Compute-Dense Architecture
The NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip combines an NVIDIA Grace CPU and NVIDIA Blackwell Ultra GPU through NVLink Chip-to-Chip.
Enterprise AI Computing for Singapore
MSI XpertStation WS300 on NVIDIA DGX Station™ Architecture
Bring data-centre-class AI capability to the deskside with a unified NVIDIA Grace Blackwell Ultra platform for model development, data science, inference and simulation.
- NVIDIA GB300 Grace Blackwell Ultra
- 748GB coherent memory
- Up to 800Gb/s networking
- NVIDIA AI software stack
What sets XpertStation WS300 apart
An integrated deskside AI system combining accelerated compute, coherent memory, high-speed networking and NVIDIA AI software.
High-Speed Networking
NVIDIA ConnectX-8 SuperNIC technology provides two 400GbE QSFP ports for up to 800Gb/s of aggregate connectivity.
Large Coherent Memory
A 748GB coherent memory pool combines up to 252GB HBM3e GPU memory with up to 496GB LPDDR5X CPU memory.
End-to-End AI Software Support
The NVIDIA AI software stack supports model fine-tuning, inference and data science workflows from desktop development to data-centre deployment.
A deskside platform for serious AI work
XpertStation WS300 combines CPU, GPU, coherent memory, high-speed networking and system management in one integrated deskside AI platform.
- One NVIDIA Blackwell Ultra GPU
- 748GB combined coherent memory
- Four NVMe M.2 storage connections
- 10GbE and dual 400GbE networking
- Qualified ASPEED AST2600 BMC module
Technical specifications
Published MSI specifications for the XpertStation WS300 platform.
| Processor and compute | Desktop Superchip | NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip |
|---|---|---|
| GPU | Architecture | 1 × NVIDIA Blackwell Ultra GPU |
| CPU memory | LPDDR5X | Up to 496GB at up to 396GB/s |
| GPU memory | HBM3e | Up to 252GB at 7.1TB/s |
| Coherent memory | Total memory pool | 748GB |
| Networking | Ethernet | 1 × 10GbE RJ45 port using Marvell AQC113 |
| High-speed networking | NVIDIA ConnectX-8 | 2 × 400GbE QSFP ports |
| CPU storage | NVMe M.2 | 2 × 2280 PCIe 5.0 x2 ports from CPU |
| Training-data storage | NVMe M.2 | 2 × 2280 PCIe 6.0 x4 ports for training data |
| Display expansion | PCIe | 1 × double-width PCIe 5.0 x16 slot for display |
| Additional expansion | PCIe | 2 × single-width PCIe 5.0 x16 slots, each with x8 signal only |
| Wireless expansion | M.2 | 1 × PCIe x1 2232 M.2 slot for Wi-Fi/Bluetooth card |
| Management | BMC | Qualified BMC module using ASPEED AST2600 BMC |
| Power | Power supply | 1600W 80 PLUS Titanium ATX PSU |
| Dimensions | W × H × D | 245 × 528.4 × 595mm (9.6 × 20.8 × 23.4in) |
Accelerate enterprise AI workloads
Local computing for development, data analysis, inference and controlled AI deployment.
AI Development
Develop and train deep learning and machine learning applications locally.
Practical use Predictive maintenance, medical imaging and language processing.
Data Science
Accelerate data ingestion, analysis and insight generation across large datasets.
Practical use Exploratory analytics and large-dataset processing.
AI Inference
Run large and complex AI models locally for responsive business workflows.
Practical use LLM generation, content analysis and enterprise chatbots.
Personal Cloud
Centralise substantial local AI compute for authorised teams and internal projects.
Practical use Team fine-tuning and on-demand internal deployment.
Compare MSI AI systems
Select the platform scale that best matches your model, memory and deployment requirements.
Swipe horizontally to view every system.
| Comparison | EdgeXpert NVIDIA DGX Spark | PRO Desk Series RTX PRO | XpertStation WS300 NVIDIA DGX Station |
|---|---|---|---|
| NVIDIA Platform | DGX Spark | RTX PRO (single GPU) | DGX Station |
| Performance | 1 PFLOP | Up to 4 PFLOPS | Up to 20 PFLOPS |
| Memory | 128GB | Up to 96GB |
748GB coherent memory
GPU: 252GB HBM3e CPU: 496GB LPDDR5X |
| Max Model Size | Up to 200B | Up to 153B | Up to 1T |
| GPU Interconnect | N/A | PCIe Gen 5 | N/A |
| Applications |
|
|
|
Comparison values follow MSI’s published XpertStation WS300 page.
MSI and NVIDIA advancing deskside enterprise AI
MSI combines workstation engineering with NVIDIA DGX Station architecture to provide an integrated platform for substantial local AI development, analysis and deployment.
Built on NVIDIA DGX Station architecture
MSI XpertStation WS300
Request a Custom Quote
Share your AI workload, model size, deployment environment, storage, networking, timeline and business requirements with Dreamcore.
Enquire About Availability
Discuss a system configuration aligned with your enterprise
AI requirements.
Loading request form...