Google Pays Households and Businesses to Cut Power Use to Free Up Grid Capacity for Data Centers

Google and energy services company Voltus have established a virtual power plant on the PJM grid in the United States, using financial incentives to encourage households and businesses to reduce electricity consumption, freeing up as much as 1,070 MW of power capacity for AI data centers—a scale comparable to a large nuclear reactor.

This initiative reflects the energy bottleneck brought about by the expansion of AI computing power. As demand for large model training and inference surges, tech companies' data center electricity needs continue to climb, while grid capacity and the pace of new power plant construction struggle to keep up, forcing companies to seek alternative solutions.

The virtual power plant model does not rely on building new power plants; instead, it aggregates distributed energy resources and manages grid load through demand-side response. This approach combines AI infrastructure energy challenges with public grid management and may become a reference path for the industry to address power shortages.

Google's move also indicates that competition in AI computing power is no longer limited to chips and algorithms—energy access is becoming one of the key factors constraining AI development.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.