Apple seems to be changing its plans for future Mac chips. Instead of rolling out the usual M6 Pro, Max, and Ultra versions, it’s reportedly jumping ahead to the M7 lineup, with much bigger Neural Engine improvements aimed at boosting AI performance.
Apple is still expected to introduce new Macs this fall with the base M6 processor. If the company follows its typical rollout, the M6 Pro and M6 Max would arrive later for higher-end MacBooks, with the M6 Ultra powering desktop models.
This time, though, Apple is reportedly changing course. Instead of following the base M6 with Pro, Max, and Ultra versions, the company has already shifted its attention to the M7. It reached the tape-out milestone just six months after the M6, showing how quickly Apple wants to improve AI performance on Macs.
According to Apple’s internal roadmap, the first Macs powered by the M7 chip are expected to arrive in the first half of 2027. M7 Pro and M7 Max models are slated to follow later in the year, with the M7 Ultra planned for 2028.
Apple has skipped an Ultra chip before; the M4 lineup never got one. But this would be the first time the company drops all of the high-end versions of a chip generation. According to people familiar with the plans, Apple believes the M7’s AI upgrades are important enough that it doesn’t make sense to keep developing the M6 lineup.
The biggest part of these changes is Apple’s Neural Engine. It’s the dedicated AI hardware that runs on-device generative AI, speeds up AI processing, and powers Apple Intelligence. Apple has improved it with every Mac chip since the M1 launched in 2020, and the M4 brought one of the biggest upgrades yet.
With the M7 generation, Apple is reportedly targeting a new level of AI capability. The M7 Ultra is expected to narrow the gap between consumer desktop processors and specialized AI accelerators used for demanding machine learning workloads.
Apple is also reportedly expanding memory capacity to support larger AI workloads. The M7 Ultra is said to accommodate up to 1.5TB of unified memory, doubling the capacity expected for the M5 Ultra server chip while matching the highest memory configuration of the 2019 Intel-powered Mac Pro.
A 1.5TB memory capacity would enable much larger AI models to remain loaded in memory, reducing bottlenecks and lowering the need for external storage or cloud-based processing. However, whether Apple releases Macs with the full configuration will likely depend on memory supply and pricing.
Apple’s desktop and server plans appear to be moving in the same direction. The company is reportedly building a new AI server using the M5 Ultra, known internally as J246, while engineers are already developing a next-generation version based on an M7 Ultra server chip. If everything stays on schedule, it could arrive around 2029 and use the same technology that powers Apple’s top-end Macs to run Apple Intelligence in the cloud.
Apple is also planning the M8 generation, with an even stronger focus on AI capabilities. Reports point to a processor code-named Soko for 2028, along with additional high-end Mac chips under the Cardinal project. The new lineup is expected to use a 1.4-nanometer manufacturing process, improving power efficiency.
Maybe you’d like some other interesting articles?