MSI EdgeXpert AI SUPERCOMPUTER GB10 DGX Spark
MSI EdgeXpert AI SUPERCOMPUTER GB10 DGX Spark
SKU:SYM-EDGEXPERT-102SAU
The MSI EdgeXpert AI SUPERCOMPUTER GB10 is a compact high-performance AI system built on the NVIDIA DGX Spark platform. Featuring NVIDIA Blackwell architecture, 64GB unified memory, and 1000 AI TOPS, it delivers exceptional performance for enterprise AI workloads and edge computing.
Couldn't load pickup availability
- See our return policy
- 36 months on-site support Manufacturer Warranty
Product Information
| Brand | MSI |
|---|---|
| Category | Systems - NUC/SFF/AIO |
| SKU | SYM-EDGEXPERT-102SAU |
| Availability | In Stock |
Overview
- The MSI EdgeXpert AI SUPERCOMPUTER GB10 is a purpose-built AI supercomputer designed for organisations requiring serious performance in a compact package. Built on the NVIDIA DGX Spark platform, it combines NVIDIA Blackwell architecture with a powerful 20-core Arm processor to deliver exceptional AI inference and computation capabilities.
With 1000 AI TOPS (tensor operations per second) and 64GB of unified LPDDR5x memory, this system handles demanding AI workloads with ease. The 2TB NVMe storage with self-encryption ensures both performance and security for sensitive data and models.
Connectivity is enterprise-grade, featuring 10 GbE Smart NIC, WiFi 7, and multiple USB 3.2 Type-C ports. The system runs NVIDIA DGX OS, ensuring optimised performance and seamless integration with the broader DGX ecosystem.
Designed for edge computing, data centres, and production environments, the EdgeXpert GB10 delivers high-performance AI computing without requiring significant physical space, making it ideal for organisations seeking scalable, efficient AI infrastructure.
Specifications
| Processor | 20-core Arm (10x Cortex-X925 + 10x Cortex-A725) |
|---|---|
| GPU Architecture | NVIDIA Blackwell |
| AI Tensor Performance | 1000 AI TOPS (FP4, Sparse) |
| System Memory | 64GB LPDDR5x unified |
| Memory Interface | 256-bit |
| Memory Bandwidth | 273 GB/s |
| Storage | 2TB NVMe M.2 with self-encryption |
| USB Ports | 4x USB 3.2 Type-C |
| Ethernet | 1x 10 GbE RJ-45 |
| Network Interface Card | ConnectX-7 Smart NIC |
| Wi-Fi | WiFi 7 |
| Bluetooth | BT 5.4 |
| Display Output | 1x HDMI 2.1a |
| Audio Output | HDMI multichannel |
| Video Encoding/Decoding | 1x NVENC, 1x NVDEC |
| Operating System | NVIDIA DGX OS |
| System Dimensions | 151mm (L) × 151mm (W) × 52mm (H) |
| System Volume | 1.19L |
| System Weight | 1.2kg |
| Warranty | 36 months on-site support |
Key Features
Ideal For
FAQs
What is the MSI EdgeXpert GB10 designed for?
The EdgeXpert GB10 is built for enterprise AI inference, edge computing, real-time AI processing, and data centre deployments. It's optimised for production AI workloads that require high performance in a compact form factor.
What does 1000 AI TOPS mean?
AI TOPS (tensor operations per second) measures the number of AI computations the system can perform per second. 1000 AI TOPS means the system can execute one trillion AI operations per second, making it suitable for demanding inference and compute tasks.
Is this system suitable for AI model training?
This system is optimised for AI inference and edge computing workloads. For large-scale model training, you may want to consider more specialised training platforms, though the EdgeXpert can handle smaller training tasks.
What connectivity options are available?
The system includes 10 GbE Smart NIC (ConnectX-7), WiFi 7, Bluetooth 5.4, 4x USB 3.2 Type-C ports, and HDMI 2.1a for display output.
Does it come with an operating system?
Yes, the EdgeXpert GB10 ships with NVIDIA DGX OS, which is specifically optimised for DGX systems and integrates seamlessly with the NVIDIA AI software ecosystem.
What warranty and support is included?
The system includes 36 months (3 years) of on-site warranty support, ensuring rapid response to any hardware issues.
How compact is this system?
At just 1.19 litres and weighing 1.2kg, the EdgeXpert GB10 is extremely compact, measuring 151mm × 151mm × 52mm. This makes it ideal for space-constrained environments and edge deployment scenarios.
What storage performance can I expect?
The system features 2TB NVMe M.2 storage with self-encryption, delivering fast data access speeds suitable for AI model loading and inference tasks.
