Cognitive dissonance is interesting. For example, tech folks generally love Japan and Hong Kong's railroads. But those systems work so well because of the exact sort of vertical integration techies get apoplectic about when it comes to the Internet:
https://www.mckinsey.com/industries/capital-projects-and-inf.... JR and MTR not only are private companies that own and operate the railroads, they own a ton of the land around the rail stations. They thus can capture value from both sides of a two-sided market--the customers going to a business near a rail station, and the business that benefits from having a rail station near it. That greatly increases incentive to invest in rail compared to a system where value can only be captured from the rider.
Interestingly, vertical integration also increases competition, as is the case in urban areas of Japan. If you can leverage your low-margin transit product to sell your high-margin real estate product, you have vastly more incentive to compete with other firms on the transit product.
Indeed, the vastly lower deployment costs enabled by 5G suggest an alternative vision of the future. We may end up with a system where you subscribe to "Google 5G" or "Facebook 5G" or even "Netflix 5G" service. It will be abhorrent to those who visualize a platonic model of infrastructure separated from content, but may in practice work a lot better than heavy government involvement in the infrastructure layer.