Some thoughts I've had rattling in my head about what I expect for the ISP market in the US assuming we continue down the current state where companies own and operate their own independent last mile infrastructure.
To begin, I think it's good to identify the current state of their infrastructure management. They are trying to bleed every last bit of revenue out of existing infrastructure. They are delaying the inevitable but the more they bleed they hope the more they have built up in the coffers for the battle ahead.
The battle that is coming will be playing a game of chicken in each market they serve between themselves and every other provider also operating on old infrastructure. However in this game of chicken it's not necessarily the first one to blink that will lose. That "blink" is when one provider finally throws in the towel and starts their overbuild for FTTP. This means they have decided that trying to bleed any more revenue out of their dated infrastructure is actually going to cost them more or lose enough customers it's in their best interest to replace it.
The problem for all of their competitors is do they jump in as well and begin the overbuild to state competitive, or wait and see if their competitor falls flat and potentially get their infrastructure for a steal. I have my doubts that many will join in on overbuilding at the same time due to the cost. What I think is most likely to happen is the competitors who waited will just do the minimum to sustain the existing infrastructure until they either lose enough customers to abandon the market, or they are able to reap that competitors infrastructure if they fail. If they are able to get the competitors infrastructure, which they didn't do the investment into building, their finances come out looking great.
What I see this heading is an even less competitive market than we see today. The current competitiveness has rarely been the result of competing Internet services. DSL and DOCSIS are both adaptations of transmitting data services over mediums that were originally intended for non-overlapping services, phone and video. This is in significant contrast to the shift which happens with the migration to fiber to the premises, because those technologies have been developed from the ground up as a data delivery service first.
This is just my current hypothesis, and its mostly based on what I've been reading in the news and what I've seen occurring where I live. Locally we have a municipal ISP who converted from DOCSIS to GPON (FTTP), and Mediacom and CenturyLink are the competitors. Working for the municipal ISP it's been pretty clear that Mediacom is doing pretty minimal work to maintain their infrastructure.
The only time Mediacom appears to try competing is when it's multi-dwelling units which sign term contracts that ensure they will get income for a while. We have had some of these come back to us after having horrible experiences with Mediacom and they realize sticking with them is painful and isn't worth the reduced price.
CenturyLink has even gotten to the point of not building out their copper plant to reach new residential developments. We don't even offer voice service yet, but we have heard they have chose not to build because they expect we will eat the market away shortly after, so investing in the infrastructure isn't going to pay off.
CenturyLink is one of the service providers that has started to overbuild their existing copper plant with FTTP in some markets. I think those will be ones to keep an eye on and see if other ISPs overbuild as well and try to compete, or if they give up and just maintain as long as it's profitable.
Another datapoint is that some of these service providers might be waiting for the right FTTP technology to mature before jumping into the market. PON technologies are at a point where a significant change is just starting with NG-PON2 and XGS-PON nearing general availability. I know Calix is supposed to be bringing these to market later this year, and CenturyLink is one of their big customers. This video demonstration shows just how significant of an upgrade it is.
I want to clarify something he says in the video. He mentions 100Gbps in the access network, and what he means by that is 100Gbps in aggregate, with 50Gbps in each direction.