Sopra Steria announced an extension of its strategic collaboration with Red Hat to industrialise sovereign-ready embedded AI.
Paris, Monday 15 June 2026 – Sopra Steria today announced an extension of its strategic collaboration with Red Hat, the world's leading provider of open source solutions, to industrialise sovereign-ready embedded AI, from the data centre to the most constrained field equipment.
Announced on the opening day of Eurosatory, the initiative addresses environments where AI is hardest to deploy and matters most. Defence systems, public services and critical infrastructure must continue to operate when connectivity drops, while keeping data and models under their own control.
For Sopra Steria, a European company anchored in critical systems and regulated sectors, this represents its first industrialised edge offering built entirely on open standards.
While many organisations have successfully experimented with AI, moving from proof of concept to reliable operation in the field remains a significant challenge. Red Hat provides the open building blocks. Sopra Steria integrates, secures and operates them as a deployable, accredited system for the field. Together, the two companies aim to help organisations transition from experimental pilots to production-scale intelligence.
The initiative brings the entire AI lifecycle into a single hybrid cloud environment. Red Hat OpenShift AI centralises model training and lifecycle management, Red Hat Device Edge enables lightweight models to run on highly constrained field equipment, and Red Hat Edge Manager automates maintenance and updates across large fleets of devices, including those operating under degraded or intermittent connectivity conditions.
Real-time intelligence across critical sectors
This collaborative initiative delivers an industrialised lifecycle for edge applications, enabling faster processing and supporting operational continuity where decisions need to be taken locally.
Examples include:
- Transport and fleet logistics: Accelerating real-time anomaly detection and predictive routing algorithms directly on embedded electronic boards within mobile transport networks, helping optimise system availability.
- Public terminal infrastructure: Enabling secure data filtering and immediate diagnostics on health and security terminals in the field, maintaining local operational continuity even when disconnected from central systems.
- Regulated communications and signal processing: Bringing distributed AI inference directly to low-power field devices and constrained hardware environments, enabling real-time processing while reducing latency.
The sovereign automated AI factory pipeline
At the heart of the collaboration is an end-to-end automated AI factory pipeline built on a consistent hybrid cloud environment based on Red Hat technologies.
- Train and govern: Red Hat OpenShift AI centralises the training and lifecycle management of AI models, allowing organisations to train, optimise and manage advanced models on a trusted foundation while ensuring transparency and architectural flexibility.
- Deploy to constrained environments: Red Hat Device Edge enables Sopra Steria to engineer these optimised models into ultra-lightweight operating system images designed specifically for resource-constrained electronic boards and field equipment, maintaining a zero-trust model without sacrificing performance.
- Operate at scale: Red Hat Edge Manager automates AI model updates and system maintenance, allowing field equipment to operate reliably and receive validated updates even under degraded, intermittent or disconnected conditions.
By building this architecture on open standards in collaboration with Red Hat, Sopra Steria is laying the groundwork for the next generation of cloud and sovereign AI innovations. This open blueprint helps organisations avoid proprietary constraints while adapting to evolving operational and regulatory requirements.
Executive quotes
Gregory Wintrebert, Director of Institutional Relations and Strategic Partnerships, Sopra Steria
"Most edge AI never leaves the lab. The hard part was never the model. It was running it reliably on a constrained board, in the field, when the network drops, across thousands of devices. That is exactly what we industrialise with Red Hat. Not another proof of concept, but a production line for embedded AI in Europe's most demanding environments, with our clients keeping full control of their data and their models."
Penny Philpot, Vice President, EMEA Ecosystem, Red Hat
"To successfully operationalize AI in highly regulated sectors like public service and defense, technology must adapt to operational realities, not the other way around. Our collaborative blueprint with Sopra Steria demonstrates the true power of the Red Hat ecosystem. Together, we are delivering a repeatable, open architecture that provides regulated industries with flexibility, compliance, and scales automated AI workflows from the data centre to resource-constrained devices at the furthest edge."