Morgan Stanley warns that a massive supply-demand imbalance is brewing in the AI sector, forecasting a 55-gigawatt electricity shortfall for US data centers by 2028.
"The market may be seriously underestimating the true explosive power and depth of the AI revolution," Morgan Stanley said in a new report, citing a widening global gap in computing power and electricity.
The bank’s research notes that demand for computing power is set to grow three times faster than Nvidia's projected annual supply growth. This power crunch is a "ticking time bomb," with the projected 55-gigawatt shortfall representing 18 to 30 percent of total deployed capacity even with alternative energy solutions.
The report signals a long-term structural tailwind for AI supply chain companies, including chip makers, data center equipment providers, and energy utilities, as tech giants race to secure power for future growth.
Morgan Stanley's forecast comes as some large language model (LLM) providers are already limiting user access due to capacity constraints. The bank cited data showing global weekly token usage surged 250% in the first quarter of 2026 to 22.7 trillion. As next-generation chips from companies like Nvidia reduce computing costs, this demand is expected to be further unleashed, intensifying the supply shortage.
The energy demand is already straining local grids. According to a report from Food & Water Watch, the largest data centers can consume as much power as 2 million households. This surge in demand led US utilities to request $31 billion in consumer rate increases for 2025, double the total from 2024, with residential rates already up 31 percent since 2020.
The AI revolution's impact is also beginning to appear in the labor market. A survey by the bank in five highly-affected industries, including retail and transportation, found a combined net layoff rate of approximately 4 percent over the past year as 11 percent of jobs were cut.
The report frames the compute and power shortage as a primary constraint on AI development, forcing companies like Meta to invest directly in energy infrastructure. Investors will be watching the capital expenditure plans of major tech and utility companies for signals on how this multi-trillion dollar gap will be filled.
This article is for informational purposes only and does not constitute investment advice.