Here is a contradiction worth sitting with. Everyone in Washington reportedly just spent eight months and more than 3,500 pages of comments arguing about how to connect data centers to the grid without sticking ordinary ratepayers with the bill. And they aimed almost all of it at the one part of the system that already gets the most scrutiny.
Start with the number that makes the abstraction concrete. Meta's data center in El Paso is forecast to draw 16 megawatts in 2026. By 2027, 286 megawatts. By 2028, 826. By 2029, 1,005 megawatts — over a gigawatt — and then it holds roughly flat through the 2040s.
Sit with the shape of that curve. This is not organic growth, not new neighborhoods filling in over a decade. It is, per the forecast, a near-vertical ramp from a rounding error to a gigawatt in three years, at a single location, that then plateaus for twenty.
And the punchline is scale relative to the host. In the 2040s, per the forecast, that one site will be more than 30% of El Paso Electric's entire projected load. One company, reshaping one utility's grid.
In the 2040s, per the forecast, that one site will be more than 30% of El Paso Electric's entire projected load. One company, reshaping one utility's grid.
There is a useful yardstick for what a gigawatt at one spot actually means. As one industry analysis put it, "connecting a 1-GW data center is equivalent to connecting 1,000 Costco stores to the grid at the same spot."
Utilities know how to plan for a Costco. They do not have a century of practice planning for a thousand of them landing simultaneously on one substation. That is reportedly the problem FERC set out to solve on June 18.
That day, under Section 206 of the Federal Power Act, the commission issued six "show cause" orders — one each to PJM, MISO, SPP, CAISO, ISO New England, and NYISO — telling them their large-load interconnection rules appear inadequate and giving them 60 days to justify the tariffs or fix them.
Chairman Laura Swett reportedly framed it as "a future of fair cost allocation, unprecedented transparency for the American ratepayer." FERC's David LaCerte was blunter: if the RTOs don't fix it, the agency will. "I say this not as a threat, but as a statement of duty."
The five issues FERC named are worth reading closely, because they reveal the boundary of the whole exercise. Transmission study processes. Cost shifts and transparency into transmission costs. Co-location and behind-the-meter generation. Transmission service for flexible loads. Studies for generators serving nearby large loads.
Every one of those lives at the transmission layer. This is a transmission order, aimed at transmission-organized markets, about how the cost of high-voltage lines gets allocated when a hyperscaler shows up.
The economics there are real. As the Brattle Group's Johannes Pfeifenberger explains it, a large load with existing headroom can just pay the average transmission rate. But when serving it means billions in new lines, "you have a situation where the incremental cost is higher than the average cost, or embedded cost, and that drives up rates for everyone."
PJM and ERCOT are reportedly both forecasting large-load growth that drives billions in new transmission, and a new line can run twice as expensive per megawatt as the depreciated 40-year-old system whose average sets today's tariff. That gap is exactly what lands on existing customers. FERC is right to chase it.
But notice what FERC's map leaves off. About a third of Americans live outside the RTOs, in vertically integrated territory the June 18 orders don't touch at all.
Analysts flagged this immediately. R Street's Devin Hartman: "I wish FERC had included non-RTO areas, which typically suffer from the worst transmission and interconnection practices." Grid Strategies' Gretchen Kershaw was sharper — in those bundled-retail regions, "there is no part of the country where there is less transparency."
El Paso Electric, incidentally, is reportedly one of those vertically integrated utilities. The single most vivid load shock in the evidence reportedly sits in exactly the regulatory blind spot FERC declined to enter.
And even that understates the gap, because the transmission layer FERC just spent eight months on is not where most of the money — or most of your rising bill — actually goes.
Here is the fact the June 18 coverage almost entirely omits. Utilities now spend more on local distribution systems — the transformers, poles, and wires in your neighborhood — than on almost anything else, "roughly a third of their capital budgets, and the biggest driver of rising electricity bills."
Distribution is the fastest-growing category. And unlike generation and transmission, "that spending faces almost no scrutiny before it's approved."
Understand how sharp that asymmetry is. For generation and transmission, Minnesota regulators Sydnie Lieb and Pete Wyckoff describe a "very robust" process: multi-year regional modeling, load and generation scenarios, sensitivities — if load grows 10% here, what buildout does that require? — plus consumer advocates, environmental groups, and states all in the room.
The whole point of that machinery is optimization at least cost. Utilities must prove their choices are the cheapest way to meet the need before regulators sign off.
Distribution gets none of it. It has traditionally been treated as pure delivery: build enough poles and transformers, replace the old ones, done. Utilities bucket it as "mandatory," and the only question regulators ask is whether they paid a reasonable price for the equipment — not whether the equipment was the right choice.
That was defensible when distribution was dumb wire. It isn't anymore. Rooftop solar, batteries, EV chargers, home energy management, virtual power plants — the distribution grid is now a resource with real design choices, not a passive conduit.
Which is Lieb and Wyckoff's argument, and their warning. Left unscrutinized, utilities "tend to prefer concrete and steel — more spending" — when newer technologies could get the same result cheaper. They want distribution subjected to the same modeling and least-cost optimization everything else survives.
Now layer the load boom onto that. A ramp like El Paso's does not just stress transmission; it drives substation upgrades, feeder rebuilds, and local reinforcement — distribution spending, the category with no scenario modeling and no optimization test standing between the utility's capital plan and your bill.
That is the door no one is guarding. FERC is watching the transmission gate closely and publicly. The RTOs will spend the next 60 days and beyond arguing over cost allocation for high-voltage lines, and the Ratepayer Protection Pledge signatories will reportedly keep insisting they want to pay their fair share.
Meanwhile the biggest single slice of the capital budget, the one growing fastest, in the layer closest to your meter, keeps getting approved on the old "mandatory" logic — with regulators checking the invoice price and nothing else.
The demand surge everyone finally agrees is real will show up on bills through whichever cost mechanism faces the least resistance. Right now that mechanism is distribution.
FERC's June 18 orders are a genuine step toward transparency where there was almost none. But they light up the part of the system that was already better lit. The dark part is where the AI-load boom will do its quietest, most expensive work — unless more regulators do what Minnesota is proposing, and force utilities to justify the black box before the ramp arrives, not after.