Apple Plans Major iPhone Memory Overhaul to Power Next-Gen AI Features

· 1 min read

article picture

Apple plans to revolutionize the iPhone's internal architecture by moving away from its long-standing memory design, aiming to boost artificial intelligence capabilities in future devices.

According to industry reports, the tech giant will abandon its current package-on-package (PoP) memory configuration - used since the iPhone 4 in 2010 - in favor of discrete memory packaging starting in 2026. Samsung, Apple's key memory supplier, has begun research to support this transition.

The current PoP design stacks memory directly on top of the main processor chip, creating a compact solution ideal for smartphones. However, this approach limits the iPhone's AI tasks due to restricted data transfer speeds and memory bandwidth.

By physically separating the memory from the processor, Apple can implement more input/output connections, enabling faster data movement and additional parallel processing channels. The new design also improves heat management, a critical factor for AI workloads.

The transition brings both advantages and challenges. While discrete packaging allows for better AI performance, it requires more physical space inside the iPhone. Apple may need to reduce the size of other components or adjust the battery capacity to accommodate the new layout.

Samsung is simultaneously developing next-generation LPDDR6 memory technology for Apple, promising two to three times faster data transfer speeds compared to current solutions. A specialized variant called LPDDR6-PIM, featuring built-in processing capabilities, is being standardized in collaboration with SK Hynix.

The redesigned memory architecture is expected to debut in the 2026 iPhone lineup, pending successful resolution of engineering challenges related to component miniaturization and internal space optimization.

This strategic shift underscores Apple's commitment to advancing on-device AI capabilities, potentially enabling more sophisticated features while maintaining user privacy through local processing.