If you end up setting Solar Farms, you're grabbing land from Nature. This way you can re-purpose the land already used (roads, footpaths, Highways) for more than moving person from point A to point B.

Also, it allows roads to be productive even when there's no traffic.

This is interesting because contrary to popular science which says that we can power the whole world with a tiny fraction of land use dedicated to solar panels, surface area is actually one of the biggest challenges we have in a 99% sustainable energy world that we have to get to.

Check out without the hot air, free book by physicist McKay at Cambridge. He's done a 1h presentation at Harvard which tells you the gist of it, and a 15 minutes TED talk which I'd skip unless you really only have 15 minutes. He covers the surface area challenges of solar and other sustainable energies quite well.

The ability then to one day put extremely cheap solar (e.g. at least 1 order of magnitude cheaper than today) in every new road (whose lifespan is a few decades, so we could on paper replace them all halfway through the century), is very interesting.

Of course there are huge, huge drawbacks. But that's not necessarily because it's impossible, but because we have path dependency. That's why you need prototypes and R&D to see how we can build roads sustainably using new materials, and whether solar panels fit into that picture.

[0] Roads are there for a reason: transport. We're now seeing for the first time ever tools on a scale that can cut down transport on a global scale. Still immature, but it's getting there. The combination of internet, 3d printing and virtual/augmented reality, means we can live global lives locally without having to physically transfer ourselves, information or products. It means we can work & study remotely better and experience entertainment and tourism more locally. And when we do move stuff, there's the option of doing it through new channels (air, with drones), or more efficient channels (self-driving vehicles that can attain higher speeds with smaller gaps safely, calculate more efficient routes and turn transport into a commodity: smaller vehicles transporting people, as opposed to cars being branded products, all of which lead to far fewer roads being necessary). It's a very bold claim but I wouldn't be surprised if roads kept explosively increasing until 2025, and then stagnating and at some point sharply reducing after a shift in human culture, manufacturing and transportation.

Prices of solar cells drop fast. Extrapolate a few years, and costs of solar installations will be dominated not by what solar cells cost, but by what it costs to install them.

In this case, something must be installed anyways to build the cycle path. It might well be that installing a (cycle path, solar cells) combination will only be marginally more expensive than installing a traditional cycle path.

Will we get there? If solar cells and the electronics needed to wire than together (which, in this case, are more complex because the road may see highly variable shading patterns) get dirt cheap, we might.