The race to harness artificial intelligence without losing control of data or violating each country’s regulations is accelerating. In this context, Atos has announced the launch of Atos Managed OpenShift AI (AMOS-AI), a new layer of services that incorporates Red Hat OpenShift AI into its managed container platform, with a clear goal: helping organizations operate AI in hybrid and multi-cloud environments while maintaining sovereignty and control over their data.
This offering is especially targeted at regulated sectors and public administrations that have been cautious about generative AI and machine learning, mainly due to fears of losing visibility over where data is processed. AMOS-AI aims precisely for the opposite: train, deploy, and manage AI models within an auditable, open, and governed platform, hosted on private clouds, public clouds, or national infrastructure depending on local rules.
A Managed PaaS for Hybrid and Multi-Cloud AI
Atos presents AMOS-AI as the evolution of Atos Managed OpenShift (AMOS), its managed as-a-service platform based on Red Hat OpenShift. The key innovation is the integration of Red Hat OpenShift AI, Red Hat’s enterprise solution for generative AI and MLOps.
Building upon this, AMOS-AI offers:
- A managed platform environment where Atos handles operations, patching, and observability.
- Capability to prepare data, train models, deploy, and monitor them through a complete and consistent cycle.
- Unified AI management across public clouds, private clouds, and on-premise environments, always within a specific country or region to ensure sovereignty compliance.
- Additional services for data governance, security, and modernization of legacy pipelines, which are often bottlenecks in large-scale AI projects.
Atos emphasizes that AI cannot operate in isolation: it requires a platform that connects data, models, and operations, while simultaneously complying with jurisdiction-specific regulations.
From Pilot to Production: The Leap Many Fail to Make
Atos acknowledges that many organizations get stuck in the pilot phase, with promising proof-of-concept tests that never fully transition to production. The reasons are well-known:
- Difficulty in industrializing models beyond data laboratories.
- Lack of a robust MLOps cycle that includes versioning, deployment, and continuous monitoring.
- Legal or reputational risks when training data or inferences leave controlled environments without sufficient guarantees.
The company claims to have validated AMOS-AI in projects with public agencies and regulated industries, demonstrating that it is possible to align business outcomes with regulatory compliance. In these cases, platform deployment is tailored to specific geography and laws, hosting data and models on controlled infrastructures.
Data Sovereignty as a Design Principle
A key highlight of this announcement is its emphasis on data sovereignty. AMOS-AI is designed so that clients can:
- Decide which country or region data is stored and processed in.
- Run AI models within the same jurisdiction where the data sources are located.
- Move workloads across clouds and data centers without losing traceability or violating location policies.
Atos underscores that the platform is cloud-agnostic: it can operate on public clouds, private clouds, national infrastructures, or hybrid setups, as long as the chain of control and auditability is maintained. The goal is to help organizations navigate an environment where AI and regulations are rapidly evolving.
AI to Automate, Secure, and Simplify
The AI component within AMOS-AI isn’t limited to running models for business use cases; it also applies to automation and operations of the platform itself:
- Process automation: orchestration of pipelines for data, provisioning environments, and continuous deployments.
- Prediction and observability: performance and usage analytics to anticipate issues and optimize costs.
- Security reinforcement: anomaly detection, event correlation, and cybersecurity support from within the infrastructure.
In this regard, AMOS-AI complements with Atos Polaris AI Platform, their platform designed to create “agentic” AI agents capable of orchestrating and executing business and software engineering workflows autonomously or semi-autonomously. Both aim to move AI beyond isolated experiments into the operational fabric of the enterprise.
A Decade of Atos–Red Hat Collaboration
The launch is built upon more than ten years of collaboration between Atos and Red Hat. AMOS, the managed OpenShift layer upon which AMOS-AI is now based, was created together with Red Hat Open Innovation Labs and is a core part of Atos’ Cloud & Modern Infrastructure portfolio.
Meanwhile, Red Hat OpenShift AI is designed to:
- Facilitate development and deployment of generative models and machine learning.
- Integrate MLOps tools (training, testing, versioning, monitoring).
- Provide a cloud-native and Kubernetes-centric foundation that can extend across public clouds and private data centers.
The combined approach aims to deliver a coherent platform both technologically (containers, Kubernetes, open source) and operationally (management, support, professional services).
Leadership Perspectives
For Michael Kollar, EVP and head of Cloud & Modern Infrastructure at Atos, this strategy combines stability and innovation:
The proven stability and scalability of Red Hat OpenShift, now enhanced with AI capabilities, provides a reliable solution for mid-size and large enterprises undertaking high-impact projects.
Kollar emphasizes that the platform:
- Helps mitigate risks during digital transformation.
- Ensures data protection within a sovereign hybrid environment.
- Guides clients toward “agentic enterprises”, where AI agents can handle complex tasks with less operational complexity and a faster time-to-value.
On the Red Hat side, Penny Philpott, VP of Ecosystems for EMEA, highlights that many organizations are torn between deploying an effective AI strategy and complying with an increasingly dense regulatory framework. She believes AMOS-AI, built on OpenShift AI, offers the hybrid, consistent platform needed to:
- Operationalize AI beyond pilots.
- Maintain the control and sovereignty required by modern operations.
What Does a Customer Gain with AMOS-AI?
For a typical organization—such as a public agency, bank, insurance company, energy firm, or industrial business operating across multiple countries—a platform like AMOS-AI can provide:
- Less friction transitioning from pilot to production: the same environment used for testing models is where they are deployed, monitored, and updated.
- Clearer data governance: from the outset, it defines where data resides, which models access what information, and under what controls.
- Cloud flexibility: the ability to combine on-premise infrastructure, sovereign country clouds, and major public clouds without rebuilding architecture.
- Reduced ad-hoc “glue”: instead of custom pipelines and tools, it relies on a standard stack (OpenShift, OpenShift AI) managed by Atos.
- Access to advanced AI agents: via integration with Polaris AI, allowing the design of agents that orchestrate business and development workflows with AI as the engine.
Frequently Asked Questions
What exactly is Atos AMOS-AI?
It is a managed AI platform offering built on Atos Managed OpenShift (AMOS) and Red Hat OpenShift AI. It enables organizations to train, deploy, and manage AI and generative models in hybrid and multi-cloud environments while maintaining control over data location and governance.
How does AMOS-AI help meet sovereignty and regulatory requirements?
AMOS-AI allows organizations to choose where data and models are hosted, running workloads on data centers and clouds compliant with local laws. The entire process—from data preparation to model monitoring—is auditable, simplifying regulatory compliance.
How does AMOS-AI differ from using OpenShift AI directly on a public cloud?
OpenShift AI provides the core technology, but AMOS-AI adds:
- Complete management and operations handled by Atos.
- Data governance, security, and pipeline modernization services.
- The ability to orchestrate complex hybrid setups (multiple providers, sovereign clouds, private data centers) under a unified operational framework.
What role does Polaris AI play in this ecosystem?
Polaris AI is Atos’ platform for “agentic” AI agents. Integrated with AMOS-AI, it allows creating agents that orchestrate and execute business and software workflows, leveraging models deployed on OpenShift AI within a secure, open, hybrid environment.
Sources
- Atos — Press Release: Atos launches Atos AMOS-AI with Red Hat OpenShift AI, a flexible hybrid and multi-cloud solution to support organizations retain data control and cloud sovereignty (Paris, 12/11/2025).