Morgan Stanley: Factors such as power and labor bottlenecks will continue to constrain computing power supply in the coming years.

Morgan Stanley's U.S. thematic research strategist, Michelle Weaver, recently stated that the application of AI by enterprises is accelerating, but the limited supply of computing power remains a bottleneck for industry growth. "Currently, the supply of computing power is severely lacking. Computing power is becoming a scarce resource," Weaver said during a media interview on Wednesday. She pointed out that the computing power bottleneck is primarily due to two reasons: the lack of labor needed to build data centers, and insufficient electricity to power the data centers. AI data centers consume a significant amount of energy, and developing new power generation facilities, transmission networks, and related infrastructure takes several years. Even considering innovative power solutions such as repurposing Bitcoin mining facilities and fuel cells, Weaver estimates there is still a power shortfall of 10% to 20%, which means that computing power will remain a limited and high-value resource in the coming years. This implies that even if companies have sufficient funds to purchase AI chips and build data centers, a lack of electricity or labor could still hinder the actual deployment of computing power. In addition to insufficient electricity and labor, political factors also pose challenges to the expansion of AI computing power supply. Weaver further noted that as the midterm elections approach, the rising anti-data center sentiment is presenting challenges for the industry.