[jcrawfordor]

1) laying fiber (either trenched or on utility poles) will likely be your biggest expense, it requires a huge, huge investment to be able to put infrastructure through a significant area. Think about feet of conduit a several-person crew can drive per day whether by trenching/trenchless methods - it's not that many. Microtrenching promised to significantly reduce this cost but Google Fiber's experiment with it went famously poorly.

2) ROWs to run cable will need to be negotiated either with the municipality (if laying underground) which can come with a lot of difficult restrictions on work quality, traffic disruption, etc, or with the electrical utility in the case of utility poles in an area with a typical franchise agreement, in which case the utilities are often uninterested in the project and will just generally make your life difficult through slow consideration of engineering proposals, requiring extensive up-front engineering work, etc. In a small town I had some involvement in the electric utility demanded over $1mm up front for engineering surveys on pole attachment - this for a market of ~8k people, and before any actual attachment fees. Completely blew the budget of the potential broadband provider which had planned a total of $3-4mm in up-front.

3) After running infrastructure, providing drops to each house is a fairly costly and disruptive up-front operation per customer (may even be trenching their front yard), which discourages customers signing up with your service when the incumbent providers already have house drops in place. You will also either have to eat this cost or pass it to the customer as an install fee or a term agreement, all of those options are bad in different ways.

4) IPv4 exhaustion has hit new ISPs hard and you are going to have to do CG-NAT. ISPs like to think customers don't care but in practice this is indeed a headache.

[erik_seaberg]

I looked up microtrenching and it looks like they're basically stopping traffic and pointing an angle grinder straight down. Directional boring machines are out there, why aren't we running a bunch of those 24/7? Do they need a lot of supervision?

[jcrawfordor]

While the range of directional trenching is limited, requiring regular access points, the bigger issue is setup. Surveys and tests need to be done to determine if the area is suitable for directional trenching, and it only works well in certain circumstances. It's definitely heavily used in telecom installation but not a panacea.

Microtrenching is extremely simple and fast, but so far I don't know that anyone has nailed durability. Google's Louisville install used microtrenching and was an absolute debacle with the sealant constantly failing and the cables ending up laying on the surface of the pavement. Google ended up shutting down service in Louisville and the cost of repairing the failed microtrenching may have been a big reason why. Certainly get them a lot of bad press and ill will from their customer base.