Anthropic Says It Will Cover Electricity Price Hikes from AI Datacenters

Artificial intelligence company Anthropic announced Wednesday it will absorb electricity price increases that consumers face due to the company's datacenter operations, amid growing concerns about AI infrastructure's impact on utility costs. It did not specify any particular datacenter in its blog post. 

This comes as AI companies race to build massive datacenters requiring unprecedented power capacity. According to Anthropic, training a single frontier AI model will soon require gigawatts of power, with the U.S. AI sector needing at least 50 gigawatts of capacity over the next several years.

Under the policy, Anthropic said it will cover 100% of grid upgrade costs needed to connect its datacenters, including infrastructure like transmission lines and substations that would typically be passed to ratepayers. The company said it will pay these costs through increases to its own monthly electricity charges.

The company also  said it will procure new power generation to match its datacenters' electricity needs. When datacenters begin operating before new power plants are built to supply them, increased demand on existing grids can drive up prices for all consumers. Anthropic stated it will work with utilities and independent experts to calculate these price increases and reimburse the difference.

The company said it plans to invest in curtailment systems that reduce power usage during peak demand periods and grid optimization tools designed to keep prices lower for consumers.

The claude maker said that Anthropic's datacenter projects are expected to create hundreds of permanent jobs and thousands of construction positions. The company stated it will deploy water-efficient cooling technologies and partner with local leaders on community initiatives.

For data centers developed with partners, Anthropic said it will directly implement these commitments. For leased capacity arrangements, the company is considering additional ways to address the potential impact of its workloads on electricity prices.

The announcement comes amid broader infrastructure investments across the AI industry, as companies work to secure computing resources while responding to concerns about energy use and electricity costs.