AI Data Centers Face Power Crunch and Rising Costs

A report by Gartner indicates that 40% of AI data centers may encounter operational challenges by 2027 due to electricity shortages, as the energy demand is expected to grow by 160%. Power shortages and rising costs threaten the sustainability and expansion of AI technologies globally.

The rapid expansion of artificial intelligence and generative AI technologies is poised to strain global power supplies significantly. A recent report from Gartner Inc. warns that 40% of AI data centers may face operational challenges by 2027 due to looming electricity shortages. This comes as the energy demand for these centers is set to grow by an estimated 160% over the next two years, pushing electricity consumption to unprecedented levels.

The power usage in AI data centers is projected to escalate from 261 terawatt-hours in 2024 to 500 terawatt-hours by 2027, effectively doubling the levels seen in 2023. The surge is attributed to the increasing need for more extensive data centers capable of handling the immense computing power required for training and deploying large language models that are essential to GenAI applications.

Despite the growing demand, utility providers are struggling to enhance their capacity swiftly enough, creating a potential energy deficit that might disrupt operations by 2026. Bob Johnson, VP Analyst at Gartner, highlighted the intent to establish new, larger data centers to manage the substantial data volumes necessary for training large language models. However, he cautioned that short-term power shortages are expected to persist for years since new power infrastructure developments could take considerable time to materialize. This shortfall is anticipated to escalate operational costs due to rising electricity prices, affecting companies relying heavily on AI technologies.

(With inputs from agencies.)