iPhone 18 Pro, iPhone 18 Pro Max and iPhone 18 Fold Said to Debut With 2nm A20 Chipset in 2026

Apple's chipsets powering the iPhone 15 Pro and 16 Pro models have been using TSMC's 3nm manufacturing process, and that could change in 2026. Apple started using the 3nm manufacturing process after its switch to the ‘Pro' branding for its chipsets, for the iPhone 15 Pro and newer models. The switch from the 4nm Apple A16 Bionic processor to the more efficient 4nm A17 Pro chipset brought several benefits to iPhone users, including better battery life and improved performance. An analyst has predicted that Apple will make the switch to a more power-efficient chipset that is also expected to bring other performance benefits.

A 9to5Mac report citing GF Securities analyst Jeff Pu claims that Apple will switch to using TSMC's 2nd-gen 2nm manufacturing process (N2) for its next generation A20 chip. The processor is expected to debut in 2026 with the iPhone 18 Pro, iPhone 18 Pro Max, and the iPhone 18 Fold, Apple's first foldable device.

With the new fabrication process, the 2nm Apple A20 processor will be smaller than its predecessor. While it could offer better power efficiency, the analyst reportedly claims that Apple will adopt a new Wafer-Level Multi-Chip Module (WMCM), which will reportedly bring the various components (SoC, DRAM) closer together at the wafer level itself.

Such a design, which reduces the need for a substrate, also provides better thermal efficiency. Overall, this design will also result in lower power consumption, which works better given the new performance requirements to run Apple Intelligence on iOS.

The analyst also predicts that TSMC's 2nd-gen 2nm (N2) production line will be up and running by 2026, and that the chipmaker will ramp up its capacity for producing more of these chips in 2027.

Apple's current iPhone 16 Pro model utilises its A18 Pro SoC, which is also manufactured by TSMC using their 3nm N3E process. It has a hexa-core architecture with the two big cores clocked at 4.05GHz and has a 6-core GPU along with a 16-core Neural Engine, which is said to run AI tasks at 35 TOPS.