AI Data Center Expansion Drives Global Energy Demand and Utility Capacity Planning
Over the past 72 hours, OSINT indicates a significant surge in energy demand driven by AI datacenter buildout, with projections showing increased capacity requirements across the US and globally. This trend highlights the growing influence of AI workloads on energy infrastructure and power consumption metrics.
The US electricity demand forecast has increased to a 4.7% CAGR from 2024 to 2030, with 60% of the incremental load attributed to AI datacenters and electrification, according to the EIA. Duke Energy reports a 10 GW increase in AI-related demand by 2030, citing AI datacenters as the fastest-growing customer class.
The ERCOT grid interconnection queue now includes 50 GW of new datacenter load requests, a fourfold increase year-over-year, with AI and crypto compute workloads responsible for most of these applications. Dominion Energy’s capacity plan projects 24 GW of new generation capacity by 2038, with 70% tied to datacenter demand, driven by the expansion of Northern Virginia’s Data Center Alley.
The IEA estimates global datacenter electricity use will reach 1,100 TWh in 2026, more than doubling from 460 TWh in 2022, with AI inference and training potentially adding Ireland-sized power demand annually. Microsoft’s datacenter pipeline indicates that 60% of new builds will incorporate on-site generation or grid partnerships to address grid stress and secure dedicated energy sources for AI workloads.
Utilities like NV Energy project 3.5 GW of new demand from AI datacenters by 2028, as local utilities double peak capacity plans in response to hyperscaler commitments. Additionally, Shell’s LNG demand revision for 2025–2027 has been revised upward by 3%, partly due to AI datacenter electricity needs in Asia, emphasizing the broader impact on energy markets.
These signals collectively demonstrate that AI datacenter expansion is now a primary driver of electricity demand growth, prompting utilities to accelerate capacity expansion and adopt on-site generation strategies to meet the increasing load requirements.
OSINT does not specify detailed capacity utilization rates or the specific mix of renewable versus conventional generation sources, nor does it include forward guidance beyond these figures.
SEO hashtags: #AIenergydemand #datacenterpower #energyinfrastructure #utilities #gridcapacity #AIworkloads #electricitygrowth #LNGdemand #renewableenergy #energymarkets
