Photo: Caitlin McGarry/Gizmodo
When it comes to chip manufacturing, smaller is better. Intel, Samsung and TSMC are in an endless race to shrink the size of their chip nodes and increase transistor density to produce faster, more efficient processors. Currently, TSMC customers are lining up to get an early taste of its latest breakthrough.
Taiwan Semiconductor Manufacturing Company (TSMC), the world’s largest chipmaker and a leading supplier to Apple, will make major improvements to its process when it begins mass production of 2-nanometer (N2) technology in late 2025. The iPhone and MacBook use 5nm processors, with 3nm planned for later this year. Unsurprisingly, Apple will be the first to use the new technology, but the company will reportedly be joining a familiar foe: Intel.
“We also see greater clarity on TSMC’s N2 expansion plans at Fab 20 (Hsinchu),” Huaxing Securities analyst Sze Ho Ng wrote in a note to clients, Tom’s Hardware reported. “According to company plans, tool move-in is expected to begin in late 2022, followed by risk production with Intel in late 2024 (graphics ‘blocks’ for client PC Lunar Lake, while CPU ‘blocks’ are manufactured using Intel’s 18A) Apple is exclusively A major customer for capacity support.”
You might wonder why Intel, which has its own chip fab, would turn to rival supplies. According to DigiTimes and UDN, Intel will be an early adopter of TSMC’s 2nm node for 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 try to compete with market leaders Nvidia and AMD.
As for Apple, the Cupertino giant has been TSMC’s biggest customer for the past decade, using the foundry for processors used in its phones, tablets and M1-powered laptops. The latest chips use 5nm processors, so there may be interim versions before N2 chips ship in early 2016. In fact, TSMC’s 3nm chips will arrive later this year and will likely power the iPhone 14 when it launches. fall.
For its part, Intel believes it can get back on track after falling far behind its rivals. The company’s chief executive, Pat Gelsinger, said last year that Intel would return to product leadership in 2025. He outlined releases over the next few years, targeting the release of the 2nm (20A) node in 2024. The next release, the 18A (codenamed Lunar, Lake) will reportedly use TSMC’s 2N for the graphics block, which is said to be ahead of schedule.
Thanks to IBM, 2nm chips may soon be on the horizon, too; the company last year unveiled a new 2nm chip using nanosheet technology — which packs 50 billion transistors into a chip “the size of a fingernail.” IBM said its 2nm process will be available to partner foundries, including Samsung and Intel, by the end of 2024, potentially giving Intel a rare lead over TSMC.
Apple and Intel aren’t the only companies taking advantage of TSMC’s advanced 5nm N5 chips. As pointed out by Tom’s Hardware, MediaTek has already released two 5G smartphone chips powered by TSMC technology – the Dimensity 8100 and the Dimensity 8000, while Nvidia is preparing to use the 4N process for its Hopper data center GPUs and possibly the upcoming Ada Lovelace (RTX 40 series) consumer graphics card.
TSMC has not disclosed any performance and efficiency gains that its N2 chips might bring. However, the upcoming 3nm process is said to bring a 15% increase in performance and a 30% reduction in power consumption, so we could see a similar upgrade for the next-generation technology.
