Deloitte warned AI data centers could drive a 30-fold surge in power demand by 2035, urging urgent innovation, investment, and cross-sector collaboration.
 |
| Deloitte’s latest report revealed AI data centers will strain U.S. energy grids, calling for $1T in investment and urgent regulatory and technological innovation. Image: Deloitte/ CH |
New York, USA, June 25, 2025:
As artificial intelligence rapidly transforms the global economy, the energy infrastructure underpinning this shift may be woefully unprepared. Deloitte’s latest report warns that AI-driven data centers in the U.S. could see their power demand skyrocket from 4 gigawatts in 2024 to a staggering 123 gigawatts by 2035—representing a more than 30-fold increase.
The study, based on Deloitte’s 2025 AI Infrastructure Survey, highlights the looming crisis: a widening gap between the exponential growth of AI and the pace at which power grids and supporting infrastructure are being developed. Seventy-two percent of surveyed power and data center executives described power and grid capacity constraints as either “very” or “extremely” challenging.
By 2035, AI is expected to account for 70% of all U.S. data center electricity consumption, up from just 12% last year. Yet, while investment in AI hardware—supercomputers, chips, and servers—is projected to surpass $1 trillion over the next four years, the energy sector is lagging behind, struggling to balance innovation with affordability and reliability.
According to Martin Stansbury, principal at Deloitte & Touche LLP, "There is an opportunity in infrastructure development to support the national strategic priorities of AI and energy dominance. However, it is a complex undertaking, and analysis reveals certain gaps are creating some challenges."
The report calls for urgent collaboration across sectors including energy, construction, and manufacturing to expand grid capacity, increase efficiency, and accelerate the deployment of renewable and smart infrastructure. The findings show that while both data center and power companies are aligned in viewing innovation and regulatory change as key solutions, confidence in current collaboration effectiveness remains low—with only 15% of data center operators and 8% of power providers calling it "highly effective."
Among other concerns cited by respondents were supply chain disruptions, resource competition, and growing risks tied to security and infrastructure delays. Deloitte’s analysis indicates that more intelligent infrastructure, new transmission technologies, and accelerated permitting processes will be essential to bridging these gaps.
Private investment in data centers hit record highs in 2024, but the energy sector has yet to fully unlock funding mechanisms to keep pace. As more AI infrastructure comes online, failure to innovate and coordinate could compromise U.S. competitiveness in the global AI race, Deloitte warns.
The report concludes that while the promise of AI is undeniable, it will only be fully realized if the underlying infrastructure is scaled efficiently, sustainably, and collaboratively.