Schneider Electric, NVIDIA and HPE launch edge AI agents for industrial monitoring

Bringing AI to the factory floor

Schneider Electric, NVIDIA, and HPE have formed a strategic partnership to deploy Physical AI across heavy and process industries. Unlike the common focus on humanoid robots, this initiative targets the complex, continuous systems that power global infrastructure. The goal is to embed intelligence directly into the environments where industrial decisions are made.

The companies are introducing Industrial AI Agents at the Edge to serve as a blueprint for this transition. These autonomous agents are designed to monitor, reason, and optimize operations in real time. By operating at the edge—the physical location where data is generated—the system provides immediate responses to industrial needs.

This collaboration combines Schneider Electric’s expertise in operational technology with NVIDIA’s hardware and software capabilities. It also leverages HPE Private Cloud AI to create a secure foundation for these agents. The partnership aims to transform raw operational data into actionable intelligence for large-scale industrial sites.

Overcoming barriers to adoption

Many industrial organizations have hesitated to adopt AI due to three primary hurdles: data sovereignty, cybersecurity, and cost. Critical operational data is often too sensitive or regulated to be moved to a public cloud. By keeping the AI at the edge, the solution allows companies to maintain strict control over their information.

Security is another major concern because increased connectivity can expand the attack surface for critical infrastructure. The HPE Private Cloud AI component helps reduce this risk by processing data close to its source. This setup ensures that sensitive systems remain protected from external threats while still benefiting from AI analysis.

Finally, the solution addresses the rising costs associated with cloud inference and data transfer. As companies try to scale AI across multiple sites and assets, the price of moving tokens and data can become prohibitive. Deploying AI locally helps manage these costs while giving operators more control over where and how their workloads are executed.

How the agents function

The first generation of Industrial AI Agents focuses on three specific pillars: OT monitoring, cybersecurity, and energy optimization. OT monitoring agents combine real-time data, equipment configurations, and maintenance history to understand the physical state of a factory. They can identify emerging issues and recommend corrective actions based on the engineering context.

Cybersecurity agents work by correlating network activity with control system behavior and asset configurations. They look for abnormal conditions that could indicate a breach or a malfunction. These agents can assess the potential impact on operations and execute response measures to return systems to a stable state.

Energy optimization agents integrate production requirements with energy consumption patterns and equipment performance. They identify specific opportunities to improve efficiency and suggest adjustments to the process. These agents verify the resulting impact on sustainability outcomes and operational performance to ensure the changes are effective.

Real-world testing and scale

The foundation for this technology has already been tested at CMPC Chile, a major producer of pulp, paper, and packaging. For the past three years, the company has used Schneider Electric’s OTMS to monitor the health of control systems across its operations. This system provides a unified view of servers, workstations, and switches to help specialists identify issues early.

The success of the OTMS deployment provides a proven base for the new AI agents. CMPC Chile plans to use this foundation to accelerate investigation and decision-making for their specialists. The goal is to move from simple monitoring to a system that helps experts move more effectively from insight to action.

Customers can manage these various agents through Schneider Electric’s EcoStruxure Industrial Insights apps. This single environment simplifies the lifecycle management and security oversight for the entire fleet of agents. It allows operators to maintain full control of their critical infrastructure while scaling the use of AI across their facilities.

The path to autonomy

This collaboration marks a shift toward what the companies call 'automating the automation.' By embedding intelligence into industrial systems, the goal is to create safer and more efficient operations. Over time, these capabilities are expected to lead to increasingly autonomous operations across the world's most critical industries.

The solution uses a mix of technologies, including NVIDIA Nemotron open AI models and the HPE AI Factory. By combining these with Schneider Electric's industrial expertise, the partners aim to provide a turnkey solution for process industries. This approach helps turn vast amounts of data into specific, real-time actions.

As the technology matures, it will likely expand beyond the initial three focus areas. The current framework provides the security and privacy required for industrial organizations to trust AI with their most sensitive systems. It lays the groundwork for a future where industrial systems can continuously reason about their own performance and optimize themselves.