Evolution happens faster than you expect.
Google Fiber is steering away from the massive capital investment required to build fiber to the home networks – even just in cherry picked fiberhoods – and going after targets of opportunity where someone else is paying for the glass. This week it’s signed a deal with Huntsville, Alabama to be the anchor service provider on a fiber to the premise system that the city will build and followed it up with an announcement that it’ll be using other people’s fiber to offer a very limited kind of service in San Francisco…
By using existing fiber to connect some apartments and condos, as we’ve done before, we can bring service to residents more quickly. This approach will allow us to serve a portion of San Francisco, complementing the City’s ongoing efforts to bring abundant, high-speed Internet to the City by the Bay…
We have a lot of work to do before we can offer details on service and timing, or identify the specific condos, apartments, and affordable housing properties that will be connected.
The “existing fiber” isn’t specifically identified, but it’s a good bet that at least some of it belongs to the City and County of San Francisco, which has an extensive muni network. But that’s not the only option – the full alphabet of fiber companies, from AT&T to Zayo, have built networks there.
Google is never forthcoming about its fiber plans, and San Francisco is no exception. In reading through the articles posted about the project, it’s clear that Google has no intention of saying anything other than wait and see. Responses to reporters’ questions were limited to boilerplate. “Before we build a brand new network in a city, we first work through a checklist process” blah blah blah one article quoted a Google spokesbot.
But it is a radically different business model, one that can be quickly expanded anywhere inexpensive fiber can be found: lease backbone fiber in a city where high capacity backhaul is sufficiently cheap, and either spend a few thousand dollars to build a lateral to subscriber-dense multi-dwelling (or multi-business) properties or convince the landlord to pick up the tab. Then it’s just a question of installing internal wiring and inexpensive electronics in the building and signing up subscribers. The capital outlay – regardless of who pays it – will be under $1,000 per unit passed and probably less than $500. That’s a bargain.