Rockwell Automation Fair 2025: AI and the Next Great Evolution of Manufacturing
Manufacturing is moving from scripted automation to adaptable autonomy. Software-defined control, embedded AI, robotics, and analytics are converging into systems that learn, adjust, and work with people-on the floor and at the edge.
Key notes
- Automation is moving toward autonomy: AI and intelligent systems augment workers rather than replace them.
- Think in systems: integrate advanced tech with human expertise to simplify complexity and anticipate needs.
- Factory of the future: digital twins, automated material handling, energy management, and workforce upskilling are live in production.
- GenAI at the industrial edge: FactoryTalk Design Studio Copilot with NVIDIA Nemotron Nano SLMs brings fast, local AI assistance to controls workflows.
- Digital trust is non-negotiable: SecureOT delivers end-to-end protection for increasingly autonomous operations.
From Automation to Autonomy
At the event in Chicago, Rockwell Automation's CTO, Cyril Perducat, was clear: the line between human intelligence and machine intelligence is blurring. The goal isn't to remove people from the loop-it's to give them greater autonomy and better decisions, faster.
Programmable systems require new code for every change. Learning systems adjust in place. That's the shift. Think adaptive parameters, self-directed AMRs, and operators who supervise outcomes instead of wrestling with scripts.
A Systems Approach, Not a Patchwork
CEO Blake Moret outlined a simple premise: advanced tech still needs to be easy to use across messy, mixed environments. No plant has a single, uniform stack. The path forward is a holistic system that anticipates needs and makes complex operations simple to run.
The takeaway for leaders: unify data flows, embed AI where decisions are made, and pair it with the know-how of people who understand the process. That's how you get speed without losing control.
The Factory of the Future (Built Inside Rockwell)
Rockwell announced a new advanced automation facility in Wisconsin to model autonomous production. It mirrors the approach used in its Singapore site and doubles as a training and customer demonstration center.
Three practical moves are driving results:
- Digital design: Emulate3D digital twins to lay out and optimize the facility before hardware hits the floor.
- Automated material flow: Lights-out warehouse and AMRs move material to and from lines, eliminating delay and rework.
- Energy management: FactoryTalk Energy Manager with dense instrumentation delivered payback in weeks, not months.
The internal AI program isn't just about throughput. It's increasing engagement because teams help build the solutions-and see wins in their day-to-day work.
GenAI at the Industrial Edge
FactoryTalk Design Studio Copilot is moving from cloud-only to edge-capable with NVIDIA's Nemotron Nano small language models. This brings natural language assistance and code-aware guidance directly into industrial environments.
By using the open-source Nemotron-Nano-9B-v2 with NVIDIA NeMo-and fine-tuning it on FactoryTalk data-Rockwell is enabling local GenAI that runs with less space and energy than a traditional data center. The intent: speed up design and maintenance while preserving predictability and control.
Digital Trust and Cybersecurity
As plants connect more assets and push intelligence to the edge, the threat surface grows. That's why Rockwell introduced SecureOT: a combined platform, professional services, and managed services suite for industrial cybersecurity.
The SecureOT Platform delivers real-time asset visibility, risk prioritization, and vulnerability management. Services cover advisory, assessments, and implementation to strengthen defenses across legacy and modern systems.
Rockwell Automation SecureOT overview
The Tech Stack That Makes Autonomy Real
Perducat pointed to three layers that have to work together: software-defined architecture, AI, and robotics. You need flexible software to let AI act quickly; you need AI to turn insights into actions; and you need physical systems that can execute without friction.
Information isn't the bottleneck anymore-closing the loop is. The plants that win will detect, decide, and act as a single motion.
What This Means for Your Team
Operations
- Standardize on a digital thread from design to maintenance. Start with a high-impact line and model it in a twin.
- Automate internal logistics. AMRs and lights-out storage remove chronic delays and variability.
- Instrument for energy. Treat kilowatt-hours as a controllable input, not a utility bill.
IT
- Prepare for edge AI. Validate hardware profiles for SLM inference and define update/rollback policies.
- Segment OT networks, enforce least privilege, and integrate SecureOT-style visibility into your SIEM.
- Create a catalog of approved models, datasets, and prompts with clear data governance.
Developers and Controls Engineers
- Adopt software-defined patterns: versioned configs, CI for PLC logic, and test benches with Emulate3D.
- Use GenAI copilots for drafting code and documentation, then enforce human review gates.
- Design for adaptability: parameters first, code second. Make change the cheap path, not the expensive one.
Leadership
- Set a 12-18 month autonomy roadmap tied to P&L: throughput, changeover time, scrap, and energy per unit.
- Fund a solid MRR cadence for cross-functional reviews. Measure value every 30 days, not once a quarter.
- Invest in upskilling. Pair operators and engineers with AI training that ties directly to plant KPIs.
Quick Start (Next 90 Days)
- Pick one pilot line. Build a digital twin, add energy sensors, and deploy AMRs for material flow.
- Roll out an edge AI proof of concept with a small language model for design or troubleshooting tasks.
- Deploy asset discovery in OT, fix top 10 vulnerabilities, and formalize incident response with OT scope.
- Launch a short, role-based AI skills track for Ops, IT, and Dev teams with clear before/after metrics.
Level Up Your Team's Skills
If you lead Ops, IT, or Dev teams and need focused upskilling tied to real outcomes, browse these role-based programs: AI courses by job.
Final Take
Autonomy isn't a distant idea. It's arriving line by line, model by model, decision by decision. Plants that combine software-defined control, AI, and secure operations-while keeping people at the center-will outpace the rest.
Your membership also unlocks:
