The rapid expansion of AI is creating a new bottleneck for tech giants, as soaring data centre demand threatens to overextend the US’s energy supply and infrastructure capabilities by 2030.
Artificial intelligence is creating a second bottleneck for the world’s biggest technology groups: not computing chips, but the power and infrastructure needed to keep them running. Alphabet, Microsoft, Amazon and Meta are spending heavily on new facilities, while the supporting systems , electricity supply, land, cooling and grid access , are becoming increasingly difficult to secure.
According to estimates highlighted by Seeking Alpha, the number of large-scale data centres in the United States could almost triple by 2030 as investment in AI infrastructure accelerates. That view is reinforced by the Electric Power Research Institute, which says data centres could account for between 9% and 17% of U.S. electricity consumption by the end of the decade, more than double today’s level.
S&P Global has separately projected that U.S. data-centre power demand will rise by 22% by the end of 2025 and nearly triple by 2030. McKinsey’s analysis also points to rapid growth, saying U.S. data-centre demand may expand by 20% to 25% a year through 2030, while global capacity requirements could more than triple over the same period.
For investors, the implication is clear. Nvidia may benefit from rising demand for accelerators and servers, but the hyperscalers face a much larger capital burden before AI revenues fully materialise. McKinsey has warned that data centres could consume 11% to 12% of total U.S. power demand by 2030, underlining how the AI boom is shifting from a chip story to an energy and infrastructure test.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
