Product Description
MSI EdgeXpert AI Supercomputer Desktop, ARM CPU, NVIDIA GB10 Grace Blackwell Architecture, 128GB LPDDR5 Unify Memory, 4TB NVMe gen5 SSD, WiFi 7, BT 5.3, NVIDIA DGX OS: 11SUS Black
Product information
Technical Details
| Screen Resolution | 1920 x 1080 |
|---|---|
| Max Screen Resolution | 1920x1080 |
| Processor | 3.25 GHz ARM_Cortex_A5 |
| RAM | 128 GB DDR5 |
| Hard Drive | 4 TB SSD |
| Graphics Coprocessor | NVIDIA Blackwell Architecture |
| Chipset Brand | NVIDIA |
| Card Description | Integrated |
| Graphics Card Ram Size | 128 GB |
| Wireless Type | 802.11ax, Bluetooth |
| Number of USB 3.0 Ports | 4 |
| Brand | msi |
|---|---|
| Series | EdgeXpert-13SUS |
| Item model number | EdgeXpert13S |
| Operating System | NVIDIA DGX OS |
| Item Weight | 2.65 pounds |
| Product Dimensions | 5.94 x 5.94 x 2.05 inches |
| Item Dimensions LxWxH | 5.94 x 5.94 x 2.05 inches |
| Color | Black |
| Processor Brand | 20 core ARM |
| Number of Processors | 1 |
| Computer Memory Type | DDR5 RAM |
| Flash Memory Size | 4 TB |
| Hard Drive Interface | Solid State |
Additional Information
| ASIN | B0G2P8WTL4 |
|---|---|
| Customer Reviews |
5.0 out of 5 stars |
| Best Sellers Rank |
|
| Date First Available | November 17, 2025 |
Warranty & Support
Feedback
From the manufacturer
Unleash Petaflop Power at Your Desk
The MSI EdgeXpert MS-C931 delivers 1 petaFLOP desktop AI performance through the NVIDIA GB10 Grace Blackwell Superchip. With 128GB memory and pre-installed AI software stack, run cutting-edge models up to 200 billion parameters from DeepSeek, Meta, and Google locally and instantly.
Unrivaled Scalability
- 200B Parameter Local Processing - Run LLMs locally for complete data security and ultra-low latency across your entire AI workflow.
- 405B Parameter Stacking - Dual-system configuration to handle larger AI models locally.
Unmatched AI Computing Power
- NVIDIA GB10 Grace Blackwell Superchip - Grace Blackwell architecture accelerates model tuning and real-time inference.
- NVLink-C2C Technology - High-performance CPU to GPU interconnect with up to 5x PCIe 5.0 bandwidth.
-
Medical Applications
NVIDIA Holoscan for medical devices
-
Retail Applications
NVIDIA Riva for retail application and efficiency.
-
Smart City
NVIDIA Metropolis for smart cities
What"s in the box
Top Brand: msi
Customer reviews


Unleash Petaflop Power at Your Desk
The MSI EdgeXpert MS-C931 delivers 1 petaFLOP desktop AI performance through the NVIDIA GB10 Grace Blackwell Superchip. With 128GB memory and pre-installed AI software stack, run cutting-edge models up to 200 billion parameters from DeepSeek, Meta, and Google locally and instantly.


Unrivaled Scalability
- 200B Parameter Local Processing - Run LLMs locally for complete data security and ultra-low latency across your entire AI workflow.
- 405B Parameter Stacking - Dual-system configuration to handle larger AI models locally.

Unmatched AI Computing Power
- NVIDIA GB10 Grace Blackwell Superchip - Grace Blackwell architecture accelerates model tuning and real-time inference.
- NVLink-C2C Technology - High-performance CPU to GPU interconnect with up to 5x PCIe 5.0 bandwidth.
-
Medical Applications
NVIDIA Holoscan for medical devices
-
Retail Applications
NVIDIA Riva for retail application and efficiency.
-
Smart City
NVIDIA Metropolis for smart cities
What's in the box
Top Brand: msi
Similar brands on Amazon
- Shop the Store on Amazon ›
GMKtec M3 Mini PC Intel i5-12450H (8C/12T, up to 4.40GHz) Desktop Computer 16GB DDR4 (8GB*2) 1TB SSD, HDMI 4K, 4X USB 3.2, USB-C, WiFi 6, BT 5.2, RJ45 2.5G Black4.44.4 out of 5 stars49$419.99$419.99 List Price:$509.99$509.99 - Shop the Store on Amazon ›
GEEKOM [2026 Edition] Mini PC AI A7 MAX with AMD Ryzen 9 7940HS 16GB DDR5(Not LPDDR) & 1TB SSD(Expandable) Desktop Mini Computers for Creation,Radeon 780M Graphics 4 Modes|Premium Support|USB4|8K|SD4.44.4 out of 5 stars385$699.00$699.00 List Price:$949.00$949.00 - Shop the Store on Amazon ›
YEYIAN Gaming PC. Core i7 14700F 5.4 GHZ, RTX 5060 8GB GDDR7, 1TB NVMe, 32GB DDR5, 650W 80+ Gold, Windows 11 Home. Model YPI-ED47F0C-5601U$1,299.99$1,299.99
Customer reviews
- 5 star4 star3 star2 star1 star5 star100%0%0%0%0%100%
- 5 star4 star3 star2 star1 star4 star100%0%0%0%0%0%
- 5 star4 star3 star2 star1 star3 star100%0%0%0%0%0%
- 5 star4 star3 star2 star1 star2 star100%0%0%0%0%0%
- 5 star4 star3 star2 star1 star1 star100%0%0%0%0%0%
Customer Reviews, including Product Star Ratings help customers to learn more about the product and decide whether it is the right product for them.
To calculate the overall star rating and percentage breakdown by star, we don’t use a simple average. Instead, our system considers things like how recent a review is and if the reviewer bought the item on Amazon. It also analyzed reviews to verify trustworthiness.
Learn more how customers reviews work on AmazonImages in this review
Top reviews from the United States
There was a problem filtering reviews. Please reload the page.

santaReviewed in the United States on February 24, 20265.0 out of 5 stars NVIDIA's Pytorch container works on EdgeXpert
Size: 4TB NVMe gen5 SSDStyle: NVIDIA GB10 Grace BlackwellVerified PurchasePros: better thermal design. This small AI box seems to work fine during heavy load without thermal throttling.
Cons: Official distribution of Pytorch does not support Blackwell GB10 yet. I could not install GB10 enabled Pytorch(I have tried some patches and how-to, but failed.) . If your task will run in a container, you can use GB10 enabled Pytorch container from NVIDIA. This cotainer works fine on EdgeXpert.
ReportHelpful
Skender KajoshajReviewed in the United States on February 20, 20265.0 out of 5 stars A great tiny AI machine
Size: 4TB NVMe gen5 SSDStyle: NVIDIA GB10 Grace BlackwellVerified Purchaseit's a great little machine. Running my models helping me optimize some applications.
ReportHelpful
Nathaniel McCleanReviewed in the United States on January 13, 20265.0 out of 5 stars Great little compute box for AI work!
Size: 4TB NVMe gen4 SSDStyle: NVIDIA GB10 Grace BlackwellVerified PurchaseFantastic compute box. Won't win any speed awards, but that massive pool of unified memory makes it possible to work with very large models without memory swapping in Comfy. This little guy keeps surprising me.
3 people found this helpfulReportHelpful
See more reviews
