As already written in my previous comment, 5G is not just about faster internet. Other use-cases are as important, if not more. Operators know they can't charge consumers extra.
You can put a mmwave basestation in a factory (a couple of cells near the ceiling) and controll everything (machines, "robots", transporters, forklifts, inventory,...) via LTE.
1) 5G improves wireless internet - though it only works over short-range and can't go through walls.
2) 5G needs a massive deployment of new devices to reach consumers - though most consumers are not willing to pay even $1 more for 5G.
3) 5G is a prerequisite for amazing new tech like IoT and self-driving cars - many of which already exist and will continue to use existing 3G and 4G bands.
"Dubious" is actually me being generous!
5G reuses existing 3G and 4G frequencies, but with better rf management (among other things). The mmwave frequencies are to address mass-crowd scenarios, which currently today already cause 4G DAS deployments to almost burst at the seams.
These are not "my" opinions. These are the current state of affairs re RF management and planning.
Telcos are blowing $26 billion on this new tech, so I think it's important to question how they're going to make that money back. If you read how the industry itself answers this question, hint: it's not self-driving cars. It is 5G network slicing. The current wireless infrastructure is built on net neutrality, so they cannot differentiate traffic without rebuilding it all from scratch. That is the real reason behind the $26 billion rollout and millions of new devices. They need to rebuild the internet to charge enterprise customers high prices and 5G is just how they decided to market this project to the public. They need to do this before the political environment changes or else they lose their opportunity.
That sounds like marketing hype from an ISP.
A self-driving car that is dependent on a radio network to operate is by defonition, not self-driving.