Schneider Electric, Nvidia and HPE launch AI agents for industrial operations

The rise of Physical AI

Schneider Electric, Nvidia, and HPE have joined forces to launch a new suite of AI agents designed specifically for industrial environments. Unlike the generative AI models used for writing text or creating images, these agents represent what the companies call "Physical AI." This technology is built to interact directly with the physical systems that power our world, such as chemical plants, energy grids, and paper mills.

The goal is to move beyond simple automation toward systems that can understand and react to the physical world. Instead of just following a set of pre-programmed instructions, these agents are designed to monitor complex operations and provide intelligent recommendations. As reported by Robotics & Automation News, the collaboration combines expertise in industrial automation, artificial intelligence, and edge computing.

The companies are focusing on industries where AI is more likely to interact with existing control systems rather than taking the form of a mobile robot. By embedding intelligence directly into the infrastructure of factories, the partners hope to help manufacturers turn raw operational data into immediate, actionable insights.

Local processing and security

One of the primary hurdles for adopting AI in heavy industry is data security. Many manufacturers are hesitant to send sensitive information about their equipment configurations or production volumes to a public cloud. To solve this, the new AI agents run locally using HPE Private Cloud AI. This means the data stays on-site at the factory or plant, which helps companies address concerns regarding cybersecurity and the high costs associated with moving large amounts of data across the internet.

By keeping the processing at the "edge"—the physical location where the work happens—the system can provide real-time responses. Schneider Electric notes that this local approach is essential for maintaining the integrity of industrial operations. It allows for a more secure environment where the AI can learn from the specific nuances of a single facility without exposing that data to outside parties.

The agents are designed to be managed through Schneider Electric's EcoStruxure Industrial Insights applications. This provides a centralized way for human operators to oversee the AI's work, ensuring that while the machines are becoming smarter, people remain in control of the most critical decisions.

How the agents function

The first wave of Industrial AI Agents at the Edge focuses on three specific pillars: operational technology (OT), cybersecurity, and energy consumption. Operational technology refers to the hardware and software that monitors and controls physical devices, such as valves, sensors, and motors. These agents analyze a wide variety of data points, including equipment configurations, historical maintenance records, and real-time alarms.

For example, the OT monitoring agents look for patterns that might indicate an emerging problem before a total breakdown occurs. They can then recommend specific corrective actions to maintenance teams. The cybersecurity agents perform a different task: they scan network activity and equipment configurations to spot abnormal behaviors that might suggest a breach or a system malfunction.

The energy optimization agents take a different approach by examining production and energy data. They look for inefficiencies in how a plant is running and identify specific opportunities to reduce power consumption. By combining these three functions, the companies aim to improve the resilience, efficiency, and sustainability of industrial sites simultaneously.

Real-world application

The technology is not starting from scratch. The underlying monitoring system is already being utilized by CMPC, a Chilean producer of pulp, paper, and packaging. CMPC has used Schneider Electric’s OT Monitoring System for three years to gain better visibility into the health and availability of their control systems. In large-scale industrial operations, identifying a small issue early is vital because a single failure can cause massive production delays and high costs.

Matías Jory, a digital and process manager at CMPC, noted that the company sees potential for AI agents to build on this existing foundation. The goal is to help specialists investigate issues faster and move more effectively from seeing a problem to fixing it. The long-term vision for this partnership is to create systems that can not only monitor and recommend actions but eventually make adjustments autonomously, while constantly checking to see if those actions produced the desired results.

The path forward

This collaboration marks a significant step toward autonomous industrial operations. By integrating Nvidia's AI capabilities, HPE's cloud infrastructure, and Schneider Electric's automation expertise, the partners are creating a blueprint for how other industries might adopt AI. The focus is on creating a foundation where intelligence is a standard part of the factory floor, rather than an add-on tool.

As these agents become more sophisticated, they will likely expand into more complex tasks. The current focus on monitoring and energy optimization provides a safe entry point for companies to see the value of AI. Once trust is established in these core areas, the technology could eventually manage even more complex sequences of industrial production, leading to a more resilient and sustainable global manufacturing landscape.

Summary of key features

The new AI agents provide a multi-layered approach to industrial management. Here is a breakdown of what they do:

• OT Monitoring: Analyzes alarms and maintenance history to predict equipment failure.
• Cybersecurity: Monitors network behavior to detect anomalies and potential threats.
• Energy Optimization: Scans production data to find ways to lower electricity and fuel use.
• Edge Computing: Runs locally on HPE Private Cloud AI to ensure data privacy and reduce latency.