The problem is the gap between the RTX3080 and RTX3080TI is actually quite large. The rated board powers are nearly the same,with a 17% reduction in shaders and only 40% of the VRAM quantity clocked lower. Unless realworld measurements show otherwise,the RTX3080 looks like it might be bottom barrel leaky silicon,pushed somewhat past its performance/watt sweatspot.
If there is a RTX3080TI that only reinforces that the RTX3080 silicon isn't the best!
So it will be interesting to see what actually reviews show. Is the RTX3080 board power rated too high,or is the RTX3090 board power optimistic??
I actually don't have much faith in Samsung,as their 14NM node was essentially used by GF to make Vega,and it was 2/2 disaster. At least Nvidia's uarch design is better!
Would be funny if it is TSMC 7NM - would be a proper jebait by Nvidia!