Some precisions and corrections:
The drag coefficient rD of the turbines aloft is also called the thrust coefficient which is used for wind turbines or propellers for planes. The thrust coefficient depends of the propeller design https://web.mit.edu/16.unified/www/FALL/thermodynamics/notes/node86.html. On the report it is 1, that is a rather high value.
On http://s3.amazonaws.com/zanran_storage/www.gl-garradhassan.com/ContentPages/107547242.pdf,
Figure 1, the generic thrust coefficient curve goes from 1.5 to 0.3 as the wind speed increases from 4 m/s to 25 m/s, and some thust coefficients for some MW range wind turbines are still lower, about 0.1 for 1-2 MW-class, about 0.2 for 0.5-1 MW-class , at 25 m/s wind speed.
So if the thrust coefficient rD of the turbines aloft is 0.2 with 26.6 m/s apparent wind speed, the kite area could be 5 times smaller, so 44 m² and as always two times the drag of the turbines aloft in order to keep the optimization. The power becomes also 5 times lesser, 75596.68 W then 49137.842 W after cosine loss. The power coefficient of the turbines aloft decreases as their thrust coefficient also decreases. This can be due to the pitch or stall control in wind turbines http://drømstørre.dk/wp-content/wind/miller/windpower%20web/en/tour/wtrb/powerreg.htm .
So Optimization of a soft wing with turbines aloft is now corrected for the part about Makani M600 of which proportions between the turbines aloft and the wing are more appropriate than what I expected. See also some observations on Makani's presentation in AWEC2017.