Red Hat has introduced the Red Hat AI Factory in collaboration with NVIDIA, a co-engineered software platform that combines Red Hat AI Enterprise and NVIDIA AI Enterprise to support organisations deploying AI at scale.
Forecasts indicate that global enterprise AI spending is expected to exceed $1 trillion by 2029, driven in part by the growth of agentic AI applications. As adoption increases, organisations are adapting their strategies to manage high-density AI workflows alongside existing infrastructure requirements. The Red Hat AI Factory with NVIDIA is designed to support IT operations teams in managing traditional systems as well as evolving AI workloads.
The platform aims to streamline the path to operational AI by providing a software foundation for AI factories running on NVIDIA GPUs. It is supported by hardware from major manufacturers including Cisco, Dell Technologies, Lenovo and Supermicro, enabling organisations to scale AI deployments with consistent operational processes.
Built on the integration of open-source expertise from Red Hat and NVIDIA, the platform is positioned as an enterprise-grade solution. It supports deployment across on-premises environments, cloud infrastructure and edge locations, with an emphasis on performance and security.
Key capabilities include:
- Accelerated Time-to-Value: Enables deployment of production AI using pre-configured models such as the IBM Granite family and NVIDIA NIM microservices, with support for workflow optimisation and model alignment through tools like NVIDIA NeMo to reduce tuning time.
- Optimised Performance and Cost: Provides a unified high-performance serving stack to maximise infrastructure utilisation, leveraging Red Hat AI inference capabilities powered by vLLM, NVIDIA TensorRT-LLM and NVIDIA Dynamo to support AI service level objectives and manage total cost of ownership.
- Intelligent GPU Orchestration: Allows on-demand access to GPU resources through orchestration and pooled infrastructure, with automated checkpointing to help maintain predictable compute costs in dynamic environments.
- Enhanced Enterprise Security: Built on Red Hat Enterprise Linux, the platform includes integrated security and compliance capabilities. It supports a security-hardened foundation for AI workloads requiring isolation and continuous verification, with NVIDIA DOCA microservices contributing to a zero-trust architecture and AI runtime security.
The launch expands the collaboration between Red Hat and NVIDIA and adds another option for organisations seeking a unified platform to deploy and manage enterprise AI workloads.
