New Delhi [India], November 21 : The rapid expansion of artificial intelligence (AI) and generative AI (GenAI) technologies is set to strain global power supplies; 40 per cent of AI data centres may face operational challenges by 2027 because of electricity shortages, says a report by Gartner Inc.
The energy demand for data centres is projected to grow by 160 per cent over the next two years, pushing electricity consumption to unprecedented levels.
Power usage in AI data centres is expected to rise from 261 terawatt-hours (TWh) in 2024 to 500 TWh by 2027, more than double the 2023 levels.
This surge is driven by the increasing need for larger data centres to handle the vast computing power required for training and deploying large language models (LLMs) that underpin GenAI applications.
However, utility providers are struggling to expand their capacity quickly enough to meet this demand, creating a potential energy shortfall that could disrupt data centre operations by 2026.
Bob Johnson, VP Analyst at Gartner, said, "New larger data centres are being planned to handle the huge amounts of data needed to train and implement the rapidly expanding large language models (LLMs) that underpin GenAI applications."
He added, "However, short-term power shortages are likely to continue for years as new power transmission, distribution, and generation capacity could take years to come online and won't alleviate current problems."
The impending power shortages are expected to lead to significant cost increases for data centre operations.
As electricity prices rise, organisations relying on AI technologies will face higher operating costs.
Gartner advises businesses to prepare for these increases by negotiating long-term contracts for stable power rates and incorporating higher energy costs into their project budgets.
In addition to rising costs, sustainability goals will also face setbacks. To meet the growing energy demand, some power suppliers are extending the operation of fossil fuel plants initially planned for retirement.
This move will likely increase CO2 emissions, making it difficult for data centres and their customers to meet zero-carbon objectives in the short term.
Furthermore, renewable energy sources like wind and solar cannot provide the consistent 24/7 power needed by data centres without alternative supply solutions, such as advanced battery storage or nuclear power.
To mitigate these challenges, Gartner recommends organisations assess the risks of power shortages on their AI operations and consider alternative strategies.
This includes using smaller AI models, edge computing, and more energy-efficient technologies. Businesses should also revise their sustainability goals and explore innovative energy solutions, such as sodium-ion batteries or small nuclear reactors, to support long-term operations.