NVIDIA is leading a new wave of AI innovation on personal computers, transforming everyday software into breakthrough experiences. From digital humans and writing assistants to intelligent agents and creative tools, NVIDIA’s RTX AI PCs are making generative AI more accessible and powerful on Windows 11.

At Microsoft Build, NVIDIA introduced several advancements aimed at helping developers and users tap into the full potential of AI. One key innovation is TensorRT for RTX, an optimized AI inference engine redesigned for RTX GPUs. It uses just-in-time, on-device engine building to create highly efficient AI execution tailored to each user’s hardware all while shrinking its package size by 8x. This results in AI workloads running more than 50% faster compared to previous frameworks.

Seamless AI Deployment with Windows ML

TensorRT for RTX is fully integrated with Microsoft’s Windows ML, a new AI inference stack that provides developers with broad hardware compatibility and state-of-the-art performance. Windows ML automatically selects the best hardware to run AI features whether GPU, CPU, or NPU simplifying deployment and ensuring users always benefit from the latest performance improvements without needing to update the app.

Industry-Leading Apps Upgrade with NVIDIA AI SDKs

Several top software makers are already incorporating NVIDIA’s AI technologies to enhance their products.

• LM Studio boosted its app’s performance by over 30% by upgrading to the latest CUDA version.
• Topaz Labs introduced a generative AI video model to improve video quality.
• Chaos Enscape and Autodesk VRED added DLSS 4 support for faster rendering and better visuals.
• Bilibili enhanced livestream quality with NVIDIA Broadcast features like Virtual Background.

These updates demonstrate how NVIDIA’s AI SDKs are reshaping creative and entertainment software to deliver smoother, smarter experiences.

Simplifying AI Development with NVIDIA NIM Microservices

Developers and enthusiasts benefit from NVIDIA NIM microservices, prepackaged AI models optimized for RTX GPUs. These containerized microservices can run on PCs or in the cloud, making it easier than ever to deploy complex AI workflows.

During COMPUTEX, NVIDIA released the FLUX.1-schnell image generation microservice, designed for fast image creation, and updated FLUX.1-dev for compatibility with more RTX 40 and 50 series GPUs. On NVIDIA’s latest Blackwell GPUs, these microservices run more than twice as fast thanks to advanced FP4 and RTX optimizations.

NVIDIA also offers AI Blueprints sample workflows using NIM microservices to help developers get started quickly and customize AI projects.

Introducing Project G-Assist: AI at Your Fingertips

NVIDIA’s Project G-Assist brings AI-powered voice and text control directly to PCs through the NVIDIA app. Users can command their system, apps, and peripherals naturally without navigating complex menus.

The G-Assist plug-in builder allows developers to create custom plug-ins without coding, using simple JSON and Python logic. Recent community plug-ins include.

• Google Gemini with real-time web search,
• IFTTT for smart home automation,
• Discord for easy sharing of gaming highlights,
• Spotify music control,
• Twitch livestream status, and
• SignalRGB for unified lighting control across devices.

This growing ecosystem lets users personalize their AI assistant and workflows with tools tailored to their needs.

Empowering Developers and Enthusiasts

NVIDIA invites AI developers and fans to join its Discord community to share ideas, collaborate on G-Assist plug-ins, and explore new AI possibilities. With the combination of powerful RTX hardware, optimized software stacks like TensorRT and Windows ML, and easy-to-use microservices, NVIDIA is lowering barriers for AI innovation on PCs.