While neuro-recovery is maximized through active engagement, it has been suggested that the use of robotic exoskeletons in neuro-rehabilitation provides passive therapy. Using oxygen consumption (VO2) as an indicator of energy expenditure, we investigated the metabolic requirements of completing exercises in a free-standing robotic exoskeleton, with 20 healthy and 12 neurologically impaired participants (six with stroke, and six with multiple sclerosis (MS)). Neurological participants were evaluated pre- and post- 12 weeks of twice weekly robotic therapy. Healthy participants were evaluated in, and out of, the exoskeleton. Both groups increased their VO2 level from baseline during exoskeleton-assisted exercise (Healthy: mean change in VO2 = 2.10 ± 1.61 ml/kg/min, p =< 0.001; Neurological: 1.38 ± 1.22, p = 0.002), with a lower predicted mean in the neurological sample (-1.08, 95%CI -2.02, -0.14, p = 0.02). Healthy participants exercised harder out of the exoskeleton than in it (difference in VO2 = 3.50, 95%CI 2.62, 4.38, p =< 0.001). There was no difference in neurological participants' predicted mean VO2 pre- and post- 12 weeks of robotic therapy 0.45, 95%CI -0.20, 1.11, p = 0.15), although subgroup analysis revealed a greater change after 12 weeks of robotic therapy in those with stroke (MS: -0.06, 95%CI -0.78, 0.66, p = 0.85; stroke: 1.00, 95%CI 0.3, 1.69, p = 0.01; difference = 1.06, p = 0.04). Exercise in a free-standing robotic exoskeleton is not passive in healthy or neurologically impaired people, and those with stroke may derive more benefit than those with MS.