The report, which was published last week, suggested that AI could use almost as much energy by the end of this decade as Japan uses today and only half will come from renewable sources. By 2030 it said that processing data would consume more electricity in the US alone than manufacturing steel, cement, chemicals and all other energy-intensive goods combined.
Global electricity demand from datacentres is expected to more than double by 2030 and AI will be the main driver, with the demand from dedicated AI datacentres set to more than quadruple. The report found that just one data centre consumes as much electricity as 100,000 households.
As a result, and with the building of even bigger data centres, the rapid adoption of AI has often been identified as posing a serious risk to meeting global climate targets but, according to the IEA, this risk is ‘over-stated’.
In fact, its report suggested that by using AI it will help to make energy use and other activities more efficient and so could help to reduce greenhouse gas emissions overall.
Using AI, it will be possible to redesign energy grids to better balance demand when most of the supply comes from renewables which tend to be from intermittent or more unpredictable sources e.g. wind and solar power.
The report also suggested that finding efficiencies within energy systems, and in industrial processes, could also become easier with AI as could the discovery and exploitation of reserves of critical minerals, vital to modern renewable energy components.
Sounds good but the report went on to warn that to achieve this it would require much greater direction from governments, without which the rapid growth of AI could prove a massive problem for energy systems and the environment going forward.
While the IEA report acknowledged these issues its report was criticised as ‘painting too rosy a picture’ and didn’t tackle the real risks associated with the growing use of AI.
Critics warned that AI has the potential to reverse all the gains made in recent years in advanced economies to reduce their energy use, and with more data centres in the offing companies could end up turning to more traditional forms of energy such as gas and even coal.
