- The demand for electricity is now growing faster than the energy suppliers can keep up
- Volatile AI workloads cause unpredictable peaks and troughs in demand
- AI could actually help predict, although it is also the cause
With three in four (77%) electricity executives now believing data center energy demand will grow faster than utilities can keep up, two-thirds (68%) expect power shortages to become more common as demand for artificial intelligence increases.
New data from a Capgemini report reveals how unpredictable AI energy needs can be, with 77% admitting they struggle to accurately predict demand amid volatile AI workloads.
This leads not only to more limited energy supply, but also more extreme and less predictable spikes in demand.
Demand for data center energy is a whole new ball game
All of this comes as local opposition continues to mount against data centers, with residents increasingly worried about power outages and rising energy costs. Last week, a Virginia county asked data centers to return to backup generators to free up grid capacity for local residents, with an ongoing heat wave causing a spike in demand for electricity for air conditioning.
Even data center companies struggle to predict how much they can consume, with 67% of electricity managers reporting speculative applications for future capacity. Around a fifth (19%) of these are not even realized, creating what Capgemini calls ‘phantom demand’, forcing utilities to either over-invest unnecessarily or under-invest and create capacity shortages.
“The challenge is no longer just how much power is needed, but whether it can be delivered reliably where and when it is required,” Capgemini Global Head of Energy and Utilities Claire Gauthier wrote, citing AI’s potential to help predict demand despite also being the cause of fluctuating and high demand. But currently, fewer than half (45%) are currently using AI for web optimization.
Looking ahead, most (87%) data center operators expect electricity consumption to increase over the next three to five years by an average of 30%.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds.
