Training robots Foxconn plans to fully digitize its factories

From Hendrik Härter | Translated by AI 2 min Reading Time

Related Vendors

At Foxconn, entire production lines are set to go fully digital, with robots being trained entirely in virtual environments. The contract manufacturer is partnering with Nvidia to achieve this.

Foxconn plans a complete factory in Mexico with a digital twin. All details right down to the robots are created and planned entirely virtually.(Image: Nvidia)
Foxconn plans a complete factory in Mexico with a digital twin. All details right down to the robots are created and planned entirely virtually.
(Image: Nvidia)

Contract manufacturers like Foxconn are driven by fast-paced electronics they have to manufacture in their factories. The company operates more than 170 factories worldwide. The newest factory is being built in Guadalajara, Mexico. What's special: it is a virtual factory. There, Nvidia wants to have its Blackwell HGX built.

But before the physical factory can be created, planners and designers need to optimally plan the robots and technical facilities. Only then does the plant operator have all the important details in view. The digital twin of the factory makes it possible to simulate and test everything exactly before the tons-heavy robots can start their work.

Digital twins help companies to efficiently estimate potential costs. "Our digital twin will lead us to new levels of automation and industrial efficiency, saving time, cost, and energy," said Young Liu, Chairman of the Board of Foxconn. The company generated revenue of almost 200 billion US dollars last year.

Collaboration with Siemens

For the digital twin, Foxconn is working with Siemens. The digital twin is powered by Siemens Xcelerator software, including Teamcenter, and Omniverse by Nvidia. The integration of Siemens' Xcelerator applications with the Omniverse cloud enables a physically accurate and visually realistic digital twin.

The developers and designers design and simulate the robots and production lines, integrating 3D CAD elements into a comprehensive virtual factory. The robots are trained with Nvidia Isaac Sim.

Foxconn is among the pioneers when it comes to training robots. Robots from manufacturers like Epson learn to see, grip, and move objects. For example, robots can learn to pick up one of the future Blackwell servers and place it on an autonomous mobile robot (AMR).

Video: Nvidia and Foxconn plan manufacturing

Navigate autonomously through the factory

Foxconn's AMRs come from the Taiwanese company FARobot. They learn to perceive their surroundings and navigate in the factory hall. To do this, Nvidia's Perceptor is used. The software helps to create a real-time 3D map with all obstacles. The routes of the robots are generated and optimized by Nvidia cuOpt. This is a microservice for route optimization. Unlike many transport robots that must adhere to carefully drawn lines in the factory hall, the intelligent AMRs can navigate around obstacles to reach their destination.

The Foxconn factory in Mexico is just the beginning. In the future, the contract manufacturer wants to design its factories worldwide with the help of digital twins. One in Taiwan, where electric buses are supposed to be built. To get even more insights into the future factory, Foxconn is working with Metropolis from Nvidia. This is an application framework for connected cities and spaces. The aim is to integrate cameras into the production process. The cameras are then connected to AI algorithms.

But not only Foxconn relies on the digital technology from Nvidia. Other global electronics manufacturers such as Delta Electronics, MediaTek, MSI and Pegatron also want to build their production lines with Nivida's technology. The goal is to make the factories more flexible and autonomous. (heh)

Subscribe to the newsletter now

Don't Miss out on Our Best Content

By clicking on „Subscribe to Newsletter“ I agree to the processing and use of my data according to the consent form (please expand for details) and accept the Terms of Use. For more information, please see our Privacy Policy. The consent declaration relates, among other things, to the sending of editorial newsletters by email and to data matching for marketing purposes with selected advertising partners (e.g., LinkedIn, Google, Meta)

Unfold for details of your consent