Logic Density (may be inaccurate, also it's not the only metric for performance ): Raipidus 2nm ≈ TSMC N2 > TSMC N3B > TSMC N3E/P > Intel 18A ≈ Samaung 3GAP
But 18A/20A already has PowerVia, while TSMC will implement Backside Power Delivery in A16 (next generation of N2)
Therefore, N2 is 2 generations ahead of 18A by definition.
Therefore power advantage you’re referring to is backside power delivery. This is a need long term but it doesn’t bring as much power benefits as you think. TSMC’s customers don’t appear to demand this feature as highly since it’s been getting pushed back to 16A. The design change for designers don’t seem worth it yet.
As for comparison between the two: According to TechInsights, Intel's 18A could offer higher performance, whereas TSMC's N2 may provide higher transistor density - [1]
[0] - https://www.tomshardware.com/pc-components/cpus/intel-announ...
[1] - https://www.tomshardware.com/tech-industry/intels-18a-and-ts...
They skipped 5nm and 3nm, and that is indeed an accomplishment.
I hope the yeilds are high.
They have Intel 7, Intel 4, Intel 3 nodes. Anyways, Intel's names do not equal to the same number on TSMC. They're usually 1 or 1.5 generation behind with the same name.
So Intel 3 would be something like TSMC N6.