Apple is developing a server based on M-series Ultra chips for AI

Apple plans to release a specialized server for artificial intelligence tasks by 2029. The device will be based on high-performance M-series Ultra processors, which will be the company's first such solution for the corporate segment in decades.

According to insider data, the server architecture is focused on performing complex calculations that require high memory bandwidth and energy efficiency, characteristic of Apple Silicon chips. The company's engineers are considering the possibility of integrating proprietary networking technologies to ensure the system's scalability when working with large language models.

The project involves creating an infrastructure that can compete with current GPU-based solutions. As part of the preparations for the launch, Apple has been consulting with market leaders, including Nvidia, to integrate advanced networking protocols. It is expected that the use of its own chips will allow the company to reduce its dependence on third-party providers of computing power.

Apple's return to the server hardware market marks a strategic shift towards developing its own AI ecosystem. The creation of specialized hardware will allow the company to optimize the operation of its cloud services and machine learning tools, providing full control over the technology stack from the chip to the software.