Particle.news
Download on the App Store

Apple Skips High-End M6 Chips to Fast-Track AI-First M7 Roadmap

Apple says major neural‑processing upgrades justify accelerating its next silicon generation to deliver stronger on‑device AI for future server use.

Overview

  • Bloomberg’s Mark Gurman reported Sunday that Apple will ship only a base M6 chip in late 2026 and will not produce M6 Pro, M6 Max, or M6 Ultra models, a change the outlets described as driven by AI priorities.
  • People familiar with the plans say Apple began finalizing the M7 design roughly six months after the M6 tape‑out, compressing the timetable so a base M7 is targeted for the first half of 2027 with M7 Pro and M7 Max later that year and an M7 Ultra in 2028.
  • The M7 Ultra is being engineered to deliver a large leap in neural processing and to support up to 1.5 terabytes of unified memory, a configuration suppliers and reporters caution may not ship widely because of current DRAM shortages and high costs.
  • Apple’s silicon push ties to a server strategy reported to include an M5 Ultra‑based AI server (codename J246) coming soon and an M7 Ultra‑derived server chip planned for around 2029, with software work such as macOS 27 and new Siri features being prepared to use the chips’ AI gains.
  • The shift builds on Apple’s decade of Neural Engine work — begun in projects like the cancelled car effort — and means professionals who wait for high‑end Mac performance will likely hold out for M7 systems while mainstream buyers will see only the base M6 in the near term.