China and the US dominate global data centre expansion, driven by AI's explosive growth. By 2030, data centre capacity is expected to hit 220 gigawatts—six times 2020 levels—per McKinsey. This surge requires massive power: a single centre consumes energy equivalent to 100,000 households, with new projects like Texas' 11GW facility pushing limits. The energy bottleneck is stark: data centre power demand is projected to more than double by 2030, reaching 945 terawatt hours (IEA), yet supply remains uneven. Climate goals and permitting delays in Europe and the US complicate scaling.

The Texas project by Fermi America, planned for 17GW private grid, highlights both ambition and uncertainty. Despite $11GW allocated, no tenant signed as of May 2026. Meanwhile, China’s state-backed strategy prioritizes renewable-rich western regions, aligning with climate targets. However, local resistance in the US—like Maine’s stalled bans—shows regulatory friction. Hyperscalers like Google and Meta race to deploy infrastructure rapidly, but experts warn this may prioritize profit over sustainability. The global energy gap between data centre concentration and renewable supply will define future winners.

Experts argue the rush isn’t just technological but economic. Jonathan Koomey notes hyperscalers’ urgency stems from “rapidly depreciating assets,” pushing them to lock in users. Rose Luckin adds that AI inference—continuous use by millions—will sustain demand. Without coordinated energy solutions, costs and emissions could soar. The key challenge remains balancing growth with sustainable power sources. The Forum’s focus on these dynamics underscores a critical juncture: data centres’ energy footprint may reshape global tech and climate agendas.