Do you have a source for your claim about the there being a point in density where infrastructure costs stop scaling linearly? Intuitively, economies of scale imply the opposite of what you're saying--that infrastructure becomes less expensive per capita as density increases.