As I was told by small-block supporters, upstream bandwidth is equally important than downstream for the Bitcoin network to work well and so it is the real bottleneck for the Bitcoin block size.
So-called 'small block supporters'* have an annoying habit of switching to some other argument when asked to provide evidence supporting their previous argument.
* The SegWit Omnibus Changeset allows blocks as large as 4MB.
And their claims are often half-truths that lead to agreement from the great unwashed, despite irrelevancy.
Case in point: of course upstream bandwidth is important. I don't think anyone disputes this. But
what evidence suggests that upstream bandwidth is either a constraint today, or the upstream bandwidth will become a constraint tomorrow?
In an era where bandwidth providers are aggressively building out to accommodate full-rate 4K video bandwidth to every home, bitcoin node traffic -- even if it were to be to every home -- is negligible.