GCP – Supporting power grids with demand response at Google data centers

Every time you search on Google, watch a YouTube video, or send an email on Gmail, a Google data center uses a small amount of electricity to complete these tasks. To enable these services, Google also runs a number of non-urgent tasks in the background, like YouTube video processing and adding new words to Google Translate.

We’re excited to share that we have developed and piloted a new way to reduce our data centers’ electricity consumption when there is high stress on the local power grid, by shifting some non-urgent compute tasks to other times and locations, without impacting the Google services you use every day.

At Google, we work to run our data centers as efficiently as possible — and we’ve taken ambitious actions to become an energy efficiency leader. We also aim to operate our data centers intelligently, applying new technology to increase the flexibility of our energy use across our data centers around the world. Since 2020, we’ve used our carbon-intelligent computing platform to shift compute tasks and their associated energy consumption to the times and places where carbon-free energy is available on the grid.

Now, we can use this task-shifting capability for demand response — temporarily reducing power consumption at our data centers to provide valuable flexibility when it is needed, to help local grids continue operating reliably and efficiently.

Why demand response?

Historically, growing energy demand has been met primarily by adding new, often carbon-intensive, resources to the grid. Reducing demand to support grid operations was mostly an emergency measure, deployed only as a last resort. But recent research shows that demand response can be a critical tool for electricity grids, helping reduce the need for investment in new fossil fuel-based resources, support the growth of variable renewable energy sources like solar and wind, and improve grid operations.

By building out a new approach for demand response across our data centers — and paving the way for others to do the same — we are helping unlock these important grid-level benefits.

How does it work?

Our new approach builds on the software that runs our carbon-intelligent computing platform, adding new capabilities that allow us to temporarily reduce the power demand of a Google data center when called on to do so by an external power system partner, such as a utility or grid operator.

When we receive notice from a grid operator of a forecasted local grid event, for example an extreme weather event that will cause a supply constraint, we can alert our global computing planning system to when and where it will take place. This alert activates an algorithm that generates hour-by-hour instructions for specified data centers to limit non-urgent compute tasks for the duration of the grid event, and allows them to be rescheduled after the grid event has passed. When feasible, some of these tasks get rerouted to a data center on a different power grid. All of this is done without additional computer hardware and without impacting the performance of Google services like Search, Maps, YouTube, Google Cloud, and Workspace (which includes Gmail, Docs, Sheets and more) that people, businesses, and public sector organizations rely on around the clock.

When we take these actions at times of high stress on local grids, we help our grid partners continue to serve customers reliably.

Our work in action

As we’ve developed this demand response approach, we’ve continuously tested it and demonstrated its impact alongside our partners in multiple regions:

Europe: In Winter 2022-23, natural gas shortages in Europe drove energy prices to historic highs. To support EU-wide efforts to reduce electricity demand and support energy security, we implemented various measures from December 2022 through March 2023, including scheduling daily power reductions during typical peak periods (5pm-9pm) across our data centers in the Netherlands, Belgium, Ireland, Finland, and Denmark. Together, these helped support the European grid during this period of energy scarcity.Asia: In Taiwan, the geographic constraints of an islanded grid can lead to grid reliability issues, especially during the summer. In the summers of 2022 and 2023, we supported a grid reliability program run by the Taiwan Power Company by reducing our data center’s power consumption daily during peak hours. These reductions helped maintain grid reliability even as local supply margins were strained.U.S.: In Oregon, Nebraska, and the Southeast, recent extreme weather events including heat waves and winter storms have increased local power demand, straining grids, and causing energy prices to surge. In collaboration with our local utility partners, we reduced our data centers’ power consumption during these periods to help ensure that these grids could operate reliably and meet the needs of local communities.

Looking ahead

Our initial pilot programs have proven out real-world use cases for our demand response approach and demonstrated how it can support grid operations. Looking ahead, we’re excited to continue evolving our capabilities to better support the grids that serve us, and to work with partners to maximize our shared impact.

As our partners highlight:

We also recognize that realizing the full potential of demand response as a reliable, cost-effective, and clean resource for the grid will require better frameworks to incentivize large energy users to operate more flexibly. We look forward to working with other large energy users, our grid partners, and policy makers to incorporate greater demand-side flexibility that will help grids become more efficient, cost-effective, and clean.