A warning has emerged that inadequate electricity infrastructure could become the biggest obstacle to AI industry growth, even as enormous sums flow into the sector. In a report released on October 19, local time, Goldman Sachs said electricity infrastructure expansion is failing to keep pace with the AI industry's rapid growth. According to the report, global data center electricity demand is expected to surge 170% from 2025 to 2030. More than 60% of the increase is expected to occur in the United States. About 2,600 gigawatts (GW) of generation and energy storage projects in the United States are currently awaiting grid connections. That is more than twice the country's existing generation capacity. However, the median time from applying for a grid connection to actual commercial operation is about 5 years. New high-voltage transmission line construction in the United States declined by about 80%, from an annual average of 1,700 miles, approximately 2,736 km, in 2010 to 2014 to 350 miles, approximately 563 km, in 2020 to 2023. Electricity infrastructure expansion is also being delayed by permitting procedures, including environmental impact assessment, which take 3 to 4 years, and lead times of 3 to 5 years for high-voltage substations and large transformers. Opposition from local residents is another obstacle to infrastructure expansion. According to Goldman Sachs, 86 related projects have been canceled or significantly delayed since January 2026, of which 76 were directly linked to local opposition. Residents were found to be concerned about noise, traffic congestion, increased water consumption and potential electricity rate increases associated with construction such as data centers. Goldman Sachs said these combined constraints could make the pace of AI industry growth dependent on actual electricity supply capacity rather than the scale of semiconductor or data center investment. Delays in electricity infrastructure expansion could push back data center construction schedules and disrupt investment recovery for companies that have committed substantial funds. Goldman Sachs estimated that about $7.6 trillion, equivalent to 10,083 trillion won, would be invested in AI computing, data centers and electricity infrastructure from 2026 to 2031. In particular, it expected major technology companies, including Amazon, Google parent Alphabet and Microsoft, to issue $420 billion in bonds in 2027. That would be more than 60% above the estimate for 2026. Jason Tofsky, head of global digital infrastructure investment banking at Goldman Sachs, said, “There are dozens of issues to resolve in building data centers, including financing, but securing electricity is likely to be the most persistent constraint.”
Original Korean article: ETNews
English brief by Korea Wind Intelligence, koreawindintel.com
