High-end desktop supercomputers for AI – Run and tune the biggest large language models locally

Summary: These desktop supercomputers enable local running and tuning of large language models using Nvidia GH200 Grace-Hopper or GB200 NVL4 Grace-Blackwell systems. They fill the gap left by Nvidia and AMD in workstation hardware for AI inferencing and tuning.

What it does

It provides high-end desktop and workstation PCs designed for AI and HPC, supporting the fastest local inferencing and tuning of frontier large language models.

Who it's for

Ideal for users needing powerful local hardware to run and optimize large language models beyond standard GPU offerings.

Why it matters

It addresses the lack of suitable desktop hardware from major vendors for AI model inferencing and tuning at scale.