NVIDIA has announced the launch of foundation models for RTX AI PCs at CES. These models, known as NVIDIA NIM microservices, are powered by the new GeForce RTX 50 Series GPUs. The GPUs offer up to 3,352 trillion operations per second of AI performance and come with 32GB of VRAM. Built on the NVIDIA Blackwell architecture, these are the first consumer GPUs to support FP4 compute, enhancing AI inference performance and enabling generative AI models to run locally with a smaller memory footprint.
GeForce has been a significant platform for AI developers since AlexNet was trained on the GeForce GTX 580 in 2012. Last year, over 30% of published AI research papers cited the use of GeForce RTX. With the introduction of generative AI and RTX AI PCs, NVIDIA aims to democratize development through low-code and no-code tools like AnythingLLM, ComfyUI, Langflow, and LM Studio.
NIM microservices connected to graphical user interfaces will simplify access and deployment of generative AI models. NVIDIA AI Blueprints offer preconfigured reference workflows for digital humans and content creation. To meet increasing demand from developers and enthusiasts, top PC manufacturers are launching NIM-ready RTX AI PCs with GeForce RTX 50 Series GPUs.
"AI is advancing at light speed," said Jensen Huang, founder and CEO of NVIDIA. "NIM microservices and AI Blueprints give PC developers and enthusiasts the building blocks to explore the magic of AI."
NVIDIA plans to release a series of NIM microservices for RTX AI PCs from leading model developers such as Black Forest Labs, Meta, Mistral, and Stability AI. These services cover various applications including large language models (LLMs), vision language models, image generation, speech processing, embedding models for retrieval-augmented generation (RAG), PDF extraction, and computer vision.
“GeForce RTX 50 Series GPUs with FP4 compute will unlock a massive range of models that can run on PC,” said Robin Rombach, CEO of Black Forest Labs.
Additionally, NVIDIA introduced the Llama Nemotron family of open models designed for high accuracy in agentic tasks. The Llama Nemotron Nano model will be available as a NIM microservice for RTX AI PCs and workstations.
NIM microservices include essential components for running AI on PCs optimized for deployment across NVIDIA GPUs in various environments. Developers can quickly download these services on Windows 11 PCs using Windows Subsystem for Linux (WSL).
“AI is driving Windows 11 PC innovation at a rapid rate,” said Pavan Davuluri from Microsoft. “NVIDIA NIM microservices give developers ready-to-integrate AI models for their Windows apps.”
The NIM microservices will be compatible with top development frameworks like AnythingLLM and LangChain. Enthusiasts can also experience these services through an upcoming release of the NVIDIA ChatRTX tech demo.
To demonstrate practical applications, NVIDIA previewed Project R2X—a vision-enabled PC avatar capable of assisting users with desktop apps and video calls using generative technology like Audio2Face-3D.
AI Blueprints allow users to create podcasts from PDFs or generate images guided by 3D scenes using local resources on RTX PCs. Initial hardware support includes several GeForce RTX series GPUs with additional support expected later.
NIM-ready RTX AI PCs will be available from major brands such as Acer, ASUS, Dell among others along with local system builders like Corsair and Origin PC.
###
