This is the opposite of the common carrier model. In the common carrier model, there are many companies that you can choose from as a consumer from the get go, because the network is unbundled from a single company.
You're missing the point: the last mile network is shared across all of those companies. Ergo, the company that actually owns the last mile network has no competition on that last mile network and no incentive to upgrade that network.
Firstly, the last mile company still has QoS agreements with the companies that buy from it, and they can plausibly choose to simply go elsewhere if the last mile company does not give them good enough QoS terms; whether that means building their own network, or simply pulling out of that area. Basically, a small group of companies can more plausibly apply joint pressure to a single company than many millions of people can, purely because of coordination costs.
Secondly, part of the common carrier model is that if the government is not convinced that the common carrier is acting in the public's best interest - for example, if they are delaying needed upgrades - then the government can simply confiscate the common carrier's assets, and sell them to a company that will actually meet its QoS and upgrade obligations, or spin the whole thing off as a SoE.