Skip to content
AI Usage3 min read

Google Expands PJM Virtual Power Plant Capacity to Over 1 GW

Google tenfold scales its Voltus virtual power plant agreement to 1,170 MW in PJM, deploying aggregated distributed energy capacity to help meet AI grid demand.

AB

Author

AUG Bot

Published

Digital representation of distributed energy storage and power grid capacity serving Google data center footprint

Google Expands PJM Virtual Power Plant Capacity to Over 1 GW

Big Tech uses distributed demand response to relieve AI grid strain

Google has dramatically expanded its virtual power plant agreement with demand response aggregator Voltus, adding 1,070 MW of flexible capacity in the PJM Interconnection to help meet the tech giant's burgeoning data center energy demands.

Key details

Under the expanded contract announced Thursday, Google is scaling up an initial 100 MW agreement signed with Voltus in June 2026 to more than 1,170 MW across the PJM region. Voltus will aggregate distributed energy resources from residential, commercial, and industrial sites, including battery energy storage, smart thermostats, on-site backup generation, and flexible industrial load controls. Voltus expects to ramp up the virtual power plant to its full 1.17 GW capacity over three to four years, maintaining the operational agreement with Google through 2040.

When grid conditions in PJM are strained, Voltus will offer the aggregated flexible demand into the wholesale power market. In PJM, demand response payments can reach up to $350,000 per megawatt-year for participating customers. Driven by surging capacity auction prices and data center load growth, total demand response revenue in PJM jumped 240% year-over-year, reaching $523.6 million in 2025, with emergency demand response accounting for 88% of all payments.

Why this matters

The tenfold expansion demonstrates how hyperscalers are leveraging "bring your own capacity" strategies to offset the immense electrical footprint of their AI computing infrastructure. By paying to aggregate and dispatch distributed assets, Google creates new flexible grid capacity without waiting for multi-year transmission upgrades or new fossil-fuel power plant construction. This approach lowers peak demand on PJM—the nation's largest wholesale power market—and helps mitigate upward pressure on electricity rates for all regional grid customers.

Context

PJM oversees the electrical grid serving 65 million people across 13 Eastern and Midwestern states plus Washington, D.C. The grid operator has faced severe capacity shortfalls and record auction price spikes driven by rapid AI data center expansion. While regulatory proposals to force large load additions to supply their own generation have met pushback at the Federal Energy Regulatory Commission, long-term corporate virtual power plant agreements effectively pioneer a commercial mechanism for data centers to fund their own grid flexibility.

What happens next

Voltus will begin recruiting and onboarding distributed energy asset owners across PJM territory to build out the 1,070 MW addition over the next 36 to 48 months. Industry analysts expect other hyperscalers to replicate Google's long-term VPP contracting model in high-density data center regions like Northern Virginia, Illinois, and Ohio to satisfy grid connection requirements and maintain power reliability.


Source: Utility Dive Published on AI Usage Global, author: AUG Bot

Related

Read more

More posts that expand on the topics, companies, and AI trends covered in this story.

Digital representation of a municipal emergency declaration and data center infrastructure in Oregon
AI Usage

Oregon City Declares Emergency Over Lost AI Data Center Revenue

The City of Umatilla declares a local state of emergency as Governor Kotek's data center permit pause threatens $5 billion in private AI investments and municipal revenue.

Former steel mill industrial site in Alton, Illinois proposed for 1 GW data center redevelopment
AI Usage

Gigawatt-scale data center proposed for steel mill in Illinois

TerraWatt proposes converting a 381-acre former steel mill in Alton, Illinois, into a 1 GW AI data center with closed-loop cooling and developer-funded grid upgrades.

Aerial render of data center facility in Floyd County Georgia
AI Usage

Microsoft files for data center campus in Floyd County, Georgia

Microsoft files DRI application for 422MW, 1.66M sq ft data center campus in Floyd County, Georgia, with 13,500 gpd water demand.