Schneider Electric on Managing AI's Energy Hunger
AI is a double-edged sword when it comes to energy — as much as automation and intelligence powered by it can drive efficiency, it's an energy-thirsty beast.
This quality paired, with its capabilities, is urging data centres at the heart of this transformation to act, particularly when it comes to managing the surge in AI adoption alongwith the escalating demand for energy — all while maintaining all-important sustainability.
“There is no doubt that AI is power hungry,” Thierry Chamayou, VP of Cloud and Service Providers at Schneider Electric said. “One estimate was that training a large language model such as GPT3 consumed in the region of 1,300MWh of electricity, about the same as 130 US domestic homes annually.
“That is not insignificant, and is likely to grow as more deployments are made with ever greater adoption.”
Meeting demand with scalable solutions
Data centres and data transmission networks play a key role in clean energy transitions, the International Energy Agency (IEA) says.
Thanks to rapid improvements in energy efficiency, energy demand growth from data centres and data transmission networks have been significantly limited.
The body urges strong government and industry efforts on energy efficiency, renewables procurement and RD&D, which it says are essential to curb energy demand and emissions growth over the next decade.
It also advises that global electricity demand from AI, data centres and crypto is set to rise to 800TWh in 2026 as a minimum, up almost 75% from 460TWh in 2022.
“Since 2010, emissions from data have grown only modestly despite rapidly growing demand for digital services, thanks to energy efficiency improvements, renewable energy purchases by information and communications technology (ICT) companies and broader decarbonisation of electricity grids in many regions,” the IEA said. “However, to get on track with the Net Zero Scenario, emissions must drop by half by 2030.”
Theirry continued: “With AI in mind, Schneider Electric has carried out a lot of research to understand the specific requirements of AI workloads in the data centre. Our white paper outlines the challenges of shifting to higher rack power densities, with recommendations on how to adapt data centre infrastructure, covering adaptations to power, cooling, white space layout and monitoring software. It also looks at the retrofitting process for liquid cooling and offers best practices to minimise the risk of failures and hazards associated with higher temperatures.
“Industry partnerships also offer an opportunity to combine strengths to achieve more. Our partnership with Nvidia has allowed us to develop reference designs, including with the new generation of Blackwell processors, to ensure efficiency and reduced consumption with vastly improved performance. This generation of processors now offers up to 25 times the processing power of previous generations of technology, while using up to 30 times less power, and with a massively reduced physical footprint for high-density deployments.”
“With great power (usage) comes great responsibility,” Sophie Phillips, Senior Associate in the construction team at law firm Bird & Bird – and part of the company's wider data centre team — told sister title Data Centre Magazine.
“Sustainability is — or should be — a major concern for all players in the industry. We have a responsibility to improve and make efficiencies in order to keep this industry resilient and to continue to be able to grow.
“There is a huge demand for data centres and a limited amount of available power. Looking to manage this challenge responsibly is key.”
**************
Make sure you check out the latest edition of Energy Digital Magazine and also sign up to our global conference series - Sustainability LIVE 2024
**************
