Samsung Electronics announced on Monday at a seminar on AI applications in industrial R&D hosted by the Korea Industry Association Forum that it plans to fully introduce AI technology, sharply compressing the R&D cycle for new semiconductor materials from the traditional five years to two years—cutting the timeline by three years.
Choi Young-sang, head of the Samsung Advanced Institute of Technology, said at the event that with deep AI-powered support, the company can not only screen more than five times as many promising candidate materials, but also reduce the number of experiments needed to find optimal process conditions by more than 80%.
The R&D of semiconductor materials has long faced high time and trial-and-error costs. Take multilayer thin-film structures as an example: a seven-layer structure alone involves more than 100 million material combinations, yet the traditional model, limited by researchers' experience and intuition, can conduct only about 50 experiments per month, leaving much failed experimental data underused.
The introduction of AI technology breaks this bottleneck, accelerating material screening and process optimization through deep learning and efficient simulation of massive data sets. Samsung also plans to further extend this AI R&D paradigm to other industrial material fields such as batteries, continuously lowering overall development costs. This strategic shift marks that semiconductor material R&D is accelerating into a new stage of intelligence and high efficiency.