The global AI revolution is at a crossroads, with the development of datacentres, the central nervous systems of AI, lagging behind the rapid advancements in AI models. This discrepancy poses significant challenges for AI firms, which rely on datacentres for training and operating their models. The Uptime Institute has identified 250 global datacentre projects exceeding 100MW in energy demand, but approximately half of these projects are at risk of cancellation or delay. This backlog is further exacerbated by the concentration of projects in 'datacentre corridors' and the lack of committed tenants, leading to supply chain issues and delays. The energy demands of these projects are vast, with planned projects announced last year alone consuming 1.3% of the world's projected electricity usage for 2025. However, the power grids in many regions, particularly in North America, are already operating under heavy strain, making it difficult to meet these demands. The Prince William Gateway project in Virginia, for example, has been halted due to local opposition and legal challenges, highlighting the challenges of securing energy supply and community support for datacentre projects. The choices involved in building massive AI datacentres, as opposed to directing those resources elsewhere, will become even more stark in a heating world with mounting geopolitical instability. The UK government's ambitions to make Britain an AI superpower appear to be underpinned by minimal attention to the tradeoffs and resources required. However, some observers are more upbeat, expecting that around 1,200 datacentres will be built globally between now and 2030, driven by AI demand. Innovations in battery storage and onsite power generation are seen as potential solutions to energy constraints. Nevertheless, the global AI revolution is at a critical juncture, with the development of datacentres lagging behind the rapid advancements in AI models, and the choices involved in building massive AI datacentres will have significant implications for the future of AI and the world.
