A group of four Mac Studio computers demonstrated the ability to handle a 1 trillion parameter AI model from a single electrical outlet during a recent product launch. Apple's new Mac mini and Mac Studio hardware is engineered to process heavy AI workloads locally rather than relying on cloud infrastructure. This shift aims to lower the total cost of AI computing by optimizing performance per watt for edge processing.

The Mac Studio's top models feature M5 Max or M5 Ultra chips with up to 512GB of unified memory and 1.2TB/s bandwidth to support large language models on-device. To further scale capacity, the units include Thunderbolt 5 with RDMA for distributed AI inference. The Mac mini starts at $899 with an M6 chip that Apple claims delivers 4 times the AI performance of previous generations.