When it comes to chip-making, smaller is better. Intel, Samsung, and TSMC fight an endless race to reduce the size of their chip nodes, increasing transistor density, which results in faster and more efficient processors. Right now, TSMC’s customers are lining up to be the first to get a taste of its latest breakthrough.
Taiwan Semiconductor Manufacturing (TSMC), the world’s largest chipmaker and Apple’s leading supplier, is set to make significant enhancements to its process when it starts high-volume production of 2-nanometer (N2) technology starting in late 2025. For context, the latest iPhones and MacBooks use a 5nm processor and 3nm is scheduled for later this year. It’s no surprise that Apple would be among the first to use the new technology, but the company will reportedly be joined by a familiar foe: Intel.
“We also see more clarity around TSMC’s N2 expansion schedule in Fab 20 (Hsinchu),” Sze Ho Ng, an analyst with China Renaissance Securities, wrote in a note for clients, reports Tom’s Hardware. “Tool move-in is expected to start by end-2022, based on company plans, ahead of risk production in late 2024E with Intel (client PC Lunar Lake’s graphic ‘tiles’, while the CPU ‘tiles’ are fabbed using Intel’s 18A) and Apple being the anchor customers for dedicated capacity support.”
You might wonder why Intel, which has its own chipmaking fabs, would turn to a competitor for supply. According to DigiTimes and UDN reports, Intel will be an early adopter of TSMC’s 2nm node for use in the company’s graphics processing units (GPUs) and other SoCs. Intel’s new Arc GPUs will reportedly use TSMC’s 3nm and 2nm process nodes as they attempt to take on market leaders Nvidia and AMD.
As for Apple, the Cupertino giant has been TSMC’s largest customer by revenue for the last decade, using the foundry for the processors used within its phones, tablets, and M1-powered laptops. The latest chips use a 5nm processor, so there will presumably be interim releases before the N2 chips are delivered in early 2016. In fact, TSMC’s 3nm chips arrive later this year, and will likely power the iPhone 14 when the phone is announced in the fall.
For its part, Intel believes it can get back on track after falling far behind competitors. The company’s CEO, Pat Gelsinger, stated last year that Intel would be returning to product leadership in 2025. He outlined releases for the upcoming years, placing the release of a 2nm (20A) node in 2024. The following release, 18A (codename Lunar Lake), will reportedly use TSMC’s 2N for graphics tiles and is said to be ahead of schedule.
A 2nm chip could also arrive soon thanks to IBM; the company unveiled a new 2nm chip with nanosheet technology last year—a technology that packs 50 billion transistors in a chip the “size of a fingernail.” IBM says its 2nm process will roll out to partner foundries—Samsung and Intel among those—in late 2024, potentially giving Intel a rare head start over TSMC.
Apple and Intel aren’t the only ones tapping TSMC for its advanced 5nm N5 chips. As Tom’s Hardware notes, MediaTek already announced two 5G smartphone chips—Dimensity 8100 and Dimensity 8000—using TSMC technology, and Nvidia is poised to use the 4N process for its Hopper data center GPU and potentially the upcoming Ada Lovelace (RTX 40-series) consumer graphics cards.
TSMC hasn’t teased any of the performance and efficiency gains its N2 chips could bring; however, the upcoming 3nm process is said to result in 15% performance improvements and 30% power reduction, so we could see similar upgrades with the next technology.