- Gartner predicts 40% of AI data centres will face operational constraints by 2027.
- Power consumption for AI-optimised servers is expected to surge to 500 TWh annually.
The rapid expansion of artificial intelligence infrastructure is setting the stage for an unprecedented power crisis. Gartner predicts that 40% of AI data centers will face operational constraints due to power availability by 2027. The looming crisis highlights the growing tension between technological advancement and energy infrastructure limitations.
According to Gartner’s latest analysis, the explosive growth in electricity consumption driven by AI and generative AI pushes data centre power demands to unprecedented levels. Forecasts project growth rates as high as 160% over the next two years.
The scale of power consumption
The numbers are staggering. Gartner estimates that power requirements for AI-optimised servers in data centres will reach 500 terawatt-hours (TWh) annually by 2027—a 2.6-fold increase from 2023 levels. This surge in demand is primarily driven by the expansion of hyperscale data centres needed to implement increasingly sophisticated GenAI systems.
“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,” warned Bob Johnson, VP Analyst at Gartner.
Infrastructure challenges and economic impact
The power crisis isn’t just a future concern—it’s already affecting industry planning. Power utilities are struggling to keep pace with the rapid expansion of AI infrastructure, and new power transmission, distribution, and generation capacity could take years to come online.
The supply-demand mismatch is expected to drive up electricity prices significantly. Major tech companies are already working to secure long-term guaranteed power sources independent of other grid demands. These increased operational costs will inevitably cascade through the AI industry, affecting providers and end-users of AI services.
Sustainability goals at risk
The power shortage crisis presents a significant challenge to corporate sustainability commitments. The urgent need for more power forces suppliers to maintain fossil fuel plants beyond their planned retirement dates, potentially derailing zero-carbon initiatives.
Johnson points out a stark reality: “The reality is that increased data centre use will lead to increased CO2 emissions to generate the needed power in the short term.” The situation is particularly challenging because data centres require 24/7 power availability, which current renewable energy sources like wind or solar cannot provide consistently without robust backup systems.
Looking ahead: solutions and adaptations
The industry is exploring various solutions to address these challenges:
- Development of improved battery storage technologies, including sodium-ion batteries,
- Investment in clean power alternatives such as small nuclear reactors,
- Focus on energy-efficient computing approaches,
- Exploration of edge computing and smaller language models as alternatives.
Gartner recommends minimising computing power requirements for organisations developing AI applications and considering alternative approaches that demand less energy. The firm also advises companies to negotiate long-term contracts for data centre services to preserve reasonable prices regardless of power costs, and factor in significant cost increases when planning new products and services.
As the AI industry continues its rapid expansion, the power crisis looms as a critical bottleneck that could reshape the future of AI infrastructure development. The coming years will test the industry’s ability to balance technological advancement with sustainable power consumption and infrastructure limitations.
