Your comparing the last few decades of internet use with the last few decades of water use, a fairer comparison would be the first few decades of water use with

Your comparing the last few decades of internet use with the last few decades of water use, a fairer comparison would be the first few decades of water use with the the first few decades of internet use. More relevant to the point though, would be to compare it with the growth in internet usage a decade or two from now: as technologies mature it would would seem reasonable to expect that the growth tails off and QoS on the internet last mile will become just as useful as QoS on water pipes.