HP Launches ZGX Nano G1n AI Workstation Powered by NVIDIA’s GB10 Superchip
HP has officially unveiled its latest compact workstation, the ZGX Nano G1n AI Station, designed to deliver cutting-edge AI performance in an ultra-small form factor. According to HP’s official page, the new system combines portability, efficiency, and power through NVIDIA’s groundbreaking GB10 Superchip, which integrates the Grace CPU with a Blackwell-based GPU into a single SoC.
The ZGX Nano G1n measures just 15 cm x 15 cm x 5 cm, making it one of HP’s most compact AI-focused workstations. Despite its size, it delivers massive computational power with up to 1000 TOPS of AI performance. The Blackwell GPU inside the GB10 Superchip features 6144 CUDA Cores and 5th Gen Tensor Cores, providing exceptional throughput for AI and machine learning workloads. These are paired with the 20-core NVIDIA Grace CPU, connected through NVLink-C2C, ensuring high-speed data transfer and optimal system efficiency.
The system also features 128 GB of unified LPDDR5X memory, shared between the CPU and GPU to eliminate bandwidth bottlenecks and improve performance consistency across complex workloads. HP highlights that the ZGX Nano G1n supports scalability via interconnect, allowing users to link two units together using a 200 Gbps QSFP/ConnectX-7 interface, effectively doubling the performance potential for large-scale AI models.
On the software side, the ZGX Nano G1n AI Station integrates the ZGX Toolkit, providing a full AI development software stack, and runs on NVIDIA’s DGX OS, enabling a smooth and optimized workflow for developers and data scientists.
In terms of connectivity, HP has equipped the workstation with USB Type-C (20 Gbps) ports, a 10 Gbps Ethernet port, and support for Wi-Fi 7 and Bluetooth 5.4. The rear panel also features a single HDMI output, although HP has not specified the version on its official product page.
The ZGX Nano G1n’s design philosophy combines portability with performance, making it ideal for AI research, on-the-go machine learning tasks, or deployment in space-limited environments such as laboratories or edge computing stations.
HP confirmed that the ZGX Nano G1n AI Station will launch this autumn, though an exact release date has not yet been announced.
What do you think of HP’s new ZGX Nano G1n? Would you consider an ultra-compact AI workstation like this for your projects?
