Apple is boosting the M7 chip for AI: what will change on the MacBook and iPad
Apple has completed the tape-out stage for the M7 just around six months after the M6 reached the same stage. This does not mean that finished devices are about to go on sale: it refers to the completion of a key stage in the chip’s design, prior to further testing, production and integration into products.
The base M7 is expected in the first half of 2027. According to preliminary reports, it could form the basis for the updated MacBook Pro and iPad Pro. The more powerful M7 Pro and M7 Max could be released in late 2027, whilst the M7 Ultra is expected in 2028.
Why Apple might skip the top-of-the-range M6
The main reason is artificial intelligence. According to Bloomberg, Apple may release only the base M6, but not produce separate M6 Pro, M6 Max and M6 Ultra models for this generation. Instead, the company wants to move on to the M7 more quickly, where significant updates for neural computing are expected.
This is a significant change for Apple. The company usually expands its chip generations in a consistent manner: a base version, followed by Pro, Max and Ultra models for more powerful Macs. If Apple does indeed skip the top-of-the-range M6 models, it will mean that the race for on-device AI has become more important than the usual product cycle.
What exactly is the M7 set to accelerate?
One of the main focuses is memory bandwidth. Tom’s Hardware, citing Bloomberg, reports that the base M7 could achieve up to 240 GB/s of memory bandwidth. By way of comparison, the M5 offers around 153 GB/s, whilst the M6 was expected to deliver approximately 200 GB/s.
To the average user, this sounds technical, but the idea is simple: the chip will be able to transfer data more quickly between the processor, graphics, neural engine and memory. This is crucial for running AI functions locally, as well as for processing video, graphics, large files and neural networks without constantly having to rely on the cloud.
What’s in store for the M7 Ultra
The M7 Ultra is set to be the most powerful chip in the range. According to Bloomberg, Apple is designing it to support up to 1.5 TB of unified memory. This is roughly double the expected maximum configuration of the M5 Ultra.
This amount of memory is important not only for video editing or 3D graphics. It could enable the device to run significantly larger language models and other AI systems. Research into running large models locally on Apple Silicon has already shown that Apple’s unified memory helps to circumvent some of the limitations faced by systems with dedicated video memory.
However, this is not yet a guarantee that Apple will actually sell Macs with 1.5 TB of memory. Some sources note that the final configuration will depend on the cost and availability of memory at the time of launch.
Why this matters for the Mac
In recent years, Apple has lagged behind its competitors in the public race for AI services, but its strength lies in its hardware. The company has long been promoting the idea of on-device computing, where some functions run directly on the device rather than in the cloud.
For users, this could mean faster AI features, lower latency, greater privacy and reduced reliance on servers. For professional Macs, it offers the ability to work with large models, video, 3D, code and complex calculations without the need for separate external accelerators.
The Verge notes that the upcoming M7 Ultra could form the basis not only for Macs but also for a new Apple server product related to AI.
What is known about the M8
Although the M7 has not yet been released, Apple, according to media reports, is already working on the next generation — the M8. Reports citing Bloomberg suggest that these chips could be the first Apple Silicon chips manufactured using TSMC’s 1.4-nanometre process.
The transition to a new manufacturing process typically offers two key advantages: greater energy efficiency and higher transistor density. In other words, future Macs, iPads and, likely, iPhones will be able to deliver greater performance without a corresponding sharp increase in energy consumption.
What to bear in mind
All this information is currently based on reports from sources rather than an official announcement from Apple. The company traditionally does not comment on upcoming chips long before their unveiling.
But if this roadmap is confirmed, the main change will not be limited to the speed of Macs. Apple is effectively redesigning its processors for the era of on-device artificial intelligence, where it is not just the CPU or GPU cores that matter, but also memory, neural units and the ability to run complex models directly on the device.
Follow us on Telegram.
