LAS VEGAS – ServiceNow Inc. and NVIDIA Corp. announced a sweeping expansion of their strategic partnership to bridge the chasm between employee desktops and massive data centers.
Centerpiece of the announcement, Project Arc, is an autonomous desktop artificial intelligence (AI) agent designed to reside on employee workstations. Unlike standard AI assistants that merely answer questions, Project Arc is built to execute complex, multi-step tasks independently. To address the security concerns inherent in giving AI the keys to a desktop, the agent is secured by NVIDIA’s OpenShell runtime and managed via the ServiceNow AI Control Tower.
“ServiceNow and NVIDIA set out to make AI real for the enterprise, and today we’re showing the proof of that work,” Joe Davis, ServiceNow’s executive vice president of AI Engineering & Delivery, said in an interview. “Whether it’s autonomous agents that can be trusted on the desktop or governance that extends to the data center… this is enterprise AI built to last.”
To ensure the new technologies are measured fairly, ServiceNow and NVIDIA are also championing a new open benchmarking standard designed to hold the industry accountable.
The NVIDIA news highlighted a raft of announcements from ServiceNow at its annual developers conference here this week.
During a 45-minute press conference with reporters and analysts, ServiceNow CEO Bill McDermott said only six in 10 enterprises have “started agentic motion” yet only one in 10 is seeing tangible results. “You have to have rules and you have to have rails” to make AI work, he said. He also revealed that half of the company’s $16 billion in annual revenue comes from consumption-based pricing” and it gleaned more than $1 billion in security revenue.
“It would not surprise me if (security) is our biggest business in 2031, ’32,” he said, crediting ServiceNow’s acquisition of Armis last month.
“The union of ServiceNow and Armis represents a fundamental shift in how we solve the complexity of the modern enterprise,” Yevgeny Dibrov, CEO and co-founder of Armis, said in an email. “By fusing real-time asset intelligence and security from Armis with automated workflows from ServiceNow, we are turning chaos into control by transforming a fragmented, invisible attack surface into a unified, governed environment. We aren’t just identifying risks; we are building the secure layer of the autonomous enterprise, closing the gap between discovery and remediation, and engineering trust into every connection.”
As AI agents gain more autonomy, the industry has grown increasingly concerned about hallucinations or agents exceeding their permissions. To that end, ServiceNow has significantly upgraded its AI Control Tower, positioning it as a centralized command center for every AI system, agent, and workflow within a company.
The updated platform introduces a literal kill switch capability. Through integration with identity security firm Veza, the system can monitor AI behavior in real-time. If an agent goes “off-script” or attempts to access unauthorized data, the AI Control Tower can detect the anomaly and shut the process down instantly.
The platform’s evolution focuses on automatically finding AI assets across AWS, Azure, Google Cloud, and apps like SAP and Oracle Corp.; using technology from the recent Traceloop acquisition to “peek inside” an agent’s reasoning process; providing out-of-the-box compliance with the EU AI Act and NIST standards; enforcing “least-privilege” access for both humans and AI bots; and tracking ROI and curbing “runaway spend” on expensive large language models (LLMs).
The announcement reflects a broader industry trend toward agentic workflows. ServiceNow is leveraging its massive data footprint — processing 7 trillion transactions annually — to give these agents the business context they need to be effective.
Beyond the NVIDIA partnership, ServiceNow also announced a multi-year agreement with Lenovo. This collaboration will integrate Lenovo’s device intelligence with ServiceNow’s AI platform to automate the entire lifecycle of IT hardware, from deployment to maintenance, aimed at slashing IT support costs.
Since taking the helm of ServiceNow in 2019, McDermott has overseen growing the company’s annual subscription revenue from $3.5 billion to nearly $16 billion this year. The rapid scaling prompted McDermott to tell reporters that the company is essentially “printing a new ServiceNow every year.”
At a recent Financial Analyst Day, leadership raised the stakes by announcing a plan to double the company again by 2030, targeting more than $30 billion in subscription revenue. Chief Financial Officer Gina Mastantuono emphasized that this goal is grounded in a proven track record, noting that the company is already on track to beat its previous five-year target of $15 billion by half a billion dollars. She framed this next milestone not as a speculative “blue sky” ambition, but as the expected output of a durable, organic growth engine that has consistently met its promises to investors.
The company’s current pipeline provides a solid foundation for these lofty projections, with $27.7 billion in remaining performance obligations — a figure roughly double its annual revenue. McDermott cited this metric as proof that ServiceNow is growing faster than any enterprise software company of its scale in history. While the company even teased an upside scenario of $32 billion driven by a 20% compound annual growth rate in its AI and platform businesses, Mastantuono remains conservative for now, asking investors to focus on the $30 billion baseline as their primary benchmark.
