Enterprise Autonomous AI Agents: NVIDIA and ServiceNow's Collaborative Leap

Enterprise artificial intelligence has evolved from generating content to reasoning through complex problems. Now, a new frontier beckons: autonomous action within real-world business workflows. At ServiceNow Knowledge 2026, industry leaders unveiled a bold partnership to bring safe, scalable autonomous AI agents to enterprises. Below, we explore the key announcements and implications through a series of focused questions.

What major partnership was unveiled at ServiceNow Knowledge 2026?

During the opening keynote at ServiceNow Knowledge 2026, NVIDIA founder and CEO Jensen Huang joined ServiceNow chairman and CEO Bill McDermott to announce an expanded collaboration. Together, they are delivering specialized autonomous AI agents designed for enterprise use. These agents are powered by NVIDIA's accelerated computing, open models, and domain-specific skills, combined with ServiceNow's Action Fabric for workflow context and AI Control Tower for governance. The partnership aims to provide a full-stack solution that makes autonomous agents safe, easy to adopt, and tightly integrated with existing enterprise systems. This marks a significant step beyond simple AI prompts toward agents that can act independently within controlled environments.

Enterprise Autonomous AI Agents: NVIDIA and ServiceNow's Collaborative Leap — Source: blogs.nvidia.com

What is Project Arc and how does it differ from standalone AI agents?

Project Arc is a long-running, self-evolving autonomous desktop agent introduced by ServiceNow. Designed for knowledge workers—such as developers, IT teams, and administrators—it connects natively to the ServiceNow AI Platform via the Action Fabric. Unlike standalone agents that operate in isolation, Project Arc brings governance, auditability, and workflow intelligence to every action. It can access local file systems, terminals, and installed applications to complete complex, multistep tasks that traditional automation cannot handle. For example, it can orchestrate software deployments or troubleshoot system issues across multiple tools. Yet it does so with the controls enterprises need to deploy AI at scale, ensuring each action is logged and compliant.

How does Project Arc ensure enterprise-grade security and governance for autonomous agents?

Security is baked into Project Arc from the start. The agent uses NVIDIA OpenShell, an open-source secure runtime that sandboxes agent execution within policy-governed environments. ServiceNow is building on and contributing to OpenShell to advance a common foundation for enterprise-grade agent security. With OpenShell, enterprises can define precisely what an agent can see, which tools it can use, and how each action is contained. Additionally, Project Arc is governed by ServiceNow's AI Control Tower, which provides centralized oversight, and is powered by the Action Fabric for workflow context. This combination ensures agents act without exposing sensitive data or systems, meeting the stringent control requirements of large organizations.

What three essential requirements must be met for deploying autonomous agents at enterprise scale?

According to the partnership's design, three requirements are critical for long-running, autonomous agents in enterprises:

Open models and domain-specific skills: Agents must be customizable and adaptable to unique business needs. Open models allow fine-tuning, while domain-specific skills enable specialized tasks.
Security without exposure: Agents must act without leaking sensitive information or compromising systems. Sandboxed runtimes and strict policy controls are essential.
Efficient tokenomics on AI factories: Running agents at scale requires cost-effective AI infrastructure. NVIDIA's AI factories deliver optimized token processing to keep operational costs manageable.

These three pillars ensure autonomous agents are both powerful and practical for enterprise deployment.

How do open models and domain-specific skills enhance enterprise AI customization?

Open models form the foundation of adaptable enterprise AI. Unlike closed, black-box systems, open models can be customized—fine-tuned on proprietary data, adjusted for industry-specific terminology, or optimized for particular workflows. Domain-specific skills, such as those for IT operations or customer service, further tailor the agent's capabilities. This modular approach allows enterprises to build agents that precisely match their operational context. For example, a financial services firm could train an agent on compliance regulations and risk assessment, while a manufacturer might focus on supply chain optimization. The combination of open models and specialized skills means agents become more effective over time as they learn from enterprise-specific interactions, all while remaining secure and governed by the underlying runtime.

Why is the NVIDIA-ServiceNow partnership significant for the future of enterprise AI?

This partnership represents a holistic approach to autonomous AI, combining cutting-edge hardware, open software, and enterprise workflow integration. By delivering a full-stack solution—from NVIDIA's accelerated computing and OpenShell runtime to ServiceNow's Action Fabric and AI Control Tower—the collaboration addresses the key barriers enterprises face: security, control, and cost. The focus on efficient tokenomics ensures that AI factories can sustain high-volume agent operations without prohibitive expenses. As companies move from experimental AI to production-grade autonomous agents, this unified platform provides the trust and scalability needed for widespread adoption. The involvement of both NVIDIA's hardware leadership and ServiceNow's enterprise workflow expertise signals that autonomous agents are ready for real business impact.