Agreed -- as someone living semi-rural (rural, but not middle-of-nowhere) I've often wondered about the economics of a fibre rollout, our only choices at the moment are very pricey wireless (for the level of service)
I have some familiarity with this from Australia's FTTH (now FTTx) rollout:
Since you mention that you live in a semi-rural area, one hugely important factor is the density of premises along a street. The less dense an area, the fewer potential customers you can "pass" by installing a new fibre cable down a street. Not all premises that you pass with a cable will want to sign up and pay for a fibre plan.
Construction can be made dramatically cheaper if it is possible to re-use existing infrastructure. This is important, as if you are laying underground fibre then the cost of digging new trenches (~AUD$100/m) is a huge component of the total construction cost. For example, when installing underground fibre, if there is existing underground duct with capacity that can be used then it is far cheaper to re-use that (~AUD$20/m) than digging new trenches. Similarly, if there are existing poles that can be rented then installing aerial cables (~AUD$30/m) is cheaper than digging trenches, but more expensive than using existing underground duct [1].
Another way to re-use existing infrastructure to reduce construction cost is to make use of existing cable, such as copper lines or HFC, instead of laying fibre over the full path from the exchange to the premises. The quality/speed of the connection will not be as good as a pure fibre connection, particularly if the existing infrastructure is old/degraded. These approaches are now being favoured in Australia's national broadband project [2].
[1] note that in the U.S. labour costs are lower than in Australia, so don't take these cost-per metre estimates too seriously. These costs exclude the cost of the fibre itself, which is typically a few dollars per metre.
[2] personally I think it is short-sighted to embark on a major infrastructure project and then try to save money by avoiding building the new infrastructure, and instead re-use the existing degraded infrastructure. These decisions were hugely influenced by political factors after a change of federal government, i.e., the old plan was necessarily bad and needed changing because the previous government started it.
We are ~10Km from the nearest village that likely has decent bandwidth, though no idea on the backhaul available from there.
Still 10K @ $30 (AUD/CAD)\m would be $300K+ to put it along existing hydro infrastructure. Wonder what the likelihood of getting 100 households to put up say $3-5K in installation costs would be? That would represent a decent fraction of those reachable along that stretch, and those nearer in may already have DSL.
Very crudely, you might want to multiply that $300k guesstimate by a factor of 2 or 3 to take into account the cost of cabling off the main stretch, splicing, splitters, installing "lead-ins" / "drops" to each household, etc. Although if you were only serving a hundred households you could plausibly serve them all from a single large fibre cable along the main stretch by "pulling off" a couple of fibres into a smaller cable per household. This wouldn't require any optical splitting gear, so that might make things a bit cheaper.
For reference, the overall construction costs for Australia's FTTH rollout was somewhere around AUD$1000 - AUD$2000 / premises -- probably toward the high end of that scale, but the order of magnitude is right. This would underestimate the cost for a semi-rural area. In Australia's project they use fixed wireless / satellite for some areas where FTTx was not cost-effective.
Another thing to keep in mind is the maximum optical path distance -- from memory this is probably somewhere around the 20km mark from the exchange, probably a bit less, will depend upon the technology used. That still gives you a ~10km buffer region of potentially reachable households around your main fibre-atop-hydro stretch.
It's pretty awesome actually. Montana's in a lucky spot; our power company in the 90's got deregulated and decided they wanted to be the next WorldCom. They threw down a Billion dollars of assets down, went bankrupt, and now Montana has an excess of backhaul fiber. As such we are able to buy bandwidth at reasonable rates. When combined with peering at the SIX in Seattle, we can provide a very nice service for rather cheap compared to other rural locations. That, and we are designing our network more like a datacenter which has greatly reduced costs.
But not in a position to move to Montana!