The world’s wealthiest corporations are investing staggering amounts of capital in data centers. In fiscal year 2025 alone, Amazon will likely spend $105 billion on data center investments, Microsoft anticipates investing approximately $80 billion to build out data center infrastructure, and Meta predicts investing between $60 and $65 billion to meet its artificial intelligence goals. To power these centers—including the computer servers, networking hardware, and cooling equipment that supports their cloud computing and other data processing applications—Big Tech will need an astonishing amount of electricity. Which raises two questions: where will they get it? And who’s paying for it?
The electric utility industry is one of monopolists. Investor-owned utility companies have monopoly service territories and enjoy state-set rates, which insulate utilities from ordinary business risks, in part by shifting risks to the public. These utilities are currently competing to serve Big Tech customers because their shareholders stand to reap enormous profits from developing the power infrastructure to serve data centers. And to provide attractive rates to secure Big Tech companies as customers, utilities are spreading data centers’ energy costs across an unwitting public.
This union of corporate power—where utilities hide subsidies for trillion-dollar companies in our power prices—could result in a massive wealth transfer to the shareholders of Big Tech and utility companies, which are already some of the richest corporations in the world. In a new paper, Ari Peskoe and I explain how utilities shift the costs of data centers’ electricity consumption to other ratepayers. Based on our review of nearly fifty proceedings about data centers’ rates, and the long history of utilities exploiting their monopolies, we conclude that utilities are exploiting their monopolies and control over rate-setting processes to subsidize Big Tech’s voracious energy appetite. Without systematic changes to prevailing regulatory practices, utility companies appear poised to continue socializing the costs of serving Big Tech.
How Data Center Costs Creep into Ratepayers’ Bills
First, a quick primer on how electricity regulation works in the United States. Government approved electricity prices reimburse utilities for their operational expenses and provide utilities an opportunity to earn a fixed rate of return on their capital investments. To protect the public from a utility’s monopoly power, while motivating the company to provide reliable and cost-effective service, state public utility commissions (PUCs) determine whether utility service is offered to all consumers within a utility’s service territory at rates and conditions that are “just and reasonable.” This legal standard requires PUCs to balance captive consumers’ interests in low prices and fair terms of service against the utility’s interest in maximizing returns to its shareholders. A utility rate case, in which a utility company makes an application to adjust its rates, is the PUC’s primary mechanism for balancing these interests.
In a rate case, after the utility establishes the amount of revenue it will require to cover its operating expenses and earn a profit on its capital investments, the utility then proposes to divide this amount among groups of consumers based on their usage patterns, infrastructure requirements, and other characteristics. If approved by regulators, all utility ratepayers end up sharing the costs of the utility’s operating expenses—including any upgrades and expansions—pursuant to a complicated cost allocation formula, approved by the PUC, that roughly tries to align consumer prices with the costs the utility incurs to provide service to that customer group. This cost causation process is fraught, as each party advocates for their own self-interest by arguing that lower rates for itself align with economic principles, fairness, and other subjective values.
For many utilities, their expectations about growth are now dominated by new data centers. Some utilities are even projecting a doubling or tripling of total energy consumption within the decade—an astonishing rate of growth that has not occurred in 70 years. The easiest way for utilities to shift data centers’ energy costs to the public is simply to follow long-standing cost-allocation practices in rate cases. If state regulators allow utilities to follow the conventional approach of socializing new expansion, utilities will impose data centers’ energy costs on the public.
However, our research has identified three additional ways that utilities can shift an even greater share of Big Tech’s energy costs onto consumers: through secret contracts, the gap between federal and state regulation, and “co-locating” data centers and existing power plants.
Shifting Costs Through Secret Contracts
Our research revealed that data centers can pay special negotiated rates that are developed by the utility outside of rate cases. While rate case decisions are lengthy documents that engage with the evidence filed by utilities and other parties, most public utility commission orders approving special contracts simply conclude that the proposed contract is reasonable without meaningfully engaging with the proposal. One challenge here is that few, if any, parties participate in these proceedings, and the PUC often reflexively grants utility requests to shield its rate proposal from public view.
When the special contract rate a utility offers to data centers is lower than the utility’s cost to serve that customer, a utility can intentionally force ratepayers to subsidize special contracts. For example, recent federal antitrust litigation against Duke Energy, one of the largest utilities, exposed that the company shifted costs of a special contract to its other ratepayers. Internal documents disclosed through litigation revealed that a new power plant developer was far more efficient than Duke, and as a result, the utility could not compete for customers based on price. Nonetheless, Duke offered one of its larger customers a new contract that amounted to a $325 million discount. In internal documents, “Duke officials disclosed a plan to shift the cost of the discount…back to its [other ratepayers] in years to come.”
To protect ratepayers, some state laws authorizing special contracts require state regulators to approve the special rates. But these laws are difficult to implement, since utilities tell PUCs what they want to hear to approve the contracts: that the deals they are offering to Big Tech isolate data center energy costs from other ratepayers’ bills and won’t increase power prices. Verifying utility claims are all but impossible, in part because state regulators act based on evidence and there are often no other parties to vigorously challenge the utility in special contract proceedings. Regulators also often face political pressure to approve the contracts because the data center investments are already touted by elected officials for their economic impacts.
Shifting Costs through the Gap Between Federal and State Regulation
When a state PUC approves a utility’s revenue requirement, it must allow the utility to include interstate transmission and wholesale power market costs that are regulated by the Federal Energy Regulatory Commission (FERC). State PUCs then apply their own formula for dividing FERC-allocated transmission costs among ratepayers. Because of the way that regional transmission costs are allocated at FERC and then by state PUCs, residential ratepayers could end up paying the majority of transmission costs that are tied to data center growth.
In December 2023, for example, the PJM Regional Transmission Organization (RTO), a utility alliance stretching from New Jersey to Chicago and south to North Carolina, approved $5.1 billion of transmission projects whose costs would be shared among PJM’s utility members. PJM identified two factors driving the need for this transmission expansion: retirement of existing generation resources and “unprecedented data center load growth,” primarily in Virginia. With FERC approval, PJM assigned approximately half of the total cost to Virginia utilities, approximately 10 percent to Maryland utilities, and the remainder to utilities across the region. Each state PUC then allocated the costs assigned by PJM to ratepayer classes of each utility it regulates. The consequence of these cost-allocation processes is that in both Virginia and Maryland, residential ratepayers are paying the majority of regional transmission costs that are tied to data center growth.
Even if PUCs fix their cost allocation methods by assigning costs to data centers, ratepayers still could be on the hook for an unfair share. For example, if a utility spends hundreds of millions of dollars building out power infrastructure for data centers that never materialize, then ratepayers will be left holding the bag. These stranded cost-risks, which would be borne by shareholders in another industry, are instead shifted to the public by the utility.
Shifting Costs by “Co-Locating” Data Centers and Existing Power Plants
Power plant owners have also developed schemes for attracting data centers that could shift energy costs from data centers to ratepayers. Under “co-location” arrangements, a data center connects directly to an existing power plant behind the plant’s point of interconnection to the utility owned transmission network.
When this happens, the power plant stops competing in interstate auction markets to supply energy to utility companies and instead directly provides its power to the data center customer. As data centers look to “co-locate” with existing generation, particularly nuclear power plants, they risk driving up prices in interstate power markets as entire swaths of existing generation resources are removed from the market. At least in the short term, as gigawatts of plants leave the market, supply will be constrained, which will drive up prices.
Extracting Value from the Public
The economic harm to ratepayers from data center discounts extends beyond rate increases. To meet data center demands, utilities are delaying opportunities to initiate power sector reforms that would benefit all ratepayers and instead proposing more of the same: expending capital on large central-station power plans and transmission reinforcements. These projects have been fueling utility profits for generations, but the power sector today can do more with advanced technologies and practices that can get more out of existing infrastructure. The facade of urgency created by utilities and data centers prevents such reimagining. And as utilities wring profits from the public, they may be developing a new alliance with Big Tech companies. Uniting utilities’ influence-peddling experience with the deep pockets of Big Tech could further entrench utility control over the regulatory process.
Utility rates have always been used as a means of achieving economic and energy policy goals. By financing favored investments through utility rates, rather than through general government revenue, policymakers avoid having to raise taxes and instead, conceal public spending through complex utility rate increases. The hidden subsidies for trillion-dollar companies in our power prices reflects an economic policy that favors Big Tech corporations at the expense of the public. This is not a new story, but the scale of public subsidies for data center companies constitutes a new threat for all of us.
