Nm size has been a horrible way or measurement for chips for years.

Why?

Hasn't really kept up with advances and now is abused in marketing material. i.e. some aren't even 2nm (classic nm measurment) even when 2nm is in the name.

I'm not in the field but my understanding is the various companies measure it slightly differently because of difficulty measuring on a nm scale. Also not all transistors are the same. Plus you can do a lot of other work which makes it faster than just shrinking the transistor size. Similar to cpu being much faster today then they were in 2015 despite having a similar clock rate(maxing out at between 3.5-4.5 ghz). Stuff like improved branch prediction, better use of simd, improvements in the address translation layer.

I'll try to find a article but a lot of the inner details are nit made public.