Apple’s new M5 Pro and M5 Max chips in MacBook Pros offer superior local AI processing capabilities, posing a significant threat to Microsoft’s AI ambitions. The M5 Max boasts up to 128GB of unified memory, exceeding competitors like Nvidia RTX 5090’s 32GB, allowing large AI models to run locally with enhanced privacy and reduced latency.
The integrated neural accelerators within GPU cores and MLX framework create a more efficient AI platform than Microsoft’s Windows ML approach. Apple’s latest MacBook Pros feature “super cores” and performance cores paired with between 20- and 40-core GPUs, each with a neural accelerator inside.
Apple’s M5 chips use the open-source array framework MLX, allowing for seamless integration of local AI models without requiring users to allocate memory manually. This enables AI models to run on either CPU or GPU without data transfer, providing a more efficient and private experience.
The new MacBook Pros include up to 128GB of available memory, with most of it dedicated to the GPU’s AI engine. This surpasses the local VRAM associated with high-end graphics cards, such as Nvidia RTX 5090. Apple’s MacBooks can now load and run local AI models without latency or subscription costs, setting a new standard for AI performance.
While AMD’s Ryzen AI Max+ is a close competitor, Apple’s M5 chips have impressed developers with their ability to handle complex AI tasks locally. The high bar set by the MacBook Pros will challenge Microsoft and other competitors to meet in terms of AI performance and efficiency.
Source: https://www.pcworld.com/article/3078457/why-microsoft-should-be-worried-about-m5-macbooks.html