Physical activity is widely promoted as a beneficial component of intervention for children with autism spectrum disorder, with numerous studies demonstrating improvements in motor coordination, balance, and strength. However, emerging evidence suggests that motor gains may not fully reflect the child’s overall experience during activity. This critical review examines potential negative or paradoxical outcomes that may accompany physical activity participation despite measurable improvements in motor skills. Drawing on sensory processing research, exercise physiology, and findings from related developmental conditions, the review identifies several pathways through which activity may impose hidden burdens. These include sensory overload, emotional dysregulation, performance-related anxiety, autonomic strain, fatigue, and increased vulnerability associated with co-occurring attention deficits, motor coordination difficulties, or dyspraxia. Methodological limitations in the existing literature, such as a focus on positive outcomes, underreporting of adverse events, and insufficient characterization of participant heterogeneity, further obscure these risks. To address this gap, the Dual-Outcome Model is proposed, emphasizing that motor competence and psychophysiological load represent distinct developmental trajectories that may diverge during intervention. Recognizing this duality underscores the importance of individualized program design, comprehensive monitoring, and multidimensional assessment to ensure that physical activity supports both motor development and overall well-being. This perspective encourages more balanced, child-centered approaches to physical activity in autism research and clinical practice.