NVIDIA showcases AI development on RTX AI PCs at Microsoft Build
From NVIDIA: 2025-05-19 12:00:00
Generative AI is revolutionizing PC software with digital humans, writing assistants, and creative tools. NVIDIA RTX AI PCs power this transformation, simplifying experimentation and boosting performance on Windows 11. TensorRT for RTX offers optimized inference on Windows ML, supported by top software applications like Autodesk and Bilibili for AI integration and acceleration. NVIDIA NIM provides prepackaged AI models for popular apps, while Project G-Assist offers a no-code AI assistant for RTX AI PCs. TensorRT for RTX delivers over 50% faster AI workload performance on PCs than DirectML, with quality-of-life benefits for developers. The library’s just-in-time engine building and smaller package size optimize AI model performance on RTX GPUs. Developers can access TensorRT for RTX through Windows ML preview, with a standalone SDK available in June. NVIDIA SDKs like CUDA and TensorRT enable AI feature integration and app performance enhancement on RTX-powered machines through Windows ML and TensorRT. Top applications are updating to leverage NVIDIA DLSS, Optix, and Riva for faster performance, better image quality, and generative AI. NVIDIA NIM microservices simplify AI development on PCs, offering optimized models for RTX GPUs with faster performance and compatibility with GeForce RTX 50 and 40 Series GPUs. NIM microservices can run on PCs or the cloud, accelerating AI development with TensorRT and quantized models. NVIDIA AI Blueprints provide sample workflows and projects using NIM microservices, enabling AI developers to jumpstart their work. NVIDIA Project G-Assist offers an AI assistant for GeForce RTX systems, allowing users to control their PC using voice and text commands. Developers can create plug-ins through Project G-Assist Plug-in Builder for natural language AI control, with new open-source plug-in samples available on GitHub for diverse AI enhancements. SignalRGB is developing a G-Assist plug-in for unified lighting control across multiple manufacturers, with plans to integrate it into Langflow for AI applications and workflows. Enthusiasts can join the NVIDIA Developer Discord channel to collaborate, share creations, and experiment with Project G-Assist plug-ins. The RTX AI Garage blog series showcases community-driven AI innovations, NIM microservices, AI Blueprints, and AI agent development for AI PCs and workstations. Stay informed by subscribing to the RTX AI PC newsletter and following NVIDIA Workstation on LinkedIn and X.
Read more at NVIDIA: AI Development on RTX AI PCs at Microsoft Build