How far have we come in using AI and digitalisation to push clean energy
Saravan It’ s about energy for AI and AI for energy. Energy and AI are two things which can’ t be separated. For AI, you need energy. For energy efficiency, there is no other technology than AI that can bring it in terms of accelerating gains than the traditional technology. We believe at AID and ADNOC it’ s not a notion but a programme for us every day. It’ s energy for AI and AI for energy.
James I think 2024 is when AI woke up to energy. It’ s a hot topic in terms of how much energy AI uses, but I think you have to be very careful of the forecasts of usage. That, for me, is exemplified by the huge range – by 2030, some people are saying it’ s a 35 % increase from AI, some people are saying it’ s 250 %. There’ s a massive range there.
WHAT IS AGENTIC AI?
Agentic AI is a type of artificial intelligence( AI) that can make decisions, act on them and adapt without much human intervention. It ' s designed to be more like a skilled employee, working to understand intricate processes and objectives.
“ It’ s a hot topic in terms of how much energy AI uses, but I think you have to be very careful of the forecasts of usage”
JAMES FORREST, EVP GLOBAL ENERGY TRANSITION AND UTILITIES INDUSTRY LEADER, CAPGEMINI
I think the data centre load will end up being built in the US up to 2030, probably about 45GW, but I think you also have to keep a close eye on efficiency. There’ s been huge advances. The latest NVIDIA processes claim a 25 % reduction in energy. There’ s Koomey’ s law which shows that over the 10 years the efficiency of compute has doubled and gotten better every 18 months. It’ s difficult to predict all of that and how there’ s so much innovation and efficiency.
Even if you get to 4 % of global energy from AI, AI will have a massive impact in reducing that energy and reducing energy demand. You have to put all of that together when it comes to the whole demand for energy question.
70 April 2025