Docker Model Runner now supports NVIDIA DGX Station GB300. Run bigger models with familiar Docker commands and effortless setup.

Docker offers insights into container technology, microservices architecture, and application deployment, providing documentation and best practices for building, deploying, and managing containerized applications. By exploring Docker's curated content, developers can learn about container orchestration, Docker Swarm, and Docker Compose for managing complex containerized environments. Whether you're deploying microservices, building CI/CD pipelines, or optimizing your development workflow, Docker offers resources to streamline your containerization journey and unlock the benefits of container-based deployment.

Docker

Docker Model Runner now supports the NVIDIA DGX Station GB300, a deskside workstation built around the GB300 Grace Blackwell Ultra Desktop Superchip. Compared to the DGX Spark, the DGX Station offers dramatically higher specs: 252GB GPU memory, 7.1 TB/s memory bandwidth, and 748GB total coherent memory. This enables running trillion-parameter models without quantization, serving multiple team members via NVIDIA MIG partitioning, and running concurrent models for agentic AI workflows. The Docker experience remains unchanged — the same commands used on DGX Spark work on DGX Station, requiring no new tooling or cloud dependencies.

Docker Model Runner on DGX Station GB300

From NVIDIA DGX Spark to DGX Station: What has changed and why does this matter?

Getting Started: It’s the Same Docker Experience