From 3:07 : comparison between reeling 3 m² plane (2.5% of potential, huge difference between average power and maximum power per cycle) and fly-gen 4 m² (25% of potential, so 10 times more). Was the fly-gen plane more optimized? Not sure.
During reel-in phase the weight of the plane (increasing more as it scales up) must be compensated by lift to allow effective control. Lift leads to some drag, and the apparent wind due to reel-in operation adds to real wind… So huge power consumption is not so surprising, and avoiding it looks difficult.
And as it seems (albeit with a high degree of approximation) that, despite years of research, the average cycle power / average generation phase ratio has hardly changed between the old and recent tests of soft wings in yo-yo mode, as I pointed out, one could deduce that the optimization for energy production of a rigid wing also in yo-yo mode could prove difficult for the less.