Because R&D isn't all that's needed. It takes time to grow an industry. It's one thing to project future economies of scale, quite another to demonstrate that those economies are already being realized.

IIRC, battery prices are on track to fall by another 50% by 2023 or so. That will be a game-changer, and there's no way it could have happened through R&D alone.

Renewable power plants/companies receive both tax incentives and R&D subsidies from state and federal government, and there is research being done directly by the DOE at national labs, as well as at universities etc. from DOE grants. There are also private companies that get DOE research grants for a range of pure through applied research.

It's unlikely that grid-scale renewables could be developed in a wholly R&D environment then deployed at once. And then what would the government do? License or auction technology that hasn't been field tested at scale? Utilities, local governments and private companies would still probably need intermediaries to set things up. So it makes sense to provide some direct funding and some tax incentives to allow companies to do their own R&D, collaborate with government researchers, and work with utilities to bring systems online. Market disruption due to sudden, untested new entrants is less appealing with utility power than with phones--it makes sense to work for a gradual transition.

As far as consumer-scale power generation (rooftop solar or whatever), I think what you've suggested is pretty much what's happening, but I don't follow it. (I'm not a homeowner but have done DOE funded geothermal research with a private R&D company, Altarock LLC.)

edit: Electric cars are also abstracted from the type of power plant that produces the energy, unlike internal combusion cars, so having more of them is more amenable to a somewhat rapidly evolving power generation landscape.