HP ZGX Nano G1n AI Station Workstation – NVIDIA GB10 Superchip 20-Core, 128GB RAM, 2TB NVMe SSD, Blackwell GPU, ConnectX – NVIDIA DGX OS

Product Overview

The HP ZGX Nano G1n AI Station brings data center-grade AI supercomputing directly to your desk in an ultra-compact mini PC footprint. Powered by the NVIDIA GB10 Grace Blackwell Superchip featuring a 20-core ARM CPU (10x Cortex-X925 and 10x Cortex-A725) and Blackwell GPU architecture, this workstation delivers an incredible 1,000 TOPS of FP4 AI performance. Designed to prototype, fine-tune, and inference complex AI models with up to 200 billion parameters locally, it eliminates data center queues and expensive cloud instance costs while keeping sensitive data completely secure.

Equipped with 128GB of coherent unified LPDDR5X RAM and a massive 2TB PCIe 4.0 NVMe SSD, the ZGX Nano G1n handles massive datasets and orchestration tasks seamlessly. Running purpose-built NVIDIA DGX OS paired with the HP ZGX Toolkit, this palm-sized system measuring just 5.9" x 5.9" x 2.1" is the ultimate local development engine for AI researchers, software engineers, and edge deployment teams.

Key Highlights

• NVIDIA GB10 Grace Blackwell Superchip: Combines a Blackwell architecture GPU with a 20-core ARM CPU to output 1,000 TOPS of FP4 AI compute for demanding AI workflows.
• 128GB Coherent Unified Memory: High-bandwidth LPDDR5X system memory allows local execution and fine-tuning of Large Language Models (LLMs) up to 200B parameters.
• 2TB High-Speed NVMe Storage: PCIe Gen 4 x4 M.2 SSD provides ultra-fast read/write speeds for massive dataset ingestion and quick model loading.
• NVIDIA DGX OS & ZGX Toolkit: Ships preloaded with NVIDIA DGX OS and integrated AI software stack for seamless model export, local serving, and prototyping.
• Ultra-Compact Footprint: Palm-sized 2.76 lb mini PC (5.9" x 5.9" x 2.1") delivers workstation-class AI performance without occupying valuable desk space.
• Cutting-Edge Networking: Includes 10 Gigabit & 200 Gigabit Ethernet support, Wi-Fi 7 wireless connectivity, and Bluetooth for ultra-low latency data transfer.
• Scalable Architecture: Pair two HP ZGX Nano units together via high-speed networking to scale local execution up to 405 billion parameter models.

Product Experience

The HP ZGX Nano G1n provides developers with a local sandbox that mirrors cloud production environments. By offloading AI development from primary workstations—whether running Windows, macOS, or Linux—developers gain dedicated compute resources without thermal throttling or system slowdowns. The preinstalled NVIDIA DGX OS offers immediate access to popular frameworks like PyTorch, TensorFlow, and CUDA acceleration straight out of the box.

Despite its minimal 2.76 lb weight and compact frame, an engineered internal air cooling system maintains thermal stability under heavy compute loads. Powered by a high-efficiency 240W USB Type-C power adapter, it supports real-time computer vision, agentic workflows, and local LLM inference with minimal latency.

Technical Specifications

• Processor Platform: NVIDIA GB10 Grace Blackwell Superchip
• NPU Performance: 1,000 TOPS (FP4 Precision)
• Processor Topology: 20-Core ARM (10x Cortex-X925 + 10x Cortex-A725)
• Processor Cache: 16MB L2 Cache
• Graphics & AI Compute: Integrated NVIDIA Blackwell GPU
• Memory Bandwidth: Up to 273 GB/s
• System RAM: 128GB LPDDR5X Coherent Unified Memory (Maximum Supported 128GB)
• Storage: 2TB PCIe NVMe 4.0 x4 Solid State Drive
• Operating System: NVIDIA DGX OS (Linux Platform, English)
• Wired Connectivity: 10 Gigabit Ethernet / 200 Gigabit Ethernet (RJ-45)
• Wireless Communications: Wi-Fi 7 (802.11be) + Bluetooth
• Video Output: 1 x HDMI Output
• USB Ports: 1 x USB Type-C Power Connector (240W), 3 x USB 3.2 Type-C (20Gbps, 30W total power output)
• Form Factor & Thermal: Air-Cooled Mini PC / Desktop Workstation
• Dimensions & Weight: 5.9" W x 5.9" D x 2.1" H (150mm x 150mm x 51mm); 2.76 lbs (1.25 kg)
• Power Supply: 240W USB Type-C Adapter (120V / 230V AC)
• Color: Black
• Compliance: TAA Compliant, RoHS, WEEE, CECP, ErP
• Condition: Brand New
• Warranty: 1-Year HP Limited Warranty (1/1/0 Parts & Labor)

Ideal For

• AI Developers & Data Scientists: Prototype, fine-tune, and run inference on large parameter models locally without incurring recurring cloud API fees.
• Edge Computing Solutions: Deploy low-latency computer vision, autonomous agents, and domain-specific AI models directly on-site where data is generated.
• Enterprise R&D Teams: Ensure strict data privacy and IP protection by keeping proprietary training datasets and proprietary models completely off cloud servers.

What's Included

• HP ZGX Nano G1n AI Station Workstation
• 240W USB Type-C Power Adapter
• Power Cord
• Quick Setup Documentation

Frequently Asked Questions

Q: What size AI models can the HP ZGX Nano G1n run locally?
A: A single ZGX Nano system can prototype, fine-tune, and inference models with up to 200 billion parameters. Connecting two units together allows execution of models up to 405 billion parameters.

Q: What operating system is installed on this unit?
A: It comes preloaded with NVIDIA DGX OS, a Linux-based platform built specifically for AI development and integrated with NVIDIA software stacks.

Q: Is the system RAM or storage upgradeable?
A: The 128GB LPDDR5X system memory is unified on-chip and represents the maximum supported memory. System storage is provided via a standard high-speed 2TB M.2 PCIe NVMe SSD.

Q: How does this system connect to existing workstations?
A: The HP ZGX Nano connects via network connection (10GbE / 200GbE Ethernet or Wi-Fi 7) to pair seamlessly with Windows, Mac, or Linux client systems.

Shop with Confidence

This product is Brand-New, factory sealed in original HP packaging, TAA Compliant, and backed by a 1-Year HP Manufacturer Limited Warranty covering parts and labor.