AI Data Centres Set to Consume More Power Than Nigeria, Pakistan and Bangladesh Combined by 2030 — UN Warns
Global electricity demand from AI-driven data centres is set to surge at an unprecedented rate, with a new United Nations-backed report warning that by 2030, they could consume nearly three times the combined annual electricity usage of Nigeria, Pakistan, and Bangladesh. The report, published by the United Nations, highlights how the rapid rise of artificial intelligence is reshaping global energy, water, and environmental systems.
According to research from the UN University Institute for Water, Environment and Health, global data centre electricity consumption is projected to rise from 448 terawatt-hours in 2025 to 945 terawatt-hours by 2030. Artificial intelligence is expected to account for around 40% of this demand, largely driven not by model training, but by everyday usage—known as “AI inference,” which powers routine user interactions and queries.
The report reveals staggering efficiency and environmental costs tied to AI usage. ChatGPT alone processes an estimated 2.5 billion prompts daily, consuming about 383 gigawatt-hours of electricity annually. Some AI tasks are significantly more energy-intensive, with image generation requiring up to 1,450 times more energy than simple text classification, while a single AI-generated video can consume electricity equivalent to 200,000 spam filters. Water usage is also rising sharply, with AI infrastructure projected to double its consumption to 9.3 trillion litres by 2030.
Beyond energy and water, the environmental footprint is expanding globally. Data centre-related carbon emissions are expected to nearly double, while their land use will more than double in size. The report also highlights inequality in AI infrastructure distribution, noting that over 90% of global computing capacity is concentrated in just the United States and China. In countries like Ireland, data centres already account for over 20% of national electricity use, forcing restrictions on new developments due to grid pressure.
In Africa, however, countries like Nigeria are positioning themselves to benefit from the AI infrastructure boom. Kasi Cloud has begun commissioning its LOS1 data centre in Lagos as part of a 100-megawatt expansion plan. The facility is expected to reduce Nigeria’s reliance on foreign cloud services and help retain an estimated $850 million currently spent annually on offshore infrastructure—signaling a growing race between development opportunities and environmental concerns.
source: nairametrics