How a Big Customer Reaches Your Bill

Mara opens her electricity bill and sees a higher total. That week, the local news is full of the proposed data center. It is easy to join the two facts with a line.

The line may be right. It may be partly right. It may be wrong.

A bill can change because a household used more electricity, because prices changed, because weather altered heating or cooling needs, or because several charges moved at once. Even when new large loads affect system costs, the effect on a particular household depends on how those costs are assigned. A nearby construction project is not enough evidence to explain a bill.

This is not an invitation to stop asking. It is an invitation to ask a question that can be answered.

Which cost changed? Why did it change? What share reached household customers? What would have happened without this project?

The last question is difficult. It is also essential. Without it, every increase can be blamed on the most visible new customer, and every new payment can be advertised as a benefit without examining the costs it brought.

We need to follow the money through the system rather than stop at the first plausible villain or the first reassuring receipt.

The customer pays and the question remains

An operator says it pays its own electricity bill. That matters. The next question is what the bill covers.

Buying energy, reserving capacity, connecting a site and strengthening a shared network can involve different charges and different decisions. A tariff is the regulated set of prices and service rules. A contract can add obligations specific to a customer or a project, within the applicable legal framework. The details determine which costs are recovered from whom.

Imagine that a customer pays for the energy it uses but has prompted equipment to be built for a much larger expected demand. If actual use remains low, the energy payment may not recover the cost of the unused commitment. Another charge, guarantee or allocation rule may address that gap. Whether it does is a question about the agreement, not the size of the customer’s monthly check.

Now reverse the situation. Imagine that suitable capacity already exists and a dependable new customer helps pay the system’s fixed costs without requiring comparably large new spending. Existing customers could benefit. Rejecting that possibility merely because the customer is large would be as careless as assuming it always happens.

The direction of the effect is conditional. Available capacity, new investment, timing, reliability and cost allocation all matter.

A good public explanation shows the conditions. It does not ask people to choose between the claims that growth always pays and growth always costs.

A small imaginary utility

Let us build a deliberately simplified example. It is an explanation of arithmetic, not a forecast for any actual utility.

Suppose a shared system has one hundred cost units to recover each year from its existing customers. A new large customer arrives. Serving it adds twenty units of annual cost. Its agreed contribution toward those costs is thirty units. If everything else remains unchanged and the accounting works as assumed, ninety units remain to be recovered from existing customers.

The new customer has improved their position by ten units in this example. Its size did not make that impossible.

Now change the agreement. The new customer prompts fifty units of additional cost but contributes thirty. One hundred and twenty units remain for existing customers. Growth has increased the amount they must cover.

Finally, change the timing. The system spends on the expectation of thirty units of new annual contribution, but the customer arrives late or uses far less than planned. The answer depends on whether an enforceable obligation supplies the missing contribution, whether the investment can serve others and how any remaining cost is treated.

These examples omit financing, different customer classes, energy purchases, taxes and many real-world details. Their purpose is narrower: revenue alone cannot tell us the effect on others. We need the added cost and the rules for any shortfall.

We also need to know whether the cost estimate was reasonable. A utility should not be able to build carelessly and make another party responsible merely by pointing to a growth story. Prudence belongs in the review. The public interest is poorly served when the only dispute is which customer group inherits an unnecessary expense.

A fair bargain requires discipline from the builder of the network as well as the user of it.

Who holds the risk when use changes?

Scroll sideways to see all columns. Keyboard: focus the table and use the left and right arrow keys.

Text equivalent: Who pays when demand changes?

Part of the figure	Meaning
Book	The Cloud Has a Power Bill
Title	Who pays when demand changes?
Starting point	Follow the obligation across time and between roles.
Question group	When to ask
Capacity ordered	What commitment backs the capacity request?
Service used	Which costs does the payment cover?
Use falls or ends	Who carries costs that remain?
Question group	Whose obligation?
Operator	Check payment duties, credit support and exit terms.
Utility	Identify prudent costs and how they will be recovered.
Regulator	Check the rules assigning approved costs.
Households	Check whether any residual cost is shifted to other customers.
Reminder	A payment in one account may leave another obligation open.
Status	Original conceptual map. No fixed outcome or complete tariff calculation.
How to read the diagram	The timing row and role grid are prompts to examine an actual agreement and regulatory decision together. They do not describe the law in every jurisdiction, prove a household bill increase, or promise that one tariff settles every cost.

Figure 2. Who pays when demand changes? An original conceptual map of questions to ask when capacity is ordered, used or left unused. The roles identify where to examine contracts, costs and decisions; they do not establish who must pay in a particular jurisdiction. This is not a tariff calculator or a validated planning model.

Follow the project through the three stages. At the order, somebody commits resources. During service, somebody pays for use and availability. If use falls or the customer leaves, the agreement decides what survives.

The people receiving the bill after that change may have had very little role in the original decision. That is the danger we should design against.

An operator can carry risk through payments for reserved service, a contribution to required works, an exit obligation or credit support. A utility can carry risk through the treatment of its investment and the standards it must satisfy. Regulators decide within their authority which costs can be recovered and from which groups. Public officials may separately make tax or infrastructure commitments.

These are not interchangeable pockets. Money collected by a town does not automatically pay a utility’s shortfall. A guarantee held by a utility does not automatically compensate a neighbor. Good policy needs to connect each protection to the exposure it is meant to cover.

We should also ask what happens if the party promising to pay cannot pay. A strong-looking clause against an empty company can provide little comfort. A guarantee needs a credible backer, a clear trigger and an amount related to the risk. It needs review as circumstances change.

The point is practical. A promise has to be collectible when it becomes inconvenient.

A real tariff with a narrow lesson

Ohio Power’s revised data-center tariff took effect on August 28, 2026. For qualifying New Load, it uses tiered minimum billing-demand rules, alongside measured demand and a prior-demand test. The minimum initial contract term includes any ramp period, capped at four years, plus eight years. Collateral and exit provisions also matter. This is not a simple obligation to buy 85 percent of expected energy, nor a universal twelve-year term. Existing Load has different rules.¹

That paragraph is a bounded description of one tariff. It is not a complete guide for a customer, a verdict on the full regulatory case or proof that every possible cost is covered. Its useful lesson is that a public rule can distinguish electricity actually taken from a commitment made available.

The wider design question is how to make the payment obligation fit the risk without becoming needlessly punitive. A minimum charge can help protect others from underuse. It can also be poorly designed if it ignores reusable assets, double-recovers costs or locks in an inefficient arrangement. The answer requires scrutiny of the whole agreement.

A reader should ask for the current filed rule, its effective date and its eligibility conditions. Summaries can become stale. A single memorable percentage can survive long after the relevant formula has changed.

The headline is not the contract.

Approved is a different date from active

Virginia’s State Corporation Commission approved Dominion Energy Virginia’s GS-5 large-user class for January 1, 2027. Its summaries describe minimum demand obligations and, for specified new contracts, fourteen-year commitments with credit-related protections and exceptions. At this book’s October 2026 cutoff, the new class was approved but not yet effective. Those summaries establish direction and timing, not every legal detail of a customer’s bill.²

The distinction matters to someone reading about protections today. A future rule can influence planning before it begins, but it should not be described as already governing every current customer. An exemption may matter as much as the main requirement. The utility territory matters too; a state is not necessarily one tariff.

Public language often rushes past these details because they seem less exciting than a promise that households have been protected. Yet the protection exists through those details. A date, a threshold, a covered service and a remedy are where the claim becomes real.

We should welcome better rules while asking how they will be applied and checked. Applause at approval is not a substitute for reviewing performance after implementation.

That is true of almost every bargain in this book.

The bill that has several causes

Suppose Mara asks her utility whether the new development explains her higher bill. A useful answer would break the change into understandable parts rather than send her a slogan.

Her own usage changed by this amount. The price of a particular component changed. A network charge reflects specified investments. A forecast or allocation affected another component. Some changes would have happened regardless of the new customer; others may be tied to its arrival.

This kind of explanation will not always be simple or perfectly precise. Some costs are shared. Some effects depend on a comparison with an alternative future we cannot directly observe. Analysts can disagree about assumptions. Those are reasons to show the method and uncertainty, not reasons to abandon the inquiry.

The same standard should apply to a claim that the development will lower household bills. Compared with what plan? Over which period? Which households? Does the estimate include the necessary investment and the chance that demand arrives late? Is a temporary benefit being presented as permanent?

Residents should not have to supply the missing caution on both sides of the argument.

An independent review can test the claimed effect. It should be funded and organized so that the reviewer can do the work without becoming dependent on producing a favorable answer. The applicant may appropriately pay the review cost under a lawful process, but payment should not buy control of the findings.

Trust needs a structure, especially when the numbers matter.

A market price is another route

So far, our examples have focused on the recovery of infrastructure costs. There is another possible channel: the price of electricity itself. A customer can meet its direct obligations and still arrive in a market where an increase in demand changes which supply is needed and what it costs.

Consider an invented hourly market with a limited amount of lower-cost supply and a more expensive next option. If extra demand requires that next option, the purchase cost can change. If additional supply arrives alongside the demand, or the new customer can avoid the tight period, the result may be different. Actual market rules determine how those changes reach particular buyers.

This example is deliberately conditional. It does not establish the cause of a real price movement or the effect on Mara’s retail bill. It explains why examining a connection charge alone may leave an important question unanswered.

A full assessment should distinguish the customer’s direct infrastructure obligations from wider changes in supply and demand, then trace the route into retail charges. It should also compare a realistic alternative: what supply, demand and investment would exist without the project? The comparison must not assume that everything else would stand still forever.

This is harder than pointing to one charge. It is also why independent analysis matters. A protective tariff may reduce one exposure while another depends on the timing of new supply, the customer’s operating pattern or regional market conditions.

We should ask whether the whole arrangement improves the system, not merely whether one carefully selected account balances. That question can recognize a project’s benefits without allowing an incomplete account to stand in for a complete one.

Fairness between old and new customers

There is another difficulty. Existing customers may have helped pay for infrastructure that a new customer will use. Yet a newcomer can also help fund improvements that others will enjoy. A fair allocation should recognize both.

It is tempting to announce that the new customer must pay for everything. But everything needs a boundary. If an upgrade replaces worn equipment, expands service for several users and accommodates one major arrival, the costs may have several causes. Assigning them all to a single party can be arbitrary. Assigning them broadly without examining the causes can be equally unfair.

The principle I favor is explainable responsibility: identify what the project requires, what would be built anyway, what benefits others and what risk the requesting party creates. Then state how each category is treated.

A person may still disagree with the rule. At least the disagreement will concern an actual choice.

This approach also protects ordinary new needs. A household adding electric equipment, a factory modernizing production and a new computing facility should all face rules that can be defended in relation to their effects. Equal treatment does not always mean an identical tariff. Very different service demands may justify different terms. The distinction should rest on the service and risk, not the political charm of the customer.

We should be able to explain a rule without relying on the company being either a hero or a villain.

A protection can be too small or badly aimed

Even a well-designed electricity agreement leaves other accounts open. It does not establish that the water plan is sustainable, that a tax concession is worthwhile or that residents can sleep comfortably. It also does not guarantee that the useful services produced will be affordable to the people supporting the infrastructure.

That is why a single protective tariff should be treated as one part of the bargain. It can solve a real problem without solving every problem.

The opposite mistake is to dismiss it because it is incomplete. Public protection often advances through specific tools. A rule that reduces an identifiable exposure is valuable even while other work remains. We should ask what it covers, check whether it works and then address the next gap.

For a household already under financial strain, the timing of protection matters. A later refund may not undo the hardship caused by an avoidable earlier charge. A credit guarantee that is checked only after the customer fails may arrive too late. A public explanation after a decision has become irreversible offers less power than the same explanation before commitment.

The order of action can be as important as the amount.

We need to move responsibility toward the moment when risk is created.

A bill is a public document in miniature

An electricity bill looks private. It arrives at one address and asks one household to pay. Behind it sit shared decisions about equipment, reliability, markets, public authority and the treatment of risk.

Most people should not have to study all of that to keep the lights on. A functioning public system does the complex work on their behalf. But doing it on their behalf creates an obligation to explain it when a major change is proposed.

The standard should be higher than telling people to trust the experts, and more useful than telling them every expert is serving someone else. Expertise is necessary. So are records, challenge and accountable judgment.

Mara may finish the inquiry and learn that the new customer is not the main cause of her bill increase. That is a useful result. She may learn that a particular cost has been shifted onto households and should be challenged. That is useful too. What she deserves is a process capable of finding either answer.

A fair power system does not promise that bills can never rise. It promises that reasons can be examined, obligations can be enforced and powerful customers cannot quietly turn their uncertain plans into everyone else’s permanent expense.

The cloud has a power bill.

The public question is which parts belong on whose bill, and whether the agreement will still make sense when the excitement has passed.