- Balancing AI’s rapid growth with sustainable energy and resource use remains a critical imperative. It requires ongoing innovation and responsible management to ensure the technology’s benefits do not come at an unsustainable environmental cost.
The escalating energy demands of artificial intelligence (AI) systems pose significant sustainability challenges for data centres worldwide.
According to Alex de Vries-Gao, founder of the Digiconomist tech sustainability website, AI will likely account for nearly half of data centre power consumption by the end of this year.
His analysis, published in the journal Joule, considers the energy usage of chips from key manufacturers, including Nvidia, Advanced Micro Devices, and Broadcom, which are instrumental in training and operating AI models.
Projections indicate that by 2025, AI could consume up to 49 per cent of total data centre power, potentially reaching 23 gigawatts—double the current energy use of the Netherlands.
Environmental impact
This substantial consumption encompasses not only the chips themselves but also the energy required for cooling the servers managing heavy AI workloads. Data centres function as the backbone of AI technology, underscoring the importance of addressing their environmental impact.
This shift underscores the substantial energy demands driven by AI workloads, which are characterised by intensive computational processes necessary for machine learning, natural language processing, and more.
The International Energy Agency has also highlighted concerns, estimating AI’s energy demand by decade’s end will rival Japan’s current consumption, with only half supplied by renewables.
The IEA estimates that all data centres – excluding mining for cryptocurrencies – consumed 415 terawatt hours (TWh) of electricity last year. De Vries-Gao argues in his research that AI could already account for 20 per cent of that total.
It projects that electricity demand from data centres worldwide is set to more than double by 2030 to around 945 terawatt-hours (TWh), slightly more than the entire electricity consumption of Japan today.
AI: Most significant driver
AI will be the most significant driver of this increase, with electricity demand from AI-optimised data centres projected to more than quadruple by 2030.
In the United States, power consumption by data centres is on course to account for almost half of the growth in electricity demand between now and 2030.
Driven by AI use, the US economy is set to consume more electricity in 2030 for processing data than for manufacturing all energy-intensive goods combined, including aluminium, steel, cement and chemicals.
In advanced economies more broadly, data centres are projected to drive more than 20 per cent of the growth in electricity demand between now and 2030, putting the power sector in those economies back on a growth footing after years of stagnating or declining demand in many of them.
While there may be a slowdown in hardware demand due to factors such as export controls or reduced popularity of certain AI applications, innovation in lower-cost, energy-efficient models, like those emerging from China, may mitigate some energy costs.
Nevertheless, these advancements could paradoxically drive increased AI usage, further elevating energy needs. Additionally, the trend toward “sovereign AI” development in various countries could amplify hardware demand.
Beyond energy, environmental sustainability challenges extend to water consumption—US data centres used over 75 billion gallons of water in 2023, raising concerns among environmentalists about resource sustainability.