Artificial intelligence (AI) systems are projected to consume nearly half of data centre power by the end of 2025, according to research by Alex de Vries-Gao. The founder of the Digiconomist tech sustainability website estimates that AI could already account for 20% of data centre power consumption, and this demand is expected to increase further.
De Vries-Gao’s calculations are based on energy consumption by chips used in training and operating AI models from companies like Nvidia, Advanced Micro Devices, and Broadcom. The International Energy Agency (IEA) forecasts that AI will require almost as much energy by the end of this decade as Japan uses today.
By 2025, De Vries-Gao estimates that energy consumption by AI systems could approach up to 49% of total data centre power consumption, excluding cryptocurrency mining. This would translate to approximately 23 gigawatts (GW) in energy demand, twice the total energy consumption of the Netherlands.
However, several factors could lead to a slowdown in hardware demand, including decreased interest in applications such as ChatGPT and geopolitical tensions that restrict AI production. De Vries-Gao notes that innovations can reduce computational and energy costs of AI, but this may encourage even more AI use.
The trend of “sovereign AI” – multiple countries building their own AI systems – could also increase hardware demand. A US data centre startup, Crusoe Energy, has secured 4.5GW of gas-powered energy capacity for its infrastructure, with potential customers including OpenAI’s ChatGPT developer.
Microsoft and Google have admitted that their AI drives are threatening their ability to meet internal environmental targets. The EU AI Act requires AI companies to disclose the energy consumption behind training a model, but not for day-to-day use.
Experts argue that more transparency is needed on how much energy is consumed by artificial intelligence systems and how much they could save by helping to make carbon-emitting industries more efficient.
Source: https://www.theguardian.com/environment/2025/may/22/ai-data-centre-power-consumption