The Jevons Paradox suggests that as AI intelligence becomes cheaper and faster, it will create millions of new tasks and workloads, leading to a surge in overall demand for AI infrastructure. A shift in market share from high-margin frontier labs to cheaper, more efficient models would increase ROI for end customers, further driving incremental token demand. This would redistribute economic value from model providers to infrastructure providers, favoring those with the lowest per-token cost and highest token efficiency.