Software
Red Hat on Tuesday unveiled Red Hat AI Enterprise, a unified platform designed to help organizations deploy, manage and govern AI models, agents and applications across the hybrid cloud, as enterprises move beyond fragmented AI pilots toward large-scale, autonomous operations.
The new platform brings together Red Hat’s AI portfolio — including Red Hat AI Inference Server, Red Hat OpenShift AI and Red Hat Enterprise Linux AI — into what the company describes as a “metal-to-agent” enterprise AI stack, integrating Linux and Kubernetes infrastructure with high-performance inference and agentic capabilities. Alongside the launch, Red Hat also introduced Red Hat AI 3.3, delivering updates across its entire AI portfolio.
The announcement comes as enterprise AI adoption shifts from basic chat interfaces to high-density agentic workflows, where AI systems independently execute multi-step tasks. However, Red Hat said many organizations remain stuck in experimentation due to fragmented tools and inconsistent infrastructure. Red Hat AI Enterprise aims to address this by unifying the AI model and application lifecycle, allowing IT teams to manage AI as a standardized enterprise system rather than isolated projects.
Built on Red Hat OpenShift, the platform supports any model on any hardware across any environment, offering high-performance inference, model tuning and agent deployment with a consistent security and governance framework. For NVIDIA-based environments, Red Hat and NVIDIA have co-engineered a new Red Hat AI Factory, combining Red Hat AI Enterprise with NVIDIA AI Enterprise to help accelerate production AI deployments.
Red Hat said the platform enables faster and more cost-effective inference using vLLM and llm-d distributed inference, while integrated observability and lifecycle management help organizations enforce governance, manage risk and monitor AI behavior in real time. The company also emphasized hybrid cloud flexibility, enabling consistent AI operations across data centers, public clouds and edge environments.
With Red Hat AI 3.3, the company expanded its model ecosystem, adding production-ready compressed models such as Mistral-Large-3, Nemotron-Nano and Apertus-8B-Instruct, while enabling deployment of newer models including Ministral 3 and DeepSeek-V3.2. The release also introduces a technology preview of Models-as-a-Service, allowing enterprises to offer self-service access to privately hosted models through a centralized API gateway.
Hardware support has been broadened to include generative AI inference on CPUs, starting with Intel processors, as well as expanded certification for NVIDIA Blackwell Ultra and AMD MI325X accelerators. Red Hat also introduced a new AI Python Index to provide hardened, enterprise-grade tools for secure data-to-model workflows, along with enhanced AI observability and safety features, including integrated NeMo Guardrails.
“For AI to deliver true business value, it must be operationalized as a core component of the enterprise software stack,” said Joe Fernandes, Vice President and General Manager of Red Hat’s AI Business Unit. He said the new platform is designed to help enterprises move from fragmented pilots to governed, repeatable and high-performance AI operations across the hybrid cloud.
Red Hat positions the launch as a strategic step toward making agentic AI a production-grade enterprise capability, rather than an experimental add-on, as organizations seek reliable, scalable and governed AI systems.
The new platform brings together Red Hat’s AI portfolio — including Red Hat AI Inference Server, Red Hat OpenShift AI and Red Hat Enterprise Linux AI — into what the company describes as a “metal-to-agent” enterprise AI stack, integrating Linux and Kubernetes infrastructure with high-performance inference and agentic capabilities. Alongside the launch, Red Hat also introduced Red Hat AI 3.3, delivering updates across its entire AI portfolio.
The announcement comes as enterprise AI adoption shifts from basic chat interfaces to high-density agentic workflows, where AI systems independently execute multi-step tasks. However, Red Hat said many organizations remain stuck in experimentation due to fragmented tools and inconsistent infrastructure. Red Hat AI Enterprise aims to address this by unifying the AI model and application lifecycle, allowing IT teams to manage AI as a standardized enterprise system rather than isolated projects.
Built on Red Hat OpenShift, the platform supports any model on any hardware across any environment, offering high-performance inference, model tuning and agent deployment with a consistent security and governance framework. For NVIDIA-based environments, Red Hat and NVIDIA have co-engineered a new Red Hat AI Factory, combining Red Hat AI Enterprise with NVIDIA AI Enterprise to help accelerate production AI deployments.
Red Hat said the platform enables faster and more cost-effective inference using vLLM and llm-d distributed inference, while integrated observability and lifecycle management help organizations enforce governance, manage risk and monitor AI behavior in real time. The company also emphasized hybrid cloud flexibility, enabling consistent AI operations across data centers, public clouds and edge environments.
With Red Hat AI 3.3, the company expanded its model ecosystem, adding production-ready compressed models such as Mistral-Large-3, Nemotron-Nano and Apertus-8B-Instruct, while enabling deployment of newer models including Ministral 3 and DeepSeek-V3.2. The release also introduces a technology preview of Models-as-a-Service, allowing enterprises to offer self-service access to privately hosted models through a centralized API gateway.
Hardware support has been broadened to include generative AI inference on CPUs, starting with Intel processors, as well as expanded certification for NVIDIA Blackwell Ultra and AMD MI325X accelerators. Red Hat also introduced a new AI Python Index to provide hardened, enterprise-grade tools for secure data-to-model workflows, along with enhanced AI observability and safety features, including integrated NeMo Guardrails.
“For AI to deliver true business value, it must be operationalized as a core component of the enterprise software stack,” said Joe Fernandes, Vice President and General Manager of Red Hat’s AI Business Unit. He said the new platform is designed to help enterprises move from fragmented pilots to governed, repeatable and high-performance AI operations across the hybrid cloud.
Red Hat positions the launch as a strategic step toward making agentic AI a production-grade enterprise capability, rather than an experimental add-on, as organizations seek reliable, scalable and governed AI systems.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
