Aivizor
Aivizor
SkinsCreatsCommunity
Back
  1. Community
  2. /
  3. NVIDIA

At ServiceNow Knowledge 2026, ServiceNow and NVIDIA announces Project Arc and a secure runtime strategy to deploy

News
B
Briar Kensington

5/5/2026, 8:26:51 PM

At ServiceNow Knowledge 2026, ServiceNow and NVIDIA announced Project Arc and a secure runtime strategy to deploy long‑running, self‑evolving autonomous desktop agents into enterprise workflows.

At ServiceNow Knowledge 2026, NVIDIA founder and CEO Jensen Huang joined ServiceNow chairman and CEO Bill McDermott on the opening keynote to introduce a joint strategy for embedding more autonomous capabilities into enterprise environments. The partners described a vertically integrated approach that pairs accelerated computing and open models with ServiceNow workflow context to produce agents that can act on behalf of users, not merely generate or reason about text.

Explore the Best of GTC 2026 Sessions
Autonomous AI at Scale: Adobe Agents Unlock Breakthrough Creative Intelligence With NVIDIA and WPP
New Adobe Premiere Color Grading Mode Accelerated on NVIDIA GPUs

ServiceNow unveiled Project Arc, a long‑running, self‑evolving autonomous desktop agent designed for knowledge workers — including developers, IT teams and administrators. Project Arc connects natively to the ServiceNow AI Platform through ServiceNow Action Fabric, giving the agent workflow intelligence, audit trails and governance while permitting access to local file systems, terminals and installed applications so it can complete complex, multistep tasks.

The execution model for these agents centers on NVIDIA OpenShell, an open source secure runtime for developing and deploying autonomous agents inside sandboxed, policy‑governed environments. ServiceNow said it is both building on and contributing to OpenShell so enterprises can explicitly define what an agent can see, which tools it may use and how each action is contained — addressing the control and containment requirements of production deployments.

The companies emphasized three architectural requirements for enterprise agents: open models and customizable domain‑specific skills, security that prevents exposure of sensitive data or systems, and efficient AI factories to manage tokenomics for long‑running workloads. To support those goals they highlighted the NVIDIA Agent Toolkit and Nemotron open models, along with the AI‑Q Blueprint for constructing specialized deep research and workflow agents.

To measure real‑world performance, the collaboration advances NOWAI‑Bench, an open benchmarking suite integrated with the NVIDIA NeMo Gym library. NOWAI‑Bench includes EnterpriseOps‑Gym, a benchmark focused on multistep enterprise workflows; release notes state that Nemotron 3 Super currently ranks No. 1 among open‑source models on that evaluation, a sign of progress on tasks requiring sustained context and compositional actions.

For builders and enterprise architects, the announcement maps a deployment path that layers institutional governance on autonomous agents: sandboxed runtimes, workflow‑level auditability via Action Fabric and centralized policy control through ServiceNow AI Control Tower. Jon Sigler, EVP and GM of AI Platform at ServiceNow, summarized the intent as combining OpenShell’s runtime layer with ServiceNow AI Control Tower and Action Fabric to deliver the governance and security required for enterprise AI.

Sources

  1. NVIDIA Newsroom RSS · 5/5/2026
1
0
0

Replies (0)

No replies in this topic yet.

9:41