Capgemini: Utilities Overlook AI Data Centre Power Demand

Utility providers are facing severe challenges in forecasting the electricity required to sustain the rapid expansion of AI within data centres. 

A recent report published by Capgemini, titled AI Meets the Grid: Shaping the Data Center Power Play, indicates that 77% of utilities are struggling to accurately predict the energy demand generated by these intensive compute operations.

The research highlights a significant shift in infrastructure requirements, noting that electricity consumption dedicated strictly to AI training and inferencing is projected to surge. Within the next three to five years, these specific processes are expected to account for 60% of total data centre electricity demand, up from the current level of 25%. 

This rapid growth is anticipated to largely displace traditional IT workloads as operators pivot their infrastructure to accommodate the intensive power draw of advanced processors.

Volatility in compute workloads

Forecasting energy requirements has become difficult for electricity providers because AI compute patterns are inherently unstable. 

Developing and operating these systems involves consumption phases that resist traditional modelling techniques. 

For instance, if a technology company trains a new foundation model, the hosting facility might experience a massive spike in compute demand that could last for several weeks. Once the initial training phase concludes, power usage might then suddenly drop to a baseline level. From that point, the energy required can fluctuate wildly in response to unpredictable real-time user queries during the inferencing stage.

This variability has emerged as a major system challenge for regional grids, forcing infrastructure operators to seek new approaches to grid planning and operational resilience.

“AI is reshaping the energy landscape, both accelerating electricity demand and creating new opportunities to improve how energy systems are planned, managed and optimised,” says Claire Gauthier, Global Head of Energy and Utilities at Capgemini.

“The challenge facing the industry extends beyond adopting AI. It is embedding intelligence into the way organisations operate, enabling them to continuously adapt, make better capital allocation decisions and respond more effectively to an increasingly volatile environment.”

AI for grid optimisation

To manage this grid volatility, the utility sector is evaluating the technology itself as a definitive tool for infrastructure management. 

Capgemini surveyed 600 electricity executives to understand their approach to network reliability. The majority views AI as a force multiplier for grid planning and reliability. Around 60% of executives expect advanced analytics to deliver improvements of over 10% in failure reduction and operational productivity. They anticipate similar gains in preventing outages and restoring systems when network failures occur.

Despite these expectations, actual deployment remains limited across the sector. Less than half of the surveyed utility leaders (45%) utilise AI for grid optimisation. Furthermore, only 16% of electricity organisations have successfully implemented advanced data-driven approaches to optimise power flows across their regional networks.

“For the first time, organisations have the opportunity to optimise the energy value chain end to end by bringing together engineering expertise, operational technology, digital capabilities and AI within a single operating model,” Claire says.

Securing a diversified energy mix

Data centre operators are altering their power procurement strategies to support heavy computing environments without sacrificing facility reliability. There is a marked transition away from relying entirely on renewable sources like wind and solar in isolation. 

Instead, operators are investing heavily in a diversified energy mix to guarantee uptime. The ability to function independently from the main utility network is valuable, with 86% of operators viewing off-grid capabilities as a major competitive advantage.

Leaders across both the technology and utility sectors agree this varied approach is essential for long-term growth. To bridge the intermittent gaps in renewable generation, these organisations are actively funding industrial-scale battery storage systems.

Gerhard Salge, CTO of Hitachi Energy, asserts that data centres require inherent flexibility in their power design.

“First of all, you need to have complementary power delivery from, for example, solar, wind or hydro,” Gerhard says.

“Then, you can play with the profiles of which complement each other and exchange power when one is high and the other might be low. The more complementary your sources of energy are, the better you can balance those changes out.”

He notes that these hardware installations provide a vital buffer against grid instability.

“Storage can also help to fill in gaps when you have an oversupply,” Gerhard says.

By capturing excess generation and deploying it during periods of clean energy scarcity, electricity executives gain the flexibility to manage unpredictable power shortages caused by sudden compute spikes.

“Those organisations that build operational agility into the core of their business will be best placed to create long-term value while strengthening resilience, competitiveness and sustainability,” Claire says.

The Capgemini findings underline the immediate requirement for electricity providers to modernise their infrastructure and adapt their planning models to match the specific power consumption patterns of modern data centres.