POTS still has alot of life left in it. It just hasn't seen any investment. Long-term it could handle greater overall capacity given each household has a physically dedicated line to an aggregation point.

Phone companies just need to invest in putting the DSL modems closer to the households, say at the neighborhood cabinet rather than the switching station. A literal mile is too far. VDSL2 can do 100Mb/s in each direction at 500 meters, and there's no bandwidth contention with your neighbor. And VDSL2 is extremely mature, hardly the state of the art. By comparison, at the office I pay $450/month for dedicated 100Mb/s fiber service from Cogent. And the POTS infrastructure is just about as reliable--which is to say, very reliable.

I live in San Francisco, which is already extremely dense. AT&T owns the POTS system, but they refuse to put in the investment. My switching station is almost 2 miles away. I finally got fiber from Sonic.net. They're investing in hanging fiber over much of the westerner half of the city at a time when AT&T is hemming and hawing about upgrading their cabinets. And interestingly, my fiber service runs straight to that switching station 2 miles away.

The only reason cable dominates is because the phone companies aren't interested in competing. AT&T and Verizon own a huge amount of the POTS infrastructure, but are entirely focused on mobile and cable. It's a ridiculous state of affairs having so much infrastructure managed by the same oligopoly. When I cancelled Comcast and told them I was only paying $40/month for Gigabit fiber there was a long pause--he was absolutely incredulous. Was it a promo price, he asked. (No.) He knew there was no chance of retaining me as a customer.

VDSL2 is great for the areas where most of the customers are already close enough to benefit. But outside of urban areas, you're still proposing almost the same degree of disruption to the neighborhood as a full fiber deployment. Instead of digging up the road and my yard and putting new cables into my house, you're now merely digging up the road and putting new boxes in every fifth yard. And all to offer downstream service that is almost as fast as what the DOCSIS provider currently delivers, and upstream capacity that most customers won't see much benefit from.

The dedicated last hop bandwidth of DSL sounds nice, but in practice the shared nature of DOCSIS usually isn't a problem and when it is the solution is no more expensive or invasive than what DSL requires up front for every neighborhood.

The vast majority of Americans live in neighborhoods far more dense than that, even deep in the suburbs. I can't find the figures for America, but according to this presentation 100% of Europeans live within 500 meters of their distribution cabinet: http://www.wik.org/fileadmin/Konferenzbeitraege/2007/VDSL-Co...

I agree DOCSIS has been good enough. But with the shift to services like Hulu and NetFlix I would argue that it's nearing a threshold. And even if it weren't, I bet if a telephone company had somebody as highly motivated as Elon Musk, they'd be able to eat their local cable company's lunch.

AT&T is not interested in delivering video over wires. They want to move customers to DirectTV as quickly as possible. There is no marginal cost to deliver video over satellite to an additional customer as far as Infrastructure.