MSI XpertStation WS300 DGX Spark Workstation

The MSI XpertStation WS300 DGX Spark is a high performance tower workstation engineered for AI development, data science, AI inference and personal cloud workloads. Powered by a state of the art NVIDIA Grace CPU Superchip with 72 Arm Neoverse V2 cores, 496GB of LPDDR5X memory, and an incredible NVIDIA Blackwell Ultra GPU with 252 GB of HBM3e Memory. The combined 748 GB of coherent memory allows for supercomputer levels of performance optimised for the most advanced AI workflows. Pre-installed with Ubuntu 24.04 LTS and NVIDIA AI Developer Tools, the XpertStation WS300 DGX Spark Workstation combines dual 2TB PCIe 5.0 NVMe M.2 SSDs configured in RAID 1 for the operating system with dual NVIDIA ConnectX-8 SuperNICs for advanced networking, alongside two additional PCIe 6.0 M.2 slots connected to the ConnectX-8 for future storage expansion. Featuring dedicated server management and security with the ASPEED AST2600, TPM 2.0 support and a 1,600W 80 PLUS Titanium power supply, the XpertStation provides a powerful foundation for advanced AI deployments.

Features

---

WS300T60L

• Tower (NVIDIA DGX Station™)
• Single NVIDIA Grace™ CPU Superchip with 72 Arm® Neoverse V2 cores
• Single NVIDIA Blackwell Ultra GPU configuration
• Supports a single additional PRO Series Blackwell GPU for display output via PCIe
• 748 GB of Coherent Memory
• (2) 2280 PCIe 5.0 x4 NVMe M.2 ports (Filled with a 2TB RAID1, connected to the CPU)
• (2) 2280 PCIe 6.0 x4 NVMe M.2 ports (Open and available for expansion, connected to the CX8)
• Ubuntu 24.04 LTS with Nvidia AI Developer Tools pre-installed
• (3) PCIe 5.0 expansion slots
• (2) 400G QSFP112 ports with NVIDIA ConnectX-8 SuperNICs
• (1) 1000Base-T Dedicated Server Management Port with:
• ASPEED AST2600 with IPMI 2.0 & DMTF Redfish Support
• Hardware Root of Trust & TPM 2.0 Supported

XpertStation WS300 on NVIDIA DGX Station™ Architecture

Powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip and 748GB of large coherent memory, MSI’s XpertStation WS300 on NVIDIA DGX Station™ architecture delivers data-center-class performance for AI training, inference, and simulation workloads, right from your deskside.

What Sets XpertStation WS300 Apart

Compute-Dense Architecture

NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip combines high-performance NVIDIA Grace CPU and NVIDIA Blackwell Ultra GPU through the NVIDIA® NVLink® Chip-to-Chip (C2C) interconnect, delivering exceptional compute density and efficient system communication.

High-Speed Networking

NVIDIA ConnectX®-8 SuperNIC™ delivers up to 800Gb/s of low-latency, high-bandwidth connectivity, enabling fast data transfer and scalable performance for large-scale AI workloads.

Large Coherent Memory

A 748GB coherent memory pool combines 252GB of high-speed HBM3e GPU memory and 496GB of LPDDR5X system memory through the C2C interconnect, enabling faster data access and efficient large-scale AI model training.

End-to-End AI Software Support

The NVIDIA AI Software Stack provides a full-stack solution for AI workloads including fine-tuning, inference, and data science, enabling seamless development and deployment from desktop to data center.

XpertStation WS300

• 245mm (9.6") Wx 528.4mm (20.8") Hx 595mm (23.4") D
• (1) NVIDIA Blackwell Ultra GPU
• CPU Memory Up to 496GB LPDDR5X | Up to 396GB/s
• GPU Memory Up to 252GB HBM3e | 7.1TB/s
• (1) 10 GbE RJ45 port (Marvell AQC113)
• (2) 400 GbE QSFP ports (NVIDIA ConnectX®-8)
• (2) 2280 PCIe 5.0 x2 NVMe M.2 ports (from CPU)
• (2) 2280 PCIe 6.0 x4 NVMe M.2 ports (for training data)
• (1) Double-Wide PCIe 5.0 x16 slot for display
• (2) Single-Wide PCIe 5.0 x16 slots (each with x8 signal only)
• (1) PCIe x1 2232 M.2 for WiFi/Bluetooth card
• Qualified BMC module from Aspeed AST2600 BMC
• 1600W 80 PLUS Titanium ATX PSU

Accelerating Every AI Workload

MSI’s XpertStation WS300 on NVIDIA DGX Station™ architecture brings data-center-class AI performance to the desktop, enabling model development, data science, and autonomous AI agents with NVIDIA OpenShell.

AI Development

Accelerate deep learning and machine learning training for AI applications ranging from predictive maintenance and medical imaging analysis to natural language processing.

Data Science

Accelerate end-to-end data science workflows to enable faster data ingestion, analysis, and insight generation across massive datasets.

AI Inference

Accelerate local inference for large and complex AI models, delivering high-speed performance for LLM token generation, data analysis, content creation, and AI chatbots.

Personal Cloud

Run advanced AI models on local data and serve as a centralised compute node for team-based fine-tuning and on-demand deployment.