The new load needs an old-fashioned bill.

A substation expansion worksheet lies on a planner's desk. Its first column lists equipment: a new transformer, two breakers, a line upgrade. Its second column gives the cost of each. Its third column is headed "payer if the project cancels," and it is empty. The worksheet is invented for this dispatch. The empty column is the subject.

Last Thursday, June 18, the Federal Energy Regulatory Commission issued tailored show-cause orders to each of the six regional grid operators under its jurisdiction, and to their transmission owners. The orders concern how data centers, factories and other large energy users connect to the transmission grid. Each operator has 60 days to justify its current tariff or file changes, across five topics: study processes, preventing cost shifting with transparency into transmission costs, co-location with generation, new services for flexible large loads, and studying generators that serve nearby loads. Within 30 days, each must also report how it intends to ensure enough generation for existing and new large loads.¹

My argument is that a connection request is also a request to allocate risk. A connection request states how much capacity a customer wants, where and when. Cost responsibility states who pays for what must be built to serve it, including when the customer arrives late, changes its plans or never arrives. Faster connections are worth wanting. Before celebrating them, name the payer in each of those cases. Clear cost responsibility need not slow investment; it can make the commitment easier to finance.

The request and the obligation

A request for capacity is easy to make and, often, easy to change. The upgrade it triggers is neither. Picture the sequence. An applicant asks for 300 megawatts. The network operator studies the request and finds that a substation must be expanded. Money is committed, equipment is ordered, construction starts. Then the applicant's plans change: the site moves, the project shrinks, the financing falls through. The transformer is still there, and so is the debt that paid for it.

If the applicant's contribution was meant to come from years of service charges, those charges never arrive. Someone else then pays for the asset, usually the other customers whose rates recover the network's costs. That is the cost shift the orders ask grid operators to address.

Shared wires, shared benefits

The answer is not to send every dollar to the newcomer. A grid is shared. An upgrade built for one customer may also relieve congestion or improve reliability for others. Charging the applicant for all of it overcharges it for benefits others receive. Hiding all of it in general rates makes existing customers underwrite someone else's plan. Both are slogans. The real allocation depends on a study.

Figure 2 makes the arithmetic concrete with invented numbers. A $10 million upgrade is assumed to be $6 million specific to the applicant and $4 million of wider benefit. The cancellation term does not change what the upgrade costs. It changes who carries the $6 million if the load never comes.

What makes the choice reviewable

The orders' preliminary direction points the same way. In a concurrence, Commissioner Judy Chang summarized a two-part solution the Commission preliminarily identified: each operator would publish the large loads seeking to connect, the network upgrades identified to serve them and their costs, and each would adopt a standard cost recovery agreement so that customers serving large loads "bear the risk and are ultimately responsible" for the costs of serving them.³ She also stressed that the findings are preliminary and that the Commission needs actionable records in each region before deciding anything.³

From that, I would propose a simple test for any large connection, whatever the final tariffs say. Can someone outside the deal see six things: the request, the upgrade attributed to it, the assumptions behind the study, who benefits, what security was posted, and what happens on cancellation? If not, the allocation cannot be reviewed, only trusted.

Flexible demand changes the question without removing it. A customer willing to be interrupted at peak times may need less new capacity, and the orders ask about new transmission services for flexible large loads.¹ Whether that lowers the cost depends on the actual service terms and engineering study, not on a promise to be flexible.

Two boundaries matter. These are show-cause orders, a stage in proceedings, not a national tariff now in force or proof that anyone's bill has fallen. And FERC's fact sheet says the orders leave retail rates, and protection against cost shifting among retail customers, to the states.²

Objections, considered

Charging new customers too much deters growth. It does. That is why the question includes shared benefits, not merely whether the newcomer can afford to pay.

Existing customers gain reliability too. Sometimes. Then quantify the benefit and disclose it when allocating the cost, rather than assuming every upgrade is wholly private or wholly shared.

Speed matters. It does, and predictable obligations can help: a developer who knows its exposure can finance it. The choice is not between delay and a blank check.

A credible connection plan names the payer in the cancellation case.

Before signing a large connection

• Which costs are specific to this request? Ask for the upgrade attributed to it and the study assumptions behind that attribution.
• Who benefits from the shared portion? Ask that the benefit be named and, where possible, measured.
• What obligation survives a changed plan? Ask what security is posted and what is owed on delay or cancellation.

Back to the worksheet. The equipment and costs are unchanged. In the third column, against each item, someone now writes a name: the applicant for its specific share, secured; the network's customers for the part that serves them, stated. The upgrade costs the same. The risk now has an owner.