: All NIM microservices for production use
  - name: Self-hosted Docker containers via NGC registry
  - name: NIM Operator for Kubernetes
  - name: TensorRT-LLM, vLLM, SGLang optimized engines
  - name: Enterprise security patches and CVE response
  - name: NVIDIA AI Workbench
  - name: NeMo framework
  - name: NVIDIA Riva, BioNeMo, Picasso, Maxine NIMs
  - name: NVIDIA AI Blueprints
  - name: Long-term support branches
  - name: Direct access to NVIDIA AI experts
  url: https://www.nvidia.com/en-us/data-center/products/ai-enterprise/
- id: dgx-cloud
  name: DGX Cloud
  description: NVIDIA-managed AI training and inference infrastructure on a hyperscaler of your choice.
  entries:
  - geo: Global
    unit: 1
    label: DGX H100 Instance
    price: 36999
    metric: instance
    timeFrame: month
    description: Reference list price; commercial deals are negotiated.
  elements:
  - name: Managed multi-node GPU clusters
  - name: Bundled NVIDIA AI Enterprise license
  - name: Hosted NIM endpoints behind integrate.api.nvidia.com
  - name: Optimized networking and storage
  - name: 24/7 NVIDIA support
  url: https://www.nvidia.com/en-us/data-center/dgx-cloud/
modified: '2026-05-25'