Energy efficiency is the key to unlock AI‘s business potential, a discussion panel at the Economist Impact Sustainability Week summit in London has suggested.
The discussion, entitled AI and energy — driving efficiency and action, featured Raquel Espada, vice-president of energy services and sustainability, Schneider Electric Advisory Services, Doug Adams, president and chief executive, NTT Global Data Centers Americas, and Sophie Graham, chief sustainability officer, IFS, and looked the role that AI can play in supporting the energy transition, provided its voracious appetite for energy can be fulfilled sustainably.
Core investment
As Espada noted, when it comes to AI, energy efficiency should be viewed as a core investment rather than an optional measure, with efficiency directly linked to cost savings.
“What we need is ‘energy intelligence’,” she noted, adding that it is “critical that efficiency isn’t just considered a ‘nice to have'”, given its transformative potential.
In her work with companies across different regions, she added, varying approaches have been adopted, with some organisations prioritising sustainability outcomes and others focusing primarily on cost reduction.
“The first thing we often ask, when we are in the boardroom of a company, is why are you not implementing energy efficiency,” she added. “It’s such a no brainer, the paybacks are fast.”
Financial and operational priority
NTT’s Adams added that improving energy efficiency is both a financial and operational priority for the business, which delivered more than 370MW of new capacity last year with ten new openings in North America, EMEA and APAC. He pointed to three pillars as being central to the group’s approach: “consume less, do better and do different”.
NTT’s ecosystem is incredibly complex, and AI has played an “invaluable” role in enabling the firm to figure out areas where it can reduce power consumption. “In the test cases we’ve done, we’ve found we’ve achieved about 20% efficiency,” Adams noted.
NTT plans to invest more than $10 billion by 2027 on additional data centre infrastructure. As Adams added, by the end of the decade, between $3 trillion and $7 trillion worth of investment will be needed globally to meet projected AI-related demand.
“I’ve seen more change in data centres the last two or three years than I’ve seen in the previous 20 years,” he added. “The grid takes a while to adjust.”
Elsewhere, IFS’ Graham highlighted the role of AI in operational optimisation, noting that scheduling and routing tools powered by AI have reduced travel distances for field engineers by an average of 37%.
“There’s a lot we can do with what we already have, rather than building out new infrastructure,” she noted.
The Economist Impact Sustainability Week summit continues on 4 March. More information can be found here.
