Apple M6 Chip Reportedly Set to Boost Mac AI Performance With Major Memory Upgrade
Quick Read
- The Apple M6 chip is reportedly moving to TSMC’s 2nm (N2) process for the first time, a major leap from the M5’s 3nm chip
- Memory bandwidth jumps from 153GB/s on the M5 to 200GB/s on the M6, aimed squarely at on-device AI performance
- The GPU core count rises from 10 to up to 12 cores, alongside a redesigned GPU architecture
- CPU performance improves across all cores, with better video encoding and decoding too
- Unlike past generations, Apple is reportedly skipping M6 Pro, Max, and Ultra chips entirely, saving those upgrades for the M7 lineup
- Launch is expected as early as Q4 2026, possibly in a new Mac Studio, iMac, or entry-level MacBook Pro
Apple’s next chip is shaping up to be one of its most unusual releases in years. According to a 9to5Mac report citing Bloomberg, the Apple M6 chip will be built on TSMC’s 2-nanometer (N2) manufacturing process, a first for Apple Silicon and a full generational jump from the 3nm node used in the current M5 family. The M6 delivers several meaningful upgrades over the current M5 generation, with the headline specification being memory bandwidth.
The M5 currently maxes out at 153 gigabytes per second, while the M6 pushes that figure to 200 gigabytes per second, a 31 percent increase. That matters more than it might sound like on paper think of memory bandwidth as the width of the highway data travels on inside your Mac. A wider highway means Apple Intelligence features, from writing tools to image generation, can pull and process information faster without bottlenecking the rest of the system.
The GPU Is Getting A Glow-up Too
The base M5 chip supports up to 10 GPU cores, while the M6 steps up to 12 graphics cores, a 20 percent increase that should translate into smoother gaming, faster video exports, and better performance for GPU-heavy creative work. The chip will also deliver faster performance across all its cores, along with improvements in video encoding and decoding, and Bloomberg reports the M6 will have a redesigned GPU.
Here’s the twist: the Apple M6 chip is breaking Apple’s usual playbook. Since the M1, Apple has always followed up with higher-end Pro and Max configurations, and two Ultra chips per generation but according to a separate 9to5Mac report, Apple will only release the base-model M6 processor, skipping Pro, Max, and Ultra variants entirely. Apple reportedly determined that the M7’s improvements are notable enough to justify skipping most of the M6 lineup, with the M7 expected to push memory bandwidth even further, up to 240 gigabytes per second.
As for timing, supply chain reports point to Apple securing more than half of TSMC’s 2nm production capacity through 2026 for the M6 and the A20 chips destined for the iPhone 18 lineup. Apple expected to launch the M6 later this year, and industry chatter suggests it could debut quietly much like the M4 first appeared in an iPad before reaching MacBooks potentially in a refreshed Mac Studio, iMac, or entry-level MacBook Pro before a fuller redesign follows.
For now, nothing is official until Apple confirms it, but if the leaks hold up, the Apple M6 chip looks less like an incremental bump and more like the foundation for Apple’s next real leap in on-device AI.
