Your utility files a rate case. A large campus is announced two counties over, with a jobs figure and a tax figure attached. Some months later your bill is higher, and nobody in the process can tell you which part of the increase is the campus and which is everything else.
There is a version of this argument that goes nowhere, and most of us have watched it. One side says the buildings raise prices. The other side produces a study showing that large customers spread fixed costs over more kilowatt-hours and so lower the average rate. Both cite numbers. Neither side is obviously lying, and the room learns nothing.
Rachel Carson spent four years assembling the documentation for a fight of exactly this shape, and the move she made is the one worth borrowing.
The question underneath the safety claim
Her instrument was not a better estimate of harm. It was the observation that behind every reassurance there is a decision structure, and that the structure is usually easier to examine than the science. Who was exposed. Who was asked. Who is doing the reassuring, and what do they gain if you accept it. Who paid for the study being waved as settlement.
Applied here, that reframes the whole dispute. The contested question is not whether large loads raise or lower average rates, because the honest answer is that it depends, and the published work genuinely points both ways: one 2026 study of retail rates found average prices modestly lower where large loads had arrived, while the largest grid operator in the country has attributed part of a steep capacity price rise to data centre demand. An average across a country is not a finding about your service territory, and neither figure tells a household in one county what it will be paying in five years.
What can be established is the structure. Three questions do most of the work.
First: which rate class carries the new load, and who is responsible if it does not arrive as forecast? Transmission gets built against a projection. If the projection is wrong, the cost does not disappear, it is allocated — and the default allocation is the residential class, because it is the class least able to negotiate.
Second: how long is the commitment, against the life of the asset? A line built to last forty years, supported by a commitment of ten or fifteen, is a bet that the demand will still be there when the contract lapses. Someone is holding that risk. It is worth knowing who, in writing.
Third: who commissioned the study? Not so you can dismiss it. So you can read it knowing what question it was paid to answer.
Watch out for
That third question is where this lens goes wrong, and it goes wrong often.
Funding tells you about incentives. It does not tell you the data are false, and treating sponsorship as a refutation is motivated dismissal wearing sceptical clothes. Carson herself was attacked in exactly those terms by an industry that preferred to discuss her motives, and what beat that attack was not counter-accusation. It was four years of documentation and a citation for nearly every claim. A lens that ends at who paid has skipped the part that actually worked.
There is a second limit, and it belongs in the open. Her subject was involuntary exposure to substances people could not see, refuse or leave. A household facing a higher bill is in a much weaker version of that position: it is a real cost, unevenly distributed and imposed through a process most people cannot follow, but it can be voted on, litigated and reversed, and a songbird could do none of those. Borrowing her decision-structure question is defensible. Borrowing the moral weight of her subject to describe an electricity bill is not, and it will lose you the argument in front of anyone who knows the difference.
Answer this next
Find your utility's most recent filing and answer one question: which rate class is the new large load assigned to, and what happens to that allocation if the load never materialises?
You can get that from the docket without a lawyer. It is also the number that decides who is actually carrying this.
