KUKA puts its new automation platform into production at Ohio auto plant

The company says the platform is already running at KUKA Toledo Production Operations, where the site turns out more than 300 vehicle bodies a day and has logged over 2 million bodies since opening.

KUKA moves AMP from concept to a working plant floor system

KUKA AG said this week that it has successfully deployed its Automation Management Platform, or AMP, at KUKA Toledo Production Operations in Ohio, giving the company a live test bed for how AI can sit on top of existing industrial automation instead of replacing it.

The platform is being used at KTPO, one of North America’s major automotive manufacturing facilities, according to KUKA Systems North America LLC, the wholly owned subsidiary that operates the site. KUKA said the implementation is intended to connect its existing automation infrastructure with next-generation AI-powered tools while production continues.

That matters to plant managers because the challenge in automation is rarely whether the hardware can move parts. It is whether new software can be layered onto a running factory without disturbing throughput, quality, or uptime. KUKA is betting AMP can do that. The company described the platform as an open orchestration layer that combines industrial automation and AI-driven decision-making to improve flexibility, productivity, and the value extracted from production data.

Marc Fleischmann, chief software and AI officer at KUKA Group, said the launch marks a shift in how manufacturers should think about automation. “Traditional automation is expanding and transforming, and the future lies in the combination of precision and reliability of proven industrial automation with the capabilities of AI-powered systems,” he said.

KTPO gives KUKA a real production benchmark

KTPO is not a greenfield pilot line. KUKA said the 335,000-square-foot factory produces more than 300 vehicle bodies per day, or about one every two minutes. Since opening, the plant has produced more than 2 million vehicle bodies and generated a large volume of operational data.

That gives AMP a high-pressure environment in which to prove itself. The facility has produced the body-in-white for every Jeep Wrangler sold worldwide since 2006 and for the Jeep Gladiator since 2019. KUKA said the site relies on expertise in welding, joining technologies, and systems integration, and that the production environment includes 285 robots and more than 60,000 connected devices.

For operators, those numbers point to the kind of integration challenge that often stalls automation projects: lots of assets, lots of data, and lots of opportunity for system drift. KUKA said AMP is meant to work as an open, modular API platform that can connect mobile robots, machines, software, and data sources without forcing a wholesale replacement of the current plant.

That is the core ROI argument KUKA is making. Rather than rip out installed equipment, the company wants to use the data already being generated to improve coordination, reduce changeover times, and increase operational flexibility. If AMP can help the plant do more with the equipment already on the floor, the value case improves fast; if it cannot, the platform becomes just another software layer competing for attention.

KUKA is pitching AMP as a way to unlock data trapped in silos

KUKA said factories generate large amounts of data, but much of it remains isolated in separate systems. AMP is designed to turn that information into actionable insights so automation can better understand production processes, coordinate activities across workflows, and improve performance over time.

That closed-loop idea is central to the company’s pitch. Fleischmann said AMP creates “a shared context layer” that lets customers integrate their own AI models and generalize them in production. By continuously collecting and analyzing operational data, KUKA said the platform can improve decision-making and help organizations scale intelligent automation.

The company also framed the platform as a foundation for what it called physical AI. In KUKA’s view, physical AI means machines and robots that learn from operational data and make more informed decisions in real time. The platform initially focuses on autonomous mobile robots, or AMRs, and was announced at NVIDIA GTC 2026.

For supply-chain leaders, the appeal is obvious: better orchestration across equipment, software, and material flow could reduce delays and improve response to demand swings. But the practical test will be whether the platform can translate data into measurable operational gains, not just dashboards and promises.

The business case depends on measurable gains, not software language

KUKA said the goal is to help manufacturers drive higher efficiency through more comprehensive automation without replacing existing assets. That phrasing is important. Most plants already carry significant sunk cost in robots, controls, conveyors, and integration work. New automation software has to beat the economics of the current system, not merely coexist with it.

The best-case business case for AMP would show up in a few places plant teams already track: throughput stability, fewer changeovers, lower downtime, and better use of existing assets. KUKA said the platform is intended to increase operational flexibility and reduce changeover times while unlocking more value from factory data. It did not disclose financial terms, implementation costs, or a payback period for the KTPO deployment.

That leaves the usual caution for automation buyers: the ROI is likely to depend on whether the platform can reduce manual coordination work and improve response time across multiple systems. In a plant producing more than 300 bodies a day, even small gains in uptime or changeover efficiency can matter. But software that touches live production also introduces its own risks, including integration complexity, data quality issues, and the possibility that a new orchestration layer becomes a bottleneck if it is not tuned correctly.

KUKA’s response is to start in a familiar environment. Fleischmann said testing the alpha version at KTPO is an important step because the site is well known to the company, allowing it to evaluate the platform under realistic conditions and improve it faster. That suggests the deployment is as much about validation as it is about immediate production payoff.

A test of how far AI can go inside installed automation

KUKA said the U.S. is a significant market for the company and that it has more than 3,400 employees in engineering, manufacturing, software development, systems integration, and customer support functions across the country. It also said AMP is being developed in the United States.

That local footprint may matter to customers who want support close to the plant and who prefer vendors that can assist with system engineering, simulation, spare parts, and robot refurbishment. KUKA said it provides those services through my.KUKA.com.

For now, the most useful signal from the KTPO deployment is not the AI label. It is that KUKA is trying to put software on top of a production system that already runs at automotive scale. The company says AMP is designed to augment the factory instead of replacing it, and the live Ohio deployment will show whether that promise survives contact with real throughput, real changeovers, and real production data.

If KUKA can demonstrate that AMP improves coordination across a plant already filled with robots and connected devices, it could give manufacturers a more practical route to AI adoption than the usual pilot-project cycle. If not, the platform will join a long list of automation concepts that looked easier on slides than on the plant floor.