The U.S. government will pursue power grid upgrades worth $5.25 billion in 26 states to respond to rapidly rising electricity demand. Of that, the federal government will cover $1.9 billion.

Tech outlet TechRadar reported on Sept. 30, local time, that the U.S. Department of Energy's Office of Electricity selected 31 grid upgrade projects through its SPARK program. Of the total $5.25 billion project cost, DOE will provide $1.9 billion, while project sponsors including utilities, cooperatives and state agencies will cover the remaining $3.35 billion. Federal funds account for about 36 percent.

DOE plans to rebuild more than 1,500 miles, or about 2,414 km, of transmission lines and apply grid performance enhancement technologies across about 21,000 miles, or about 33,796 km, to secure more than 23 gigawatts of additional transmission capacity. It expected about 100 million Americans to benefit in terms of grid reliability and costs.

The announcement is a selection stage, not final grant agreements. DOE expects to sign agreements between October and January next year and may adjust or halt support depending on project progress. The funding comes from the third round of the Grid Resilience and Innovation Partnerships, or GRIP, created under the 2021 Infrastructure Investment and Jobs Act.

DOE's official announcement did not put AI or data centres at the forefront of the projects. But a higher-level policy, Speed to Power, aims to secure AI competitiveness and respond to large electricity demand. Of the 31 projects, three explicitly named AI or data centres as direct beneficiaries, totalling about $371 million. Some other large projects also focused on expanding transmission networks to accommodate "new large loads" such as data centres and factories.

A key issue is cost sharing. Federal support could reduce the burden on utilities, but it is not clear from the announcement how much large users such as data centres will bear in additional transmission network costs. The impact on household electricity rates is also expected to vary depending on future local rate regulation and cost allocation methods.