1. The rapid development of AI software is outpacing the construction and powering of data centers, leading to uncertainty about future AI demand and potential supply glut.
2. Major tech companies like Oracle, Meta, and others have made substantial financial commitments towards AI infrastructure, with billions pledged in investments. However, there's mixed evidence on actual AI adoption by businesses.
3. There are concerns regarding the capacity of existing data centers to handle the power demands of advanced AI chips, and potential infrastructure bottlenecks due to limited data center space and slow grid adaptability.