Red Hat Inc. said today it’s gearing up its artificial intelligence ambitions with the launch of a new platform called Red Hat AI Enterprise that’s meant to make it easier to deploy and manage models, AI agents and applications in hybrid cloud environments.

It debuts alongside the latest version of Red Hat AI and a new, co-engineered software platform called the Red Hat AI Factory with Nvidia.

The Red Hat AI Enterprise and Red Hat AI platforms form part of a comprehensive new “metal-to-agent” development stack, the company said, while the Red Hat AI factory is all about creating and managing the most efficient environment for deploying AI agents.

The IBM Corp. unit said its latest innovations are designed to help enterprises move their AI projects past the “pilot phase.” It said far too many enterprises get stuck, unable to deploy and scale up their AI projects due to the use of fragmented tools and inconsistent infrastructure. To get around this, Red Hat AI Enterprise unifies model and application lifecycles so AI can be managed as a regular enterprise system. That way, it said, AI delivery will become as repeatable and reliable as traditional software deployment.

The company is positioning Red Hat AI Enterprise as a “foundation for AI production” that provides capabilities including AI inference, model tuning, customization, deployment and management tools in a single package. It’s meant to support any kind of AI model in any environment, including the cloud or on-premises. Red Hat’s cloud application platform OpenShift sits at the core of Red Hat AI Enterprise, which means developers will be using familiar development and deployment tools and frameworks, it said.

Using Red Hat AI Enterprise, organizations will benefit from fast, scalable and cost-effective AI inference powered by Red Hat’s vLLM inference engine, integrated observability and lifecycle management tools and flexible deployment options for any environment,

Red Hat AI Vice President and General Manager Joe Fernandes said AI needs to be operationalized as a core component of enterprise software stacks, rather than a standalone silo. “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,” he said.

Architecting AI factories

Red Hat AI Enterprise will also serve as the hybrid cloud foundation of the new Red Hat AI Factory with Nvidia, which combines Red Hat’s model management and deployment tools with Nvidia’s accelerated computing software. It’s meant to simplify the management of both traditional infrastructure and complex AI computing stacks, Red Hat said, so teams can accelerate their path from pilot to production AI.

The new platform takes care of things such as provisioning the underlying infrastructure for AI workloads and optimizing it to enhance its performance. It provides access to dozens of preconfigured AI models, including IBM’s Granite family and Nvidia’s Nemotron and Nvidia Cosmos models, enhancing flexibility for developers. Because it’s built on Red Hat, users will also benefit from AI that inherits Red Hat’s security and compliance capabilities, reducing risk and mitigating downtime.

“We’re accelerating the path to deploy AI and move quickly to production using Red Hat AI Factory with Nvidia,” said Red Hat Chief Technology Officer Chris Wright. “With a stable, high-performance foundation driven by our proven hybrid cloud offerings, we’re enabling our customers to own their AI strategy and scale with the same rigor they apply to their core IT platforms.”

More models and Model-as-a-Service access

Somewhat confusingly, Red Hat also offers a popular platform known as Red Hat AI, which is receiving a major upgrade with arrival of version 3.3.

Red Hat AI can be considered as the broader portfolio of tools and services used for AI development in hybrid cloud environments, while Red Hat AI Enterprise is the foundation for running models on flexible infrastructure platforms.

With Red Hat AI 3.3, developers are getting access to an expanded library of AI models to work with, including compressed, production-ready versions of Mistral-Large-3, Nemotron-Nano and Apertus-8B-Instruct, as well as new foundational models such as Ministral 3 and DeepSeek-V3.2 with sparse attention. There’s also a technology preview of Model-as-a-Service that’s meant to facilitate self-service access to privately-hosted models through an application programming interface gateway. Moreover, Red Hat is expanding its hardware support with a new technology preview of generative AI support on Intel Corp.’s central processing units, which can now be used to run more cost-effective small language models.

Other new features include the Red Hat AI Python Index, which gives developers the option to use hardened, enterprise-grade versions of tools such as Docling, Training Hub and SDG Hub, on-demand access to GPU resources, and enhanced observability and security features.

Image: SiliconANGLE/Microsoft Designer