Apple is working on M7 and M8 chips

Apple is accelerating the development of its in-house processors. In order to focus more on artificial intelligence, a chip generation is partially skipped. At the same time, major upgrades in RAM are on the horizon.

Apple is pushing AI chip development

Apple is apparently accelerating its processor roadmap to catch up in the area of ​​artificial intelligence. As the company’s plans have already suggested, the company will severely cut back on the upcoming M6 chip generation. Only a basic chip will be released, while more powerful variants such as Pro, Max and Ultra will be completely omitted. Instead, the resources flow directly into the development of the M7 and M8 processors.

The primary reason for the unusual approach is the architecture of the new processors, which is specifically designed for complex AI calculations. The basic M7 is scheduled to come onto the market in the first half of 2027. It is expected to offer a memory bandwidth of 240 gigabytes per second, which should noticeably speed up local data processing.

Extended RAM planned

Apple is planning major hardware upgrades, especially for professional users. As Bloomberg reported, the M7 Ultra expected for 2028 will support up to 1.5 terabytes of RAM. This doubles the capacity of the upcoming M5 Ultra and for the first time reaches the level of the Intel-based Mac Pro from 2019. Such a large amount of RAM brings clear advantages for developers. It makes it possible to run large AI models locally on a Mac Studio without relying on cloud servers. Since Apple solders the RAM directly onto the chip, users benefit from very high transfer rates.

At the same time, however, this eliminates the possibility of subsequently upgrading the memory. The concept also has other disadvantages. The current shortage in the memory market could limit the availability of the chips. This is likely to drive prices up further. According to estimates, the maximum upgraded version of the Mac Studio with 1.5 terabytes of RAM could cost around 30,000 euros, provided Apple maintains its current surcharges for memory expansion.

The jump to 1.4 nanometers

At the same time, the hardware engineers are reportedly already working on the next generation. The M8 chip, developed internally under the code name “Soko”, is scheduled to be produced at TSMC from 2028 using the new 1.4 nanometer process. This smaller structure width promises greater energy efficiency while simultaneously increasing computing power. The iPhone 2028 should also benefit from the new technology with the A22 Pro.

The newfound performance potential of the M7 and M8 chips is not only used in Macs. Apple reportedly plans to use modified versions of the M7 Ultra for its own AI server infrastructure starting in 2029. In this way, the group is positioning itself in the long term against established products from competitors such as Nvidia.