It is for now; come back in 10 years and lets see. Keeping fibre in the ground working in the face of storms, other people digging it up and tech churn (ok, what is after 10Gbe ? How are you going to run that round the mainstreet?) isn't rocket science, but it is a technical maintenance task and new build infrastructures have a great upfront advantage. For a while.
Buried fiber works just fine regardless of storms. Aerial fiber is a different matter. Patching up fiber cables after a cut is no big deal. Tech churn, i.e. what active equipment to use, is just run of the mill stuff. You can upgrade the optics to 100G if you need it, but if you don't nobody is forcing you to move past 10G.
I've built and ran fiber networks for over 10 years and operations and maintenance is no big issue.
Super interesting: what fault rate (per 1000 per week) do you see and what kind of geography are you in (urban?)
On the past 10Gbe point : contention is the killer, there is a rise of "super users" for example video editors shipping to and from the cloud, which drives the need for > backhaul - are you not seeing that?
Also on the bigger points; small providers have to peer, like everyone, but do they get good deals? How about prices on kit? Strategic deals with Huweii / Cisco are hard enough for tier one, do you not think that you'll get squeezed ?
Buried fiber doesn't really develop faults once it's put in the ground and doesn't really require any maintenance either.
The deeper you put it the safer it is and when we put in fiber we put it in deep. The rate faults that occur are fiber cuts and these fiber cuts are usually the result of some contractor not using call before you dig or being sloppy about it.
Bandwidth usage is always growing, but then again the cost of IP transit is constantly going down too. Heavy users aren't really an issue and end user 10G connections are rather rare. Those that really need dedicated 10G connections buy leased lines or wavelengths and pay accordingly. More bandwidth usage isn't a problem, it's a good thing that drives more business.
Peering is easy and cheap to small and large providers alike. Transit is also so cheap that it often is cheaper to buy more IP transit than to build out more peering capacity.
Access to networking hardware is not a problem and there are products for every need and price point. The largest issues are access to capital, barriers to entry and competition.