Red Hat has announced Red Hat AI Enterprise, an integrated AI platform for deploying and managing AI models, agents and applications across the hybrid cloud. It joins the Red Hat AI portfolio which includes Red Hat AI Inference Server, Red Hat OpenShift AI and Red Hat Enterprise Linux AI.
Red Hat is also introducing Red Hat AI 3.3, bringing significant updates and enhancements across the company’s entire AI portfolio.
Together, these solutions provide a comprehensive “metal-to-agent” stack, integrating the underlying Linux and Kubernetes infrastructure with advanced inference and agentic capabilities to help organisations move from fragmented experimentation to governed, autonomous operations.
The enterprise AI landscape is rapidly evolving from simple chat interfaces toward high-density, autonomous agentic workflows that require deeper integration across the entire technology stack.
However, many organisations remain stuck in the “pilot phase” due to fragmented tools and inconsistent infrastructure. Red Hat AI Enterprise addresses this by unifying the model and application lifecycles allowing IT teams to manage AI as a standardised enterprise system rather than a siloed project – making AI delivery as reliable and repeatable as traditional enterprise software.
Red Hat AI Enterprise: The foundation for AI production
Red Hat AI Enterprise provides core capabilities, including high-performance AI inference, model tuning and customisation and agent deployment and management, with the flexibility to support any model and any hardware across any environment.
Fueled by Red Hat OpenShift – the industry’s leading hybrid cloud application platform powered by Kubernetes – at its core, Red Hat AI Enterprise delivers a highly scalable and more consistent experience with a stronger security footprint, anywhere, using familiar tools and frameworks.
For NVIDIA AI infrastructure, NVIDIA and Red Hat co-engineered the new Red Hat AI Factory with NVIDIA, combining Red Hat AI Enterprise and NVIDIA AI Enterprise to help speed and scale production AI for enterprises.
Key benefits of Red Hat AI Enterprise include:
- Faster, more cost-effective and scalable AI inference using the vLLM inference engine and llm-d distributed inference framework for optimised generative AI model deployments across hybrid hardware environments.
- Integrated observability and lifecycle management to help drive AI lifecycle governance and mitigate risk with an integrated, tested and interoperable enterprise-ready AI stack.
- Flexibility across the hybrid cloud by empowering organisations to deploy and manage AI models, agents and applications with greater consistency wherever their business needs to run backed by trusted Red Hat platforms.
Extending strategic flexibility and full-stack efficiency with Red Hat AI 3.3
Red Hat’s strategy centers on bridging the gap between mission-critical stability and frontier innovation through a unified platform.
The latest software release expands model choice, deepens full-stack optimisation for next-generation silicon and hardens operational consistency for frontier models.
New features and enhancements include:
- Expanded model ecosystem with validated, production-ready compressed versions of Mistral-Large-3, Nemotron-Nano and Apertus-8B-Instruct, available via the OpenShift AI Catalog. Additionally, the release enables deployment of state-of-the-art models like Ministral 3 and DeepSeek-V3.2 with sparse attention, while delivering multimodal enhancements including 3x Whisper speedup, geospatial support, improved EAGLE speculative decoding and enhanced tool calling for agentic workflows.
- Self-service access to AI models with a technology preview of Models-as-a-Service (MaaS). IT teams can provide self-service access to privately hosted models via an API gateway. This centralised approach ensures that AI is available on-demand for internal users, fostering a ready-to-go AI foundation that promotes private and scalable AI adoption within the enterprise.
- Expanded hardware support including a technology preview of generative AI support on CPUs, starting with Intel CPUs for more cost-effective small language model (SLM) inference. Additionally, the platform has expanded its hardware certification for NVIDIA’s Blackwell Ultra and support for AMD MI325X accelerators.
- Unified data-to-model lifecycle secured by the new Red Hat AI Python Index. This trusted repository delivers hardened, enterprise-grade versions of critical tools—including Docling, SDG Hub, and Training Hub—enabling teams to move from fragmented experimentation to repeatable, security-focused production pipelines.
- Comprehensive AI observability and safety with greater visibility into model health, performance and behavior. This provides real-time telemetry across AI workloads, llm-d deployments and Models-as-a-Service (MaaS) cluster and model usage and is paired with a technology preview of integrated NeMo Guardrails, enabling developers to enforce operational safety and alignment across AI interactions.
- Provide on-demand access to GPUs resources by empowering organisations to deploy their own internal GPU-as-a-Service capabilities through intelligent orchestration and pooled hardware access with automatic checkpointing to save the state of long-running training jobs, preventing work loss and maintaining more predictable compute costs, even in highly dynamic or preemptible environments.
Joe Fernandes, vice-president and GM: AI business unit at Red Hat, comments: “For AI to deliver true business value, it must be operationalised as a core component of the enterprise software stack, not as a standalone silo.
“Red Hat AI Enterprise is designed to bridge the gap between infrastructure and innovation by providing a unified metal to agent platform.
“By integrating advanced tuning and agentic capabilities with the industry-leading foundation of Red Hat Enterprise Linux and Red Hat OpenShift, we are providing the complete stack – from the GPU-accelerated hardware to the models and agents that drive business logic.
“Additionally, with Red Hat AI 3.3 organisations can move beyond fragmented pilots to governed, repeatable and high-performance AI operations across the hybrid cloud.”