SUSE Announces NVIDIA OpenShell Integration and SLES for NVIDIA for Enterprise AI

The following announcement was issued by SUSE on October 21, 2026. Product capabilities, benefits and customer statements are company-reported. The release schedules general availability of SLES for NVIDIA for November 2026; it does not establish a specific release date within that month.

SUSE Brings Sovereign Control to Enterprise Agentic AI with NVIDIA

Updates give enterprise IT sovereign control over agentic AI workflows while accelerating AI workloads with NVIDIA OpenShell Reference Architecture Integration and SLES for NVIDIA.

India — Oct. 21, 2026 — SUSE®, an open infrastructure company, today introduced new capabilities enabling enterprises to adopt Private Enterprise AI while maintaining operational control, security, and governance over their systems. Available with SUSE AI Factory with NVIDIA, the latest release introduces zero-trust agent security via the NVIDIA OpenShell Reference Architecture alongside SUSE Linux Enterprise Server (SLES) for NVIDIA. The solution is designed to deliver immediate support for next-generation NVIDIA silicon. Together, these innovations provide a unified, enterprise-ready foundation for developers to build production AI applications and autonomous agents, platform engineering teams to deploy them, and enterprise IT to govern them across any footprint.

“As autonomous AI agents and next-generation silicon advancements reshape enterprise operations, AI sovereignty becomes non-negotiable,” said Dr. Thomas Di Giacomo, Chief Technology and Product Officer at SUSE. “By combining enterprise Linux engineered for bleeding-edge hardware with active agent security, SUSE AI Factory with NVIDIA OpenShell equips IT teams to scale AI rapidly while keeping proprietary data, system availability, and budgets strictly under their control.”

Purpose-Built Infrastructure: SLES for NVIDIA

SLES for NVIDIA empowers enterprises to accelerate AI workloads with immediate access to NVIDIA’s latest hardware innovations, including the NVIDIA Vera Rubin infrastructure at the heart of the NVIDIA OpenShell Reference Architecture. As a proven, trusted, enterprise-grade Linux foundation, SLES for NVIDIA is designed to scale seamlessly to rack-scale architectures, bridging rapid silicon innovation with strict operational stability and security. By keeping the underlying infrastructure in lockstep with NVIDIA’s fast-paced release cycles, enterprises gain instant access to peak performance without integration friction or risking system stability.

Key benefits include:

Accelerates Day-Zero Hardware Adoption: Runs AI workloads on cutting-edge NVIDIA silicon immediately while leveraging a trusted enterprise SLES 16 base, with the stated aim of supporting productivity, security and compliance.
Eliminates Production Deployment Friction: Bridges development and production into a unified enterprise footprint, with compatibility and reduced replatforming requirements as stated objectives.
Consolidates Enterprise Support: Delivers single-source technical support backed by continuous automated validation across the operating system, Linux kernel, GPU driver stack, and Kubernetes runtime, with the stated objective of maintaining stack compatibility and reducing support friction.

SUSE Integrates NVIDIA OpenShell Reference Architecture to Secure Autonomous AI Agents

SUSE is enhancing enterprise security for autonomous AI agents by integrating the new NVIDIA OpenShell Reference Architecture into SUSE AI Factory with NVIDIA. The architecture brings together NVIDIA Vera Rubin for model inference, NVIDIA Vera CPUs running OpenShell for agent execution, and NVIDIA BlueField-4 DPUs running NVIDIA DOCA-powered NVIDIA Sentry for host-independent monitoring and in-silicon policy enforcement. As a launch partner for the OpenShell Reference Architecture, SUSE is helping deliver an open, industrial-scale architecture for building, isolating, observing, and governing autonomous AI agents with enterprise-grade security. As businesses increasingly deploy autonomous AI agents to execute complex workflows, they need security policies that control access, protect credentials, and restrict untrusted network activity. By leveraging Kubernetes sandboxing and API gateways alongside OpenShell, SUSE AI Factory with NVIDIA provides a governed environment that controls agent execution while supporting AI workload performance, including:

Active Zero-Trust Governance: Supervisor proxies continually inspect agent traffic, authorize access according to defined policies, and block unauthorized network requests to help prevent code or data exfiltration.
Data Sovereignty & Compliance: Supports agent code execution within the corporate perimeter, with access policies that restrict transfers of enterprise IP and proprietary datasets outside private infrastructure.
Predictable Cost Model: Private model serving, leveraging NVIDIA NIM microservices, eliminates third-party per-token public API fees with predictable capacity-based infrastructure investments.

Securing Public Sector Innovation with Sovereign AI

According to GartnerⓇ, “by 2027, 35% of countries will be locked into region-specific AI platforms using proprietary contextual data.”¹. Building on its existing relationship with SUSE, Mexico’s Guanajuato Secretariat of Finance is leveraging SUSE AI Factory with NVIDIA to validate sovereign AI capabilities to build a cognitive government for its citizens.

Cognitive Government POC: The Secretariat is actively testing SUSE AI Factory with NVIDIA to validate smart, AI-driven public services and streamline internal workflows.
Workload & Hardware Acceleration: The POC evaluates AI workload deployment on high-performance GPU infrastructure, bridging the gap between initial software testing and enterprise-grade production.
Sovereign Infrastructure Foundation: By integrating SUSE AI Factory with NVIDIA into its broader SLES and SUSE Multi-Linux Manager strategy, Guanajuato maintains full data sovereignty, centralized management, and compliance across hybrid cloud and on-premises environments.

“Using SUSE AI Factory with NVIDIA in our environment gives us a clear path toward becoming a truly cognitive government,” said José Luis Huerta Santos, Planning and IT Architecture, Secretariat of Finance of Guanajuato. “Having a sovereign, open source foundation allows us to rapidly evaluate AI workloads while ensuring our citizens’ data, infrastructure availability, and operational costs remain strictly under our control.”

Availability & Event Information

General availability for SLES for NVIDIA is scheduled for November 2026 as an integrated capability of SUSE AI Factory with NVIDIA. Partner certification specifications and blueprint documentation are available at suse.com/product-certification/ready/ai-certification-requirements. NVIDIA Open Agent Safety Platform was announced in September.

SUSE is demonstrating NVIDIA OpenShell and various integrations, governed MCP agentic workflows, SLES for NVIDIA 16.1, and partner solutions live at NVIDIA GTC Berlin (Booth #3053 Gold Sponsor, Oct. 20–22, 2026).

About SUSE

SUSE describes itself as an open infrastructure company for hybrid AI. Backed by 30 years of community collaboration, we transform open source innovation into secure, enterprise-grade solutions for mission-critical environments. Built on choice and neutrality, SUSE says its platform supports management across Linux and Kubernetes distributions. Through our open infrastructure platform and partner ecosystem, IT leaders gain control over their data, budgets, and sovereign AI strategies. Our solutions deploy across hybrid environments, from developer workstations and on-premises data centers to multi-cloud and the farthest edge, and are optimized and certified for SAP. With SUSE, choice happens. Learn more at suse.com.

¹ Gartner, Inc., Predicts 2026: AI Sovereignty, Gaurav Gupta, Katja Ruud, Ran Xu, Mukul Saha, October 16, 2025.

What the announcement means for enterprise deployment

The update brings together an operating-system layer and controls for agent execution. Those are related but distinct parts of a deployment: hardware compatibility supports running a workload, while access policies address what an agent is permitted to do. The announcement does not provide independent penetration-test results, measured workload benchmarks or a detailed pricing schedule. Its security and performance statements should therefore be read as vendor descriptions of the intended architecture.

The Guanajuato example is described as a proof of concept. It is evidence of an evaluation discussed in the announcement, rather than confirmation of a completed government-wide production rollout. Similarly, the proposed capacity-based infrastructure model changes the basis of expenditure but does not establish that operating AI privately has no costs. Hardware, software support and operational requirements remain relevant to any deployment assessment.