NVIDIA Launches AI Foundation Models for RTX AI PCs

NVIDIA NIM Microservices and AI Blueprints - NVIDIA announced NIM microservices and AI Blueprints, enabling developers and enthusiasts to build AI agents and creative workflows on PCs [1] - NIM microservices are accelerated by GeForce RTX 50 Series GPUs, offering up to 3,352 trillion operations per second of AI performance and 32GB of VRAM [1] - The RTX 50 Series GPUs, built on NVIDIA Blackwell architecture, support FP4 compute, boosting AI inference performance by 2x and enabling generative AI models to run locally with a smaller memory footprint [1] GeForce RTX and AI Development - GeForce has been a key platform for AI developers, with over 30% of published AI research papers citing the use of GeForce RTX last year [2] - New low-code and no-code tools like AnythingLLM, ComfyUI, Langflow, and LM Studio allow enthusiasts to use AI models in complex workflows via simple graphical user interfaces [2] NIM Microservices and AI Blueprints Integration - NIM microservices connected to GUIs make it effortless to access and deploy the latest generative AI models [3] - NVIDIA AI Blueprints, built on NIM microservices, provide preconfigured reference workflows for digital humans, content creation, and more [3] - Top PC manufacturers and system builders are launching NIM-ready RTX AI PCs with GeForce RTX 50 Series GPUs to meet growing demand [3] NVIDIA's Vision for AI Development - NVIDIA's CEO Jensen Huang emphasized that NIM microservices and AI Blueprints provide building blocks for PC developers and enthusiasts to explore AI [4] - Foundation models, trained on immense amounts of raw data, are the building blocks for generative AI [4] NIM Microservices Pipeline - NVIDIA will release a pipeline of NIM microservices for RTX AI PCs from top model developers like Black Forest Labs, Meta, Mistral, and Stability AI [5] - Use cases include large language models (LLMs), vision language models, image generation, speech, embedding models for retrieval-augmented generation (RAG), PDF extraction, and computer vision [5] GeForce RTX 50 Series GPUs and AI Models - GeForce RTX 50 Series GPUs with FP4 compute will unlock a massive range of models that can run on PCs, previously limited to large data centers [6] - NVIDIA announced the Llama Nemotron family of open models, with the Llama Nemotron Nano model offered as a NIM microservice for RTX AI PCs and workstations [6] NIM Microservices Deployment - NIM microservices are optimized for deployment across NVIDIA GPUs, whether in RTX PCs, workstations, or the cloud [7] - Developers and enthusiasts can quickly download, set up, and run NIM microservices on Windows 11 PCs with Windows Subsystem for Linux (WSL) [7] Windows 11 and AI Development - AI is driving Windows 11 PC innovation, with Windows Subsystem for Linux (WSL) offering a cross-platform environment for AI development alongside Windows Copilot Runtime [8] - NIM microservices, optimized for Windows PCs, provide ready-to-integrate AI models for Windows apps, accelerating AI deployment to Windows users [8] Project R2X and AI Agents - NVIDIA previewed Project R2X, a vision-enabled PC avatar that assists with desktop apps, video conference calls, document reading, and summarization [10] - The avatar uses NVIDIA RTX Neural Faces and NVIDIA Audio2Face-3D model for realistic rendering and animation [11] AI Blueprints for PC Users - AI Blueprints provide reference AI workflows that can run locally on RTX PCs, enabling developers to create podcasts from PDF documents and generate images guided by 3D scenes [12] - The PDF to podcast blueprint extracts text, images, and tables from PDFs to create podcast scripts and audio recordings [15] - The 3D-guided generative AI blueprint allows artists to control image generation using 3D objects in tools like Blender [17] Availability and Hardware Support - NIM microservices and AI Blueprints will be available starting in February, with initial hardware support for GeForce RTX 50 Series, GeForce RTX 4090 and 4080, and NVIDIA RTX 6000 and 5000 professional GPUs [18] - NIM-ready RTX AI PCs will be available from top PC manufacturers and system builders [18]