Power Transmission

Schneider Electric, NVIDIA and HPE Bundle Edge AI Agents for Process Industries

Tomas BerglundOctober 25, 20265 min read18 views

At a glance

  • The three partners combine HPE Private Cloud AI, NVIDIA Nemotron models and Schneider Electric's operational technology software into edge-deployed agents for OT monitoring, cybersecurity and energy optimization.
  • Schneider Electric has launched a collaboration with NVIDIA and HPE to bring Physical AI into process and heavy industries, starting with three Industrial AI Agents that run at the industrial edge on HPE Private Cloud AI. The partners say the edge approach addresses data sovereignty, cybersecurity and AI cost barriers, building on three years of OTMS monitoring at CMPC Chile.
A cutaway industrial plant with a compact edge-computing rack on the floor, linked by glowing filaments to processing nodes inside pipelines and control cabinets.

Schneider Electric has announced a strategic collaboration with NVIDIA and HPE intended to accelerate the adoption of Physical AI across process and heavy industries. The first concrete result is a set of Industrial AI Agents designed to run at the industrial edge, closest to where operational decisions are made.

The companies argue that while much of today's Physical AI discussion centers on robotics, humanoids and autonomous machines, continuous process operations pose a different problem. Operators must keep complex, round-the-clock processes running while meeting demands for reliability, safety, cybersecurity and energy efficiency. Until now, the partners say, there has been no clear blueprint for applying Physical AI in those environments.

What the offering combines

The solution pairs HPE Private Cloud AI — co-engineered with NVIDIA as part of the HPE AI Factory with NVIDIA portfolio — with Schneider Electric's operational technology expertise. NVIDIA's Nemotron open AI models form part of the stack. The stated aim is to let organizations deploy and scale AI applications at the industrial edge, keeping real-time intelligence near critical operations while retaining control over sensitive data, security and spending.

Schneider Electric's EcoStruxure Industrial Insights applications provide one environment through which customers can deploy, manage and govern multiple agents, simplifying lifecycle management and security oversight. The first agents concentrate on operational technology monitoring, cybersecurity and energy optimization.

Three barriers the edge approach targets

  • Data sovereignty. Operational data is often too valuable, sensitive or regulated to leave operational boundaries. Running AI close to the plant is presented as a way to preserve control, security and compliance.
  • Cybersecurity. Greater connectivity widens exposure across industrial and critical infrastructure sites. Processing data near its source is intended to reduce the attack surface.
  • AI consumption costs. Cloud inference, data transfer and token-related charges can climb with usage, making site-by-site scaling expensive. Edge deployment is offered as a way to curb some of those costs and to govern where workloads run.

The first three agents

OT Monitoring Agents draw on real-time operational data, equipment configurations, process topology, alarms, maintenance history and engineering context to interpret the physical state of operations. They are described as identifying emerging issues, reasoning about likely operational impact, recommending or initiating corrective actions and verifying that an intervention produced the expected physical outcome.

Cybersecurity Agents correlate network activity, control system behavior, asset configurations, operational events and process context to detect abnormal conditions, assess potential effects on operations, recommend or execute responses and validate that systems have returned to a secure, stable state.

Energy Optimization Agents combine process data, energy consumption patterns, equipment performance, operating conditions and production requirements to find optimization opportunities, propose or implement adjustments and verify the resulting effect on energy efficiency, operational performance and sustainability outcomes.

Proven base at CMPC Chile

The groundwork is already running at CMPC Chile, part of Empresas CMPC S.A., one of Latin America's largest pulp, paper and packaging producers. For three years the company has used Schneider Electric's OTMS in production to monitor the health and availability of control systems across continuous pulp and paper operations. OTMS gives teams a unified view of control systems, servers, workstations and switches, helping specialists identify emerging issues earlier, diagnose problems faster and set clearer maintenance priorities.

Matías Jory, digital and process manager in CMPC's industrial operations vice-presidency, said the platform has improved visibility into control system health and that early identification of issues is critical in continuous operations. He sees potential for agents to help specialists investigate faster and move from insight to action while people stay in control of critical decisions.

Company positions

Gwenaelle Huet, executive vice president for industrial automation at Schneider Electric, framed the next wave of industrial innovation as intelligence embedded directly into operations rather than confined to autonomous machines. She said the three partners are pooling expertise in automation, AI and edge computing to help customers turn operational data into action, improving resilience, efficiency, cybersecurity and sustainability.

Sri Kodali, NVIDIA's global head of Developer Relations for Energy and Utilities, said industrial operations generate vast amounts of data but converting it into action remains one of the sector's biggest difficulties, and that combining Nemotron models and HPE Private Cloud AI with Schneider Electric's OTMS platform lets agents monitor, reason and optimize in real time.

Alexia Clements, SVP of Hybrid Cloud Solution Sales at HPE, said industrial organizations want to expand AI use without putting sensitive data or critical systems at risk, and that bringing AI closer to where data is generated and decisions are made supports that goal.

The partners present the collaboration as a foundation for progressively more autonomous industrial operations, with systems that continuously monitor, reason about and optimize performance. That remains a stated ambition rather than a delivered capability; the current release covers the first three agent types.

Reported by Tomas Berglund, Distribution and protection analyst.