Summary
- Virginia regulators have directed Dominion Energy to develop a method for directly assigning dedicated transmission costs to data centres and other large loads.
- The intervention follows a dispute over nearly $1bn of transmission costs and how much should fall on ordinary customers.
- The model provides a useful comparator as European regulators wrestle with large-load connections, reinforcement costs, and queue reform.
Virginia’s State Corporation Commission has ordered Dominion Energy to develop a tariff approach that places more of the cost of dedicated transmission infrastructure directly on data centres and other large electricity users.
The decision follows arguments over how the utility should recover $998.5 million of transmission costs. Dominion’s initial allocation would have increased a benchmark residential bill by $2.90, prompting state officials and consumer advocates to push for a larger share of infrastructure built for large loads to be assigned to those customers.
The commission has directed Dominion to develop a policy for directly assigning transmission costs where facilities are built specifically to connect data centres and comparable users. Virginia Governor Abigail Spanberger’s administration had formally intervened in favour of shifting those costs away from households and small businesses.
The order does not make every grid investment triggered by data centre growth mechanically separable. Transmission networks are interconnected systems, and new substations or lines can provide wider capacity, resilience, or network benefits alongside serving a particular development.
Dedicated assets are easier to price than shared reinforcement
The clearest cases are infrastructure whose purpose can be traced to a specific customer: dedicated connection lines, substations, transformers, or upstream works that would not have been constructed without the large load. Assigning those costs directly reduces the risk that speculative or unusually large connection requests leave other customers paying for assets built around them.
The harder question is how to allocate shared network investment. A new transmission line may initially be driven by one cluster of data centres but later serve residential growth, industrial loads, renewable generation, or another substation. Regulators then have to determine how much of the cost is genuinely attributable to the data centre class.
Virginia provides an unusually large test case because data centre demand has moved from a specialist utility load into a material driver of system planning. The state contains one of the world’s largest concentrations of data centre capacity, meaning transmission requirements can be measured in gigawatts rather than individual megawatts.
The commission’s approach adds cost allocation to other measures intended to prevent large-load development from transferring risk to existing customers. Utilities are increasingly seeking longer contract terms, minimum-demand commitments, security arrangements, and reimbursement mechanisms before committing capital to projects whose ultimate construction remains uncertain.
Europe is facing the same question through different rules
European connection regimes differ substantially from Virginia’s regulated utility model, but the underlying infrastructure problem is familiar. Developers are requesting grid capacity faster than networks can build substations and transmission reinforcement, leaving regulators and system operators to decide which projects are credible and who should fund the works required to connect them.
Britain is reforming large-load connections while transmission and distribution operators work through queues containing data centres, industrial electrification, batteries, and other major demand. Spain, Ireland, Portugal, and several Nordic markets are dealing with their own versions of the same constraint.
Cost signals are one way of separating serious projects from speculative reservations. If a developer must commit significant capital to dedicated grid works, the incentive to hold connection capacity without a mature project is reduced. Poorly designed charges, however, can also discourage viable investment or create inconsistent treatment between locations.
DataCentral has tracked that effect in Spain, where connection economics and grid rights are already shaping which proposed campuses advance. The wider question is moving from whether data centres should pay for their electricity towards how much of the network built to supply them should sit on their side of the commercial boundary.
Virginia’s ruling is therefore useful less as a template than as evidence of the direction of travel. Regulators are increasingly unwilling to treat very large data centre loads as ordinary customers when their connection can trigger infrastructure expenditure measured in hundreds of millions of dollars.
The next step falls to Dominion, which must develop the tariff mechanism required by the commission. The detail will determine which facilities are treated as dedicated, how costs are calculated, and how shared assets are allocated.
Those questions are likely to recur wherever hyperscale campuses meet constrained networks. Power availability determines whether a project can be built, but the politics of who pays for that availability can determine whether regulators allow the necessary grid investment to proceed.
