HP Launches ZGX Fury AI Station With Nvidia GB300

The desktop AI workstation starts at $132,000 and ships with Nvidia's GB300 Grace Blackwell Ultra chipset.

Updated on Nov. 10, 2026 in Semiconductors

HP Launches ZGX Fury AI Station With Nvidia GB300

HP has released the ZGX Fury AI Station, a desktop built around an Nvidia GB300 Grace Blackwell Ultra chipset. The machine ships with 496 GB of LPDDR5X RAM and supports up to 276 GB of VRAM and 32 TB of storage.

Why it matters

The system brings data-center-class AI hardware into a desktop form factor, with memory and bandwidth specifications aimed at large model workloads. Its price places it well above consumer workstations.

The GB300 pairs 72 ARM Neoverse V2 cores with 252 GB of HBM3e memory at 7.1 TB/s of bandwidth, alongside 496 GB of LPDDR5x-6400 RAM and dual 2 TB SSDs. Independent benchmark results were not available in the source material.

The players

HP

HP is a major American computer hardware manufacturer that sells personal computers, workstations, and printers.

Nvidia

Nvidia is a semiconductor company that designs GPUs and AI accelerator chips widely used in data centers and AI development.

The details

The ZGX Fury AI Station includes three PCIe Gen 5 ports and QSFP112 400 Gbps ports for linking two machines, plus two additional M.2 slots for storage expansion. Buyers can configure it with an RTX Pro 2000 or RTX Pro 4000 Blackwell GPU.

Timeline

  1. Released on November 10, 2026.

The Tech Race

The ZGX Fury AI Station packages Nvidia's GB300 Grace Blackwell Ultra superchip into a desktop chassis, moving data-center AI hardware onto a single workstation. The move positions HP alongside other OEMs competing to supply on-premises AI systems built on Nvidia silicon.

The ZGX Fury AI Station is priced at $132,000 before any GPU upgrade, placing it out of reach for most individual buyers. Its primary practical relevance is to organizations that need on-premises capacity to develop and run large AI models locally.

The takeaway

Organizations evaluating on-premises AI infrastructure should compare the ZGX Fury's memory and bandwidth specifications against their actual model sizes before committing to a six-figure purchase. Buyers should also confirm GPU configuration options, since the base price excludes a dedicated GPU.

Further reading

For more on the chips powering machines like this, explore our Semiconductors coverage.