The But-For Doctrine and Who Pays for Data Center Power
By Keith Reynolds | Publisher & Editor, ChargedUp!
The But-For Doctrine and the New Fight Over Who Pays for Data Center Power
Pennsylvania and Virginia are becoming the clearest laboratories in the country for a simple but consequential question: when a large new load drives the need for new infrastructure, why should everyone else pay?
Key takeaways
Pennsylvania’s framework centers on cost causation tied to the infrastructure a large load requires.
Virginia’s debate makes the rate-shift consequences more visible to households and data centers alike.
Owners should assume interconnection and upgrade costs are becoming front-end deal variables.
The next major fight over data-center growth is not only about whether communities want more campuses or how quickly utilities can interconnect them. It is also about a colder question, one that reaches straight into rates, leases, and site economics: when a project drives the need for new wires, substations, or other upgrades, who should pay?
Pennsylvania and Virginia are offering two of the clearest answers so far, and both point in the same direction. The era in which existing ratepayers quietly absorbed a meaningful share of the infrastructure burden tied to very large new loads is coming under direct challenge.
Pennsylvania’s approach is the cleaner expression of the principle. The state’s large-load tariff framework centers on a “but-for” idea. If the infrastructure would not be built but for a specific customer, then that customer should bear the cost. This may sound obvious, but in practice it is a major shift. It moves the debate away from broad claims about regional growth and system benefit and back toward a more legible concept of cost causation.
That matters because the commercial real estate market can usually adapt to a known cost faster than it can adapt to vague risk. A developer can model a defined infrastructure obligation. An owner can negotiate around it. A planner can evaluate it. What is much harder to price is a situation in which the infrastructure need is obvious, the beneficiaries are specific, but the costs are spread through a system in a way that leaves everyone else exposed to speculative overbuild.
Virginia’s version of the debate is more politically vivid because the trade-off has been framed in household terms. Reporting around Senate Bill 253 and the State Corporation Commission’s analysis has focused on the possibility that shifting more costs onto data centers and other large GS5 customers could reduce the typical residential bill by about $5.52 per month while increasing rates for those large users by roughly 15.8 percent. In the world’s largest data-center market, those are not background numbers. They are a statement about how far the state is willing to go in separating the economics of hyperscale growth from the bills paid by everyone else.
For commercial real estate, the consequences run in two directions at once. Owners of non-data-center property may benefit if states become more aggressive about insulating general ratepayers from the infrastructure bill created by large loads. Owners courting data-center users or other power-intensive tenants face a different reality. Energy infrastructure cost becomes more of a site-specific negotiation, and the tenant’s power profile becomes more central to the property’s economic story.
This is where the Pennsylvania and Virginia moves become more than local utility policy. They are early versions of a broader national argument that will increasingly shape real estate. The central grid cannot expand everywhere, at every speed, under yesterday’s socialized assumptions. Once that becomes obvious, regulators have three basic choices. They can socialize more of the cost. They can slow the projects. Or they can assign cost more directly to the customers causing it. Pennsylvania and Virginia are leaning toward the third option.
That changes how deals get done. Interconnection is no longer just an engineering workstream. It is becoming part of land valuation, lease structure, project feasibility, and lender confidence. In markets where large-load demand is crowding the system, owners and developers should expect more scrutiny of minimum demand commitments, contract duration, exit fees, and contribution requirements. In other words, power economics are moving upstream into the deal itself.
There is also a strategic implication for the broader built environment. If the cost of utility upgrades becomes more explicit and more project-specific, then on-site generation, storage, microgrids, and staged power strategies become more valuable. Not because they replace the grid outright, but because they can shrink the contested part of the equation. In a world of sharper cost allocation, reducing dependence on expensive external upgrades becomes a form of financial risk management.
That is why this story matters well beyond the data-center sector. It represents a shift in regulatory mood. States are becoming less willing to treat huge new loads as simply another chapter in economic development and more willing to ask what incumbent customers should be protected from. For owners and planners, that is a useful change in one respect and a more complicated one in another. It may reduce cost leakage onto the rest of the market, but it also makes power-intensive development a more exacting business.
The old utility-era language of diffuse system benefit worked best when load growth was gradual and infrastructure expansion could be digested over time. Today’s load growth is faster, more concentrated, and more politically visible. That makes the “but-for” doctrine more than a tariff concept. It is becoming a broader logic for how states think about growth under constraint.
For commercial real estate, the conclusion is straightforward. Owners should assume that upgrade cost allocation is becoming more explicit, more local, and more relevant to site value. Developers should expect more front-end work on power economics. Planners should expect public pressure to keep incumbent customers from subsidizing speculative demand. And everyone should get more comfortable talking about electricity not as a utility afterthought but as a central part of the project capital stack.
Why it matters for CRE
The infrastructure bill attached to a large-load project can now determine whether the deal pencils, whether nearby ratepayers push back, and whether the site remains politically viable.
