As enterprise AI systems increasingly operate without direct human oversight, infrastructure governance is emerging as a new challenge. Last week, Quali unveiled an expansion of its Torque platform aimed squarely at that problem, introducing what it calls an “Agentic Control Plane” to govern autonomous, GPU-intensive AI workloads.
The move reflects a broader shift in how AI is deployed inside organizations. AI workloads are no longer limited to predefined pipelines that run predictable tasks. Instead, they are evolving into systems that plan, adapt, and reconfigure infrastructure on the fly, often consuming large amounts of compute across cloud, hybrid, and on-premises environments.
Quali’s new control layer is designed to operate in that context. Rather than relying on static automation rules triggered by specific events, the Agentic Control Plane continuously evaluates system behavior and enforces governance policies in real time.
Join The European Business Briefing
New subscribers this quarter are entered into a draw to win a Rolex Submariner. Join 40,000+ founders, investors and executives who read EBM every day.
Subscribe“AI workloads today behave less like scheduled jobs and more like autonomous assistants operating at machine speed,” said Quali CEO Lior Koriat. “Traditional automation wasn’t built to govern systems that are constantly adapting and making decisions on their own.”
In conventional IT environments, automation typically executes predefined workflows. By contrast, agentic AI systems can initiate actions independently, responding to changing data, resource availability, and performance requirements. This autonomy has accelerated deployment but has also complicated oversight around cost, compliance, and operational risk.
Quali says Torque’s expanded capabilities are intended to serve as a unifying governance layer across infrastructure-as-code frameworks, Kubernetes environments, GPU clusters, cloud platforms, and on-premises systems. The platform is designed to coordinate both human-defined automation and machine-initiated actions without replacing existing tooling.
Among the new features are agent-driven operations that allow modular software agents to manage tasks such as environment provisioning, cost modeling, lifecycle management, and compliance checks based on workload intent rather than fixed rules. The platform also applies continuous runtime governance, adapting security, compliance, and financial controls as workloads scale or change configuration.
Industry observers note that governance has become one of the least mature aspects of large-scale AI deployment. As enterprises experiment with autonomous AI systems, the lack of visibility and control across sprawling infrastructure environments has become a growing concern.
Quali positions Torque’s Agentic Control Plane as an attempt to address that gap, enabling organizations to impose continuous oversight without constraining AI systems designed to operate independently.
The new capabilities are available immediately as part of Quali’s Torque platform.
