You have a computer with servos and winches fly parafoil kites to generate power from wind. This is definitely the theoretically optimal answer. It's also very low capital intensity.
Lots of omitted details there. A kite can pull its tether from one location to another location of lower potential energy, but how does it get back to the higher-energy location? If we're talking about a circular train on a circular track, half of the kites must continually have either a lower-drag configuration or a lower-wind altitude. That's plausible, but that's a lot of additional mechanisms, plus a lot of "cancelled-out" wind traction, plus a lot of friction as the train rolls around the track. Will there be power left over for electricity generation? Probably, but it isn't clear that it will be worth all the trouble. If a reasonable prototype exists, it should have been in the video.