logo
Here's how we all ended up paying Big Tech's power bill

Here's how we all ended up paying Big Tech's power bill

In several states, public utilities are asking power-hungry data centers to cover more of their mounting electricity costs.
Utilities have long enforced special terms and conditions, called tariffs, on their largest customers, to ensure they are paying their fair share.
Facing a historic surge in electricity demand amid Big Tech's AI race, utilities in states like Indiana and Ohio are now turning to tariffs to protect residents and small businesses from higher power bills as more data centers come online.
Data centers consumed about 4.4% of all electricity in the US in 2023, according to research from the US Department of Energy. That amount could triple by 2028.
That level of demand growth hasn't been seen since the 1960s, when gains in efficiency tempered the power industry's costs of growth, protecting consumers from hefty rate hikes.
Now, power plants are more expensive to build, especially the natural gas plants that some utilities are proposing to serve new data centers.
Ratepayer advocates and regulatory experts say that using tariffs for data centers is a good start, but they may not be enough to prevent the full costs of the AI boom from hitting consumer wallets.
The demand surge coming from a handful of very large customers, combined with the rising costs of building new infrastructure, could mean the old rules of utility pricing don't apply in the age of AI.
"Data center growth is overwhelming long-standing approaches to approving utility rates," researchers at Harvard Law School's Electricity Law Initiative wrote in a paper earlier this year, "causing the public to subsidize Big Tech's power bills."
Earlier this month, Ohio regulators approved American Electric Power's plan to create a tariff specifically for its data center customers. In February, regulators in Indiana approved changes to Indiana Michigan Power's large customer tariffs, aimed at shielding the utility's individual and small business customers from higher costs associated with data centers.
Several other states, including Virginia, Texas, Kansas, and California, are considering or have proposed implementing special tariffs for data centers.
The tariffs are a response to the unprecedented demand for electricity coming from Big Tech's AI data centers.
Investor-owned utilities in the US are on track to build $1.1 trillion in new infrastructure over the next five years, largely to keep pace with that demand, according to industry group Edison Electric Institute. About 72% of electricity customers in the US are served by investor-owned utilities, which profit by building new infrastructure.
This business model was born when population growth and economic development drove electricity demand, and everyone was thought to benefit from system-wide upgrades.
With Big Tech companies driving most of today's electricity demand, some are questioning which customers exactly should pay for upgrades.
"We really need to think about who should pay for this stuff," Ari Peskoe, Director of Harvard Law School's Electricity Law Initiative, told Business Insider.
"I don't think it's fair to just have these very wealthy corporations come in and force utilities to build more infrastructure and then just socialize the costs the way we normally do."
Why am I subsidizing Big Tech's power bill?
Historically, state regulators have allowed public utilities to recover the cost of investment, plus a rate of return, of new transmission lines and power plants by raising rates for all customers.
The utility business model was based on the assumption that society benefited from the grid's expansion; therefore, "we should all pay for the infrastructure needed to allow for its growth," said Peskoe.
Staggering data center power demand has flipped this assumption on its head. At a time when residential electricity use is going down, Big Tech companies stand to gain the most from major utility expansion. Under the old rules of cost allocation, everyone else might have to pay for it.
In Louisiana, customers of the state's major public utility could be responsible for $5 billion in costs associated with new natural gas plants and transmission lines needed to serve a Meta data center.
Wisconsin ratepayers could pay for a $2 million transmission project for Microsoft's data center campus outside Milwaukee. In Colorado, Xcel Energy customers could end up paying $1.7 billion for new transmission lines to serve the state's growing data center load.
What are states doing about it?
In February, Indiana regulators approved changes to Indiana Michigan Power Company's tariff for large industrial customers, setting the terms for connecting data center customers to the grid.
The changes were part of a settlement agreement between I&M, its three largest data center customers — Amazon, Microsoft, and Google — consumer advocacy groups, and the Data Center Coalition.
The revised tariff requires "large load" customers that use a lot of electricity to make "long-term financial commitments proportional to their size," I&M said in a press release.
Data centers are driving record load growth in the utility's Indiana service territory. Peak load — the highest amount of electricity used by customers at a time — is on track to more than double by 2030.
Ben Inskeep, program director at Citizens Action Coalition Indiana, a participant in the IMP settlement, said the revised tariff will help protect ratepayers from the cost of "stranded assets" — infrastructure that gets built, but ultimately doesn't get used. But it doesn't address several other aspects of data center growth that could show up in consumer electricity bills, he said.
Many factors account for the final number that customers see on a monthly electric bill, and the cost of new infrastructure is just one of them.
"We're seeing really high load growth from data centers driving the cost of capacity super high and driving the cost of building new resources super high, and those costs could still end up in consumer bills," said Inskeep.
For its part, Big Tech is fighting to keep its own electric bills from rising.
In Ohio, a lawyer for Amazon told regulators that AEP had "singled out" data centers for a "discriminatory tariff." A lawyer for Google called the tariff a "departure from the fundamental rules" of utility regulation.
That's exactly the point, said Peskoe.
"What we really need is a new approach to allocating the utilities' costs," said Peskoe. "That's what a lot of this comes down to — making sure that the data centers are paying for infrastructure that's serving them, that's being built for them."
Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

US Charges Chinese Nationals With Nvidia Chips Export Breach
US Charges Chinese Nationals With Nvidia Chips Export Breach

Bloomberg

time3 minutes ago

  • Bloomberg

US Charges Chinese Nationals With Nvidia Chips Export Breach

Two Chinese nationals were arrested this week on charges that they sent tens of millions of dollars worth of advanced AI chips made by Nvidia Corp. to China in violation of US export restrictions, according to authorities. The defendants used a company based in El Monte, California, to export sensitive technology, including graphics processing units, used in artificial intelligence without obtaining the necessary government licenses, the Justice Department said in a statement Tuesday.

Despite legal options, many in China still prefer banned GPUs for AI workloads and pay for repairs
Despite legal options, many in China still prefer banned GPUs for AI workloads and pay for repairs

Yahoo

time20 minutes ago

  • Yahoo

Despite legal options, many in China still prefer banned GPUs for AI workloads and pay for repairs

When you buy through links on our articles, Future and its syndication partners may earn a commission. AI chip repair firms in China are thriving due to heavy demand surge Smuggled GPUs fuel a booming underground repair market amid US export restrictions Chinese shops simulate data centers and fix hundreds of chips monthly at scale A quiet but growing business has emerged in China that focuses on repairing Nvidia's advanced AI chips, despite strict US export controls. Reports from Reuters found around a dozen small firms, mainly based in Shenzhen, claiming they service large numbers of Nvidia's H100 and A100 GPUs, even though these chips were officially banned from sale to China in 2022. One company told the news agency it repairs up to 500 Nvidia AI chips every month - and with roughly 12 similar firms operating year-round, that could amount to tens of thousands of chips annually. Significant demand Many of these units are worn down from heavy use, especially as some have run around the clock for years in AI training workloads. 'There is really significant repair demand,' Reuters was told by the co-owner of a Shenzhen firm that moved into AI hardware in late 2024. That demand led to the creation of a second company purely to handle AI chip repair. Their facility includes a server room that can simulate data center conditions with up to 256 servers. Another shop that shifted from GPU rentals to repairs told Reuters it fixes around 200 chips per month, typically charging about 10% of the original purchase price. Repairs can include fan replacement, circuit board fixes, memory diagnostics, and software testing. Nvidia cannot legally support or replace restricted GPUs within China. An Nvidia spokesperson said only the company and approved partners are authorized to offer the necessary service and support, adding that running restricted chips without full infrastructure is not viable long-term. The potentially high failure rate raises concerns about what will happen to tens of thousands of aging A100s and earlier GPUs once they fail. The existence of such a repair sector is down to the widespread smuggling of banned chips into China, something we've reported on previously. While Nvidia recently began offering the H20 GPU in China to comply with export restrictions, many customers there still prefer the banned H100 for training LLMs. You might also like Two million people watched this video of a mini PC being built from a slab of aluminum China has spent billions of dollars building far too many data centers for AI Microsoft to stop using China-based engineers for US military tech support

HSBC Reiterates Buy on Amazon (AMZN), $256 PT
HSBC Reiterates Buy on Amazon (AMZN), $256 PT

Yahoo

time30 minutes ago

  • Yahoo

HSBC Reiterates Buy on Amazon (AMZN), $256 PT

Inc. (NASDAQ:AMZN) is one of the . On August 4, HSBC analyst Paul Rossington reiterated a Buy rating on the stock with a $256.00 price target. Amazon has had a strong second quarter due to its growth in North America and International divisions. AWS growth was not as hoped, highlighting the need to spend more money to allow AI and cloud growth which are currently at early stages. 'While AWS growth disappointed vs Microsoft's 4QFY25 result, both updates outline the increased investment required to support growth at what remains the early stages of cloud and AI cycles, for which AMZN, with a leading share of the cloud market, looks well placed." christian-wiediger-rymh7EZPqRs-unsplash "As a result, 2Q25 capex of USD32.2bn was 25% higher than consensus of USD25.6bn. We believe this should now be considered the quarterly run-rate going forward. Cash & cash equivalents of USD61.5bn were broadly in line after taking higher capex into account.' Inc. (NASDAQ:AMZN) is an American technology company offering e-commerce, cloud computing, and other services, including digital streaming and artificial intelligence solutions. While we acknowledge the potential of AMZN as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock. READ NEXT: 10 Must-Watch AI Stocks on Wall Street and Disclosure: None.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store