How NVIDIA and SAP Are Building Trustworthy AI Agents for Enterprise Operations

From finance and procurement to supply chain and manufacturing, specialized AI agents are rapidly entering the enterprise systems where business decisions are made, data is accessed, and workflows run at scale. But with this shift comes a fundamental challenge: how to ensure these autonomous agents operate securely and within strict governance boundaries. At SAP Sapphire, NVIDIA and SAP announced an expanded collaboration to address exactly this—bringing trust to specialized AI agents through robust runtime security and policy enforcement.

The Challenge of Autonomous AI Agents in the Enterprise

Moving beyond simple AI assistants, autonomous agents can now touch systems of record, cross application boundaries, and operate without human review at every step. This change in the trust equation demands new safeguards: agents need clearly defined boundaries, enforced policy controls, and a comprehensive audit trail before being unleashed into production environments. Enterprise leaders need confidence that agentic AI can handle critical tasks without compromising data integrity or violating security protocols.

Why the Application Layer Matters

NVIDIA CEO Jensen Huang has described AI as a five-layer cake: energy, chips, infrastructure, models, and applications. The application layer sits on top, where AI creates economic value and drives productivity for knowledge workers. SAP, as a global leader in enterprise applications and business AI, is a key catalyst at this level—running finance, procurement, supply chain, and manufacturing workflows where agents must operate within existing policy, identity, and process controls. This makes SAP's position at the core of enterprise operations a critical driver for the adoption of agentic AI. Business agents need to understand roles, permissions, data boundaries, and the execution environment that limits what an agent sees and does.

NVIDIA and SAP Collaboration: OpenShell as the Runtime Guardian

At the heart of the collaboration is OpenShell, an open source runtime for securely developing and deploying autonomous AI agents. SAP is embedding OpenShell into its SAP Business AI Platform, providing a security guardrail for all SAP AI agents—including custom agents built in Joule Studio, SAP's environment for building and managing end-to-end enterprise agents. OpenShell offers isolated execution environments, policy enforcement at the filesystem and network layers, and infrastructure-level containment that protects against damage when agent logic fails.

Joule Studio: Building Secure Custom Agents

Within the SAP Business AI Platform, OpenShell serves as the runtime security layer. This means developers using Joule Studio to create specialized agents can now rely on built-in safeguards. The open source nature of OpenShell also allows the community to contribute and audit the code, fostering transparency and continual improvement in enterprise agent security.

SAP's Unique Position in Enterprise Agentic AI

SAP's applications are deeply woven into the fabric of global enterprises—managing financials, supply chains, human resources, and more. This makes SAP an essential partner in deploying trustworthy AI agents. By integrating OpenShell directly into its AI platform, SAP ensures that agents operate within the same policy, identity, and process controls that govern existing enterprise systems. This alignment minimizes risk and accelerates enterprise adoption of agentic AI, because agents inherit the trust infrastructure already in place.

Shared Context: NVIDIA as an SAP Customer

NVIDIA brings a unique perspective to this collaboration: it is itself a longstanding SAP customer, running finance, supply chain, and logistics on SAP systems. This gives both companies shared context for what enterprise-grade governance requires in practice. The lessons NVIDIA has learned from its own use of SAP inform the development of features that matter most to large organizations—such as runtime hardening, policy modeling, and enterprise identity integration.

Codeveloping the Future of Enterprise Agent Governance

SAP engineers are now working alongside NVIDIA to further develop OpenShell's open source codebase. Their contributions focus on what enterprises need to run agentic AI in production: runtime hardening, policy modeling, enterprise identity integration, and auditing and governance hooks. This codevelopment ensures that the runtime evolves to meet real-world security and compliance requirements. Both companies are also contributing back to the open source project, benefiting the broader AI ecosystem.

For enterprises, the result is a trusted foundation for deploying AI agents that can touch critical business data while respecting strict boundaries. The combination of SAP's application-layer expertise and NVIDIA's AI and security know-how creates a powerful solution for the next generation of autonomous enterprise systems.

Learn more about how these technologies work together for secure AI deployment in our section on the application layer.