A Deloitte analysis of the growth in data center infrastructure for AI workloads and its implications for the U.S. power system. It uses DC Byte and Wood Mackenzie data, estimates of AI data center penetration, and projections through 2035 to assess capacity. The focus is on interconnection and generation shortfalls, the impact on utility capital expenditures, and potential sources of and approaches to accelerating power supply for large AI sites.
Key Takeaways
AI workloads are becoming a significant driver of data center capacity additions and electricity-demand growth in North America through 2035.
Grid interconnection, generation availability, and transmission-development timelines are becoming critical constraints on the siting and commissioning of large AI data centers.
In forecasting capacity, the authors adjust growth rates for expected efficiency improvements, applying different CAGRs through 2030 and through 2035.
For large campuses, the analysis considers not only conventional grid connections but also alternative power-supply arrangements, including on-site generation and clean energy sources.
Growing data center loads require coordinated action by developers, utilities, system operators, and regulators; investments in the power system cannot always be attributed solely to data centers.
Key Figures
Growth rate used for the calculation through 2030
24 % CAGR
Deloitte used this rate to derive the capacity forecast, accounting for efficiency.
Growth rate used for the calculation through 2035
8 % CAGR
Applied after 2030 to reflect expected efficiency gains.
DC Byte North American forecast horizon
2030 year
North American data center capacity forecasts were used.
Operational capacity assessment horizon
2035 year
Deloitte derived data center operational capacity for this year.
Deloitte AI Infrastructure Survey respondents
98 respondents
The share of respondents answering the question on AI's impact on electricity demand was considered.
Stated scale of the AI campus in Abu Dhabi
5 GW
The source mentions Phase 1 of a new AI campus with a total stated capacity of 5 GW.
Utility capital expenditure forecast period
2025–2029 years
S&P Global estimate for a sample of 47 publicly traded US electric and gas utilities.
Forecast utility capital expenditure
over 1 trillion US$
S&P Global forecast for 2025–2029; the source specifically notes that the figure does not isolate spending driven exclusively by data centers.
Number of utilities in the S&P Global sample
47 utilities
Representative sample of publicly traded US electric and gas utilities.
Practical Value for Data Center Owners
For data center owners and investors, the analysis supports incorporating power supply into early investment and site due diligence: assessing not only stated available capacity, but also interconnection timelines, generation and transmission plans, competition for resources, and the potential for on-site power supply. When modeling demand from AI data centers, expected equipment-efficiency improvement scenarios should be explicitly considered rather than extrapolating peak growth rates unchanged through 2035.
A special World Energy Outlook report by the International Energy Agency examining the relationship between AI development and energy. It analyses growth in data-centre electricity consumption and capital expenditure, the physical constraints on scaling AI infrastructure, power-supply options, and implications for energy systems, prices and emissions. It also considers AI’s impact on energy and industrial efficiency, as well as demand scenarios through 2030.
AI data centersData center power supplyGrid interconnectionLiquid cooling