NVIDIA AI Workbench Locations
Type: Concept Tags: NVIDIA, AI Workbench, locations, local development, remote development, SSH, Brev, cloud GPU, container runtime Related: NVIDIA-AI-Workbench, NVIDIA-AI-Workbench-Projects, NVIDIA-AI-Workbench-Applications, NVIDIA-Brev, NVIDIA-DGX-Spark, NVIDIA-DGX-Station, NVIDIA-DGX-Cloud, NVIDIA-Jetson-Platform, NVIDIA-AI-Enterprise Sources: https://docs.nvidia.com/ai-workbench/user-guide/latest/concepts/location-concept.html, https://docs.nvidia.com/ai-workbench/user-guide/latest/how-to/add-existing-location.html, https://docs.nvidia.com/ai-workbench/user-guide/latest/overview/introduction.html Last Updated: 2026-04-29
Summary
NVIDIA AI Workbench Locations are machines with AI Workbench installed where users can create, clone, build, and run Workbench projects. A location can be a local machine, a remote Ubuntu system reachable over SSH, a cloud instance, a workstation, a server, or an NVIDIA Brev-provisioned GPU environment.
Detail
Purpose
AI developers often move between local laptops, RTX workstations, cloud GPUs, data center servers, and shared lab machines. AI Workbench Locations provide a consistent management layer so projects can move across machines while retaining the same user experience and container-driven environment model.
Current scope
- The Desktop App is a lightweight UI, not itself a location.
- A full local install turns the local machine into a Workbench location.
- Remote-only mode lets users add remote locations without a full local install.
- Remote locations are machines reachable by SSH; current docs specify Ubuntu-based remote systems and key-based authentication.
- Locations are independent single machines, not Kubernetes clusters.
- Location Manager shows local and remote location cards.
- A Location Window manages projects on a specific location.
- AI Workbench integrates with NVIDIA-Brev to provision GPU cloud instances and add them as locations.
- When projects move across locations, AI Workbench combines versioned project configuration with host and user information such as architecture, mounts, and credentials.
NVIDIA context
Locations are the compute abstraction behind NVIDIA-AI-Workbench. They connect project portability to NVIDIA local systems such as NVIDIA-DGX-Spark, deskside and data center systems such as NVIDIA-DGX-Station and NVIDIA-DGX-Cloud, and developer cloud environments such as NVIDIA-Brev.
Connections
- NVIDIA-AI-Workbench - Workbench manages local and remote locations.
- NVIDIA-AI-Workbench-Projects - projects can exist in different states on different locations.
- NVIDIA-AI-Workbench-Applications - applications run in project containers on locations and are proxied back to the user.
- NVIDIA-Brev - cloud GPU provisioning path integrated with Workbench locations.
- NVIDIA-DGX-Spark and NVIDIA-DGX-Station - local/deskside systems suitable for Workbench locations.
- NVIDIA-DGX-Cloud - cloud GPU infrastructure target for remote development and scale-up workflows.
- NVIDIA-Jetson-Platform - Jetson OS is listed in current remote-location guidance.
- NVIDIA-AI-Enterprise - enterprise development workflows can use Workbench locations as reproducible project environments.
Source Excerpts
- NVIDIA docs define a location as a machine with AI Workbench installed.
- Current docs state that Workbench locations provide the same UI/UX across supported operating systems.