Google agrees to cut AI data center power use during peak demand

Reports indicate that Google has become the first U.S. Big Tech company to employ demand-response programs as a resolution to the energy concerns resulting from high AI demand.

The rapid growth of AI infrastructure is happening faster than the power grid can adapt, leading to concerns about energy shortages and higher electric bills. As a result, Google has entered into an agreement with two U.S. utilities, Indiana Michigan Power and the Tennessee Valley Authority, to cut power to its data centers during periods of peak demand.

Google agrees to cut AI data center power use during peak demand

Google entered into agreements with two U.S. electric utilities, Indiana Michigan Power and the Tennessee Valley Authority, to curb power usage from its artificial intelligence (AI) data centers during periods of peak electricity demand. The agreement was announced on Monday.

The demand for AI-driven services like large language models, real-time data analysis, and machine learning, and the power required to support these functions have increased substantially.

Tech companies have ramped up requests for new electricity connections to power AI data centers, especially in regions where available supply is already lagging. Some areas have seen a higher demand for electricity from tech firms than the available electricity capacity.

The resulting shortage of energy has raised alarms among grid operators and local communities. Power grids are nearing their limits and concerns have emerged over the potential for higher electricity bills for homes and businesses, increased risk of blackouts, and delays in approving new data center projects.

How Google’s new demand-response agreement works

Demand-response programs involve temporarily reducing electricity consumption during periods of high demand in exchange for financial compensation or reduced utility rates. They help ease pressure on the power grid, lower emissions, and delay the need to build new transmission infrastructure or power plants.

Google’s participation is particularly notable because it involves its machine learning workloads that typically run continuously in massive data centers. Under the agreements, Google will reduce or defer some of these workloads during periods when the utilities request it, allowing more electricity to flow to critical systems and households.

According to Google, the flexibility offered by these demand-response efforts “allows large electricity loads like data centers to be interconnected more quickly, helps reduce the need to build new transmission and power plants, and helps grid operators more effectively and efficiently manage power grids.”

The specific commercial details of Google’s agreements have not been made public, but the company may set a precedent for other major tech firms like Microsoft, Amazon, and Meta, all of which are also expanding their AI data center footprints.

Get seen where it counts. Advertise in Cryptopolitan Research and reach crypto’s sharpest investors and builders.