• AI-enabled data centre electricity use could triple by 2030
  • Tech firms are driving 40% of renewable PPAs and expanding nuclear investments

What happened

The International Energy Agency has warned that AI-driven data centres are rapidly increasing global electricity demand, with consumption set to surge sharply this decade. In its latest report, the agency said electricity use from data centres rose by 17% in 2025, with AI-focused facilities growing even faster than overall demand.

The IEA found that electricity consumption from AI-enabled data centres could triple by 2030, as adoption expands and use cases such as AI agents increase energy intensity. Overall data centre demand is expected to double in the same period.

Investment is also accelerating. Capital expenditure by five major technology companies exceeded $400bn in 2025 and is projected to rise by a further 75% in 2026, reflecting the scale of AI infrastructure buildout.

However, expansion faces constraints. Grid connection delays, planning bottlenecks and supply chain pressure on components such as transformers and advanced chips are slowing deployment. Developers are increasingly turning to onsite power, including gas generation, particularly in the US.

At the same time, the sector is shifting towards cleaner energy. Technology companies accounted for around 40% of corporate renewable power purchase agreements in 2025, while interest in nuclear is rising. The small modular reactor pipeline linked to data centres has grown from 25GW in 2024 to 45GW.

Why it’s important

The IEA’s findings show that energy is becoming a defining constraint on AI growth. Power availability and grid access now directly shape how fast data centre capacity can expand.

At the same time, AI is reshaping the energy system itself. While it drives demand as an “energy taker”, it is also accelerating innovation as an “energy maker”. Advances in nuclear, long-duration storage and flexible grid design are being driven partly by AI requirements.

Sustainability is no longer optional. Operators are adopting liquid cooling, waste heat reuse and grid-interactive designs to improve efficiency and manage rising thermal loads. These engineering-led approaches are becoming essential for high-density AI deployments.

The growing role of renewables and nuclear highlights a structural shift towards low-carbon power. If supported by policy and grid investment, AI demand could help scale clean energy deployment rather than undermine climate goals.

However, without faster infrastructure upgrades and regulatory reform, bottlenecks could delay projects and increase costs. The IEA suggests that countries able to align AI expansion with reliable, flexible and sustainable energy systems will gain a competitive edge.

Also read:   https://btw.media/en/tia-expands-beyond-telecom-with-first-ai-data-centre-standards

Also read:   https://btw.media/en/aws-lumen-launch-last-mile-cloud-connectivity-service