Apple's Supplier TSMC to Start Mass Production of 3nm Chip Soon

TSMC, Apple's key chip supplier, will begin mass production of next-generation chips using 3nm process technology from the last week of December 2022. Apple is the main customer of the new process, which could initially be applied to future M2 Pro chips that are anticipated to power updated MacBook Pro and Mac mini models.

The A16 Bionic processor in the iPhone 14 Pro series is currently manufactured by Apple using TSMC's 4nm process, although the company may switch to 3nm. According to a report from August, the future M2 Pro chips will be the first to use the 3nm manufacturing process. In early 2023, revised 14-inch and 16-inch MacBook Pro models are anticipated to be the first to have the M2 Pro chip, along with possibly upgraded Mac Studio and Mac mini versions.

TSMC and Samsung are the only two companies worldwide that will make 3nm chips. The chips will aid in creating integrated circuits that increase gadget performance without consuming battery power. It will also boost the market for 5G smartphones and other electronic goods. TSMC's 3nm technology would also be used in modem chips for 5G and 6G wireless devices. High-performance computing, which refers to central processors, graphics processors, and AI accelerators, would leverage TSMC's technology. 

The 3nm technology provides 250 million transistors per square nanometre, compared to the 2nm process will give over 310 million transistors in the same area. A higher transistor counts lower power consumption while offering better performance with the same power usage.

The M3 chip and the A17 Bionic for the iPhone 15 will be built on TSMC's enhanced 3nm technology, which is yet to be made available later in 2023, according to another report. The DigiTimes report states that 3nm process chip production is "unlikely to ramp up" until the manufacture of the improved version begins. Unlike Samsung, TSMC's 3nm process technology would stick to FinFET transistors and instead rely on "innovative features" to achieve the full-node scaling.