The M5 Ultra serves as a performance powerhouse, utilizing a quad-die architecture that fuses two M5 Max chips. With an up-to-36 core CPU, an 80-core GPU, and 1.2 TB/s of unified memory bandwidth, the chip offers a 50% increase in memory speed compared to the previous M3 Ultra. Designed for complex 3D rendering and visual effects, the hardware is positioned to handle frontier AI models with significant throughput.
In section Startups & Technology
Apple Unveils M5 Ultra and M6 Chips for On-Device AI
Apple has expanded its custom silicon lineup with the M5 Ultra and M6 chips, signaling a strategic push to migrate heavy artificial intelligence workloads directly to Mac hardware. The launch aims to challenge cloud-dependent AI models by prioritizing local processing power, security, and energy efficiency for both developers and high-end creative professionals.
Conversely, the M6 focuses on the broader user base, leveraging a 2 nm manufacturing process. According to Apple vice president of Silicon Engineering Sri Santhanam, the chip integrates a new CPU complex and a dual 16-core Neural Engine to enhance energy efficiency. The M6 features a 12-core CPU, delivering a 30 percent boost in peak GPU compute over the M5. This architecture is specifically tuned to accelerate on-device large language model processing, allowing developers to fine-tune and execute proprietary AI models without relying on remote servers. By keeping workloads local, Apple intends to replicate the privacy-first success of its Face ID and Touch ID technologies, offering a secure environment for AI experimentation. The latest Mac Mini models featuring these processors are available for pre-order starting at $899, with shipping scheduled after September 22.
Comments (0)
No comments yet. Be the first!