Software#

AI software is critical to the NVIDIA HGX AI Factory, providing the foundation for deploying, managing, and orchestrating AI workloads at scale. The following NVIDIA software is part of the NVIDIA HGX AI Factory:

  • NVIDIA AI Enterprise brings together microservices, frameworks, and libraries for AI development with advanced GPU orchestration and infrastructure management:

    • NVIDIA Agent Toolkit is an open-source development platform that gives developers the building blocks—reasoning models, an agentic harness, tools and CUDA-X skills, and the OpenShell secure runtime—to build long-running agents. Security is built into the infrastructure, so enterprises can deploy agents more safely inside the systems where work is done.

    • NVIDIA Run:ai is an AI workload and GPU orchestration platform that helps enterprises accelerate the AI lifecycle through dynamic resource allocation and policy-driven scheduling to maximize throughput and utilization.

  • NVIDIA CUDA-X libraries, built on NVIDIA CUDA™, deliver dramatically higher performance across application domains, including AI and HPC. Through domain-specific optimizations, they enable breakthroughs in engineering, research, and scientific discovery. By leveraging the parallel processing power of GPUs, CUDA-X libraries accelerate computations significantly faster than traditional methods, with continuous NVIDIA optimization further boosting performance.

  • NVIDIA NetQ, a highly scalable, modern network operations toolset that provides visibility, troubleshooting, and validation of the fabric in real time, ensuring the AI factory is operating smoothly without interruption and delivering optimal performance.

The capabilities of NVIDIA software are complemented by reference designs which include the NVIDIA software ecosystem: