Samsung Electronics is collaborating with NVIDIA to build what the Korean tech giant is calling an “AI Megafactory”. If the name doesn’t already make it obvious, this facility will focus on AI-driven production.
Powered by more than 50,000 NVIDIA GPUs, this AI Factory will integrate every element of semiconductor manufacturing into a single network. This includes aspects like design and process, as well as equipment, operations and quality control. AI will oversee the production flow by continuously analysing, predicting, and optimising environments in real time.
This initiative builds on the longstanding partnership between the two companies, which spans 25 years. The two are currently working on HBM4, a high-bandwidth memory chip. Samsung will also continue to collaborate with the US-based firm in various areas including robotics and mobile devices.
In a statement, Samsung declared its plans to leverage NVIDIA’s accelerated computing to scale its AI Factory “over the next several years”. The company relies on the NVIDIA Omniverse platform for its digital twins, which provide virtual environments for optimising production. Aside from that, Samsung is integrating the NVIDIA cuLitho library to improve its computational lithography performance.
According to Samsung, it will use these technologies to expand its AI Factory infrastructure. Aside from that, it plans to bring this infrastructure to its global manufacturing hubs, including in the US.
(Source: Samsung newsroom)
