Red Hat is sharpening its focus on enterprise AI. The company has introduced Red Hat AI Factory with NVIDIA, a jointly developed platform designed to help businesses turn AI ideas into fully operational, production-ready systems.
For many organisations, the challenge is no longer experimenting with AI — it is scaling it. Red Hat’s new platform aims to simplify that shift.
The AI Factory combines Red Hat AI Enterprise and NVIDIA AI Enterprise into a single, integrated software stack. Together, they provide tools for model fine-tuning, inference, customisation and the deployment of AI agents across on-premises infrastructure, public clouds and edge environments. The goal is to give IT teams a more structured and reliable way to run AI workloads at scale.
According to Red Hat, the platform offers Day 0 support for NVIDIA’s latest hardware architectures, allowing enterprises to immediately leverage new GPU technologies. This is particularly important as companies increasingly move toward agentic AI systems — applications capable of reasoning, acting and making decisions with minimal human intervention. Industry projections suggest enterprise AI spending could cross the $1 trillion mark by 2029, largely driven by these advanced AI use cases.
The platform is supported on infrastructure from Cisco, Dell Technologies, Lenovo and Supermicro, making it easier for enterprises to deploy within familiar data centre environments. Built on Red Hat Enterprise Linux, it is designed to manage AI workloads alongside traditional business applications while maintaining enterprise-level security and compliance standards.
Red Hat AI Factory also comes with pre-configured models, including IBM’s Granite family, NVIDIA Nemotron and NVIDIA Cosmos open models. These are delivered as NVIDIA NIM microservices, helping organisations deploy AI services more quickly. The integration of NVIDIA NeMo enables businesses to adapt and align models using their own proprietary data, improving relevance and performance.
For inference optimisation and observability, the serving stack incorporates technologies such as vLLM, NVIDIA TensorRT-LLM and NVIDIA Dynamo. Together, these components aim to improve performance while giving IT teams better insight into how AI workloads behave in production environments.
Chris Wright, Red Hat’s CTO and Senior Vice President of Global Engineering, said the move from AI experimentation to enterprise-wide production requires a fundamental shift in managing the AI computing stack. He emphasised that the collaboration with NVIDIA is intended to speed up AI deployment and help customers move confidently into production.
Justin Boitano, Vice President of Enterprise AI at NVIDIA, noted that enterprises are increasingly building “AI factories” that transform data into intelligence at scale. He said such environments demand production-grade infrastructure and software that works seamlessly across hybrid cloud setups, adding that the joint platform provides a dependable foundation for developing and deploying next-generation agentic AI applications.
With this launch, Red Hat is reinforcing its ambition to play a larger role in the enterprise AI landscape — offering businesses a clearer path from pilot projects to scalable, real-world AI systems.
Also Read: NVIDIA Partners With Activate Fund to Support Indian AI Startups
