In a June 12, 2024 interview, JPMorgan analyst Michael Cembalest warned that U.S. data‑center build‑outs could stall after the November 2024 midterm elections, leaving a projected 30% shortfall in capacity needed for the AI boom expected to peak in 2028. He cited a $450 billion projected spend on AI compute infrastructure through 2028 and noted that three major cloud providers—Amazon, Microsoft, and Google—have already earmarked $120 billion for new hyperscale facilities, but permitting delays could push completion dates past the critical 2028 demand window.
A capacity gap would force AI startups and enterprises to pay premium rates for limited compute, inflating operating costs and potentially slowing AI product rollouts across sectors like fintech, healthcare, and autonomous vehicles. The bottleneck also creates a strategic opening for alternative providers, such as edge‑compute firms and sovereign cloud operators, to capture market share if they can navigate regulatory hurdles faster.