AIInfrastructure – Data Centre Power Needs Set for Sharp Global Rise by 2026
AIInfrastructure – The rapid expansion of artificial intelligence is expected to significantly increase electricity consumption at data centres worldwide, with global power usage projected to climb sharply over the next year. According to a new industry assessment released on Wednesday, electricity demand from data centres is forecast to reach 565 terawatt hours (TWh) in 2026, compared with an estimated 447TWh in 2025, representing an increase of more than 26%.
AI Workloads Driving Unprecedented Energy Growth
The report attributes the surge primarily to the growing adoption of advanced artificial intelligence applications that require extensive computing resources. As businesses and governments increasingly rely on AI-powered systems, data centres are being pushed to support larger workloads, resulting in a substantial rise in energy requirements.
Analysts noted that modern AI models demand far greater processing capacity than traditional computing tasks. This shift is placing pressure on existing infrastructure and creating new challenges for data centre operators seeking to expand operations while maintaining efficiency.
Power Capacity Expected to Increase Rapidly
Industry projections indicate that data centre power demand will rise from 104 gigawatts in 2025 to approximately 132 gigawatts in 2026. Looking further ahead, demand could reach 290 gigawatts by 2030, highlighting the extraordinary pace at which AI-driven infrastructure is expanding.
Experts believe that access to reliable electricity is becoming one of the most critical factors influencing the growth of the global AI industry. The ability to secure sufficient power resources may determine how quickly companies can scale their operations and remain competitive in the rapidly evolving technology landscape.
AI-Optimised Servers Becoming Major Energy Consumers
A key contributor to increasing electricity usage is the growing deployment of AI-focused servers. These specialised systems are designed to handle complex machine-learning tasks and large-scale data processing operations.
The report estimates that AI-optimised servers will account for nearly one-third of total data centre electricity consumption in 2026. By 2027, their power usage is expected to exceed that of conventional servers, reflecting the accelerating adoption of AI technologies across industries.
Concerns Over Future Electricity Supply
Researchers also warned that future demand could place considerable strain on power grids worldwide. Data centre electricity consumption is projected to surpass 1,200TWh annually by 2030, raising concerns about whether existing energy infrastructure will be capable of supporting continued expansion.
As new facilities are built to meet growing digital requirements, competition for electricity resources is expected to intensify. Industry leaders are being encouraged to develop long-term strategies that balance operational growth with sustainable energy management.
Efficiency Measures Becoming Increasingly Important
To address these challenges, experts recommend that infrastructure and operations teams focus on improving energy efficiency and securing access to reliable power supplies. Investments in advanced cooling technologies, energy-efficient systems and edge computing solutions are viewed as important steps toward reducing pressure on electricity networks.
The report also highlighted rising demand linked to cooling systems and supporting infrastructure. Energy requirements for these functions are expected to increase by 22.6% in 2026 and by a further 24.6% in 2027 as facilities expand to accommodate more powerful computing equipment.
Environmental Impact Remains a Key Concern
Growing electricity consumption has also renewed concerns about the environmental footprint of data centres. A separate study recently reported that data centres consumed an amount of electricity comparable to the annual usage of Saudi Arabia during the previous year.
That analysis further suggested that if global data centre power consumption doubles by 2030, offsetting the resulting carbon emissions would require approximately 6.7 billion trees to be grown over a ten-year period. The findings underscore the importance of balancing technological advancement with sustainability goals as the AI era continues to accelerate.