Back when LTE was rolled out initially the competitor was wimax.

It was around that same timeframe that "TV Whitespace" was going to become the next big thing.

Anyway, LTE should be the literal last option. It requires more than 2x as many towers as fixed wireless, with gear more than 20x more expensive. That's also more than 2x-3x the required amount of of battery backup systems, networking equipment, and land / tower leases.

If you have extreme density, you NEED fiber and you need WiFi. You extend from the fiber network with extremely high quality ngFW. To fill gaps, use satellite.

Fiber requires a certain density of subscriber/mile(km), the same as any technology.

Even with 0 labor cost, you still need to get conduit in the ground (materials), fiber, terminations, switching, routing, OLT/ONT cost, handholes, any permitting or utilities location, horizontal boring equipment , jackhammers, splicers, etc. The upfront cost is many, many, many times higher for fiber and if you're okay with your cost-per passing being more than you would ever make on customer ARPU, then sure do that. Even if labor cost was 0. And it will take YEARS longer to deploy and see a return on investment from, of ever.

It doesn't matter if there's broadband to the location if nobody at the location can afford it.

If you want broadband, LTE is the worst option.

Nowhere with even just an "improved" road (i.e., gravel road, not "only" a path cleared of tall vegetation) is too low density for fiber.

Unless local conditions make you want to use aerial cable, you'll just cable plow a speed pipe and put in a small access riser every 2~3 miles. You blow the cable in segment-by-segment, either splicing at these locations or spooling the ongoing length up before moving the blower and doing the next segment.

If the cable is damaged you measure with OTDR where the break is, walk there with a shovel, some spare speedpipe, and two speedpipe connectors. You dig out the damage, cut it out, put good pipe in, join it to the open ends where you cut the damaged section out, and bury it while taking more care to make it last better this time. You pull/blow out the section of cable and blow in a fresh one, splice it to the existing cable and both ends of the segment, and the connection is fixed.

AFAIK cable plow for fiber in not-very-hard ground is cheaper than planting "telegraph" poles like they did in the old days.

The only expensive parts about fiber optic Internet are the machine that allows you to splice (about 1k$, unlike the 5$ LSA tool for attaching RJ-45 sockets to Cat.5/6/7 cable; this only blocks DIYers from easily doing it) and digging up developed area with more finely controlled tools than a literal plow if you forgot to put in speed pipe the last time the ground was dug up for any infrastructure at all (say, piped water).

Oh, and arguably the optics if you expect to be cheaper than copper on distances within a building at speeds under 10 Gbit/s.

Are we talking about Mumbai or an area w/ 0.2 homes per 10sq km? Because I'm talking about how to do both. Vastly different challenges and economic viability, and I have experience doing both types of environments.

Mumbai doesn't have issues of arguably too low density to split digging costs across enough users.

It might have issues with the cost of digging, though.

It's the "homestead in Maine" type of situation where the fiber plow enables economic viability of burial.