Apple skipping M6 Pro chips, fast-tracking M7 for AI

Key Takeaways

• Apple will skip M6 Pro and Max chips entirely, releasing only a base M6 for entry-level Macs this year
• The M7 generation is being accelerated by six months, with Pro and Max parts due late 2027
• Memory bandwidth jumps from 153 GB/s (M5) to 200 GB/s (M6) to 240 GB/s (M7), targeting on-device AI workloads

Apple is breaking its four-year silicon playbook. According to Bloomberg, the company will release only a base M6 chip for entry-level Macs this year, skipping the Pro and Max variants entirely. Instead, Apple is accelerating development of the M7 generation, built specifically for on-device AI, pulling its release forward by roughly six months to late 2027.

This marks the first time since the original M1 that Apple has shipped a generation without higher-end derivatives. Every family from M1 through M5 paired its base silicon with Pro and Max variants, and three of those generations added an Ultra. The M6, codenamed Komodo, breaks that pattern.

What's actually shipping in the M6?

The base M6 targets entry-level machines, including a refreshed 14-inch MacBook Pro that has seen its starting price climb to $1,999. Bloomberg's sources say the chip reaches around 200 GB/s of memory bandwidth, up from 153 GB/s on the M5. The GPU gets a redesign with up to 12 cores, compared to 10 on the current generation.

Apple is also reportedly pairing an upgraded Neural Engine with faster GPU and CPU cores. But without Pro, Max, or Ultra variants, the M6 won't power Apple's professional Mac Studio or high-end MacBook Pro configurations.

Why is Apple fast-tracking the M7?

The answer is AI inference. Memory bandwidth dictates how fast a chip can move the large data blocks that AI models require. The climb from 153 GB/s (M5) to 200 GB/s (M6) to a planned 240 GB/s (M7) represents a 56% bandwidth increase from M5 to base M7. That's a direct play for running larger local models without cloud dependency.

The base M7, codenamed Delos, could arrive in the first half of 2027. Pro, Max, and Ultra variants, grouped internally under the Andros codename, would follow in late 2027 and 2028. Apple has declined to comment, and none of these specifications or dates are confirmed.

For context, the current M3 Ultra already delivers up to 819 GB/s by fusing two Max dies. The high-end tiers, not base chips, carry the heaviest local AI workloads. Skipping M6 Pro and Max means Apple's pro users will wait until M7 for meaningful AI silicon upgrades.

The M5 Ultra is still coming

Apple hasn't abandoned the current generation. Bloomberg reports an M5 Ultra, codenamed Sotra, remains on track for this year. It would pack around 36 CPU cores, 80 GPU cores, and tested support for up to 768GB of unified memory. That's a significant jump from the current M3 Ultra Mac Studio, which Apple capped at 96GB after pulling its 512GB tier and dropping the 128GB option.

The M5 Ultra would refresh the Mac Studio, which currently runs M3 Ultra silicon. Apple skipped an M4 Ultra entirely. Whether the highest M5 Ultra memory configurations actually reach buyers depends on supply constraints that, according to the report, haven't eased.

Memory supply and the price problem

Apple raised prices across its Mac and iPad lines last week. In an interview with the Wall Street Journal, outgoing CEO Tim Cook called the situation a "hundred-year flood," blaming AI server demand for pulling high-bandwidth memory away from consumer hardware.

That explains the supply constraints, but not entirely the strategy shift. By condensing the M6 to a single base chip, Apple can focus manufacturing capacity on the M7 generation where AI performance gains matter most. It's resource triage disguised as a product decision.

What this means for Mac buyers

If you're in the market for a Mac Studio or high-end MacBook Pro, the timeline just stretched. The M5 Ultra offers a refresh this year, but the next major jump in professional silicon won't arrive until the M7 Pro and Max in late 2027. That's a two-year gap for users who bought into M3 Max or M3 Ultra machines.

Entry-level buyers get the base M6 this year, with a claimed 30% bandwidth improvement over M5 and a more capable GPU. But the AI-first M7 base chip, due in early 2027, might be worth the wait if on-device inference is your target workload.

Frequently Asked Questions

When is the Apple M7 chip releasing?

According to Bloomberg, the base M7 (codenamed Delos) could arrive in the first half of 2027, with Pro, Max, and Ultra variants following in late 2027 and 2028.

Is Apple skipping the M6 chip entirely?

No. Apple will release a base M6 chip (codenamed Komodo) for entry-level Macs this year, but will skip the Pro and Max variants of that generation.

What is the memory bandwidth of the M7 chip?

The base M7 targets approximately 240 GB/s of memory bandwidth, up from 200 GB/s on the M6 and 153 GB/s on the M5.

Why is Apple accelerating the M7 development?

Apple is fast-tracking the M7 to meet demand for on-device AI and heavier graphics workloads. The company needs higher memory bandwidth to compete in local AI inference.

Will there be an M5 Ultra chip?

Yes. Bloomberg reports an M5 Ultra (codenamed Sotra) is still planned for this year, with 36 CPU cores, 80 GPU cores, and support for up to 768GB unified memory.

Source: Latest from Tom's Hardware