Gartner warns of AI data centre power shortages by 2027
Gartner forecasts that 40% of existing AI data centres will face operational constraints due to power shortages by 2027.
The prediction stems from a report by Gartner, which states that the rapid increase in electricity consumption for AI and generative AI (GenAI) solutions will surpass utility providers' capacity to expand their infrastructure promptly. This shortfall could lead to significant power shortages affecting data centre growth.
Bob Johnson, Vice President Analyst at Gartner, explained the situation, "The explosive growth of new hyperscale data centres to implement GenAI is creating an insatiable demand for power that will exceed the ability of utility providers to expand their capacity fast enough. In turn, this threatens to disrupt energy availability and lead to shortages, which will limit the growth of new data centres for GenAI and other uses from 2026."
Gartner's estimates indicate that the power requirements for operating AI-optimised servers within these centres could reach 500 terawatt-hours (TWh) annually by 2027, marking a 2.6 times increase from the levels recorded in 2023.
New, larger data centres are planned to manage the volume of data necessary for training and implementing large language models (LLMs) crucial for GenAI applications. However, Johnson noted, "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 availability of power is set to dictate the number of new data centres and the growth trajectory of GenAI developments in the near future. Gartner advises organisations to assess the potential risks that power shortages could pose across their product and service offerings.
As a consequence of looming shortages, electricity prices are expected to rise, further inflating the costs of maintaining LLMs, as per Gartner. Johnson added, "Significant power users are working with major producers to secure long-term guaranteed sources of power independent of other grid demands. 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."
As organisations plan for the future, they should factor in these anticipated higher power costs and seek long-term contracts with data centre services at favourable power rates. Moving forward, consideration should also be given to alternative methods that demand less power.
Gartner also points out that sustainability objectives are likely to be compromised. Zero-carbon goals are at risk as suppliers may resort to extending the operation of fossil fuel plants beyond their planned decommissioning to meet demand surges.
Johnson noted the environmental impact: "The reality is that increased data centre use will lead to increased CO2 emissions to generate the needed power in the short-term. This, in turn, will make it more difficult for data centre operators and their customers to meet aggressive sustainability goals relating to CO2 emissions."
Gartner highlights the need for 24/7 power availability for data centres, which renewable sources cannot provide without alternative solutions during inactive periods. At present, hydroelectric, fossil fuel, or nuclear power plants offer reliable solutions. In the future, advancements in battery storage technologies and clean energy sources like small nuclear reactors could help achieve sustainability targets.
Organisations are encouraged to reassess sustainability goals concerning CO2 emissions, keeping future data centre requirements and power sources in mind. When developing GenAI applications, it is advised to focus on minimising computing power usage and considering other options like edge computing and smaller language models.