Enterprise Agentic AI workflows can now operate within private corporate infrastructure through an architecture integration introduced by SUSE and NVIDIA. The deployment combines the NVIDIA OpenShell Reference Architecture with SUSE Linux Enterprise Server to support autonomous software agents while enforcing strict data controls.
Governing Enterprise Agentic AI Workflows
The system connects NVIDIA Vera Rubin systems for inference with NVIDIA Vera central processing units running OpenShell for agent execution. Furthermore, NVIDIA BlueField-4 data processing units run NVIDIA Sentry software to manage host-independent monitoring and policy enforcement directly in hardware.
Security policies run inside the private corporate perimeter to prevent unauthorized transfers of enterprise intellectual property. Supervisor proxies actively inspect agent traffic, authorize requests according to defined rules, and block network access attempts that could expose internal datasets to public networks.
“As autonomous AI agents and next-generation silicon advancements reshape enterprise operations, AI sovereignty becomes non-negotiable.”
Dr. Thomas Di Giacomo, Chief Technology and Product Officer at SUSE
Dedicated Operating System for Next-Generation Hardware
To support advanced processors, SUSE introduced SLES for NVIDIA. Specifically, this platform provides immediate operational support for new hardware while maintaining a standard SLES 16 base across corporate data centers.
The Linux environment includes unified technical support covering the operating system, kernel, graphics processing unit drivers, and Kubernetes container runtimes. Consequently, engineering teams can deploy workloads directly without rewriting application code or restructuring operational stacks.
Sovereign Deployments in Public Administration
Public sector organizations are testing the framework to support sovereign administrative operations. In Mexico, the Guanajuato Secretariat of Finance initiated a proof-of-concept using the system to evaluate automated public services and internal workflows.
The initiative evaluates high-performance computing hardware under local governance requirements. By maintaining models on premises, the agency ensures public records remain under internal management while preventing third-party platform lock-in.
Partner Certifications and Quantum Integration
SUSE also established a two-tier certification program for independent software vendors. The framework includes partners such as ClearML, Tensormesh, Traefik Labs, AlchemiStudio, Wiser Technology, and Catalogic to protect Enterprise Agentic AI systems with standardized backup, caching, and network gateways.
In addition, SUSE collaborated with Fujitsu to deploy an enterprise hybrid quantum platform combining SUSE software with NVIDIA CUDA-Q libraries on Fujitsu hardware. The system links quantum simulation algorithms directly to production applications, providing secure execution for high-performance computing tasks.
