Apple doesn't rely on NVIDIA semiconductors to build AI infrastructure

July 30, 2024 9:45 AM GMT+9

In constructing a new artificial intelligence (AI) software infrastructure, it became clear in Apple's research materials published on the 29th that it is not a semiconductor of NVIDIA, the largest semiconductor company for AI, and relies on Google's cloud infrastructure under the umbrella of Alphabet.
The survey materials do not clearly state that Apple does not use NVIDIA semiconductors at all, but there was no description of NVIDIA hardware in the explanation of AI tool software and hardware infrastructure.

According to research data, Apple uses an integrated circuit called Google's “tensor processing unit (TPU)” to train AI models.
NVIDIA sells semiconductors and systems as products to customers, and Google makes it possible for customers to use TPU through the company's cloud platform.