Good point, but I think it is interesting to wonder why P2P so rarely succeeds at delivering the huge efficiency that's suggested by theory.
As their service grows, Spotify should be liking P2P superlinearly more. Since they and other services are apparently not, it something is wrong with my conception of the economics of the internet.
> As their service grows, Spotify should be liking P2P superlinearly more.
Only if you assume that the price of bandwidth is linear. It is suggested in other comments that when you buy bandwidth in bulk you can get large discounts. If I don't misremember, back in the early days of Youtube, they came to a point where the network operators wanted to get access to Youtube rather than the other way around.
Also, as a company grows and get access to more capital they might be able to invest in other infrastructure that is needed (other than just bandwidth) that combined with the bulk prices, make the opportunity cost of developing the p2p system larger than the opportunity gains.
Unless there are hidden costs to maintaining a P2P architecture. Support costs may grow superlinearly, or the company might decide they want (or need) more central control.
As their service grows, Spotify should be liking P2P superlinearly more. Since they and other services are apparently not, it something is wrong with my conception of the economics of the internet.