The world’s data centres are entering a new era of extreme energy demand, with global electricity consumption projected to grow 16% in 2025 and double by 2030, according to new estimates from Gartner. The sharp rise is being driven primarily by the explosive growth of AI-optimised servers, which are rapidly reshaping power consumption patterns across digital infrastructure.
Gartner forecasts that global data centre electricity usage will jump from 448 terawatt-hours (TWh) in 2025 to 980 TWh by 2030. For context, this increase is equivalent to adding the entire power consumption of several mid-sized countries to the grid.
AI Servers Becoming the Dominant Power Consumer
The most significant shift lies in the accelerating deployment of AI workloads. Gartner expects electricity usage from AI-optimised servers to increase nearly fivefold, from 93 TWh in 2025 to 432 TWh by 2030.
By 2030, AI servers alone will account for:
44% of total data centre energy consumption, and
64% of the incremental power demand added between 2025 and 2030
Linglan Wang, Research Director at Gartner, noted that AI is reshaping both the physical and operational footprint of data centres. High-density racks, advanced cooling systems, and specialised accelerators are now the primary drivers of power and thermal load.
U.S. and China Lead the AI Infrastructure Buildout
Regionally, the United States and China will remain the largest contributors to global data centre electricity usage. Together, they will account for more than two-thirds of global power demand from data centres by 2030.
By Gartner’s projections:
U.S. data centre electricity usage will rise from 4% to 7.8% of regional power consumption
Europe will increase from 2.7% to 5%
China and Asia-Pacific will grow at a comparatively moderate pace due to more efficient server hardware and proactive energy planning
China’s earlier investments in on-grid efficiency and advanced cooling infrastructure give it a structural advantage over many Western markets.
New Energy Models Will Be Needed to Power AI
As fossil-fuel-dependent grids strain under rising demand, the industry is exploring new clean energy models. Gartner expects the next decade to accelerate the shift toward:
Green hydrogen
Geothermal power
Small modular nuclear reactors (SMRs)
Battery Energy Storage Systems (BESS)
Tony Harvey, VP Analyst at Gartner, said natural gas will remain the dominant source for near-term on-site generation, but large battery storage systems will grow rapidly to stabilise solar and wind fluctuations.
Geothermal remains promising but challenged by high upfront costs and slow permitting cycles.
A Critical Moment for the Data Infrastructure Ecosystem
Gartner’s outlook underscores the scale of transformation underway. AI is no longer an incremental workload but a structural force reshaping global energy systems, data centre architecture, and sustainability planning.
Enterprises expanding AI workloads will need to account for rising operational costs, supply constraints, and the urgent requirement for power-efficient design and clean energy sources. Governments too will face growing pressure to upgrade grid capacity for a world where compute demand is doubling every few years.
