The demand for artificial intelligence is leading to increased energy usage by data centres and creating a link between the two assets, says Matthew Mendes, managing director and head of infrastructure at the Investment Management Corp. of Ontario.
“The growth is different than it was before — [there’s] more hyperscale companies coming into the mix. It’s accelerating the growth trajectory of [date centre] businesses pretty materially.”
Data centres — already notorious for being an electricity-intensive asset — are projected to more than double their current energy consumption in the U.S. by 2030, according to a 2024 report by Goldman Sachs. It noted data centre power demand is expected to have a compound annual growth rate of 15 per cent. By 2030, data centres will comprise eight per cent of total U.S. energy demand, an increase from about three per cent in 2024.
“As an investor, we think, that’s pretty exciting,” adds Mendes. “There’s going to be more clean energy that’s going to be needed to accommodate the new data centres.”
Questions around data centres’ increasing energy consumption are leading institutional investors to take a different view of the asset class, requiring them to have some expertise in energy as part of their due diligence process. “We can see a world where the energy and the digital businesses actually start to become integrated.”
Read: Caisse invests up to €300 million in European data centre operator, CPPIB sells Amitra Capital
Goldman Sachs’ report noted the growth of data centres in the U.S. will require roughly US$50 billion in new power generation capacity, noting institutional investors’ interest in AI could also open opportunities in utilities, renewable generation and industrials.
Looking ahead, there will be questions around technological solutions that improve the energy consumption of data centres and whether alternative energy sources could play a bigger role in fuelling these businesses, says Mendes.
A report by the Global Risk Institute suggests that it’s possible for synergistic advancements in both renewables and energy efficiency measures, which could then help offset elevated electricity demands from increasing use of AI. However, the report also envisions a scenario in which these efforts fall short, sustaining a dependence on fossil fuel-based power generation to match the ongoing interest and investment in AI solutions.
Read: Survey finds Canadian employers looking at AI to enhance worker productivity, but concerns remain