TL;DR

The UK Ministry of Defence has selected Red Hat to build a unified AI and hybrid cloud platform across its entire estate. The agreement standardises MLOps for the Royal Navy, British Army, and RAF through the Defence Digital Foundry, aiming to close the gap between AI development and operational deployment.

A platform approach to defence AI

The MOD is moving away from fragmented AI pilots toward a centralised platform engineering approach. By standardising on Red Hat’s infrastructure, including Red Hat OpenShift AI, the department aims to let developers build AI models once and deploy them anywhere — on-premise, in the cloud, or on disconnected field devices.

The agreement centres on the Defence Digital Foundry, the MOD’s software delivery hub. It will now provide a consistent MLOps environment to all service branches, addressing what Red Hat calls the “inference gap” between data science teams and operational infrastructure.

Mivy James, CTO at the UK MOD, said the move was about “rapid adoption, replicating good practice, and the ability to scale” in the era of AI.

Bridging legacy and modern systems

A significant element of the deal is managing the coexistence of legacy virtualised workloads alongside modern containerised AI applications. Red Hat OpenShift Virtualisation provides a migration path for existing systems, allowing the MOD to run traditional virtual machines and neural networks on the same control plane.

The platform also incorporates Red Hat Ansible Automation to enforce governance as models are retrained and redeployed, ensuring compliance with defence security standards throughout the software supply chain.

Looking forward

The deployment signals a broader shift in defence AI maturity — from individual model performance toward the infrastructure needed to reliably deliver, update, and govern AI at scale. For UK defence, success depends less on any single algorithm and more on the ability to move AI capability from the data centre to the front line.