HP Japan has announced the ZGX Fury G1n AI Station for organizations developing and running AI models locally. Development tools and model-serving software accompany the NVIDIA-based workstation. Manufacturing uses include working with design data and internal technical documents. HP expects customers to begin with trials before moving validated projects into production, and has installed a demonstration machine in Tokyo.
HP Japan introduced the ZGX Fury G1n AI Station on October 26 for organizations working with large AI models on their own premises. The workstation follows NVIDIA’s DGX Station design. It brings development, validation and inference, the production of answers by a trained model, onto the same machine. The approach is intended to keep sensitive company data within a local computing environment while models are developed and used.[1], [2]
HP ZGX Toolkit helps set up the development environment. HP Z Runtime then provides deployment and management functions for serving the resulting models locally. The workstation supports Red Hat Enterprise Linux, an operating system used in existing organizational infrastructure, and NVIDIA OpenShell, a runtime for keeping AI agents in separate sandbox environments. HP describes the latter as a way to control where agent activity and inference take place.[1]
HP’s product representative points to manufacturing applications such as developing models around design data, examining analysis results and answering workers’ questions by consulting operating manuals. Internal technical documents could also be searched when preparing an answer. HP expects initial users to test pilot projects and validate them before production deployment. A machine in its Tokyo office is available for such trials, and shared access elsewhere in the Asia-Pacific region is being prepared.[1]
