This study aimed to investigate if hip, knee and ankle peak power is influenced by the relative load lifted, altering the joint and system load-power relationship during a countermovement jump (CMJ). Twenty-Three male national representative athletes (age: 20.3 ± 3.1y, squat 1RM: 133.8 ± 24.8kg) completed three CMJs at relative barbell loads of 0, 10, 20, 30 and 40% of an athlete's estimated back squat 1RM. Ground reaction force and joint kinematics were captured using a 16 camera motion capture array integrated with two in ground tri-axial force plates. Hip (x~ = 20%, range 0 > 40%), knee (x~ = 0%, 0 > 20%) and ankle (x~ = 40%, 0 > 40%) peak power was maximized at different percentages of absolute strength, with an athlete dependent variation in load-power profiles observed across all lower-body joints. A decrease in system (body + barbell mass) peak power was significantly (P<0.05, r = 0.45) correlated with a reduction in knee peak power. Timing of instantaneous system and hip peak power occurred significantly closer to toe-off as load increased. The findings highlight that the generation and translation of lower-body joint power is influenced by external load and athlete-dependent traits. This subsequently alters the power-load profile at a system level, explaining the broad spectrums of loads reported to optimize system power during a CMJ. When training, we recommend that a combination of barbell loads based on assorted percentages of the estimated 1RM be prescribed to optimize joint and system power during a CMJ.