Schneider Electric, NVIDIA and HPE have announced a partnership to bring physical AI to process industries, starting with a concrete product: industrial AI agents that run at the edge, inside the plant itself, on HPE Private Cloud AI and NVIDIA’s open Nemotron models. The first three agents monitor operational technology (OT), cybersecurity and energy consumption, and are managed through Schneider Electric’s EcoStruxure Industrial Insights applications.
Schneider, NVIDIA and HPE’s industrial AI agents in 20 seconds
- Partnership to apply physical AI in process industries.
- Combines HPE Private Cloud AI, NVIDIA Nemotron and Schneider’s OT expertise.
- The first agents cover OT monitoring, cybersecurity and energy optimization.
- They run on-site to keep data local and control costs.
The partnership was announced on October 22, 2026, as reported by Robotics & Automation News, and Schneider Electric has now shared further details. The approach starts from an idea the three companies keep repeating: much of the conversation around physical AI focuses on robots and autonomous machines, but a refinery, a paper mill or a chemical plant face a different challenge, running continuous processes where reliability, safety, cybersecurity and energy efficiency are critical.
Three agents for the plant floor
According to Schneider Electric, each agent follows the same cycle: it combines data from multiple sources, detects a problem or an opportunity, recommends or carries out an action, and then checks that the outcome matches what was expected in the physical process.
- OT monitoring: cross-references real-time operational data, equipment configuration, process topology, alarms, maintenance history and engineering context to catch emerging issues and assess their effect on operations.
- Cybersecurity: correlates network activity, control-system behavior and operational events to detect anomalous conditions, proposes or applies response measures, and validates that systems return to a safe state.
- Energy optimization: analyzes process data, consumption patterns, equipment performance and production requirements to find adjustments and measure their effect on efficiency and sustainability.
The companies have not disclosed pricing or availability dates, and present these agents as the first demonstration of the partnership. The longer-term goal, according to the information shared, is more autonomous industrial operations, though with critical decisions left in human hands. Matias Jory, head of digitalization and processes at Chilean paper company CMPC, says Schneider’s OT monitoring system (OTMS) has given him greater visibility into the state of his control systems for three years, and sees the agents as a way for specialists to investigate problems sooner.
Why the edge instead of the public cloud
The infrastructure piece is HPE Private Cloud AI, the private AI cloud platform that HPE develops with NVIDIA as part of its HPE AI Factory offering. cloudnews.tech already reported on HPE and NVIDIA opening Private Cloud AI to agents with MCP. Gwenaelle Huet, Schneider Electric’s executive vice president of Industrial Automation, sums up the case for running the agents inside the plant in three reasons:
- Data sovereignty: operational data is often too sensitive, valuable or regulated to move outside the facility.
- Cybersecurity: processing data close to its source reduces the attack surface on critical infrastructure.
- Cost: cloud inference, data transfer and token consumption all grow with usage, and running AI on-site gives more control over where and how workloads are processed.
These are the companies’ own arguments and worth reading as such: deploying AI at the edge also means buying, operating and updating hardware at every plant. Alexia Clements, HPE’s senior vice president of Hybrid Cloud Solutions Sales, argues that industrial customers see concrete opportunities in AI but need assurances that their data and critical systems will stay protected.
On NVIDIA’s side, Sri Kodali, global head of Developer Relations for Energy and Utilities, notes that industry generates enormous amounts of data and that turning it into action remains one of its biggest challenges. Schneider Electric, for its part, already works with other chipmakers on AI infrastructure, as shown by the reference design it unveiled with AMD for AI factories built on Helios.
Frequently asked questions
What are edge industrial AI agents?
AI agents that run inside the plant, on HPE Private Cloud AI and NVIDIA Nemotron models, to monitor operational technology, cybersecurity and energy consumption.
What does each company bring?
Schneider Electric contributes its operational-technology expertise and EcoStruxure applications; HPE provides the Private Cloud AI platform; and NVIDIA supplies its open Nemotron models and AI technology.
Why don’t they run in the public cloud?
According to the companies, to keep sensitive data on-site, reduce the attack surface, and control inference and data-transfer costs.
When will they be available?
The companies have not given dates or pricing; they present these three agents as the first demonstration of their partnership.
Sources:
