Efficiency 0.00001%
Quite pessimistic. Given that consumer grade laser diodes are ~30% efficient and solar cells ~20% efficient, and specialized products (like solar cells working at narrow wavelength) are much better than this.
And the losses though absorption and scattering caused by gas molecules and aerosols, atmospheric turbulence, thermal blooming etc?
We all know that during good weather, loss of light in typical drone distances is some single-digit percents, maybe just barely double-digit. Otherwise we could not
see the drone properly, it would appear darker or fuzzy. If you can see it, you totally can power it.
I don't know enough to say what happens when you significantly exceed the power density of sunlight (1000W/m^2), i.e., thermal blooming. Does it become non-negligible at 10kW/m^2? 100kW/m^2? 1MW/m^2?
During colossally bad weather, the problem with absorption and scattering in water droplets is obvious. On the other hand, it is not set in stone that a drone that can only fly during good weather is useless. Quite the opposite, most small commercial aircraft is only allowed to fly during so-called visual flying conditions, and yet we do have those airplanes. Whenever visual flying is possible, powering a drone is, too. Line-of-sight is necessary, but I think that is obvious.
Then it's matter of if we have really good weather (say, >90% transfer medium (air) efficiency), or marginal weather (maybe 10% efficiency).
Saying that efficiency is 0.000001% unconditionally, is just bollocks. Maybe that would be the worst case during a rainstorm.