The rapid expansion of artificial intelligence is increasingly being constrained by delays and cancellations of large-scale data centre projects around the world. According to industry estimates, around 250 hyperscale data centre projects were announced globally between 2021 and 2024, but nearly half could be delayed or canceled due to mounting obstacles such as power shortages, environmental concerns, planning disputes, and rising construction costs. These facilities are essential for training and deploying advanced AI models, making them a critical part of the global AI ecosystem.
One of the biggest challenges is access to reliable electricity. AI data centres consume enormous amounts of power, and many regions already have overstretched electricity grids with long waiting times for new connections. In addition, developers are facing community opposition over concerns about land use, water consumption, noise, carbon emissions, and the limited number of permanent jobs these projects create. Legal challenges and stricter environmental reviews are further slowing approvals in several countries.
These infrastructure bottlenecks threaten national AI ambitions. Governments in the UK, the US, and elsewhere are investing heavily in AI to drive economic growth, but without sufficient data centre capacity, companies may struggle to train next-generation AI models or expand AI-powered services. The delays also risk disrupting semiconductor demand, cloud computing expansion, and broader digital transformation efforts that depend on high-performance computing infrastructure.
Despite the setbacks, industry experts remain optimistic that new energy solutions, including battery storage, onsite power generation, and grid modernization, can help ease these constraints over time. However, the report concludes that balancing AI growth with sustainable energy infrastructure and community interests will be one of the defining challenges of the global AI revolution in the years ahead.
