By estherdidsang@gmail.com 
- XpertStation WS300 supports trillion-parameter models without relying on cloud infrastructure
- Dual 400GbE LAN ports enable high-speed distributed multi-node AI workloads
- Unified HBM3e GPU and LPDDR5X CPU memory maximizes bandwidth for AI
MSI has officially launched the XpertStation WS300, a deskside AI workstation based on Nvidia’s DGX Station architecture.
This system is designed to handle demanding large language models, generative AI, and advanced data science workloads.
The platform is powered by the Nvidia GB300 Grace Blackwell Ultra Desktop Superchip and supports up to 748GB of unified, large coherent memory.
Article continues below
Unified memory architecture for high-bandwidth AI processing
The XpertStation WS300 combines HBM3e GPU memory with LPDDR5X CPU memory for high-bandwidth data sharing.
This configuration allows local processing of trillion-parameter models and supports extensive AI workflows without relying on cloud infrastructure.
The workstation includes dual 400GbE LAN ports, which enable multi-node distributed computing with up to 800Gbps aggregate bandwidth.
MSI claims that the XpertStation WS300 delivers data center class performance directly to the desktop environment, with its setup intended to help organizations move from experimentation to production while maintaining consistent compute reliability.
The XpertStation WS300 supports the full AI lifecycle, including large-scale model training, data-intensive analytics, and real-time inference.
By functioning as a centralized AI compute node, the platform enables collaborative fine-tuning and on-demand deployment, but maintains control over its data and intellectual property.
High-speed PCIe Gen5 and Gen6 NVMe storage accelerates dataset ingestion and AI pipelines, ensuring sustained utilization during compute-intensive operations.
Combined with the Nvidia AI Software Stack, the workstation integrates hardware and software to allow seamless workflow transitions from research to production environments.
MSI also integrated Nvidia NemoClaw, an open-source stack that runs OpenShell within a policy-controlled sandbox.
This allows autonomous AI agents to operate continuously and safely at the deskside, using the workstation’s 20petaFLOPS compute potential.
The configuration supports always-on AI processes locally, enabling experiments with advanced AI and robotics applications without transferring sensitive workloads to cloud servers.
“MSI has a strategic vision to advance AI-first computing,” said Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions.
“With Nvidia, we are defining the next era of AI infrastructure, bridging centralized performance and distributed innovation, and enabling organizations to move from experimentation to production with greater speed, scale, and confidence.”
The platform provides extensive capabilities for advanced AI workflows, but its $84,999.99 price tag raises concerns about cost efficiency.
Organizations that do not require maximum memory or continuous trillion-parameter model operation may find the investment difficult to justify.
The system delivers unprecedented local AI performance, enabling demanding computations at the desk.
However, the practical value of this workstation is likely limited to enterprises with high-throughput AI workloads and specific infrastructure requirements.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.