HP
New to our range

Now available: "HP ZGX Nano G1n AI Station" for professional AI applications

Debora Pape
2.2.2026
Translation: machine translated

This computer in mini PC design is aimed at professional users. The workstation can perform AI calculations of up to 1000 TOPS.

The «ZGX Nano G1n AI Station» from HP is not for people who bake small buns. However, it could be suitable for projects that require local AI calculations with a high workload and powerful hardware.

The system is based on the «GB10 Grace-Blackwell Superchip» from Nvidia, which combines an ARM CPU with 20 cores (ten Cortex-X925 plus ten Cortex-A725) and a fast GPU from the Blackwell generation. According to the manufacturer, this enables AI performance of up to 1000 TOPS (one quadrillion operations per second). 128 gigabytes of standardised LPDDR5X RAM back up the chip.

The computing power is packed into a housing with a compact form factor of around 15 × 15 × 5.5 centimetres (with feet). The workstation is now available from us with 1000 or 4000 gigabytes of SSD storage space.

HP ZGX Nano G1n AI Station (1000 GB, 128 GB, nVidia GB10)
PC
CHF3428.–

HP ZGX Nano G1n AI Station

1000 GB, 128 GB, nVidia GB10

HP ZGX Nano G1n AI Station (4000 GB, 128 GB, nVidia GB10)
PC
CHF4950.–

HP ZGX Nano G1n AI Station

4000 GB, 128 GB, nVidia GB10

The connections are located at the rear:

  • 1 USB-C for the power connection
  • 3 USB-C 20 Gbit/s
  • 1 Ethernet 10 Gbit/s
  • 2 QSFP 200 Gbit/s
  • 1 HDMI 2.1a

The QSFP connections are a highlight for such a small device. They enable a very high data throughput rate, which is essential for cluster connections or high-speed storage connections in AI development.

Bluetooth 5.4 and Wi-Fi 7 are on board for wireless communication. The Nvidia DGX operating system, which is based on Linux, is pre-installed. It is optimised for maximum GPU acceleration.

For professional AI use cases

The «ZGX Nano G1n AI Station» is designed to run AI workloads such as model prototyping, fine-tuning and inference locally. This makes the device particularly suitable for scenarios in which low latency, data protection or offline work capability play a role.

Developers, data scientists, researchers and teams that test or further develop AI models are the target group for this workstation.

Header image: HP

4 people like this article


User Avatar
User Avatar

Feels just as comfortable in front of a gaming PC as she does in a hammock in the garden. Likes the Roman Empire, container ships and science fiction books. Focuses mostly on unearthing news stories about IT and smart products.


Computing
Follow topics and stay updated on your areas of interest

New to our range

Presentations of select new products at Digitec and Galaxus. Tested, to the point, no marketing blah.

Show all

These articles might also interest you

  • Product test

    Back to the roots: the Asus ROG NUC tested

    by Kevin Hofer

  • Background information

    Pure Sleeper-PC p*rn: Sleepy is ready!

    by Kevin Hofer

  • Product test

    M4 Mac Mini tested: The smallest is the biggest

    by Samuel Buchmann

1 comment

Avatar
later