This project investigates how competitive and cooperative social dynamics influence user performance, motivation, and engagement in virtual reality exergaming. A bespoke VR cycling application was developed, featuring three distinct cycling scenarios (1) Baseline (solo cycling), (2) Cooperative (cycling with teammates), (3) Competitive (cycling with opponents). Scenarios incorporated adaptive AI and real-time feedback mechanisms to simulate different social environments whilst maintaining consistent exercise parameters. A user study was conducted with 12 participants, capturing quantitative data (power output, heart rate, cadence) and qualitative feedback (intrinsic motivation, perceived exertion, user experience). The findings indicate that competitive dynamics significantly enhance exertion, intrinsic motivation, and performance compared to solo and cooperative modes, though cooperative scenarios also showed benefits over baseline. The project highlights the potential of social mechanics and adaptive AI in improving engagement and personalisation in VRbased fitness, while also identifying areas for refinement in AI behaviour, environmental design, and motion sickness mitigation. These results provide actionable insights for the future development of immersive, engaging, and inclusive VR fitness solutions.