Enterprise AI has evolved from simple generation to reasoning, and now it's learning to take action. At ServiceNow Knowledge 2026, NVIDIA founder Jensen Huang and ServiceNow CEO Bill McDermott announced a major expansion of their partnership, introducing Project Arc—an autonomous desktop agent designed for knowledge workers. This collaboration combines NVIDIA's accelerated computing, open models, and secure runtime with ServiceNow's workflow context and governance tools. Here are 10 essential things you need to know about this groundbreaking partnership and the future of autonomous AI in enterprises.
1. The Shift from Generative to Autonomous AI
For years, enterprise AI focused on generating content and reasoning through data. Now, the next frontier is action: AI systems that can independently execute complex, multi-step tasks. This shift requires agents that understand context, operate within defined boundaries, and integrate seamlessly into existing workflows. NVIDIA and ServiceNow are addressing this by building agents that don't just respond to prompts but actively perform work, from managing IT tickets to handling administrative processes, all while maintaining enterprise-grade control and auditability.
2. NVIDIA and ServiceNow Deepen Their Partnership
The two companies are expanding their collaboration across the entire AI stack—from accelerated computing hardware and open-source models to domain-specific skills and secure execution layers. This full-stack approach ensures that autonomous agents are not only powerful but also safe and easy to adopt. By combining NVIDIA's AI infrastructure with ServiceNow's enterprise workflow platform, they aim to deliver agents that can be deployed at scale without compromising security or consistency.
3. Introducing Project Arc: The Desktop Agent
Project Arc is a long-running, self-evolving autonomous desktop agent built for knowledge workers like developers, IT teams, and administrators. Unlike simple chatbots, Project Arc can access local file systems, terminals, and installed applications to complete tasks that traditional automation cannot handle. It performs actions such as deploying code, updating configurations, or managing system settings—all while being governed by enterprise policies. The agent learns from its environment and improves over time, making it a true assistant for complex workflows.
4. How Project Arc Connects to the ServiceNow Platform
Project Arc isn't a standalone tool; it integrates natively with the ServiceNow AI Platform through ServiceNow Action Fabric. This connection provides governance, auditability, and workflow intelligence for every action the agent takes. Through Action Fabric, the agent can trigger processes, retrieve data, and enforce approvals, ensuring that all actions are traceable and compliant. The AI Control Tower then monitors and manages these actions, giving enterprises a centralized view of agent behavior.
5. Three Pillars for Enterprise Autonomous Agents
According to NVIDIA and ServiceNow, successful autonomous agents rely on three requirements: open models that can be customized, domain-specific skills that align with business needs, and robust security that prevents data exposure or unauthorized actions. These pillars underpin Project Arc's design, allowing enterprises to tailor agents to their unique environments while maintaining strict controls. The agents run on AI factories that deliver efficient tokenomics, optimizing cost and performance at scale.
6. NVIDIA OpenShell: Secure Agent Runtime
Project Arc utilizes NVIDIA OpenShell, an open-source secure runtime for developing and deploying autonomous agents in sandboxed, policy-governed environments. OpenShell defines what an agent can see, which tools it can use, and how each action is contained, preventing accidental or malicious harm. ServiceNow is both building on and contributing to OpenShell to advance a common foundation for enterprise-grade agent execution. This collaboration ensures that security is built in from the start, not added as an afterthought.
7. Domain-Specific Skills and Open Models
To be effective, enterprise AI must be adaptable to specific industries and roles. NVIDIA and ServiceNow are providing access to open models that can be fine-tuned with proprietary data, along with a library of domain-specific skills—pre-built capabilities for tasks like IT operations, customer service, and finance. These skills allow agents to hit the ground running, reducing the need for extensive custom development. The open model strategy also avoids vendor lock-in, giving enterprises flexibility in their AI stack.
8. AI Control Tower for Governance
The ServiceNow AI Control Tower provides a centralized platform for managing and governing autonomous agents. It enforces policies, monitors actions in real-time, and provides audit trails for compliance. This ensures that even as agents become more autonomous, human oversight remains possible. The Control Tower works hand-in-hand with Action Fabric to ensure that every agent action is aligned with enterprise rules, from data access limits to approval workflows.
9. Action Fabric: Orchestrating Workflows
ServiceNow Action Fabric is the backbone that connects Project Arc to the broader enterprise ecosystem. It enables the agent to trigger and participate in complex multi-step workflows, seamlessly integrating with existing tools and processes. For example, an agent could automatically escalate a support ticket, update a database, and send a notification—all without human intervention. The fabric also logs every action for later analysis, providing insights into how work gets done.
10. The Future of Enterprise AI Agent Deployments
NVIDIA and ServiceNow are setting a new standard for how enterprises can adopt autonomous AI safely and effectively. By combining open models, secure runtimes, and robust governance, they are addressing the key concerns that have held back wider adoption. As Project Arc evolves, we can expect more intelligent, adaptable agents that handle increasingly complex tasks across industries. The partnership between NVIDIA and ServiceNow signals a future where AI doesn't just assist—it takes action, with full accountability.
In summary, the collaboration between NVIDIA and ServiceNow at Knowledge 2026 marks a significant step forward for enterprise AI. Project Arc, powered by OpenShell and integrated with ServiceNow's platform, offers a blueprint for autonomous agents that are both powerful and safe. As these technologies mature, businesses will be able to automate more complex processes while maintaining the control and transparency they need.