The source is this paywalled report, the headline of which notably doesn’t go as far as the techpowerup one, though that might just be to bait people into buying access to the full report.
I know very little about fabrication itself, but it seems like the gist is it’s impressive, but pretty expensive and pushing the limits of what’s realistic with DUV. I wonder what the plan is if China still can’t source (or develop) EUV tech soon?
they should make free electron lasers (i have asked around, seems synchrotron are not bright enough, but lasers have a couple of start ups and stuff doing feasibility studies already) and or do funny stuff with electrons, but those are even more baby-tech level.
it’s impressive if they can scale it, duv requires more and more patterning steps to overcome resolution limits, and fuck up on each one brings pain. (there were some claims when euv was arriving that it was a step like from 30 patterns to 3-5 at same resolution)
I saw someone on reddit claim you can get 0.5x smaller node with 2x multi-patterning steps
probably, i remember numbers floating around 7-5 nm transition of like 25 to 35 patterns change (without euv), so in that area