According to reports from Bloomberg, Apple is gearing up to delve into generative AI, starting with a strategy to utilize its M2 Ultra chips in data centers for handling complex queries. The plan is to gradually transition to the more advanced M4 chips as they become available.
Initially, Apple intended to develop custom chips for data centers under a project dubbed Project ACDC, or Apple Chips in Data Center. However, it’s now believed that the existing processors, such as the M2 Ultra, already possess the necessary security and privacy features.
These chips will be integrated into Apple’s data centers and eventually deployed to servers operated by third-party entities. Apple manages its own servers across the United States and has been constructing a new facility in Waukee, Iowa, as part of its expansion plans announced back in 2017.
While Apple may not be moving as swiftly as its competitors like Google, Meta, and Microsoft in the realm of generative AI, the company has been actively involved in research on the subject. In December, Apple’s machine learning research team unveiled MLX, a machine learning framework designed to optimize AI model performance on Apple silicon. Additionally, Apple has been publishing research papers on AI models, providing insights into the potential applications of AI on its devices and potential upgrades for existing products like Siri.
Apple has highlighted the significance of AI performance in its recent announcement of the M4 chip, boasting about its new neural engine as “an outrageously powerful chip for AI.”
