AI could account for nearly half of datacentre power usage ‘by end of year’ | Green economy

The analysis revealed that artificial intelligence systems can explain nearly half of the data consumption in the data stage by the end of this year.

Alex de VRIES-GAO estimates, founder of the Sustainability site for technology in Digiconomist, came at a time when the International Energy Agency expects to be artificial intelligence By the end of this contract Japan is also used today.

De Fis Gao’s accounts, to be Posted in the Sustainable Energy Magazine GulIt is based on the energy consumed by the NVIDIA chips and advanced fine devices that are used to train and operate artificial intelligence models. The paper also takes into account the energy consumption of the chips used by other companies, such as Broadcom.

IEA estimates that all data centers – except for mining cryptocurrencies – consumed 415 TERWATT (TWH) hours of electricity last year. De vries-gao argues in his research that artificial intelligence can actually represent 20 % of this total.

He said that a number of variables have entered his accounts, such as energy efficiency in the data center and the electricity consumption related to the refrigeration of servers that deal with the crowded work burdens of the artificial intelligence system. Datacentres is the central nervous system of artificial intelligence technology, with its high -energy demands, making sustainability a major concern in the development and use of artificial intelligence systems.

By the end of 2025, De VRIES-GAO can approach energy consumption by artificial intelligence systems up to 49 % of total energy consumption, except for encryption mining again. Artificial intelligence consumption can reach 23 GW, research estimates, twice the total energy consumption in the Netherlands.

However, de VRIES-GAO said that a number of factors can lead to slowing the demand for devices, such as declining demand for applications such as ChatGPT. Another problem is geopolitical tensions that lead to restrictions on the production of artificial intelligence devices, such as export controls. De vries-gao cited an example of barriers on Chinese access to chips, which contributed to the release of Deepsek R1 The artificial intelligence model is said to have used fewer chips.

“These innovations can reduce the costs of arithmetic and energy for Amnesty International,” said De Fais.

But he said that any efficiency gains can encourage more artificial intelligence. Multiple countries trying to build their artificial intelligence systems – a trend known as “sovereign artificial intelligence” – can increase demand for devices. De Park-Gao also pointed out the start of the US Data Data Center, Crusoe Energy, 4.5GW Energy Overflow capacity, with infrastructure, with Chatgpt Openai developer among potential customers through the joint Stargate project.

“There are early indications that this [Stargate] Datacentres can exacerbate dependence on fossil fuels.

Openai announced on Thursday the launch of the Stargate project in the United Arab Emirates, the first outside the United States.

Microsoft and Google Confess last year The artificial intelligence drive engines were at risk to achieve internal environmental goals.

De Fais Gao said that information related to the requirements of the power of artificial intelligence have become increasingly rare, as the analyst described it as a “transparent industry.” Amnesty International’s European Union Law requires artificial intelligence companies to detect energy consumption behind a model but not for daily use.

وقال البروفيسور آدم سوبي ، مدير المهمة للاستدامة في معهد آلان تورينج في المملكة المتحدة ، وهو هيئة أبحاث منظمة العفو الدولية ، إن هناك حاجة إلى مزيد من الشفافية حول مقدار الطاقة التي تستهلكها أنظمة الذكاء الاصطناعي ، ومقدار ما يمكن أن يوفرهم من خلال المساعدة في جعل الصناعات التي تنبعث منها الكربون مثل النقل والطاقة أكثر كفاءة.

“I think we do not need many very good use cases [of AI] To replace the energy used in the front end. “