The AI power crisis: a growing challenge for the tech industry

Goldman Sachs has raised concerns about the potential for a power shortage to derail the rapid expansion of AI infrastructure, warning that global data center power demand could rise by 170% between 2025 and 2030, with over 60% of that growth expected in the U.S. The report highlights that the physical infrastructure required to support AI is lagging behind the pace of technological advancement, with the U.S. electrical grid moving at a slower pace than Silicon Valley's innovation. Goldman Sachs argues that the so-called AI race is fundamentally about securing enough power, rather than just developing advanced models. The report also notes that since January, 86 projects have been canceled or delayed, with 76 linked to local opposition. Despite these challenges, slower deployment could help prevent over-investment by aligning spending with actual demand. Meanwhile, utilities are benefiting from the AI boom, transforming from slow-growing businesses into major beneficiaries of the sector's investment. The report underscores the financial risks associated with the massive expansion of AI infrastructure, which is largely financed through debt. If construction timelines slip, the stakes could rise significantly. Goldman Sachs warns that without sufficient power, the growth of AI systems could be hampered, as users will only stay if there is enough energy to support their needs.