The rapid expansion of AI-driven data centres is making electricity demand more volatile and difficult to predict, according to a new report by the Capgemini Research Institute. As AI workloads grow, utility providers are facing unprecedented pressure to supply reliable power while managing increasingly unpredictable spikes in energy consumption. The report highlights that traditional power grids must evolve to support the next generation of AI infrastructure.
Capgemini's study found that 80% of utility executives expect AI-powered data centres to create more extreme and less predictable electricity demand patterns. Many energy companies are therefore investing in AI-based grid management, demand forecasting, and smart energy systems to balance supply and demand more effectively. The report also points to growing interest in on-site power generation, renewable energy, natural gas, and small modular nuclear reactors (SMRs) to meet future electricity needs.
The report emphasizes that utilities must modernize their infrastructure to accommodate the rapid growth of AI computing. This includes expanding transmission networks, deploying digital grid technologies, and improving coordination between power providers and data centre operators. AI itself is expected to play a key role in optimizing electricity distribution, detecting grid disruptions, and improving operational efficiency across power systems.
Capgemini concludes that the AI revolution is reshaping not only the technology industry but also the global energy sector. As demand for AI computing continues to accelerate, collaboration between technology companies, utilities, and policymakers will be essential to ensure reliable, sustainable, and resilient electricity infrastructure capable of supporting the next generation of AI-driven innovation.
