Residential broadband is a business model that depends heavily on successful bandwidth utilization speculation. However since Netflix is extremely popular (close to half of all internet traffic in the US), the pattern gets extremely predictable and so there is little advantage to be had in the core business of bandwidth arbitrage, since it's being arbitraged "upstream" of residential ISPs and baked into the larger pipe prices.<p>If any service becomes sufficiently popular, the price of providing bandwidth for it to a large subscriber base approaches the characteristics of a dedicated bandwidth circuit for more and more nodes on the network.<p>Of course, Comcast is using it's last mile power as leverage, but the core issue the economics of bandwidth speculation.