McKinsey: How to Sate AI’s Hunger for Energy

AI is on everybody’s mind.

From improving energy efficiency to customer service chatbots, just about every industry is getting involved. 

While AI has the potential to make huge improvements to everyday life, it is one of the most power hungry inventions of the 21st century because of its demand on data centres.

A report from McKinsey, How data centres and the energy sector can sate AI’s hunger for power, says that to keep pace with the current rate of adoption, the power requirements of data centres are expected to grow around three times higher than current capacity by 2030.

This would take power needs from between 3-4% of total US power demand today to 11-12% by 2030 and require an investment of more than US$500bn.

Karen Fang, Managing Director and Global Head of Sustainable Finance at Bank of America, says: “As AI scales at an unprecedented rate, its impact on energy systems and the entire value chain becomes more complex.

“Data centres and AI now play a crucial role across nearly every sector of the economy, making it essential not only to ensure their sustainability but also to harness their potential to drive sustainable solutions. 

“These solutions include but are not limited to further improving energy efficiency, better forecasting renewable energy supply and optimising distribution, speeding up interconnection permitting, as well as more targeted planning for resilience and adaptation efforts.”

Escalating demand for data centres

McKinsey says the US is expected to be the fastest-growing market for data centres, fuelled by a continued increase in cloud migration and the scaling of new technologies including AI.

It says that generative AI could help to create between US$2.6tn and US$4.4tn in economic value throughout the global economy.

However, achieving a quarter of this by the end of the decade would require up to 60GW of additional data centre infrastructure in the US alone.

McKinsey says this will require considerably more electricity than is currently produced in the US and data centre load may make up 30-40% of all new net demand added until 2030.

Melanie Nakagawa, Chief Sustainability Officer at Microsoft, says: “Together technology, policy and investment can help catalyse innovation at the scale we need at this moment.”

Opportunities for investors in data centre growth

McKinsey highlights that this growth brings opportunities for investors across the energy sector.

Transmission and distribution investments into utility companies should grow alongside data centre demand and allow for better compute and storage systems.

Secondary markets with access to cheap and reliable power can bridge the timing gap between data centre development, taking 18-24 months, and power infrastructure development that can reach ten years. 

McKinsey says 70% of data centre growth is expected to be fulfilled directly or indirectly by hyperscalers by 2030, so investors can seek opportunities to fuel growth through data centre developers at company or site level.

Data centres and AI’s impact on the environment

If renewable energy solutions don’t scale up in line with data centre power demand, there’s potential for an increase in global emissions.

“According to the IEA, energy has been responsible for more than three quarters of global greenhouse gas emissions in recent years,” says Aiman Ezzat, CEO at Capgemini.

“Organisations moving to clean electricity, combined with digital technologies and data, have incredible untapped potential to accelerate the energy transition.”

Kate Brandt, Chief Sustainability Officer at Google, says: “As we look at the role of AI to accelerate clean energy solutions, it's important to consider three key areas: Identifying and scaling solutions such as providing helpful information like rooftop solar data, optimising for energy efficiency solutions and prediction to support renewable energy integration through wind forecasting.

“Managing AI's footprint by gaining a better understanding of the potential forecast for related electricity demand, and continuing to drive energy efficiency in data centres and AI chips, and run data centres on 24/7 carbon free energy by 2030.

“Policy can maximise these benefits by unlocking clean energy access and supporting education, while also leveraging AI to fast-track the implementation of climate policies as new NDCs are established next year.”

Make sure you check out the latest edition of Energy Digital Magazine and also sign up to our global conference series - Sustainability LIVE 2024​​​​​​​

Energy Digital is a BizClik brand.