The proliferation of data centers across the United States represents new “loads” (i.e., sources of demand) on the electrical grid. Data centers require enormous amounts of energy to power and cool their computing systems that operate continuously or near-continuously. To meet this demand, new energy infrastructure—both generation and transmission—will need to be developed.
For local governments and their residents, the addition of “large load customers” like data centers to the grid poses many risks. Among other things, community members may be on the hook for the capital investments needed to accommodate data-center build-out, which could increase energy bills. Additionally, the increased electricity demanded by data centers may be generated using fossil fuels, contributing to climate change, the impacts of which are often felt particularly acutely by local governments and their residents.
Protecting both residential ratepayers and the energy transition is a tall order that requires creative regulatory solutions. There are several pathways being considered and implemented, but large-load tariffs have gained widespread traction among state utility regulators. This blog post seeks to provide information useful to local leaders and their communities about large-load tariffs as they decide whether and how to engage with their state Public Utility Commission (PUC), Public Service Commission (PSC), or equivalent body around data center development and energy use.
Utility Regulatory Structure in the U.S.
The U.S. electricity grid has been called the most complex machine ever built and is the second largest source of greenhouse gas (GHG) emissions in the country. As the energy law scholar William Boyd explained in 2014, “[i]t goes without saying that the power sector must be a vital part of any effort to build a low-carbon future” in the U.S. State PUCs and PSCs hold a tremendous amount of power over that effort. Their primary responsibility is to ensure that customers have access to safe, reliable, and adequate electric (and sometimes other) service at just and reasonable prices. As part of their broad oversight and regulatory authority over electric utilities, PUCs / PSCs determine, for example, what costs utilities are allowed to recover from customers and the rate of return that utilities can earn on their investments. This oversight is critically important to the development of data centers because PUCs / PSCs are directly responsible for approving the capital investments, like new generation and transmission, that will serve data centers as well as the cost allocation for those investments. It is PUCs / PSCs, in other words, who ultimately determine who pays for the energy infrastructure needed to serve new data center load.
Large-Load Tariffs
PUCs / PSCs and utilities can, and some already do, use large-load tariffs to protect ratepayers from data center growth. Large-load tariffs are specialized, legally binding rate and service rules for large-load customers like data centers that “determine how much they [must] pay and the conditions for how they connect to the grid.” These types of tariffs can help mitigate risks by requiring new data centers to bear the incremental costs from capital expenditures needed to serve them.
PUCs / PSCs have the authority to establish large-load tariffs pursuant to various legal bases, while state legislatures can also specifically direct PUCs to do so. According to the Edison Electric Institute, as of May 2026, twenty-three states have approved at least one large load tariff, and another seven states have pending large load tariffs. For example, in 2025, Ohio’s PUC approved a proposed tariff filed by the utility AEP Ohio, targeting new, large data centers. The tariff created a new customer class for data centers with a monthly maximum energy demand of 25 megawatts (MW) or greater. The tariff requires data centers in that class to pay at least 85% of their contracted capacity or billing demand (even if they use less) and enter into initial contracts for a minimum of twelve years. After the tariff was approved, AEP’s large-load forecast decreased by half, suggesting that their introduction shapes cost recovery and market behavior by discouraging speculative or inflated interconnection requests.
In Virginia, the state with the largest concentration of data centers, the State Corporation Commission approved a new large-load tariff in 2025 for Dominion Energy to ensure long-term cost recovery by shifting risk from ratepayers to data centers and other large loads. Customers that demand 25 MW or more will automatically be subject to the new “GS-5” tariff beginning on January 1, 2027. These customers will be required to enter into contracts with minimum 14-year terms; pay 85% of contracted transmission demand and 60% of contracted generation demand; and put up collateral worth $1.5 million per MW of capacity. Moreover, the GS-5 tariff provides for “campus aggregation,” which allows certain geographically proximate facilities to be treated as a single load.
On the legislative side, in June 2025, Minnesota enacted HF 16 which, among other things, requires the state PUC to create a new “very large customer” rate class and allocate all attributable costs of service to that class to those customers. Other states have enacted, or are considering, legislation that directs PUCs to regulate large loads, including Oregon, Texas, and Virginia.
Moreover, when combined with another type of tariff—Clean Transition Tariffs (“CTTs”)—large load tariffs can have added benefits. Large-load tariffs typically do not include aspects that help bring new resources online or ensure that those resources produce renewable energy. CTTs can, however, achieve those results. A CTT is a type of tariff designed to allocate responsibility for procuring new energy resources to serve large loads like data centers. These tariffs can operate similarly to power-purchase agreements, as we’ve explained in a prior post on Google’s proposed CTT with NV Energy. Depending on the program, CTTs might involve three parties—the data center developer-owner, the utility, and an independent energy generator—unless the data center is being developed in a state that permits retail electricity purchases from an independent, non-utility supplier or if the new energy resources will be co-located with the data center.
Under CTT arrangements, several dynamics converge to help protect other ratepayers from financial risks associated with rapid data center development while also ensuring that renewable energy resources are developed to meet the additional electricity demand:
- First, the independent power producer agrees to develop new renewable energy resources to meet the expected electricity demand of the data center.
- Second, the utility enters into a long-term agreement to purchase power from the generator, often at a fixed and stable price.
- Third, the utility sells the electricity to the data center developer-owner pursuant to the tariff structure, with the data center operator assuming the financial responsibility for the long-term costs associated with the new generation resources.
Since CTTs result in the deployment of new renewable generation, they can help to reduce the emissions associated with data center development. And, as noted in our previous post, “[t]he CTT structure also resolves regulatory concerns over stranded costs by ensuring that the [utility’s] agreement to purchase the output of [the independent power producer] is backed by [the utility’s] sale of energy and capacity” to the data center operator.
Google is pursuing CTTs in other states, like Minnesota, where it is developing a data center in the City of Pine Island. In addition to being subject to Xcel Energy’s super large customer tariff, Google will pay for 1,900 MW of clean energyassets to be added to the grid (1,400 MW of wind, 200 MW of solar, and 300 MW of long-duration energy storage) over the fifteen-year contract term as part of the CTT with the utility Xcel Energy. Minnesota’s Public Utilities Commission has not yet approved the petition, in which Xcel observes that “[n]ot only will Google pay all costs attributable to adding its expected load, but the [a]greements include firm protections—like a Minimum Monthly Bill, credit support requirements, and an Exit Fee—that all serve to insulate other customers from potential impacts of adding this load to the system.” Local governments, then, should be aware that large-load tariffs can work in tandem with CTTs to reduce both the financial and climate risks associated with data center development. And local governments should consider getting involved in the PUC / PSC proceedings where these sorts of tariffs are being developed and implemented.
Local Government Involvement in PUC Proceedings
With respect to data center issues generally, there appears to be a significant unmet need at PUCs / PSCs for city-specific information and analysis. For example, Pennsylvania’s PUC opened a docket “concerning interconnection and tariffs for large loads customers” in March 2025, scheduled a hearing so that the PUC could “receive testimony and written comments from all interested parties” on several topics related to a large-load tariff, and directed PUC staff to develop a model tariff for large load customers available for use by utilities. Though many utilities, corporations, nonprofit organizations, and individuals provided testimony and comments, it appears that not a single city participated.
Some cities have entire offices dedicated to utility consumer advocacy, while others, whether due to resource constraints, differing priorities, or other circumstances, may not be able to devote that many employees toward it. But having some level of dedicated resources for PUC / PSC engagement can have many benefits. Among other things, it helps funnel community and resident input into a city’s engagement strategy, which in turn brings community-specific information into PUC / PSC proceedings. If resources are scarce for individual cities but there is consensus on data center regulation across jurisdictions, participating as coalitions may alleviate that burden while possibly elevating the impact of their advocacy.
For PUC / PSC proceedings considering large-load tariffs, cities aiming to protect their residents’ energy bills and mitigate climate impacts of data center development may argue that a well-designed, ratepayer-protecting large-load tariff should (1) fully allocate incremental system costs to new large-load customers and (2) ensure that load growth does not induce new fossil generation. These principles can be reinforced through additional safeguards that cities might advocate for, including minimum contract terms, exit fees, collateral requirements, take or pay requirements, and CTTs. Requiring minimum contract lengths between utilities and data center owners can create stability for the grid by ensuring customer buy-in over a longer period of time. Exit fees, meanwhile, are payments to a utility that penalize data center owners for prematurely withdrawing from a contract with a utility. To further protect ratepayers, collateral requirements like letters of credit or actual cash deposits, help ensure that large loads can cover certain costs, reducing the risk of default and stranded assets. Take or pay requirements, as reflected in AEP Ohio’s tariff, require the data center to pay for a certain amount of transmission or generation demand, regardless of whether the facility uses it. This measure is intended to avoid a “build[] it and they don’t come” situation. Finally, electricity demand has been increasing for the first time in more than a decade at the same time that the climate crisis is intensifying, meaning that cities should consider advocating for investments that can both meet additional demand and strengthen grid resilience. CTTs can be a useful tool for this.
Conclusion
With data center development surging, state utility regulators will increasingly determine how the costs and environmental impacts of that growth are distributed. Large-load tariffs are mechanisms that can help protect residential ratepayers from unwarranted cost-shifting while CTTs can reduce the likelihood that new electricity demand will be met with expanded fossil generation. Decisions about cost allocation and resource planning are being made right now, and without municipal involvement in state-level proceedings, local perspectives may be underrepresented. Cities, therefore, should continue tracking the evolving data center regulatory landscape, and look for opportunities to engage with utility regulators to ensure that these decisions reflect the interests of their communities.
Vincent M. Nolette is the Sabin Center's Equitable Cities Climate Law Fellow.
