Wildfire liability and mitigation programs, plus heavy distribution spending, pounded bills; Lawrence Berkeley’s case study traces about 40% of the five-year increase to wildfire-related costs—insurance, hardening, and the price of failure.² Pull back and the national pattern snaps into focus. Generation costs, on average, have drifted down, while “delivery”—especially distribution—has marched upward; for investor-owned utilities, inflation-adjusted distribution CapEx rose ~35% from 2019–2024.¹
Then the hardware crunch hit. Distribution and large power transformers became the chokepoint throttling electrification—prices up ~60–80% since 2020, with lead times stretching from about a year to several in the hardest-hit regions.³ Disasters compound it. Florida’s storm run pushed hardening and restoration riders onto bills; in Houston, repairs after Hurricane Beryl (July 2024) sparked bruising fights over cost recovery and securitization as residents told commissioners about spoiled groceries, medical-equipment worries, and week-long outages.⁴ In other words, the bill most families see is increasingly the cost of street-level infrastructure and the risks around it—not a simple tally of electrons.
Wholesale markets add their own weather vane. When supply tightens, prices remind everyone who’s boss. PJM’s 2025/26 capacity auction cleared high across constrained zones, a warning shot for retail rates in places where load growth outruns new steel in the ground.⁵
Closer to New England’s kitchen tables, the drivers are now printed right on the bill. In New Hampshire, consumer advocate Don Kreis warned that lifting fixed customer charges—nearly $20 today with escalators aimed much higher—“hits harder for people who can least afford it,” a shift the commission agreed to reconsider after a public backlash.¹³ In Maine, the Office of the Public Advocate itemized storm-recovery costs that now show up as both a flat and per-kWh rider—evidence that the most reliable way to shrink bills is to need fewer repairs, not fewer headlines.¹⁴
So is it true that data centers aren’t causing price spikes? The honest answer is it depends. Where utilities plan ahead, properly assign upgrade costs, and make big loads flexible, growth can hold prices steady—or even lower them—by spreading fixed costs and keeping capacity markets calm.¹ ¹² Where capacity, wires, and weather run hot, new mega-loads can lift bills unless they bring their own supply, their own flexibility, or direct payments for the upgrades they trigger. That’s not a culture-war verdict; it’s an engineering and accounting one. Blame is a blunt tool; planning is a scalpel.
The reform path runs through three wires: pricing, efficiency, and flexibility. Start by pricing upgrades where they happen. Virginia’s nonpartisan JLARC panel lays out a toolkit: directly assign grid upgrades that primarily serve hyperscale campuses; create dedicated rate classes; require minimum bills tied to contracted demand; and steer siting toward headroom that already exists.⁶
Next, wring more out of the wires we already have. Grid-enhancing technologies—dynamic line ratings that reflect real weather, topology optimization, and modern power-flow controllers—unlock headroom quickly and cheaply; FERC Order 881 requires ambient-adjusted ratings nationwide by July 2025, with more dynamic approaches under active consideration.⁷
Then make big loads flexible on purpose. Google’s new agreements with Indiana Michigan Power and TVA commit data centers to shift AI workloads at peak, effectively turning server halls into dispatchable demand-response resources instead of pure liabilities.⁸
Stop paying more than we need to for capacity that can be deferred.