- US senators accuse Big Tech of leaving households to pay soaring electricity bills
- Data centers consume hundreds of megawatts, severely straining regional power grids
- Private contracts hide which companies are truly paying for power expansion
Three U.S. Democratic Senators – Elizabeth Warren, Chris Van Hollen, and Richard Blumenthal – are pressing major technology firms to explain why electricity bills continue to rise in regions packed with large data facilities.
Their letters target companies deeply invested in cloud hosting and large-scale artificial intelligence infrastructure.
The lawmakers argue that public assurances about absorbing power-related costs do not align with what consumers are experiencing through higher utility rates.
Tech firms under fire for power bill failures
“Tech companies have paid lip service in support of covering their data centers’ energy costs, but their actions have shown the opposite,” the trio wrote.
“When utilities expand their grid infrastructure, they incorporate the cost of expansion into their utility rates, passing the extra costs onto their customers,” they added.
On the same day the letters became public, Amazon released a study it had commissioned from Energy and Environmental Economics.
The report claims that data center hosting facilities generate enough revenue for utilities to offset the cost of serving them.
In some scenarios, the study suggests surplus revenue could even benefit other ratepayers.
However, the analysis relies heavily on projections and modeled outcomes rather than verified historical billing data.
There is little disagreement that modern data centers consume a large amount of electricity.
Facilities supporting AI workloads often require hundreds of megawatts, with some approaching gigawatt-scale demand.
Many regional grids were not built for this level of sustained consumption, forcing utilities to invest billions in new generation, transmission lines, and local upgrades to keep servers online reliably.
According to the Senators, utility companies typically recover infrastructure expansion costs by raising rates across their customer base.
This means residential and small business users absorb expenses tied to industrial-scale computing projects.
Research cited in the letters points to electricity prices potentially rising 8% nationwide by 2030, with far steeper increases in data center-dense states such as Virginia.
A recurring concern involves private contracts between utilities and technology companies.
Studies referenced by lawmakers indicate that many firms successfully negotiate favorable rates while avoiding direct responsibility for grid upgrades.
Confidentiality clauses prevent regulators and the public from clearly seeing how costs are distributed.
This lack of transparency makes it difficult to reconcile corporate claims with documented increases in wholesale and retail electricity prices.
“Contracts between data centers and utility companies that set electricity prices and other terms are typically confidential,” the Senators wrote.
“Tech companies searching for a site for a new data center reportedly use hard-nosed tactics to achieve lower rates … and then [pressure] utilities to give them favorable rates by suggesting they may build elsewhere instead.”
Amazon maintains that its facilities help rather than harm ratepayers, despite anecdotal evidence and regulatory records suggesting otherwise.
Some regions with substantial data center activity have reportedly seen wholesale power prices rise sharply over recent years.
Projections about potential benefits remain difficult to square with current billing trends, leaving open questions about who ultimately pays for the rapid expansion of AI-driven infrastructure.
Via The Register
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
Add Comment