Increased use of AI use may be driving up energy demand, but it also offers some of the solutions needed to tackle the challenges by making the energy industry more efficient, according to a detailed report just published by the International Energy Agency (IEA).
In its Energy and AI report, the Paris-based think tank for OECD countries underscores the huge rise in demand in energy consumption that is likely to hit all regions in the next few years, as more power-hungry data centres are built.
Electricity demand from all types of data centres worldwide is set to more than double by 2030 to around 945 terawatt-hours (TWh). That figure, the IEA notes, is larger than the entire electricity consumption of Japan today.
Meanwhile, for those data centres optimised to handle AI processing, the challenge is even bigger. Power demand for these globally is likely to quadruple by 2030. In the world’s advanced economies as a whole, data centres are projected to drive more than 20% of growth in electricity demand between now and 2030.
In the US – a global hotspot for AI and cloud computing – data centre power consumption is forecast to account for almost 50% of electricity demand growth between now and 2030. If that happens, the US economy could be using more electricity in data centres by then than in manufacturing all energy-intensive goods, such as aluminium, steel, cement and chemicals, put together.
Fatih Birol, the IEA’s Executive Director, said that, with the rise of AI, the energy sector was at the forefront of a technological revolution.
“AI is a tool, potentially an incredibly powerful one, but it is up to us – our societies, governments and companies – how we use it,” he said.
The report, which builds on the IEA’s previous work and also its input to the February’s AI Action Summit held in Paris, aims to inform future strategies. The IEA also plans to launch a new Observatory on Energy, AI and Data Centres soon, which will gather comprehensive data globally on AI’s electricity needs and track cutting-edge AI applications across the energy sector.
The IEA is keen to stress that significant uncertainties remain from the macroeconomic outlook to the speed at which AI will be adopted, how fast efficiency improvements will occur, and whether bottlenecks in the energy sector can be resolved.
Coping with a world where power demand may rise faster – after years of stagnation in some regions – may not be quite as daunting as it sounds, in part because AI could also provide some of the answers to the challenges it is setting.
For example, AI has made cyberattacks on energy utilities more sophisticated and common – tripling in the last four years – but AI is also now a critical tool for energy companies to defend against such attacks, the IEA notes.
AI breakthroughs
AI is also driving technology breakthroughs that could accelerate innovation in energy technologies such as batteries and solar photovoltaic power.
The IEA sees natural gas partnering renewables in catering for much of the increased energy demand from AI-focused data centres.
Birol told the report’s launch event that the technology companies with which the author’s had discussed the issue said their priorities were to have electricity supply that are secure 24/7 using low emissions technologies.
“When we look at their plants, their projects, their strategies, we see that the current preferences lie with renewables coupled with the batteries and strong grids and natural gas – these are the two main options for many companies,” he said.
That will push up carbon emissions, but the agency forecasts levels would still be small compared to those from the energy sector overall. In the US, carbon emissions from data centres are forecast to be no higher than around 4.5% of total combustion emissions by 2035, while for China it is over 3% and in the EU the figure is less than 0.5%.
Despite the low share of the total, data centres could still record the highest emissions growth of any sector by 2030. But the IEA notes that these centres could also help cut emissions in the broader energy system regions, for example by providing waste heat as an input to district heating systems in parts of Europe and China.
Future Digital Twin has reported previously on efforts already underway to power data centres more effectively, for example by making existing grid infrastructure more efficient. Furthermore, emissions growth could be curtailed by using AI, if adoption of the technology is widespread, driving energy efficiency gains and other benefits.
The IEA report’s authors of the IEA report warn that time is of the essence, if the growth of data centres is to be integrated successfully with the energy transition.“Countries that want to benefit from the potential of AI need to quickly accelerate new investments in electricity generation and grids, improve the efficiency and flexibility of data centres, and strengthen the dialogue between policy makers, the tech sector and the energy industry,” it says.
(Photo: Shutterstock)
