NVIDIA is set to redefine the landscape of local AI computing with its latest innovations, introducing NIM Microservices and AI Blueprints. These advancements promise to enhance generative AI capabilities on RTX AI PCs and workstations, as highlighted in the inaugural session of the RTX AI Garage series, following key announcements at CES 2025.
The past year has witnessed generative AI's transformative impact across various sectors, pushing the boundary on writing, gaming, productivity, and more. As PC enthusiasts and developers increasingly embrace this groundbreaking technology, NVIDIA is committed to facilitating the AI evolution with its state-of-the-art offerings.
Echoing the innovative spirit of many technological breakthroughs that began in a simple garage, NVIDIA introduces the RTX AI Garage series. This series will provide continuous insights for developers and enthusiasts interested in learning about NVIDIA NIM microservices and AI Blueprints, enabling the construction of advanced AI agents, creative workflows, and productivity applications on AI PCs.
During CES, NVIDIA unveiled new foundational AI models specifically designed for RTX AI PCs, elevating digital human interactions, content creation, and productivity. Powered by GeForce RTX 50 Series GPUs based on the NVIDIA Blackwell architecture, these models showcase up to 3,352 trillion AI operations per second, 32GB of VRAM, and FP4 compute to boost local generative AI performance.
To facilitate rapid AI integration, NVIDIA presents AI Blueprints — ready-to-deploy workflows built on NIM microservices. These allow enthusiasts and developers to swiftly build, test, and launch AI-driven experiences, heralding a new era of practical AI applications for PC users.
“NVIDIA NIM and AI Blueprints enable developers to swiftly bring AI experiences to life on PCs, driving innovation like never before.”
Key challenges in advancing AI on PCs include keeping pace with rapid AI research and the complex task of model adaptation for PC hardware. NVIDIA's NIM microservices offer a streamlined solution with prepackaged AI models optimized for PC deployment, enhancing integration and performance through APIs and RTX GPU capabilities.
CES highlighted NIM microservices geared towards applications such as large language models (LLMs), vision-language integration, and image generation. The upcoming Llama Nemotron family of open models promises exceptional accuracy in agentic tasks, with the Llama Nemotron Nano model set to excel in areas like chat, function calling, and coding.
Developers will soon download and operate these microservices on Windows 11 PCs using Windows Subsystem for Linux (WSL), minimizing backend complexities and prioritizing innovation.
Traditional AI APIs often impose extensive setup demands, hindering innovation. In contrast, NIM microservices provide user-friendly, intuitive APIs designed around distinct model inputs and outputs, facilitating seamless integration with leading AI development frameworks.
“With NIM’s APIs, NVIDIA accelerates AI innovation on PCs, simplifying integration for developers and driving forward the capabilities of AI applications.”
NVIDIA AI Blueprints offer comprehensive frameworks for creating sophisticated AI applications, such as digital humans and podcast generators. At CES, NVIDIA introduced two AI Blueprints: PDF to podcast conversion and 3D-guided generative AI, both designed to streamline AI development on RTX PCs.
These Blueprints simplify the transition from AI experimentation to production, providing resources to build, run, and customize innovative workflows.
The new GeForce RTX 50 Series GPUs are engineered for generative AI challenges, incorporating advanced Tensor Cores, FP4 support, and an AI-management processor to manage multitasking efficiently. This results in enhanced performance and efficiency, broadening generative AI possibilities on PCs.
“The GeForce RTX 50 Series revolutionizes AI performance, supporting more models with improved efficiency thanks to FP4 technology.”
NVIDIA NIM microservices and AI Blueprints are expected to be available next month, with product support extending across key GPU models, including the GeForce RTX 50 Series. Enthusiasts can anticipate exploring these advancements on NIM-ready RTX AI PCs from leading manufacturers like Acer, ASUS, and Dell.
Revisit NVIDIA CEO Jensen Huang’s CES keynote to explore more about these groundbreaking releases in AI technology.
```
