Apple unveils 2nm M6 and Quad-Die M5 Ultra chips focused on on-device AI

Apple has officially announced its next generation of Mac processors, the entry-level 2nm M6 and the high-end quad-die M5 Ultra.

The next generation of Apple Silicon

Focus on on-device AI?

The announcement marks a major milestone in hardware engineering, debuting tiny physical shrinkages, new architectural layouts, and significant speed boosts aimed at both everyday consumers and creative professionals.

The new base M6 - Efficiency and On-device AI for everyone

Designed to power everyday computing and light creative projects, the M6 serves as Apple's new foundational desktop chip.

• Moving to 2-Nanometer Technology - The new M6 chip is Apple's first processor built on a cutting-edge 2-nm manufacturing process. By shrinking the physical transistors on the silicon, engineers can pack significantly more computing power into a smaller footprint while keeping heat and power consumption low.
• Expanded Core Count - It features a new 12-core CPU (with 2 super cores, 4 performance cores, and 6 efficiency cores) - a two-core bump over the previous generation.
• Dual 16-Core Neural Engine - To accelerate local artificial intelligence tasks, the M6 introduces two 16-core Neural Engines working in tandem, delivering double the peak AI performance of its predecessor.
• Integrated GPU Accelerators - The 12-core graphics engine now includes Neural Accelerators inside every single GPU core for faster matrix calculations and prompt processing in Large Language Models (LLMs).
• Upgraded Memory & Bandwidth - Supports up to 32GB of unified memory with an expanded bandwidth of up to 170GB/s.

The M5 Ultra - Quad-Die Power for Heavy Pro Workloads

It is targeted at software engineers, 3D artists, and AI researchers; the M5 Ultra scales Apple Silicon performance to the absolute limit.

• First Quad-Die Design - Utilizing Apple's UltraFusion interconnect technology, the chip fuses four separate dies into a single unified processor, yielding over 4.4TB/s of inter-die bandwidth
• 36-Core CPU & 80-Core GPU - Scalable up to 36 cores (12 super cores & 24 performance cores) and up to an 80-core GPU, complete with embedded Neural Accelerators in every graphics core.
• Up to 512GB Unified Memory - Features a massive memory pool with up to 1.2 TB/s of bandwidth - a massive 50 percent increase over previous Ultra Generations, allowing developers to run massive, open-weight AI models directly on the machine without cloud processing.
• Thunderbolt 5 & RDMA Clustering - Built-in remote direct memory access allows up to four M5 Ultra desktop systems to be networked together via Thunderbolt 5, pooling memory across machines to handle massive enterprise-scale AI inference.

For consumers, this should yield benefits such as faster day-to-day responsiveness, private AI without subscriptions (queries run privately on-device without sending data to the cloud), and uninterrupted 3D and Video workflows.

At launch, the M6 will be available on the new Mac Mini base configuration with a 12-core CPU, 12-Core GPU, and Dual 16-Core Neural Engines. The M5 Ultra will come with a 30-core CPU, 64-core GPU, and 32-core Neural Engine. 

Apple is also offering a 512GB Unified memory variant of the new Apple Mac Studio with the M5 Ultra. It will be available later in October 2026.

Price and availability

Apple Mac Mini M6 - starts at PHP 64,990

Apple Mac Studio M5 Ultra - starts at PHP 379,990

The new M6 Mac Mini and M5 Ultra Mac Studio are currently listed on Apple's PH website but are not yet available for purchase.

Stay tuned for updates.

What do you guys think?

Source: Apple