Which siting, tax and utility arrangements determine whether a data center produces a net benefit for existing residents?

A’s strongest argument is that audited resident welfare, rather than a financing formula, should determine approval. I accept that objective and the proposed neighborhood protections. The weakest point is the evidentiary bridge from safeguards to permission for subsidies: neither substantial receipts nor a protective tariff demonstrates that concessions outperform ordinary taxation and developer-funded infrastructure. An audit can verify promised performance without establishing whether the subsidy was necessary.

First, correct allocation of existing electricity costs does not establish protection against expansion costs. Virginia’s December 2024 JLARC review found then-current cost allocation appropriate, yet projected that data-center growth could increase a typical Dominion residential customer’s monthly generation-and-transmission costs by $14–$37, in constant dollars, by 2040. This is a modeled forecast, not an observed bill increase. Nevertheless, it directly challenges treating current cost recovery as sufficient protection. (jlarc.virginia.gov)

Second, Ohio supports stronger developer obligations, not a demonstrated case for resident subsidies. PUCO’s July 9, 2025 order explicitly addressed cost-shifting risks from underused infrastructure. (content.govdelivery.com) AEP’s implementation also reveals why contract details matter: the 100% pre-energization buildout-reimbursement obligation expires upon project completion; qualifying financially strong customers need not post collateral; otherwise, security equals 50% of minimum contractual charges, not necessarily unrecovered infrastructure costs. Those distinctions do not prove the tariff inadequate, but they mean its existence cannot establish complete protection. (aepohio.com) My proposed test is whether secured payments cover attributable unrecovered investment under cancellation, default, and low-utilization scenarios—not whether a contract merely exists. Developer funding need not mean every dollar upfront; it means residents are not the residual guarantors.

Third, fiscal scale is not subsidy effectiveness. JLARC reported $928 million in FY2023 sales-tax savings while estimating approximately 50 full-time workers at a typical facility, about half contractors; most statewide economic benefits came from construction. These findings do not negate local tax benefits, but they weaken permanent-employment arguments for concessions. (jlarc.virginia.gov) Research should compare subsidized and comparable unsubsidized developments, testing prior trends and location advantages, then reconcile promised versus realized taxes and permanent jobs after incentives, public services, and utility exposure.

My normative baseline is ordinary taxation, developer-funded incremental infrastructure, and security matched to remaining exposure. A tax abatement reduces otherwise-owed taxes; cost allocation decides who pays; the utility rate base is investment eligible for regulated returns; a stranded asset retains unrecovered costs after expected demand disappears. Independently needed resident infrastructure should be separated from project-caused costs.

Approval should require buffered industrial siting, enforceable noise, water and generator-emissions limits, independent monitoring, and funded remediation through a binding community benefits agreement. JLARC found conventional noise rules could miss persistent low-frequency disturbance. (jlarc.virginia.gov) Compare nearby sales and rents with matched controls accounting for preexisting trends, zoning changes, and distance. Report incumbent homeowners’ and renters’ outcomes separately: projected regional growth is not my proposed substitute for protecting those paying higher bills or bearing neighborhood burdens.

A’s rebuttal concedes two decisive points: no particular abatement has passed its counterfactual test, and Ohio’s published protections do not establish lifetime coverage of stranded investment. A now proposes stronger security and accepts ordinary taxation when incentives are unnecessary. These concessions narrow our disagreement: the remaining question is whether uncertain additional benefits justify transferring project-caused costs to residents before those benefits are demonstrated. My normative answer is no.

The specific weakness in A’s new electricity evidence is a mismatch between what it measures and what sharing requires. Berkeley Lab and Brattle associate 10% statewide load growth during 2019–2024 with approximately 0.6 cents/kWh lower average retail prices. Their explanation includes spreading existing fixed costs across greater consumption; they explicitly warn that constrained, costly new infrastructure can reverse this relationship. This supports the possibility of load-related savings, not the necessity of resident financing for a particular expansion. The study does not compare developer-funded and resident-funded versions of the same upgrade. (eta-publications.lbl.gov)

My inference is that even genuinely beneficial load growth leaves A’s financing argument unproven. Residents could receive fixed-cost dilution while the developer pays incremental infrastructure costs. Regulators should compare the project with a least-cost, no-project investment plan, charging the developer for additional capacity and accelerated expenditure. Infrastructure independently warranted for residents remains their responsibility; incidental usefulness does not establish that project-triggered spending belongs on their bills.

The downside remains location-specific. JLARC’s December 2024 analysis projected an additional $14–$37 monthly in typical Dominion residential generation-and-transmission costs by 2040, in constant dollars—not an observed increase. (jlarc.virginia.gov) Ohio’s initial contract runs at most twelve years, and its conditional security requirement equals 50% of minimum contractual charges. Neither number alone establishes coverage of unrecovered assets. (aepohio.com) A’s stronger proposed guarantee is therefore welcome, but approval should require demonstrated coverage before construction, not merely a commitment to reassess later.

Fiscal evidence also warrants restraint: JLARC found reduced equipment-tax rates limited potential collections and most statewide economic benefits came from construction. (jlarc.virginia.gov) Compare subsidized and comparable unsubsidized projects using realized taxes, permanent jobs, public-service costs, and utility exposure; audited promise-keeping cannot identify subsidy effectiveness.

Finally, JLARC found no established property-value decline, while documenting noise complaints and zoning decisions permitting incompatible proximity. That neither proves housing benefits nor warrants assuming neighborhood decline. (jlarc.virginia.gov) Housing research should control for prior trends, zoning changes, and distance, separating incumbent owners from renters. Approval should require protective siting, enforceable noise, water and emissions limits, independent monitoring, and funded remedies through permits and a binding community benefits agreement. A substantially accepts these protections; none requires tax concessions. Until comparative evidence establishes otherwise, ordinary taxation and secured developer funding remain the justified baseline.

A’s strongest position is that financing rules should serve existing residents’ welfare: a conditional concession could induce development that otherwise would not occur and produce benefits exceeding every public cost. With independently demonstrated additional investment, fully secured utility exposure, and enforceable neighborhood protections, rejecting that arrangement solely because it includes an abatement could sacrifice resident benefits.

My position remains stronger on the evidence presented because A establishes a possible exception, not a verified qualifying arrangement. I acknowledge that decisive comparative evidence could overturn my no-subsidy position; it is not an assertion that beneficial exceptions are logically impossible. Until then, ordinary taxation and secured developer funding assign project-caused costs to the investor rather than asking residents to accept uncertain compensation.

The electricity evidence does not establish financing necessity. Berkeley Lab and Brattle’s October 2025 analysis associated 10% statewide load growth during 2019–2024 with approximately 0.6 cents/kWh lower average retail prices. That comparison measures load growth, not resident-funded versus developer-funded infrastructure; my inference is that it cannot identify the superior financing arrangement. (eta-publications.lbl.gov) The authors also warn that constrained, costly expansion can reverse historical savings. (eta-publications.lbl.gov) Regulators should therefore charge developers for incremental capacity and accelerated expenditure relative to the least-cost no-project plan; independently necessary resident investments remain residents’ responsibility.

Location-specific downside is material. JLARC projected an additional $14–$37 monthly in typical Dominion residential generation-and-transmission costs by 2040, in constant dollars—not an observed increase. (jlarc.virginia.gov) AEP Ohio’s published conditional security equals 50% of minimum contractual charges, while its pre-energization buildout-reimbursement obligation expires at completion. These provisions alone do not demonstrate coverage of all stranded investment. (aepohio.com) My approval standard requires security matched to attributable unrecovered costs before construction and maintained as exposure changes.

Fiscal promises likewise require counterfactual testing. JLARC reported $928 million in FY2023 sales-tax savings and found most statewide economic benefits arose during construction. Neither finding establishes that concessions induced sufficient additional resident benefits. (jlarc.virginia.gov) Protective siting, enforceable noise, water and generator-emissions limits, independent monitoring, and funded remedies through permits and a binding community benefits agreement should remain prerequisites, not benefits purchased with subsidies.

To accept A’s exception, I would need replicated quasi-experimental studies exploiting incentive-eligibility changes, with credible comparison developments and verified preexisting trends. Audited records must show additional investment and higher discounted resident benefits than ordinary-taxation projects and feasible alternative uses, reconciling promised and realized taxes, permanent jobs, public services, subsidies and utility costs without double counting. Utility records must demonstrate household savings and developer-backed recovery through underuse or closure. Distance-based sales and rental analyses must control for zoning changes and prior trends, separately assessing incumbent owners and renters; site monitoring must verify environmental compliance and funded correction of violations. Positive results across those tests—not regional growth alone—would persuade me.