'Tell me about body defects': Hyundai Motor, Kia accelerate push toward self-evolving factories

E-FOREST Tech Day 2026

Over 200 new manufacturing technologies unveiled

Manufacturing AI 'Polaris' deployed at Ulsan EV factory

Defect analysis to corrective action, end to end

Virtual commissioning developed with Nvidia using physical AI

Autonomous manufacturing system upgrade at Singapore by 2030

Full humanoid robot deployment planned from 2032

An automated door garnish installation system unveiled by Hyundai Motor and Kia at their Uiwang research center in Gyeonggi Province on Friday. An industrial robot automatically installs components on the lower section of a vehicle door and inspects the assembly. [Hyundai Motor and Kia]

"Show me the top three body quality defects from July, and tell me the cause and corrective action for the No. 1 defect."

The moment the question was spoken, an AI system began sorting the major quality defects recorded at an automotive production factory by type. Within seconds, a screen displayed the defect categories and their occurrence counts — and the AI went further, recommending that "mold contamination management be strengthened and regular inspections conducted to prevent recurrence." No manual search through complex production data was needed; a single spoken query was enough to grasp the factory's quality status and identify a path to improvement.

The demonstration took place Friday at the E-FOREST Tech Day 2026 media day at Hyundai Motor and Kia's Uiwang research center in Gyeonggi Province. The system on display was Polaris, a manufacturing-specialized AI agent developed by Hyundai Motor and Kia. The demo drew on live data from the company's new Ulsan electric vehicle factory, which began production of the GV90 — its first mass-production model — this month, illustrating how AI is reshaping the automotive production floor.

An automated body weatherstrip installation system demonstrated by Hyundai Motor and Kia at their Uiwang research center in Gyeonggi Province on Friday. An industrial robot automatically installs the rubber weatherstrip that runs along the edge of a vehicle door. [Hyundai Motor and Kia]

Hyundai Motor and Kia will hold E-FOREST Tech Day from Monday through Wednesday at the Uiwang research center. Now in its seventh year, the event will showcase more than 200 new manufacturing technologies spanning software-defined factories, automation, energy solutions and future mobility. The central theme is a shift beyond simply adding more robots to production lines — toward factories that analyze data, make decisions and optimize their own processes autonomously.

Choi Young-tae, executive director and head of the E-FOREST Center at Hyundai Motor and Kia's Manufacturing Solutions Division, said the factory of the future is "not simply a factory with a high level of automation," but rather "a manufacturing environment where people and robots collaborate, and where data and AI continuously learn and optimize the factory on their own."

The foundation for that transformation is software-defined factory technology that connects data previously siloed across individual plants. Under the old model, each factory developed its own separate systems, leading to duplicated investment in similar functions and making data sharing difficult. Hyundai Motor and Kia are now consolidating those systems into a common manufacturing platform, building a framework in which software covering quality, equipment maintenance and production operations can be shared across multiple factories.

An automated robot path generation system unveiled by Hyundai Motor and Kia at their Uiwang research center in Gyeonggi Province on Friday. AI generates movement paths for body welding robots and automatically corrects positional errors that arise during actual operations. [Hyundai Motor and Kia]

Virtual commissioning with Nvidia; AI inspects vehicles using 5,000 images

The company's collaboration with Nvidia is also taking concrete shape on the manufacturing side. At the event, Hyundai Motor and Kia introduced a physical AI-based "virtual commissioning" technology being co-developed with Nvidia — a system that verifies the movements of robots and equipment in a virtual environment before actual production machinery is installed.

Conventional digital simulations can render robot movements on screen but have struggled to fully account for real-world physical phenomena such as gravity, friction and collision. As a result, even after virtual verification was complete, engineers still had to re-adjust robot behavior on the actual production floor.

A Hyundai Motor and Kia researcher demonstrates digital twin-based robot learning technology while wearing a virtual reality headset at the Uiwang research center in Gyeonggi Province on Friday. [Hyundai Motor and Kia]

Hyundai Motor and Kia are addressing that limitation using Nvidia's Isaac Sim and Omniverse-based physics simulation technology. The approach verifies not only robot movements but also equipment and control programs together in a virtual environment, minimizing the gap between the virtual and real factory.

A company official said the technology "is the first concrete example of our collaboration with Nvidia," adding that the two companies are developing technology that combines Hyundai Motor's manufacturing floor data with Nvidia's physical AI capabilities "to predict and verify in a virtual environment the problems that could occur in an actual factory."

The companies also unveiled a "Virtual R&D Lab" — a digital twin of an actual research and development space — where 12 virtual laboratories allow engineers to pre-verify a range of technologies, from robot movement paths and wireless communication environments to painting processes. Hyundai Motor and Kia plan to extend the technologies validated in the virtual space to the preparation process for actual new-vehicle mass production.

A tunnel-type finished-vehicle exterior inspection system unveiled by Hyundai Motor and Kia at their Uiwang research center in Gyeonggi Province on Friday. As a Kia PV5 enters the inspection equipment, cameras capture about 5,000 ultra-high-resolution images and AI analyzes the exterior for defects. [Hyundai Motor and Kia]

AI is also being applied to vehicle quality inspection. A tunnel-type finished-vehicle exterior inspection system demonstrated at the event works by having cameras photograph a vehicle's exterior as it passes through a tunnel-shaped structure, with AI then scanning the images for defects. Each inspection involves capturing and analyzing about 5,000 ultra-high-resolution images to check for scratches and other exterior abnormalities, with the location and size of any defect displayed on screen.

Lee Dong-hyun, a senior researcher on the Vision Task Force at the E-FOREST Center, said more than 95 percent of the vision systems deployed at domestic and overseas factories currently rely on external solutions — a dependency the in-house development is designed to reduce, cutting costs that arise each time a new model is added or a design is changed while building the company's own technological competitiveness.

Hyundai Motor and Kia developed the optics, inspection algorithms and software in-house and plan to deploy the technology at the Kia Hwaseong EVO Plant West factory.

A standardized logistics robot platform unveiled by Hyundai Motor and Kia at their Uiwang research center in Gyeonggi Province on Friday. By integrating proprietary driving software and hardware, the platform cuts manufacturing costs by about 30 percent and reduces charging time to as little as 12 minutes. [Hyundai Motor and Kia]

AI generates welding robot paths; humanoid robots to be deployed from 2032

Robots on the production line are also getting smarter. A standout example is technology that uses AI to automatically generate movement paths for welding robots. A car body contains more than 1,700 weld points, and until now workers had to manually program the path each robot takes to reach them.

Hyundai Motor and Kia are developing technology that automatically generates robot movement paths using AI and uses sensors to correct positional discrepancies between actual equipment and the virtual model. The key target is guiding the robot's end effector to each weld point with an error margin of no more than 1 millimeter. Development is targeted for completion in June 2027.

The event also featured a robot technology that mimics human hand movements. An industrial robot fitted with finger-shaped grippers was demonstrated picking up and detaching wires attached to electric vehicle chassis components — a technology capable of recognizing and grasping parts of varying shapes that is currently in early development. The aim is to extend the reach of robotics into precise assembly tasks that conventional industrial robots have struggled to automate.

Changes are also coming to repetitive logistics work. A standardized logistics robot platform developed in-house integrates hardware and driving software to cut manufacturing costs by about 30 percent. Charging performance has also been improved, reducing what previously took about two hours down to as little as 12 minutes. The technology is slated for deployment at the Hwaseong EVO West factory and the Ulsan factory.

Choi Young-tae, executive director and head of the E-FOREST Center at Hyundai Motor and Kia's Manufacturing Solutions Division, presents Hyundai Motor Group's future factory vision at the E-FOREST Tech Day 2026 media day held at the Uiwang research center in Gyeonggi Province on Friday. [Hyundai Motor and Kia]

The ultimate goal for Hyundai Motor and Kia is to connect these disparate automation technologies into a single integrated manufacturing system. Hyundai Motor Group is using its Singapore Global Innovation Center, known as HMGICS, as a proving ground for next-generation manufacturing technologies, with plans to roll out what is developed there in stages to global production sites including Hyundai Motor Group Metaplant America in Georgia and the new Ulsan EV factory.

By 2030, the group aims to advance a virtual environment-based autonomous manufacturing system at HMGICS, and from 2032 onward plans to deploy humanoid robots and full-scale autonomous manufacturing technology at new factories.

"The purpose of automation is not to reduce the number of people," Choi said, "but to create a manufacturing environment where people can focus on safer and more important work."

kwater@heraldcorp.com