TLDR: Red Hat has launched Red Hat AI 3, a significant evolution of its enterprise AI platform, designed to streamline the transition of AI workloads from proof-of-concept to production. The platform integrates Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI), and Red Hat OpenShift AI, focusing on high-performance distributed inference and building a robust foundation for next-generation AI agents across hybrid cloud environments.
RALEIGH, N.C. – October 14, 2025 – Red Hat, the world’s leading provider of open source solutions, today announced the release of Red Hat AI 3, marking a pivotal advancement in its enterprise AI platform. This new iteration is specifically engineered to address the complexities and hurdles organizations face in moving AI initiatives from experimental stages to full-scale production, with a strong emphasis on high-performance inference and the burgeoning field of AI agents.
The platform unifies the latest innovations from Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI), and Red Hat OpenShift AI. This integration aims to simplify the deployment and management of AI workloads at scale, fostering greater collaboration between platform engineers and AI engineers. According to a study by the Massachusetts Institute of Technology NANDA project, ‘The GenAI Divide: State of AI in Business,’ approximately 95% of organizations fail to achieve measurable financial returns from their substantial investments in enterprise AI, highlighting the critical need for solutions like Red Hat AI 3.
Joe Fernandes, vice president and general manager of Red Hat’s AI Business Unit, stated, “With Red Hat AI 3, we are providing an enterprise-grade, open source platform that minimizes these hurdles.” He emphasized that the platform enables IT teams to operationalize next-generation AI with confidence, on their own terms, across any infrastructure.
Key Innovations and Features of Red Hat AI 3:
Distributed AI Inference: Red Hat AI 3 introduces enhanced capabilities for distributed inference, leveraging `llm-d`, an extension of the vLLM project, to run large language models natively on Kubernetes. This innovation promises lower costs, faster response times, and more efficient utilization of accelerators such as GPUs from Nvidia and AMD.
Foundation for AI Agents: The Red Hat OpenShift AI 3.0 release is central to building scalable agentic AI systems. It not ably enhances inference capabilities but also includes new features focused on agent management. To accelerate the creation and deployment of agents, Red Hat has implemented a Unified API layer based on Llama Stack, ensuring alignment with industry standards like OpenAI-compatible LLM interface protocols.
Model Context Protocol (MCP) Adoption: Red Hat is an early adopter of the Model Context Protocol (MCP), an emerging standard vital for streamlining how AI models interact with external tools – a fundamental requirement for modern AI agents.
Modular Model Customization Toolkit: The platform includes a new modular and extensible toolkit for model customization, built on existing InstructLab functionality. It provides specialized Python libraries and utilizes open-source projects like Docling for efficient data processing.
Unified Platform for Collaborative AI: Red Hat AI 3 delivers a cohesive and flexible experience designed for the collaborative demands of building production-ready generative AI solutions, unifying workflows across diverse teams.
Model as a Service (MaaS): This feature allows for the central management and on-demand availability of AI models across an organization, giving companies control over their data and infrastructure.
AI Hub: Users gain access to a curated catalog of models and tools to manage the entire lifecycle of models, including deployment via OpenShift AI.
Gen AI Studio: A dedicated environment for developers to experiment with large language models and build applications based on techniques like retrieval-augmented generation (RAG).
Optimized LLMs: The platform also incorporates new optimized Large Language Models, including OpenAI’s gpt-oss, DeepSeek-R1, and speech models such as Whisper and Voxtral Mini.
Also Read:
- Oracle Unveils Global AI Data Platform, Bolstered by $1.5 Billion in Partner Investments
- Microsoft Unveils Expansive Vision for AI Agents in Enterprise Customer Service and Operations
Red Hat AI 3 is designed to support any model on any hardware accelerator, from data centers to public clouds, sovereign AI environments, and the farthest edge, providing a consistent, unified experience for CIOs and IT leaders to maximize their investments in accelerated computing technologies.
