At a joint event in San Francisco on Wednesday, Microsoft and NVIDIA announced co-engineered hardware and software updates designed to run local, background AI agents on Windows PCs. The announcements include new operating system capabilities from Microsoft to govern agent execution, alongside dedicated hardware configurations from NVIDIA capable of running large models directly on desktop and laptop hardware.
OS Infrastructure for Persistent Agents
To support persistent AI agents operating on desktop systems, Microsoft announced the general availability of Microsoft Execution Containers (MXC). According to Microsoft EVP Pavan Davuluri, MXC provides OS-level infrastructure that allows agents to execute continuously and securely in the background under operating system control. Microsoft CEO Satya Nadella emphasized during the event that securing desktop execution environments is vital as AI interactions shift toward personalized, background-assisted workflows integrated into daily system operations.
RTX Spark Hardware Brings Models Local
Complementing the operating system changes, NVIDIA introduced RTX Spark, a hardware module combining an NVIDIA Blackwell RTX GPU with up to 6,144 cores and a 20-core Grace CPU connected at 600 GB/s. The unit delivers one petaflop of FP4 AI performance and up to 128GB of unified memory. Microsoft highlighted the Surface Laptop Ultra built around the architecture, while partners including Dell, HP, Lenovo, ASUS, and Acer plan systems ranging from thin laptops to 24/7 compact desktops. Preorders for RTX Spark laptops opened Wednesday with shipments starting October 16, followed by compact desktop availability in November. The architecture can run models such as Qwen 3.8 Flash Next locally without relying on cloud processing.
Deskside Supercomputing Comes to Windows
NVIDIA also previewed DGX Station for Windows, bringing its enterprise desktop AI infrastructure into the Windows ecosystem. Previously available only on Linux, the desktop supercomputer runs on a GB300 Grace Blackwell Ultra Desktop Superchip, providing 748GB of coherent memory and up to 20 petaFLOPS of FP4 AI compute. The platform allows enterprise developers to fine-tune and run models up to one trillion parameters locally while maintaining access to standard Windows productivity tools and software workflows.
AI-generated editorial illustration; not a documentary photograph.
Source: NVIDIA, Microsoft Kick Off a New Beginning for Windows PCs With RTX Spark and AI Agents
SIGNAL TITANS
Leave a Reply