Anthropic vows to protect consumers from rising electricity costs – News

Anthropic PBC has become the latest artificial intelligence model maker to pledge that its vast energy consumption will not drive up the cost of electricity for regular citizens.

The company said today that it’s going to work with utilities providers to “estimate and cover” the likely price increases that will result from its data centers’ massive consumption. The pledge applies specifically to facilities located in areas where it cannot generate enough new energy on its own. It will also foot the bill for the infrastructure upgrades that electricity providers need to make to connect its energy-intensive data centers to the power grid.

In a blog post, the company pointed out that the country needs to build new data centers quickly in order for it to remain competitive in the AI race. “The U.S. AI sector will need at least 50 gigawatts of capacity over the next several years,” the company said. “But AI companies shouldn’t leave American ratepayers to pick up the tab.”

Because AI data centers consume such enormous amounts of energy, there’s a very real risk that demand will outstrip available capacity, and if that happens, then the cost of electricity will increase. That’s because of the expensive infrastructure that must be put in place to support this increased demand. Expanding the grid requires intricate planning studies and the purchase and installation of new hardware such as transmission lines and substations, and this all costs money.

Traditionally, when electricity providers need to expand capacity, they increase their rates to fund these efforts. But in the case of AI, the expansion is going to be so fast that it isn’t fair to leave regular Americans to foot the bill, Anthropic says.

“We will pay for 100% of the grid upgrades needed to interconnect our data centers, paid through increases to our monthly electricity charges,” the company said. “This includes the shares of these costs that would otherwise be passed onto consumers.”

In addition, Anthropic promised to try to procure new power sources to match its electricity needs, so it’s not putting additional strain on the grid. That includes investing in “curtailment systems” that will optimize its data centers’ energy usage during periods of high demand, in order to minimize the impact of its consumption. Finally, it said it will invest in local communities wherever it builds new data centers, while striving to address any environmental impact its facilities might have.

AI firms under pressure to step up

Though Anthropic’s pledge is commendable, it’s likely that the company felt pressured to do so, given that rivals including OpenAI Group PBC and Microsoft Corp. have made similar commitments to reduce the cost impact of their data centers. AI companies are facing heavy political pressure to take action. Last week, New York senators introduced a bill that would temporarily halt any new permits for data centers in the state. It also calls for a new state commission to be set up to look at ways to minimize the impact of data centers on electricity and gas rates.

Last month, Sen. Chris Van Hollen, D-Maryland, and a number of other Democrats proposed a bill that would oblige AI companies to foot the bill for any increases in electricity prices and also pay for the infrastructure if required to add capacity to existing grids.

It’s not just Democrats who are taking issue with electricity costs. The White House last month appealed directly to electricity providers in the Mid-Atlantic region to stop the rising costs of electricity for consumers. The Trump administration is reportedly drafting a voluntary agreement that would compel AI firms to offset any increases in household electricity costs.

Data centers used about 4.4% of all electricity generated in the U.S. in 2024, according to research by the Lawrence Berkeley National Laboratory. However, that number is expected to rise to 12% by 2028. A later report by Carnegie Mellon University and North Carolina State University said this demand increase could lead to a 25% increase in consumer electricity costs by 2030.

Image: News/Microsoft Designer