TSMC schedules 1.4nm mass production for 2028, targets sub-1nm pilots in 2029
submitted by
edited
https://biz.chosun.com/en/en-it/2026/04/20/LBMB76R2VZHFBMTGDJJFQ25PBE/
Taiwan’s TSMC, the world’s largest foundry (contract semiconductor manufacturer), plans to begin mass production of 1.4-nanometer (nanometer, one-billionth of a meter) process semiconductors in 2028, Taiwanese media including the China Times reported on the 20th.
ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86
Share on Mastodon
Reminder that the “nm"-node names do not correlate to any physical feature size directly and have been mostly marketing terms for the last ~20 years. Nowadays, “1.4nm” just means vaguely better than the “2nm” process.
On the “3nm"-node, wikipedia
Still impressive, of course, but quite misleading.
Can’t wait to never be able to afford it
jeez like aren’t we in just the 3 figure atom point then. Like they might as well measure with number of atoms at this point.
1nm is approximately 5 silicon atoms
I don’t think they are anywhere near 1nm in terms of semiconductor feature size:
https://techlevated.com/evolution-of-metal-pitch-in-semiconductor-transistors/
Wikipedia seems to imply that that 1 nm class fabrication processes will have a 16 nm metal pitch:
https://en.wikipedia.org/wiki/2_nm_process
Don’t get me wrong, it’s mind boggling that we operate on such scales and that semiconductors that use these leading edge processes are somewhat accessible (even with the massive increase in price due to data centre demand). But we’re not anywhere near actual 1nm features.
thats one direction. Presumably this is at least a circle or square if not a cube or sphere.
Right, it’s linear just like the headline. It didn’t say 1.4 nm^2 or nm^3
So its no big deal…. /s