People who make chips should try to make them as small as they can. Intel, Apple Supplier Expected to Begin 2nm Chip Production in 2025 Samsung, and TSMC are always trying to make their chip nodes smaller. This makes their processors faster and more efficient. If you want to be the first to try out TSMC’s latest invention, you have to wait in line right now.
The world’s biggest chipmaker, Taiwan Semiconductor Manufacturing (TSMC), plans to make big changes to its process when it starts making 2-nanometer (N2) technology in late 2025. When you think about it, the new iPhone and MacBooks have a processor that is 5nm, and 3nm is coming later this year. If true, Apple will be one of the first to use the new technology, but Intel is also said to be there.
It’s also getting more clear about when TSMC will expand its N2 line in Fab 20 (Hsinchu), an analyst with China Renaissance Securities wrote in a note for clients. This is what the company says: Tool move-in is expected to start in late 2022 and risk production will start later that year. Intel and Apple will be the anchor customers for dedicated capacity support. Intel will be making Lunar Lake client PCs, and Apple will be making iPhones.
Why would Intel, which makes its chips, need to get them from another company? People say Intel is going to use TSMC’s 2nm node in its graphics processing units (GPUs) and other SoCs, and they say Intel will be an early adopter. People say Intel will use TSMC’s 3nm and 2nm process nodes to make its new Arc GPUs, which are aimed at taking on the market leaders Nvidia and AMD.
By revenue, Cupertino company Apple has been the most important customer for TSMC at this time. Apple uses TSMC processors in its phones and tablets as well as M1-powered laptops, which are all made by the company. The new chips have a processor that runs at 5nm, so there will probably be a few more releases before the N2 chips come out in early 2016. Indeed, TSMC’s 3nm chips will arrive later this year, and they’ll likely power the iPhone 14 when it comes out in the fall.
Besides, Intel thinks it can get back to where it was when it was a long way behind other companies. Pat Gelsinger, the CEO of the company, said last year that Intel would be back at the top of the market in 2025. He talked about what was coming out in the next few years, and he put the release of a 2nm (20A) node in 2024. 18A (codename Lunar Lake) is the next release, and it’s said that TSMC’s 2N graphics tiles will be used in it. It’s also said to be on time.
IBM has also made it possible for a 2nm chip to come out soon. Last year, the business announced the introduction of a new 2nm chip using nanosheet technology. This technology packs 50 billion transistors into a chip the size of a fingernail. IBM says that its 2nm process will be available to partner foundries, including Samsung and Intel, by the end of 2024. This could give Intel the chance to get ahead of the game for a short time.
You don’t have to be an Apple or Intel customer to get the new 5nm N5 chips from TSMC, though. TSMC has already made two chips for 5G phones, the Density 8100 and the Density 8000, and Nvidia is planning to use the 4N process for its Hopper data center GPU and the RTX 40-series consumer graphics cards.
TSMC hasn’t said much about the performance and efficiency gains its N2 chips could make. However, the 3nm process, which is coming soon, is said to make 15% more performance and 30% less power, so we could see similar improvements with the next technology.