Commenting on the findings, Tom Winstanley, CTO and head of new ventures at NTT DATA UK & Ireland, warned that this trajectory is unsustainable without major innovation.
“It is clear that this growth trajectory of data centre energy usage cannot continue unchecked,” Winstanley said.
“As AI models continue to grow in scale and sophistication, so too must our data centre infrastructure strategies.
“Rather than simply expanding power capacity, we have an opportunity to innovate and reimagine how digital infrastructure is built and operates.”
The IEA report estimates that global data centre energy consumption could hit 945 terawatt hours (TWh) annually by 2030 — more than three times the UK’s current total consumption.
A major factor in this surge is the development of powerful AI models, such as OpenAI’s GPT-4, whose training alone is thought to have used 42 gigawatt hours (GWh) of electricity — equivalent to the daily energy use of 28,500 homes in developed nations.
Winstanley highlighted the need for new solutions, from advanced cooling systems to software optimisation and photonics-based technologies, which use light rather than electricity to process and transmit data.
“Photonics-based technologies offer a compelling path forward,” he said. “They have the potential to dramatically reduce latency and energy consumption at scale – two of the biggest pressure points for next-generation data centres.”
He added that energy efficiency is “extending from a technical concern to a business imperative”, urging companies to embrace low-footprint, high-performance solutions to ensure the digital economy remains both “resilient and responsible.”
RELATED STORIES
Pioneering digital transformation in golf and beyond
