Google has signed agreements with Kairos Power to supply small modular reactors for its AI data centers, securing low-carbon electricity capacity. Industry forecasts predict global data center power demand will rise 165% by 2030, pushing Alphabet to lock in reliable power to support its AI model training needs.
Global data center electricity demand is projected to increase 165% by 2030, driven by energy-intensive AI model training. A single AI query consumes roughly ten times the power of a standard web search, highlighting the scale of the bottleneck facing hyperscalers.
Google has finalized agreements with Kairos Power to deploy small modular reactors that will deliver a stable, low-carbon power supply to its AI data center network. This move secures long-term electricity capacity critical for large-scale model training and hyperscaler expansion.
By locking in SMR capacity, Google aims to reduce operational energy costs and ensure uninterrupted power for its AI workloads. This strategic initiative strengthens its competitive position against other cloud providers and supports future AI infrastructure investments.
