GoKawiilHP opens orders for ZGX Fury AI station built on Nvidia's GB300 superchip
HP has begun taking orders for its ZGX Fury AI workstation, a system based on Nvidia's GB300 Grace Blackwell Ultra Desktop Superchip pairing a Blackwell Ultra GPU with a 72-core Grace CPU. The machine offers 748GB of unified, coherent memory and up to 20 petaFLOPS of FP4 compute, and ships alongside a new partnership between HP, Red Hat, and Nvidia to run Red Hat AI Factory software on the hardware.
GoKawiil's interpretation of the reporting above, not reported fact.
By combining massive unified memory with dense compute in a desk-sized box, HP is pitching an alternative to cloud or data-center GPU clusters for running large AI models locally, targeting departments, factories, or branch offices that need on-premises inference without full data-center infrastructure. The Red Hat software bundle signals a push toward turnkey enterprise AI deployment rather than leaving customers to assemble their own stack.
- ZGX Fury combines Nvidia's GB300 superchip with 748GB of unified CPU-GPU memory
- System targets on-premises, department-level AI inference rather than personal workstation use
- HP, Red Hat, and Nvidia are bundling Red Hat AI Factory software with the hardware at launch
Source: storagereview.com — Brian Beeler, 2026-09-14
Published there as: “HP ZGX Fury Is Now Orderable: GB300 Superchip, 748GB Unified Memory”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.