Enterprise AI Agents Enter Production: NVIDIA and SAP Collaboration for Trust and Governance

At SAP Sapphire 2025, NVIDIA founder and CEO Jensen Huang joined SAP CEO Christian Klein via video to announce an expanded collaboration between the two technology leaders. This partnership focuses on enabling enterprises to deploy specialized AI agents with robust security and governance controls, particularly in critical business domains like finance, procurement, supply chain, and manufacturing.

What was announced at SAP Sapphire between NVIDIA and SAP?

The key announcement is the integration of NVIDIA OpenShell—an open source runtime for securely developing and deploying autonomous AI agents—into the SAP Business AI Platform. SAP engineers are actively codesigning OpenShell alongside NVIDIA, contributing back to the open source project. This collaboration ensures that every AI agent built on SAP's platform, including custom agents in Joule Studio, benefits from a security layer that enforces policies at the filesystem, network, and infrastructure levels. As a result, enterprises can run agentic AI with confidence, knowing that failures are contained and audit trails are maintained.

What is NVIDIA OpenShell and how does it enhance safety?

OpenShell is an open source runtime designed to provide isolated execution environments for autonomous AI agents. It enforces policies at the filesystem and network layers, and offers infrastructure-level containment that guards against damage when agent logic fails. Within the SAP Business AI Platform, OpenShell acts as the runtime security layer for all AI agents, including custom ones built in Joule Studio. This means that if an agent behaves unexpectedly—for example, trying to access restricted data or execute unauthorized actions—OpenShell immediately blocks the action and logs the event, providing a clear audit trail. This is essential for production environments where agents touch systems of record and cross application boundaries without human review at every step.

Why is trust a critical issue for autonomous AI agents in enterprises?

The shift from AI assistants to autonomous agents fundamentally changes the trust equation. An agent that can access financial systems, update procurement records, or modify supply chain data must operate within strict boundaries. Without proper governance, an agent's failure could lead to data breaches, incorrect transactions, or operational disruptions. To become part of production work, agents need enforced policies, role-based access controls, and comprehensive audit trails. SAP and NVIDIA's collaboration addresses this by embedding OpenShell into SAP's platform, ensuring every agent operates within defined limits. The partnership also leverages NVIDIA's experience as a long-time SAP customer, providing real-world insights into what enterprise-grade governance requires in practice.

How does the application layer matter for agentic AI?

According to NVIDIA's Jensen Huang, AI is like a five-layer cake: energy, chips, infrastructure, models, and applications. The application layer sits at the top, where AI creates economic value and drives productivity. SAP, as a global leader in enterprise applications, runs critical workflows in finance, procurement, supply chain, and manufacturing. For AI agents to be effective in these domains, they must understand roles, processes, permissions, and data boundaries. They also need an execution environment that limits what an agent can see, do, and where inference runs. SAP's position at the core of enterprise operations makes it a key driver for agentic AI adoption, and the partnership with NVIDIA ensures that agents are both powerful and compliant.

What specific contributions are SAP engineers making to OpenShell?

SAP engineers are 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 to prevent exploits, policy modeling to define granular permissions, enterprise identity integration so agents respect user roles, and auditing and governance hooks for compliance. By contributing back to the open source project, SAP ensures that the entire community benefits from these enterprise-grade enhancements. This codesigning approach gives both companies shared context from NVIDIA's own use of SAP systems for finance, supply chain, and logistics, ensuring the solution meets real-world needs.

How do SAP and NVIDIA address enterprise requirements for trusted agent deployment?

Together, SAP and NVIDIA technologies address key enterprise requirements for deploying trustworthy agentic AI. OpenShell provides the runtime security layer, while SAP's identity and governance framework integrates with it to enforce policies consistently. The collaboration ensures that agents can be used in production with confidence: they have isolated execution environments, policy enforcement at every layer, and detailed audit trails. Additionally, because NVIDIA itself runs on SAP systems, the partnership benefits from practical knowledge of what governance means in a large-scale enterprise. This shared experience helps shape OpenShell's development to meet the highest standards of security, compliance, and operational reliability.