A dedicated Windows PC for local AI
NVIDIA has started selling RTX Spark compact desktops built by ASUS, Dell and MSI. The systems are designed to bring AI agents, image and video generation, and model fine-tuning together on a single Windows PC. A dedicated desktop can keep agents and scheduled tasks running while a laptop travels or handles other work.
Hardware and capacity
RTX Spark desktops pack a 20-core NVIDIA Grace CPU, a Blackwell RTX GPU with 6,144 CUDA cores and up to 128GB of unified memory. NVIDIA rates AI performance at up to 1 petaflop. Ten-gigabit Ethernet enables fast transfers of large datasets and model files to networked storage. According to the company, the hardware can run 120-billion-parameter large language models locally, render 90GB+ 3D scenes in real time and smoothly edit 4:2:2 video.
Software side
Native CUDA support and Linux development tools through Windows Subsystem for Linux give developers a foundation for local AI workflows. Agents can be set up with Hermes Agent or NVIDIA NemoClaw blueprints, and supported models can be fine-tuned with Unsloth. NVIDIA Switchyard, now available in Hermes, automatically routes agent tasks between local and cloud models: everyday tasks stay on device while advanced reasoning can draw on cloud models.



