• Red Hat launches RHEL AI, offering scalable AI integration for businesses without the need for advanced ML expertise.
• RHEL AI simplifies AI development with scalable solutions and IBM’s Granite models, expanding soon to major cloud platforms.

Red Hat has officially launched to general availability its Red Hat Enterprise Linux (RHEL) AI.

As reported by ZDNet, this is a crucial move for enterprises that want a streamlined, easily scalable solution for AI development and deployment across hybrid cloud environments. RHEL AI has been designed to simplify the implementation of AI technology by providing a completely optimised, bootable RHEL image explicitly created for server deployments.

One of the main features of RHEL AI is its compatibility with Granite models and InstructLab tooling. Such components include enhanced PyTorch runtime libraries and GPU accelerators for AMD Instinct MI300X, Intel, and NVIDIA GPUs, as well as the NeMo framework. By combining these components, Red Hat provides enterprises with the infrastructure required to run complex AI tasks.

RHEL AI is Red Hat’s foundational platform for generative AI model development, testing and production. It combines IBM Research’s Granite large language model family with InstructLab alignment tools, which are based on the LAB methodology. The two-company collaboration introduces a new way to build AI models—developers can contribute to the models in the same way they would to any other open-source project.

The IBM LAB methodology utilises synthetic data generation and multiphase tuning to properly align AI/ML models. IBM LAB accelerates learning while reducing the time and manual effort previously required for model creation. The InstructLab project, an open-source initiative, enables the broader community to further engineer these models, making AI development more accessible. Additionally, IBM has released some of its Granite models, such as the Granite 7B English language model, and various code models under the Apache licence.

Scaling AI with Red Hat OpenShift AI

Apart from the integration of Granite and InstructLab, RHEL AI operates on Red Hat’s OpenShift AI, which is the company’s machine learning operations (MLOps) platform. OpenShift AI lets enterprises scale and adopt machine learning models wherever Kubernetes is operational, deploying models on a large scale.

AI development is an expensive affair, especially the training of large language models (LLMs), which can cost millions. Red Hat’s approach is particularly cost-effective because it employs Retrieval-Augmented Generation (RAG) that allows LLMs to access knowledge from approved sources such as documents and databases. This way, the RAG implementation provides a more accurate and efficient solution for businesses.

Another key benefit of RHEL AI is its ability to be trained using data from a company’s subject matter experts, eliminating the need for advanced machine learning expertise. This practical, domain-specific training makes RHEL AI a valuable tool for companies looking to integrate AI into their workflows without requiring extensive AI knowledge.

“RHEL AI provides the ability for domain experts, not just data scientists, to contribute to purpose-built generative AI models across the hybrid cloud, while also enabling IT organisations to scale these models through Red Hat OpenShift AI,” said Joe Fernandes, Red Hat’s vice president of the Foundation Model Platform.

The flexibility of RHEL AI is one of its key advantages. It can be used wherever data is stored, whether on-premise, at the edge or in public cloud environments, requiring very few changes to the current infrastructure of any organisation to implement AI.

Future availability and collaborations

RHEL AI is currently available on Amazon Web Services, and IBM Cloud through a bring-your-own subscription model. Red Hat has announced plans to expand availability to services like Google Cloud Platform and Microsoft Azure in the coming months, making it accessible to a broader range of enterprises.

In addition, Dell Technologies, together with Red Hat, will bring RHEL AI to its PowerEdge servers – an alternative high-performance server system optimised specifically for AI deployment with NVIDIA-accelerated computing. This partnership aims to simplify AI deployment for businesses by providing pre-validated solutions.

It is difficult to say now what the future of RHEL AI will be. However, Red Hat’s commitment with combining open-source innovation with enterprise-grade support may help the company become one of the key players in the world of artificial intelligence. RHEL AI could well become one of the breakthroughs that allows businesses of all sizes to focus on scalable and cost-efficient AI tools.