40% of AI data centres to face power constraints by 2027
Written by Nicole Cappella Thu 14 Nov 2024
A total of 40% of the world’s artificial intelligence (AI) data centres could suffer operational limitations due to power constraints, according to a new report from Gartner. This is based on the rapid rise of AI and generative AI, which is driving unprecedented energy demands across the globe.
The report projected the growth in power consumption needed to support AI applications could surgy by 160% over the next two years. By 2027, AI data centres could consume 500TWh (Terrawatt hours) annually. That is two times the amount consumed in 2023.
“The explosive growth of new hyperscale data centers to implement GenAI is creating an insatiable demand for power that will exceed the ability of utility providers to expand their capacity fast enough,” said Bob Johnson, VP Analyst at Gartner.
This escalating demand is projected to disrupt power availability on the grid, limiting data centre construction and expansion, and the scalability of GenAI projects by 2026.
The anticipated energy demand is also sparking concern across the tech sector. Industry players, including Gartner and IDC, have raised alarms about how AI workloads could exacerbate greenhouse gas emissions, potentially affecting sustainability goals and corporate social responsibility commitments.
In 2023, according to IDC, a standard 1MWA data centre consumed 6.6GWh of electricity. With capacity expansion, increased utilisation, and higher-density computing, that amount could grow to as much as 16GWh. This aligns with the prediction that the energy consumption for data centres will see a compound annual growth rate (CAGR) of 40.5% through 2027.
“Significant power users are working with major producers to secure long-term guaranteed sources of power independent of other grid demands,” said Johnson. This could spell trouble for smaller businesses and residences, who may have to contend with higher prices and even instability on the power grid.
Johnsons stressed in the meantime, the cost of power to operate data centres will increase significantly as operators use economic leverage to secure needed power. These costs will be passed on to AI/GenAI product and service providers as well.
The report follows concerns raised last week that data centres are rapidly multiplying worldwide to support the booming growth of artificial intelligence (AI). However, AI technologies consume significantly more energy than other data centre services, meaning that not only will more data centres be needed, but each will also have higher energy demands than before.
Written by Nicole Cappella Thu 14 Nov 2024