Google has used its carbon-intelligent computing platform, built to shift energy use at its data centers to times when clean energy is more readily available, to respond to energy crises around the world, according to an October blog post by the company.
The tech giant plans to continue piloting the software as a demand response mechanism in partnership with grid operators, said Savannah Goodman, data and software climate solutions lead for Google.
Demand management and other efforts to improve energy efficiency at data centers could have a significant impact on the grid, according to Baosen Zhang, an associate professor of electrical and computer engineering at the University of Washington.
Google's efforts to achieve 24/7 clean energy could have benefits beyond its goal of eliminating carbon emissions from its operations, according to the company's latest dispatch.
Google reports that it used its carbon-intelligent computing platform to respond to extreme weather events in Oregon, Nebraska and in the Southeastern U.S. in collaboration with utility partners in those areas. The company also used the software to reduce electrical demand and support energy security in Europe by reducing its power consumption from 5 p.m. to 9 p.m. during Europe's December 2022 to March 2023 energy crisis.
Developed in 2020, Google's carbon-intelligent computing platform allows the company to direct its data centers to perform non-urgent computing tasks when renewable energy resources are more widely available to reduce its carbon footprint. But as demonstrated over the past year, the platform can also be used to respond to local grid events to maintain reliability, the company said.
“The demand response capability leverages the technical foundations of this platform to deliver targeted energy reductions at specific data centers based on an event request from a local grid partner — typically during the system wide peak load,” Goodman said. “In other words, demand response is a new capability that leverages existing technical infrastructure to serve a new set of use cases and objectives.”
Google continues to work with grid operators to test and expand its demand response capabilities, Goodman said.
Google isn't the only large tech company testing demand response initiatives, Zhang said. With more utilities implementing rate structures that penalize power use during peak hours and reward customers for having more flexible power needs, data centers have begun to optimize their operations in response.
And while it can be difficult to calculate exactly what this will mean in terms of carbon emissions — a question that multiple companies are tackling — Zhang said he sees significant potential in demand management programs and other initiatives to reduce electrical use at data centers.
Data centers may represent a mere fraction of the nation's total energy use, he said, but energy use by data centers is growing rapidly while demand from the rest of the economy remains relatively flat. And data centers typically have more flexibility in terms of when they need energy compared to other large, industrial customers, Zhang said.