Nvidia Debuts Platform for Enterprise AI Agents

Nvidia announced NemoClaw at its annual GTC conference, an enterprise-ready version of the viral artificial intelligence agent platform OpenClaw that adds security, privacy controls and policy enforcement, giving companies a way to deploy self-operating AI assistants without exposing sensitive business data.

OpenClaw, created by developer Peter Steinberger, is an open-source software tool that lets an AI assistant act on a user’s behalf across digital systems. It can read and send email, browse the web, access files and initiate transactions without a human approving each individual step.

According to Nvidia, OpenClaw became the fastest-growing open-source project in history after its release. Its problem was straightforward: It was built for individual users, not companies. It had no controls over what data the agent could access, where it could send information, or how its actions could be audited.

NemoClaw solves that problem.

What NemoClaw Does, Why It Matters

According to Nvidia’s announcement, NemoClaw can be installed in a single command and pairs OpenClaw agents with Nvidia’s Nemotron AI models and the newly announced OpenShell runtime. In plain terms, OpenClaw is the AI worker. OpenShell is the walled environment in which a worker operates, one where a company can specify what the agent is allowed to do, what it cannot touch, and what requires a human to sign off.

As reported by TechCrunch, the platform does not require Nvidia’s own hardware and connects to Nvidia’s existing business AI software suite. Agents can use AI models stored locally on company systems or pull from cloud-based models through a connection that keeps internal data from being exposed externally. Nvidia is treating NemoClaw as an early-stage product, acknowledging it is not yet production-ready.

“For the CEOs, the question is, what’s your OpenClaw strategy?” Nvidia CEO Jensen Huang said on stage at GTC. “We all needed a Linux strategy. We all needed an HTTP strategy. Every company in the world today needs to have an OpenClaw strategy, an agentic systems strategy.”

Multiple Competitors, One Unsolved Problem

NemoClaw arrives in a market that is filling fast. OpenAI launched its own enterprise agent platform, OpenAI Frontier, in February. Perplexity AI has also moved into the space. According to PYMNTS, Perplexity introduced a product called Computer, designed to take a broad instruction, such as preparing a research report, break it into smaller tasks and complete the work with minimal human involvement.

Perplexity’s Computer is fully managed by the company, which controls the infrastructure, the AI models used and the rules governing how the system interacts with outside services. OpenClaw hands that responsibility to whoever installs it. NemoClaw sits between those two models, preserving the flexibility of OpenClaw while allowing companies to set and enforce their own rules.

That middle ground is exactly what regulated industries have been waiting for.

Finance Is Ready, but Only With Guardrails

CFOs are already testing the waters. According to PYMNTS Intelligence research, close to 7% of U.S. enterprise CFOs have already deployed AI agents in live finance workflows, with an additional 5% running pilots. The same research found that companies using AI agents capable of autonomous action have automated up to 95% of their accounts receivable work, compared to 38% at firms without that capability. A separate PYMNTS Intelligence study reported that 43% of CFOs expect a significant impact from AI agents that can automatically shift budget dollars based on real-time spending data, with another 47% expecting some impact.

As reported by TechCrunch, Gartner identified governance tools for AI agents in a December report as the essential infrastructure companies need before enterprise adoption of the technology can scale.

For executives weighing when and how to deploy AI agents across their organizations, NemoClaw represents a meaningful shift in what is actually available. The technology to automate complex business tasks has existed for months.

What has been missing is a way to do it without handing over the keys. A system that keeps AI agents productive while enforcing company-defined limits on what they can access, execute and report removes the central objection that has kept procurement, legal and compliance teams on the sidelines.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

The post Nvidia Debuts Platform for Enterprise AI Agents appeared first on PYMNTS.com.