The investment surge in the AI computing power sector shows no signs of slowing down, with total investments in AI infrastructure projected to soar to nearly $7 trillion in the coming years. Yet, following the current trajectory, by 2030, over $3.5 trillion of these investments in AI infrastructure could be at risk of asset impairment, a figure that might even surpass $4 trillion by 2032. The industry predicts that achieving a 12% return on invested capital may not be feasible until at least the end of 2032. Concurrently, due to cost pressures and the prevalence of similar business models, more than 80% of emerging AI cloud service providers may find it impossible to sustain independent operations by 2030. Nevertheless, the demand for AI computing power remains strong, and related investments are unlikely to cease abruptly, nor will there be a wave of industry consolidation in the near future. However, this scenario may lead to future challenges, such as volatility in AI service prices and a restricted range of supplier options.