Build Personal AI Agents on Windows PCs with New Tools from Microsoft and NVIDIA | NVIDIA Technical Blog

Technical Blog

Subscribe

Related Resources

Agentic AI / Generative AI

Build Personal AI Agents on Windows PCs with New Tools from Microsoft and NVIDIA

Turnkey agent sandboxing on native Windows is now available, plus 2x faster agentic inference, new agent apps, and more

Jun 02, 2026

By Annamalai Chockalingam and Gerardo Delgado

Like

Discuss (0)

AI-Generated Summary

Like

Dislike

NVIDIA and Microsoft have introduced new tools at COMPUTEX 2026 to support the development of secure, on-device AI agents on Windows, including Microsoft eXecution Containers (MXC) for enhanced security and NVIDIA OpenShell for runtime integration.<br>NVIDIA RTX Spark desktops and laptops deliver powerful AI performance with 1 petaflop capability and up to 128 GB memory, while Microsoft offers a developer edition preloaded with tools for AI development.<br>Updates to NVIDIA NemoClaw, Hermes Agent, and H Companys Holo 3.1 models expand agent capabilities across Windows and Linux, improving ease of setup, native app integration, and performance on NVIDIA GPUs.

AI-generated content may summarize information incompletely. Verify important information. Learn more

AI agents are changing how you interact with your PC. Creators, developers, and AI enthusiasts are already using these agents extensively to assist with day-to-day tasks such as coding, video editing, and content management.

NVIDIA and Microsoft are teaming up to enable the next generation of developers to build on-device agents on the Windows platform, with easier setup, native security, and integration with the apps and tools developers already use.

This post details new tools NVIDIA and Microsoft unveiled at NVIDIA GTC Taipei at COMPUTEX 2026 and Microsoft Build 2026 to meet the exploding demand for agents. These tools include turnkey agent sandboxing on native Windows, 2x faster agentic inference, new agent apps and tools from Nous Research and H Company, and enhanced multi-GPU support across llama.cpp and ComfyUI. The local AI development stack is now ready to run complex agentic AI workflows alongside users.

How to secure local agents with Microsoft eXecution Containers and NVIDIA OpenShell

At Microsoft Build, Microsoft announced a set of security primitives to allow agents to execute code, operate on files, and orchestrate tasks across systems with built-in identity and policy execution. The Microsoft eXecution Containers (MXC) form the policy layer, defining and instrumenting isolation and containment while relying on native Windows operating system constructs to apply these policies.

For developers, this lowers a critical barrier: agents interacting with personal files and apps pose real prompt injection risks, and MXC ensures they can’t access the full system.

NVIDIA is also collaborating with Microsoft to bring NVIDIA OpenShell runtime to Windows, built on MXC. Integrating MXC through OpenShell provides an easy-to-integrate package for developers to deploy autonomous, always-on agents safely, while providing additional capabilities such as policy creation and management, inference routing, and personally identifiable information (PII) obfuscation.

Top agentic apps are looking to leverage MXC and OpenShell to strengthen their security in Windows, including the popular open source agents OpenClaw and Hermes Agent.

How does NVIDIA RTX Spark power personal AI agents?

Earlier this week at GTC Taipei, NVIDIA unveiled the NVIDIA RTX Spark product family, including small form factor desktops and laptops built for the age of personal assistants. These desktops and laptops deliver 1 petaflop of AI power, up to 128 GB of memory, and CUDA-accelerated AI frameworks for running large models alongside everyday work.

Microsoft is creating an RTX Spark special developer edition—the Surface RTX Spark Dev Box—preloaded with a modified Windows configured for developers and the top developer tools you need to get started. To learn more, see Building the next generation of devices for developers: Surface RTX Spark Dev Box.

How are NVIDIA NemoClaw, Hermes Agent, and H Company expanding agent capabilities?

NVIDIA NemoClaw for building autonomous AI agents now supports all NVIDIA client systems—GeForce RTX, NVIDIA RTX PRO, NVIDIA DGX Spark, and NVIDIA DGX Station for Windows—through Linux and Windows Subsystem for Linux (WSL). This enables you to easily set up and sandbox an agent, with optimized local models handpicked for your hardware. The update also includes enhancements to the installer to make it easier and more seamless. NemoClaw also now supports running Hermes Agent as an option.

This week, Hermes Agent also released native Windows support, including both a command-line interface, alongside a sleek, new desktop application. This streamlines the user experience, while making it easier for the agent to interact with and use native Windows apps, APIs, and...