AI power demand volatility damages data center equipment
AI workloads cause rapid power demand fluctuations in data centers that spike up to 50% above design capacity, damaging equipment like batteries, generators and cooling systems faster than expected and causing potential revenue losses from outages and grid instability.
What happened
Rapid swings in AI data centers' power demands are straining vital equipment, causing batteries, generators and cooling systems to malfunction or wear out far sooner than expected.
AI power usage spikes at times as much as 50% above design capacity.
Reported impact
These technical problems suggest added costs and unforeseen reliability problems, with even a few minutes of lost uptime hitting data-center developers' revenue.
The problems are a potential source of wider instability in power grids that are already straining to keep the lights on.
Index analysis
Amber Villegas-Williamson, principal consultant at the Uptime Institute, stated that AI creates very unusual power demand patterns, comparing the effect to over-revving an engine which wears it out faster than maintaining constant speed.
Facilities designed for AI computing differ from traditional data centers because their demand is so large and swings much more dramatically.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.