The ability to support is an upper limit on penetration, not a penetration estimate. Compare "US homes passed by cable TV" to "US homes connected to cable TV" (the former figure is - ballparking since I do not have my reference handy - about twice as large as the latter).

Depending on local geography, I could easily imagine that the difference in infrastructure rollout required for supporting 20 % is not materially greater than that required to support 5 %. In that case, any demand forecast between 5 and 20 % would result in essentially the same infrastructure rollout.

- Jake

Friends come and go. Enemies accumulate.