AI data centers After Virginia’s Cost Decision

AI data centers in Virginia now operate under a more explicit cost-allocation model than they did before the State Corporation Commission’s 2026 infrastructure cost decisions. The central regulatory question is no longer only whether large-load projects can connect to the grid. It is also who pays when those projects require substations, transmission lines, distribution upgrades, and capacity commitments that may extend well beyond ordinary commercial demand.

For developers, utilities, local governments, and smaller ratepayers, the change is practical rather than abstract. Virginia has not banned large-load growth, and the available record does not show a single statewide rule that settles every future project. Instead, the state has moved toward tariffs, contract terms, collateral requirements, direct cost assignment, and a temporary electricity consumption tax. Those tools shift more financial risk toward the facilities that cause the load, while leaving some implementation details to future proceedings and utility-specific filings.

Why AI Data Centers Face A New Cost Test

What AI Data Centers Changed In Grid Planning

Virginia’s case is notable because the scale of data center demand has become large enough to affect transmission and distribution planning. Research notes for September 22, 2026, identify 371 operating data centers across Virginia as of August 2026 and 438 more planned. That level of activity helps explain why regulators have treated large-load service as a distinct cost-recovery problem rather than a routine extension of ordinary commercial service.

The SCC’s public materials state that it created a GS-5 rate class for “large load customers,” including hyperscale data centers. Under that class, affected customers must pay 85% of transmission and distribution costs incurred to serve them each month, even if they do not use all of their contracted capacity. The rule applies to new and existing large-load customers, except those that began service before January 1, 2016, according to the Virginia SCC facts.

That structure addresses a specific regulatory risk: a utility may build infrastructure for a large contracted load, but the customer may later consume less electricity than expected, delay a facility, or abandon capacity. Without minimum charges, other customers can be exposed to stranded or underused infrastructure costs. The 85% requirement does not eliminate every risk, but it creates a clearer cost floor tied to the service capacity the customer requested.

The Contract Term Is A Cost-Recovery Tool

Starting January 1, 2027, new large-load customers in Virginia will be required under SCC orders to commit to at least 14-year contracts for electric service for cost-recovery purposes. That long term matters because grid assets are planned and financed over multi-year periods. A short or uncertain service commitment can leave utilities and regulators with a mismatch between the expected life of infrastructure and the customer behavior that justified building it.

The SCC has also approved collateral requirements for large-load customers lacking sufficient credit, with collateral up to 60% of minimum charges over the contract term. In regulatory terms, this is a credit-risk control. It does not determine whether a specific project is socially beneficial, and it does not answer land-use objections. It does, however, gives utilities a mechanism to reduce exposure if a high-load customer cannot support the financial obligations associated with its requested capacity.

How Virginia Shifted Infrastructure Costs

Direct Assignment Narrows The Subsidy Question

On July 31, 2026, the SCC ordered Dominion Energy to directly assign the cost of certain transmission infrastructure to the large-load facilities that made those assets necessary. Research notes identify substations and lines as examples of infrastructure covered by this approach. The policy direction is clear: when a specific large-load project drives a specific upgrade, regulators are less willing to spread that cost broadly across all ratepayers.

This is a significant adjustment because prior cost recovery could place some data-center-related transmission expenses into broader rates. Research notes indicate that since 2021 Virginia ratepayers have covered approximately US$2.8 billion in transmission line costs related to data centers. The SCC’s direct-assignment approach responds to that concern by linking project-caused infrastructure more closely to the project’s bill responsibility.

There are limits to what can be inferred from that shift. Direct assignment is clearest where a facility-specific upgrade can be identified. Broader system reinforcements may be harder to attribute cleanly because transmission networks are shared assets. Future proceedings may still need to decide whether a cost is project-specific, system-wide, or some combination of both. That uncertainty is why the decision is better understood as a stronger allocation framework, not a complete settlement of every future grid-cost dispute.

The Temporary Power Tax Adds A Separate Charge

Virginia also enacted a data center electricity consumption tax that took effect on July 1, 2026, and runs until July 1, 2028. The enacted budget language sets the charge at US$0.011 per kilowatt-hour for all electricity consumed at each data center per month, as shown in Item 3-5.24.

This tax is distinct from the SCC’s tariff and direct-assignment tools. A tariff can recover utility costs associated with service. Direct assignment can attach infrastructure costs to the customer that caused them. The consumption tax applies to electricity use at the data center during the specified period. It therefore operates as a usage-based public charge rather than a project-by-project infrastructure reimbursement mechanism.

For operators, the combined effect is more material than any one provision viewed alone. A facility may face minimum monthly transmission and distribution payments, long-term contract obligations, collateral requirements, directly assigned upgrade costs, and the temporary consumption tax. The exact financial impact will depend on contracted capacity, actual usage, credit profile, location, and the infrastructure needed to connect the project.

Permitting, Local Review, And Utility Planning

Local Procedure Still Matters

Virginia’s regulatory shift is not limited to utility ratemaking. Local approvals and court review remain part of the project risk profile. Research notes state that the Virginia Court of Appeals voided the Prince William County Digital Gateway Project in March 2026 over a procedural technicality, and that the matter was being appealed to the Virginia Supreme Court. That example shows that a project can face risk even after policy discussions focus on electricity costs.

Local opposition can influence the timing, conditions, or legal durability of project approvals. For counties, the challenge is to evaluate land use, water, noise, tax base, transmission corridors, and community impacts without relying on incomplete assumptions about the utility bill. For developers, the practical lesson is that energy-cost allocation and land-use procedure must be managed together. A project that solves its grid-cost exposure may still encounter procedural or zoning problems.

Project-Specific Upgrades Are Becoming More Visible

Research notes identify a planned Google data center campus in Botetourt County that would require transmission-line and substation upgrades estimated at US$264 million. The utility asserted that those costs would be borne by Google rather than passed to other customers. This example aligns with the broader state direction: regulators and utilities are making facility-caused grid costs more visible and, where supportable, assigning them to the customer that needs the upgrade.

That visibility is useful, but it does not remove uncertainty. Cost estimates can change as engineering, permitting, procurement, and interconnection studies advance. A public statement that a customer will bear costs should still be examined against tariff language, contract terms, and final regulatory approvals. For related technology coverage across the same publishing network, visit the platform provided by Abacus News.

Who Is Affected By Virginia’s Rules

Residential meters and commercial power cabinets lined along a utility wall

Ratepayers Gain A Stronger Protection Theory

Residential and smaller commercial customers are central to the policy debate because they can be affected when large-load infrastructure costs are socialized. The SCC’s approved tariffs, minimum contract terms, collateral rules, and alternative cost-allocation methods are designed to reduce cost-shifting. They do not promise that ordinary customers will never pay for shared grid investments. They do create more tools for regulators to ask whether a cost was caused by a large-load customer and whether that customer should carry more of it.

Utilities are also affected. They must serve load reliably, plan infrastructure before demand materializes, and avoid under-recovery of prudent costs. A stricter large-load framework may reduce stranded-cost risk, but it can also add contract negotiation, credit review, billing, and regulatory complexity. The evidence available in the research notes supports that direction, but it does not establish how each utility will implement every requirement in every service territory.

Developers Face Earlier Financial Scrutiny

Data center developers now have stronger incentives to validate power capacity, usage forecasts, financing, and phasing before seeking service. A request for large contracted capacity can carry obligations even if the facility ramps slowly. That changes the commercial calculus for speculative campuses and phased builds because the cost of reserving capacity may become harder to defer or transfer to the general customer base.

This connects with broader questions about AI energy requirements, especially where compute demand, grid interconnection, and electricity procurement must be evaluated together. AI data centers may still be developed in Virginia, but the rules now make capacity reservation and project-caused infrastructure a more explicit part of the business case.

Virginia’s AI Data Centers Regulatory Case

Virginia’s 2026 decisions show a state moving from general concern about data-center load toward more defined cost-allocation mechanisms. The most concrete pieces are the GS-5 large-load rate class, the 85% monthly transmission and distribution cost requirement, the 14-year contract term for new large-load customers beginning January 1, 2027, collateral requirements for customers with weaker credit, direct assignment of certain project-caused infrastructure, and the temporary US$0.011 per kilowatt-hour data center electricity consumption tax.

The cautious reading is that Virginia has not resolved every regulatory question. Shared transmission costs, local approval disputes, changing project scopes, and future market rules can still affect outcomes. Yet the direction is measurable: AI data centers are being asked to carry more of the costs and risks associated with the infrastructure they require. For businesses evaluating Virginia projects, the key due-diligence task is to model power obligations as a long-term regulatory exposure, not just an operating expense line.