Nobody yet really knows how fixed wireless enabled by a 5G network will compare, in terms of deployment cost, with fiber to home costs, except to say virtually everyone expects that cost to be less than FTTH.
The issue is “how much lower, per potential passing,” those costs will prove to be. Many of the potential data points (fixed wireless to high rise buildings; mesh networks for business or consumers) are so different from potential ubiquitous 5G small cell deployments that those other examples are not so useful. Nor can the other deployments fully capture the costs of dealing with line of sight impediments, when small cells might be deployed very densely, perhaps on every other light pole.
What is really different about dense small cell networks is that, for the first time, the total cost of the infrastructure might be dominated by the cost of the trunking network, not the radio access network.
It might not be unreasonable to assume that a fiber-deep network of the sort Verizon is building, which might be called a “fiber to light pole” deployment, would be less than, or close to, the cost of a fiber to node architecture. According to Nokia estimates, the trunking network should cost less than half the cost of fiber to the home, and conceivably just a quarter of FTTH cost.