Governing the Machine: Quali Unveils Control Plane for NVIDIA's Autonomous AI Agents

AUSTIN, Texas – April 29, 2026 – By Sarah Hughes

The era of the autonomous AI agent is rapidly moving from developer sandboxes to the corporate data center, bringing with it immense potential for productivity and a host of new challenges around security, cost, and control. Addressing this critical inflection point, AI infrastructure management firm Quali today announced that its Torque platform will provide an enterprise-grade governance and control plane for NVIDIA NemoClaw, the new reference stack for deploying powerful AI agents.

The integration aims to solve a problem that grows more urgent with each AI breakthrough: how to scale powerful autonomous agents from a single developer's pilot project to a secure, policy-driven, and cost-effective service across an entire organization. While NVIDIA provides the engine to make agents run, Quali is delivering the guardrails to keep them from running wild.

The Rise of the Enterprise AI Agent

At the heart of this announcement is NVIDIA NemoClaw, an open-source reference stack designed to bring enterprise-grade security to the burgeoning world of autonomous agents. NemoClaw bundles several cutting-edge technologies into a single, deployable package. It is built upon OpenClaw, the wildly popular open-source framework that enables AI models to execute real-world tasks and has amassed over 145,000 stars on GitHub, signaling massive developer interest.

NemoClaw packages this agent framework with NVIDIA's own powerful technologies. The model backbone is Nemotron 3 Super, a highly capable open-weight large language model designed for agentic workloads. For security, it integrates NVIDIA OpenShell, a specialized runtime that creates sandboxed execution environments. This critical component wraps the AI agent in layers of security, enforcing policies on what files it can access, what networks it can connect to, and what actions it can perform, with a deny-by-default posture to prevent unauthorized activity.

While NemoClaw makes AI agents secure and production-ready on NVIDIA infrastructure at the individual or team level, it doesn't solve the broader enterprise operations challenge. How does a large company manage dozens or even hundreds of these powerful agent environments across different business units, each with its own security requirements and budget constraints?

Bridging the Governance Gap from Pilot to Production

This is precisely the gap Quali's Torque platform is designed to fill. Torque acts as a centralized control plane that sits above the individual NemoClaw deployments, transforming them from isolated instances into a cohesive, manageable, and multi-tenant infrastructure service.

"NVIDIA NemoClaw is a significant step. It makes powerful autonomous AI agents accessible and secure at the individual deployment level. Torque picks up exactly where that ends," said Lior Koriat, CEO of Quali, in the announcement. "When an organization wants to scale NemoClaw across ten teams, fifty environments, multiple DGX systems, and a hybrid cloud estate, they need governance, policy enforcement, lifecycle automation, and cost control. That is what Torque delivers. NemoClaw gets agents running. Torque keeps them governed."

Torque's capabilities directly address the primary concerns of CIOs and IT leaders. It automates the provisioning and lifecycle management of full NemoClaw environments from version-controlled blueprints, ensuring consistency and compliance. When an agent's task is complete, the environment can be automatically torn down to conserve expensive GPU resources.

Critically, the platform enforces strict multi-tenancy, creating hard isolation between teams to ensure that one team's AI agent cannot interfere with or access data from another. This extends to financial governance, with Torque providing the ability to attribute GPU costs directly to the specific team or project using the environment, a crucial feature for managing the notoriously high expense of AI infrastructure.

Empowering Developers Without Sacrificing Control

While IT leaders gain control, the integration promises to accelerate innovation by empowering developers and data scientists. Instead of filing tickets and waiting weeks for infrastructure to be provisioned, technical teams can use Torque's self-service portal to spin up governed NemoClaw environments on demand.

This self-service model is not a free-for-all. It operates within the policy-driven guardrails established by IT and security teams. A developer can request an AI agent environment, but the type of GPU, the network access policies, and the data permissions are all pre-defined and enforced by the platform. This approach eliminates the traditional bottleneck of IT fulfillment without creating what the company calls "infrastructure chaos."

By providing developers with immediate, secure access to the tools they need, organizations can dramatically shorten the development cycle for new AI-powered applications and services. This balance of agility and control is essential for enterprises looking to leverage the power of autonomous agents without exposing themselves to unacceptable levels of risk or unpredictable costs.

A Strategic Play in the NVIDIA Ecosystem

This announcement is more than a simple product integration; it represents a key piece of NVIDIA's broader strategy to build a comprehensive, end-to-end AI ecosystem. NVIDIA provides the foundational hardware (GPUs) and software (CUDA, NemoClaw), but it relies on a robust network of partners like Quali to provide the specialized management and operational layers required by large enterprises.

Quali's support for NemoClaw extends its already deep integration with the NVIDIA ecosystem, which includes support for DGX Spark, DGX Station, and the NVIDIA GPU Operator. This long-standing relationship ensures that the Torque platform is finely tuned to the nuances of managing NVIDIA's high-performance infrastructure.

By combining NVIDIA's secure agent runtime with Quali's enterprise control plane, the two companies are offering a complete stack that takes autonomous AI from a promising concept to a practical, scalable, and governable reality for businesses. This allows organizations to confidently deploy agents for complex tasks, knowing that a framework is in place to manage their lifecycle, enforce security policies, and control the associated costs.