40% the cost is hardly nothing, and it's quite possible Google has arranged for lower costs. Also, the fiber stays even after the current owners leave, so Google might just want ubiquity and be willing to take a small hit to get it.
FTTH install cost is entirely dependent on the territory.
Overhead vs underground, apartment buildings vs. single family homes, number of directional bores (to get under streets), miles of sidewalk torn up to lay cable, etc. etc.
It's expensive and how expensive all depends on the neighborhood.
Consider this: you can't do $700 to just one home. That's the AVERAGE cost once you've decided to do a neighborhood. You sometimes need to do those homes 1000 at a time. SO: network build >= $700K minimum.
Which Google makes in... about ten minutes (according to 2011 revenue of ~$38B). That's a rounding error if I've ever seen one, never mind the future revenue it will bring them: some customers will be a straight $300, the rest will be $70-120/m recurring on the fiber service alone, plus the added use of their services bringing in advertising dollars, etc.
Heh. Well you've obviously got it all figured out.
Just keep this in mind: running a telecom network that requires lots of field personnel isn't something that GOOG has a proven capability in. For their money, their BEST bet would have been to BUY a cable company for the network. They get all of the customers in one fell swoop along with a network that can be upgraded to FTTH or low-split HFC.
One city doesn't a telecom company make. I would be VERY surprised if the monthly ad revenue pulled in by a broadband subscriber will cover the lifetime customer acquisition cost.
