This is correctly observed. The KM1 was designed now more than four years ago. Although we have made incremental improvements, these are mainly related to robustness. Also the software is far from complete in terms of producing full power. The efficiency (power per wing area) should improve with better kite aero design and more refined software. A new kite design is just around the corner with KM2, though depending on the funding situation
Though 3 MW is a hard target for sure, and we did not yet produce 20 kW with KM1 on average. The goal of our flights are currently focused on flight endurance and low/high wind handling, not power output. So expect a lot of low hanging fruit to be sorted out once we are flying for longer durations
Related: Flying a rigid kite with a single tether attachment point Tallak Tveide, KiteMill