AI needed by struggling electricty infrastructure

16 January 2025

Artificial intelligence (AI) will play a key role in implementing cohesive and scalable electricity systems that integrate a mix of energy sources, including fossil fuels, renewables, water and nuclear.

Trenton Thornock, founder and CEO of US-based data centre technology company Prometheus Hysperscale, told a panel at the ongoing Abu Dhabi Sustainability Week that AI advancements can help build a cohesive and scalable ecosystem to address growing fragmentation that results from integrating renewable energy such as solar and nuclear power into electricity systems designed for fossil fuels.

“AI will play a key role in building and maintaining resilient electricity systems.

“A resilient infrastructure will require a modular and adaptable platform that can integrate various energy inputs,” Thornock added.

Former BP CEO Bernard Looney, who stepped down in September 2023, is chairman of Prometheus Hyperscale. The company is building sustainable AI infrastructure and data centres across several US states, including its flagship 1GW data centre in Evanston, Wyoming.

In addition to being powered 100% by renewable energy, the data centre will tap advanced fission technology through a partnership with Oklo, a nuclear technology startup funded in part by OpenAI chief executive Sam Altmann’s Hydrazine Capital.

Speaking at a separate panel at ADSW, Looney said it is important to amplify the positive effects of AI-based technologies across industries despite the power density they require.

He stressed ongoing and rapid innovations in microprocessor and liquid cooling technologies that can help reduce the energy consumption of data centres over time.

“We can minimise power usage and amplify [a data centre infrastructure’s] positives,” Looney said. 

AI computing platforms such as Google and Amazon Web Services rely on so-called hyperscale data centres, which apart from being massive, have extreme scalability capabilities and are engineered for large-scale workloads.

Typically, they are built with optimised network infrastructure, streamlined network connectivity and minimised latency enabling.

https://image.digitalinsightresearch.in/uploads/NewsArticle/13268530/main.jpg
Jennifer Aguinaldo
Related Articles