Apple unveils M6 chip 2nm, M5 Ultra pushes AI compute

Apple today announced its next-generation silicon, the M6 chip, marking a significant advancement with its first 2-nanometer process technology. The M6, debuting in the new Mac mini, features a 12-core CPU, 12-core GPU, and a Dual 16-core Neural Engine, promising substantial gains in everyday performance and power efficiency. Alongside the M6, Apple also unveiled the M5 Ultra in the new Mac Studio. This chip represents Apple's most powerful to date, utilizing a quad-die architecture for extreme desktop-class performance and massive unified memory bandwidth, targeting professional and AI workloads.

StartupHub.ai data shows Apple with a strong score of 83/100, reflecting its consistent innovation in custom silicon. This latest announcement places further pressure on competitors like Google (NASDAQ:GOOGL), which has a StartupHub score of 74/100, and Samsung Electronics (NASDAQ:SSNLF) (score 37/100) in the high-performance, AI-focused chip market.

The M6 chip's 2nm architecture is engineered for enhanced transistor density, leading to improved performance per watt. Its 12-core CPU, featuring 2 super cores, 4 performance cores, and 6 efficiency cores, aims to deliver the world's fastest single-threaded performance. The integrated 12-core GPU includes Neural Accelerators, boosting AI compute by nearly 30 percent over the M5 generation.

For demanding professional tasks, the M5 Ultra stands out with its quad-die architecture, achieved through Apple's UltraFusion technology connecting two M5 Max chips. This design enables an up-to-36-core CPU and an up-to-80-core GPU, coupled with a massive 1.2TB/s of unified memory bandwidth. This configuration is designed to handle complex 3D rendering, visual effects, and large-scale AI model training and inference directly on device.

Apple's continued investment in custom silicon, particularly in advanced process nodes like 2nm for the Apple M6 chip 2nm, underscores its strategy to maintain a performance lead. The M5 Ultra's quad-die approach, while complex, positions Apple to compete directly with high-end discrete GPUs from companies like NVIDIA (NASDAQ:NVDA) for AI-centric workloads, leveraging its unified memory architecture for efficiency and speed.

The enhanced Neural Engines and Neural Accelerators across both chips signal a strong push towards on-device AI capabilities, aligning with broader industry trends. Developers utilizing Apple's frameworks like Core ML and Metal will gain access to significantly increased compute power for AI model execution and fine-tuning, potentially accelerating the adoption of sophisticated AI applications across the Mac ecosystem.