6 Key Insights on How NVIDIA and SAP Are Securing AI Agents for Enterprise Success
As specialized AI agents move from experimental labs into the core systems that run finance, supply chain, and manufacturing, enterprises face a new challenge: how to trust these autonomous decision-makers. NVIDIA and SAP have teamed up to solve this, unveiling a collaboration that embeds security, governance, and policy controls directly into the runtime environment. Here are six crucial insights from their partnership—covering why trust matters, how OpenShell works, and what it means for the future of business AI.
1. The Rise of Specialized AI Agents in Core Business Systems
Enterprises are no longer just piloting AI assistants; they are deploying autonomous agents that directly interact with systems of record. From finance and procurement to supply chain and manufacturing, these specialized agents make decisions, access sensitive data, and execute workflows at scale. But this shift from simple assistants to autonomous actors changes the trust equation entirely. An agent that can cross application boundaries and operate without human review at every step needs boundaries, policy enforcement, and a complete audit trail before it can safely enter production. This is exactly the challenge NVIDIA and SAP are tackling together—ensuring that as agentic AI becomes mainstream, it does so with the security and governance enterprises demand.
2. A Major Announcement at SAP Sapphire 2025
At SAP Sapphire 2025, NVIDIA founder and CEO Jensen Huang joined SAP CEO Christian Klein’s keynote via video to unveil an expanded collaboration between the two tech giants. The core of the announcement: SAP is embedding NVIDIA’s OpenShell—an open-source runtime designed for secure development and deployment of autonomous AI agents—into the SAP Business AI Platform. This integration means every AI agent built on SAP, whether custom or pre-built, will run inside a security-hardened environment from the start. The move signals that both companies are betting on agentic AI as the next wave of enterprise innovation, but only if trust and safety are built into the foundation.
3. OpenShell: The Runtime Security Foundation
OpenShell provides isolated execution environments for AI agents, enforcing policies at the filesystem and network layers. This infrastructure-level containment guards against damage when agent logic fails or behaves unexpectedly. Within SAP Business AI Platform, OpenShell serves as the runtime security layer for all SAP AI agents, including custom agents built in Joule Studio—SAP’s environment for building and managing end-to-end enterprise agents. By giving each agent a sandboxed space with strict controls, OpenShell ensures that even if an agent makes a mistake, it cannot impact other systems or data. This is the safety net enterprises need before letting autonomous agents touch critical business processes.
4. Reinforcing Trust in Autonomous Agent Operations
Trust is the cornerstone of agentic AI adoption. An agent that can initiate transactions, update records, or trigger workflows without human oversight must be trusted to act correctly. To earn that trust, NVIDIA and SAP are focusing on three pillars: boundaries, policy enforcement, and auditability. OpenShell enforces what an agent can see and do, integrated with enterprise identity and permissions. Every action is logged, creating a transparent audit trail. This approach means that compliance teams can verify agent behavior, and IT can set role-based limits. By addressing trust head-on, the partnership clears a major roadblock for enterprises hesitant to deploy autonomous agents in production environments.
5. Why the Application Layer Is Critical for Agentic AI
NVIDIA CEO Jensen Huang famously describes AI as a five-layer cake: energy, chips, infrastructure, models, and applications. At the top sits the application layer—where AI creates economic value and drives productivity for knowledge workers. SAP, as a global leader in enterprise applications, occupies a unique position at this layer, running the core workflows of finance, procurement, supply chain, and manufacturing. Agents operating within these workflows must understand roles, permissions, and business processes. SAP’s deep integration into enterprise operations makes it the natural platform for agentic AI, ensuring that agents operate within policy and process controls from day one. This application-layer focus is what makes the NVIDIA-SAP collaboration so impactful.
6. Co-Developing the Future of Enterprise Agentic AI
The collaboration goes beyond integration: SAP engineers are codesigning OpenShell alongside NVIDIA, actively contributing to the open-source project. Their focus is on what enterprises need to run agentic AI in production: runtime hardening, policy modeling, enterprise identity integration, and auditing and governance hooks. This co-development ensures that the runtime evolves to meet real-world enterprise requirements, not just academic use cases. Notably, NVIDIA itself is a long-standing SAP customer, running its own finance, supply chain, and logistics on SAP. This shared context gives both companies firsthand insight into what enterprise-grade governance requires in practice. The result is a trust framework built from both sides—technology provider and customer.
The partnership between NVIDIA and SAP represents a pragmatic, security-first approach to agentic AI. By embedding OpenShell into SAP’s Business AI Platform, they give enterprises the confidence to deploy autonomous agents where they matter most: in the core systems that run global businesses. As agentic AI continues to evolve, having a secure, governed runtime will be the difference between a successful deployment and a costly failure. For CIOs and AI leaders, this collaboration offers a blueprint for bringing trust to the next generation of enterprise automation.