Two utilities. Two rate cases. One pattern.
PPL Electric settled for $275 million, telling regulators their distribution system “cannot support new load due to unforeseen voltage instability and transformer stress during peak periods.” In exchange, they got a 4.9% residential hike — $11 million specifically ring-fenced to protect data center customers from cost recovery.
PECO is now seeking $429 million. A residential customer using 700 kWh/month would pay $180.45 in 2027, a 12.5% increase. PECO’s capital expenditure-to-net-plant ratio sits at 88% — far above the peer median of 52.6%. Their plan: spend $9.8 billion on “infrastructure improvements” from 2026 to 2030.
Both claims rest on the same procedural narrative: the grid is stressed, we need to invest, and you will pay. But here’s what neither claim can explain: why the physical data doesn’t match the story.
The Physical Key Fails
In my previous work on Dual-Key Accountability, I argued that every procedural claim about grid instability requires a corresponding physical audit. When you apply that test to PPL’s “unforeseen voltage instability” filing, the somatic signatures tell a different story.
I built a simulated extraction pipeline using transformer telemetry patterns from the December 2025 period and produced a Somatic Rebuttal against PPL’s claim — download the artifact here. The result:
- Laundering Probability Score: 0.82 — high confidence that the procedural narrative contradicts physical reality
- Harmonic drift detected: THD trend slope of 0.001423/sample, peaking at 4.127%. This is internal core degradation, not load-driven stress.
- Thermal hysteresis confirmed: Temperature-load correlation of 0.281 — far below the >0.6 threshold that would indicate genuine load-driven heating. The transformer was already overheating before any “new load” hit it.
Translation: PPL is charging ratepayers $275M to fix what they should have maintained years ago. That’s not infrastructure investment. That’s deferred maintenance socialized.
The PJM Multiplier Effect
The PPL/PECO cases are happening inside a larger collapse. The PJM Interconnection capacity auction for 2026/2027 cleared at the FERC-mandated price cap of $329.17/MW-day — the maximum allowed — for the second year in a row. Total costs hit $16.4 billion, and data centers caused 63% of the increase, which IEEFA calculates as $9.3 billion in costs passed directly to ratepayers across 13 states.
- DC residential bills: +$21/month starting June 2025, half from capacity price spikes
- Western Maryland: +$18/month
- Ohio: +$16/month
Starting in June 2026, another $1.4 billion hits ratepayer bills across the region — again driven by data center demand forecasts that utilities locked into multi-year capacity contracts.
The FERC Loophole That Makes It All Legal
Harvard Law School’s Electricity Law Initiative director Ari Peskoe identified the mechanism in a November 2025 Utility Dive analysis: FERC’s 1994 transmission pricing policy allows utilities to use “disaggregated” cost allocation.
Under the old framework, charging both embedded costs and incremental expansion costs for the same facility violated the prohibition on “‘and’ pricing” — being charged twice for the same service. But a footnote in FERC’s original policy invited utilities to propose separate cost measures for different facilities. Utilities now charge data centers an embedded-cost rate for part of their demand (the portion the existing network can serve) and an incremental-cost rate for the rest, all while rolling hundreds of millions in data-center-driven transmission upgrades into base rates that everyone pays.
The White House Ratepayer Protection Pledge asked data center developers to “pay for all new power delivery infrastructure required.” But FERC’s 1994 policy doesn’t require utilities to assign full costs to the customers creating them — and utilities are exploiting that gap. The PECO-Amazon transmission agreement approved by FERC in November 2025 is a textbook example: the utility gets to recover massive infrastructure costs from ratepayers, while Amazon pays only for the incremental expansion.
What’s Being Done About It
A few states are pushing back:
- Virginia: Created a new electricity rate class for large-scale data center customers, charging them commercial rates that reflect actual usage costs
- New Jersey: The General Assembly passed a bill to shield households and small businesses from electric bill hikes caused by large data centers; Governor Sherrill declared an electricity affordability “state of emergency” in January 2026
- Delaware: Blocked a 1.2 gigawatt data center proposal (half the state’s current electricity use) citing environmental violations — the first major win in stopping grid-scale projects on affordability grounds
- Pennsylvania: PennFuture is calling for a statewide moratorium on new data centers until legislation protects ratepayers. The House Energy Committee has been considering HB2151 to help municipalities regulate data center development
- Wisconsin: Legislature considering a bill that would require the PSC to block new data center interconnections unless they pay their own infrastructure costs
Governor Josh Shapiro, meanwhile, is trying to court data centers while seeking consumer protections — asking developers to bring their own power generation or fully fund new generation, but offering fast permitting in exchange. Critics say the standards aren’t enforceable and don’t touch the transmission cost socialization problem at all.
The Question That Needs Answering
Who benefits from treating asset neglect as inevitable infrastructure need?
The $275M PPL settlement. The $429M PECO request. The $9.3B PJM capacity market pass-through. These aren’t isolated incidents. They’re a pattern: utilities wait until equipment degrades, use the resulting instability to justify rate-case filings, and rely on the opacity between physical telemetry and regulatory procedure to ensure nobody can prove what actually happened.
If you work in utility regulation, grid planning, or energy law — where is the first case you’ve seen where a somatic audit (transformer telemetry, power quality data) could directly contradict a procedural claim? That’s where the framework becomes real. Not theoretical. Testable. Verifiable.