Apple skips M6 Pro and Max to fast-track AI-focused M7 chips
Apple reportedly plans to skip Pro and Max variants for its upcoming M6 chip, focusing the M7 generation on on-device AI processing and enhanced video encoding. The M6 chip will feature improved memory bandwidth up to 200 GB/s, an updated neural engine, and a redesigned GPU with up to 12 cores. These architectural changes are aimed at accelerating local video processing, rendering, and AI workloads, with the M5 Ultra Mac Studio also expected later this year.
Key Takeaways
- M6 memory bandwidth reaches 200 GB/s, up from 153 GB/s on the M5, with a redesigned GPU scaling to 12 cores (vs. 10 on M5).
- Apple will not release M6 Pro or M6 Max — the first generation since 2020 to lack these variants; Pro and Max return with M7 in late 2027.
- M7 base chip targets 240 GB/s memory bandwidth and is designed primarily around on-device AI processing, per Bloomberg.
- M5 Ultra Mac Studio expected as early as this year with ~36 CPU cores, 80 GPU cores, and tested support for up to 768 GB of unified memory.
- M7 Ultra is slated for 2028, completing the full M7 Pro/Max/Ultra lineup.
Why It Matters
Apple's decision to skip M6 Pro and Max creates a gap in its high-end Mac lineup through late 2027, leaving the M5 Ultra Mac Studio as the only near-term upgrade path for professional video and AI workloads. The M7 generation's focus on on-device AI processing signals Apple is prioritizing local inference capabilities — directly relevant to streaming video pipelines that increasingly use AI for encoding, upscaling, and real-time processing. The M6's improved video encoding/decoding and redesigned GPU for concurrent AI and graphics tasks also matter for production hardware buyers. Watch whether the M5 Ultra Mac Studio launches by October 2026 and whether component constraints limit its 768 GB memory configuration.
Additional Context
The M6 chip will be Apple's first built on TSMC's 2-nanometer N2 process, replacing the 3nm node used since the M3 generation, per MacRumors, June 2026. The transition also shifts from InFo packaging to Wafer-Level Multi-Chip Module (WMCM) packaging, which integrates CPU, GPU, DRAM, and Neural Engine more closely to improve inter-component communication. These architectural changes underpin the M6's jump to 200 GB/s memory bandwidth and its redesigned 12-core GPU optimized for concurrent AI and graphics workloads. The competitive landscape for on-device AI chips has intensified. Macworld reported in June 2026 that Nvidia's RTX Spark chips are coming to Windows laptops starting late 2026, while AMD, Intel, and Qualcomm are all racing to improve consumer AI processing. Apple's decision to compress the M6-to-M7 timeline — moving M7 up by as much as six months — appears to be a direct response, with the M7's 240 GB/s bandwidth representing a 57% increase over the current M5's 153 GB/s. At WWDC 2026, Apple introduced Core AI, a new framework for on-device AI model deployment that powers Apple Intelligence and is now available to developers, per Apple Developer documentation from June 2026. Core AI supports inference across CPU, GPU, and Neural Engine with ahead-of-time model compilation and Swift APIs. The M7's AI-focused architecture aligns with this software push, suggesting Apple is building a vertically integrated on-device AI stack from silicon through framework. The M5 Ultra Mac Studio faces headwinds from a global DRAM shortage. Bloomberg reported in April 2026 that supply chain issues had already delayed the Mac Studio launch from early 2026 to around October 2026. Apple removed the 512GB memory option from the M3 Ultra Mac Studio in March 2026 and raised the 96GB configuration's price by $1,300, per MacRumors, June 2026. Whether the M5 Ultra's tested 768GB memory configuration reaches consumers remains uncertain given these constraints.
Read full article at 9to5mac.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source