In a groundbreaking move, chip manufacturer Nvidia and the world’s largest contract producer, Foxconn, have joined forces to initiate the global construction of AI factories, as revealed in their joint announcement on Tuesday.

AI factories, as described by Nvidia founder and CEO Jensen Huang and Foxconn Chairman and CEO Young Liu during a fireside chat at Hon Hai Tech Day in Taipei, are specialized data centers designed explicitly for the processing, refining, and transformation of extensive datasets into valuable AI models and tokens.

Huang emphasized the emergence of a novel manufacturing paradigm—intelligence production—and highlighted that these AI factories’ data center infrastructure would integrate Nvidia’s advanced computing platform, featuring the latest GH200 Grace Hopper Superchip and Nvidia’s AI enterprise software.

Foxconn is poised to manufacture a significant number of systems based on Nvidia’s hardware, software, and networking stack, catering to its global clientele seeking to establish and operate their own AI factories.

“With these systems, Foxconn customers can harness Nvidia accelerated computing to provide generative AI services and utilize simulation for expediting the training of autonomous machines, including industrial robots and self-driving cars,” the joint statement from the companies noted.

Foxconn is also contemplating the integration of Nvidia’s stack into its manufacturing facilities, foreseeing the potential benefits. “An AI factory with the Nvidia stack can empower Foxconn to conduct AI training and inference, optimize factory workflows, and conduct simulations in the virtual realm before real-world deployment,” the statement explained. The comprehensive simulation of the entire robotics and automation pipeline provides Foxconn with a pathway to operational efficiency gains, resulting in time and cost savings.

Currently, Nvidia holds a dominant position in the AI chip market, driven by the escalating demand for generative AI.