A rate case that is really about who pays for AI
On June 4, 2026, the Pennsylvania Public Utility Commission approved a 275 million dollar annual revenue increase for PPL Electric Utilities, the company's first base rate hike since 2016. Residential customers will see a 4.9 percent average increase, pushing typical monthly bills to roughly 184 dollars. That headline number is not what makes this settlement significant to anyone running technology infrastructure. Buried in the filing is a new large-load customer tariff built specifically to make data centers pay for the grid capacity they require, rather than spreading those costs across every ratepayer in eastern and central Pennsylvania.
PPL Electric President Christine Martin credited the process rather than the policy shift itself, saying the company appreciated the parties' ongoing commitment to working constructively throughout an extensive review and for representing their constituents so effectively. That diplomatic framing understates what actually happened. Pennsylvania regulators looked at a service territory facing 20 gigawatts of contracted large-load interconnection requests against a current peak of 7.8 gigawatts and built a rate structure that explicitly separates data center risk from residential risk for the first time.
The mechanics of the new tariff
The large-load class applies to any facility with 50 megawatts or more of peak demand at a single location, or 75 megawatts aggregated across multiple facilities on 69 kilovolt or higher lines within a 10-mile radius. That threshold is specifically sized to catch hyperscale data center campuses without pulling in smaller industrial customers who were never the source of the grid strain. Facilities in that class now face minimum 10-year service agreements, a load ramp schedule that can stretch up to five years, minimum load guarantees with exit fees if a project is cancelled or scaled back, and security deposits that cover the utility's system upgrade costs upfront.
That last provision is the one worth underlining for anyone evaluating a new data center lease or build. Security deposits covering system upgrade costs mean a data center operator, not PPL's existing residential and commercial customers, now bears the financial risk if a facility never reaches its projected load or gets cancelled mid-build. Combined with the mandatory interconnection agreements for high-voltage transmission and the 11 million dollars per year the settlement dedicates to residential low-income assistance programs, the structure is designed to make large-load growth self-funding rather than subsidized.
Pennsylvania is reacting to a regional problem, not a local one
PPL sits inside PJM Interconnection, the grid operator spanning 13 mid-Atlantic and Midwestern states plus Washington, D.C. PJM's capacity auction price jumped from 28.92 dollars per megawatt-day for the 2024/2025 delivery year to 269.92 dollars for 2025/2026, a nearly ninefold increase, and it is set to rise another 22 percent to 329.17 dollars for 2026/2027. According to analysis from the Institute for Energy Economics and Financial Analysis, data centers accounted for 63 percent of that 2025/2026 auction price increase, translating to 9.3 billion dollars in additional costs that flowed through to ratepayers across the region in a single year, with another 1.4 billion dollars in added capacity market costs starting in June 2026.
The regional impact is uneven but real everywhere it lands. Washington, D.C. customers on Pepco's system saw roughly 21 dollars added to monthly bills, with about 10 dollars of that directly attributable to capacity price increases. Western Maryland customers saw an 18 dollar monthly increase, and Ohio customers saw roughly 16 dollars added. Virginia remains the center of PJM's data center concentration, but growth is actively expanding into Ohio and Maryland as operators chase available grid capacity and cheaper land, carrying this same cost dynamic with them.
Why this becomes a template, not an outlier
Pennsylvania is not the first state to act, and it will not be the last. Regulators across the PJM footprint and beyond are under direct political pressure from ratepayers who have started connecting rising electric bills to AI data center announcements in their region, and utilities have every incentive to shift cost allocation onto the customers actually driving new infrastructure spending rather than defend rate increases that hit every voter equally. The PPL model, with its explicit demand threshold, minimum contract terms, exit fees, and upfront security deposits, is a clean template other state utility commissions can adapt without reinventing the underlying mechanics.
That matters directly for site selection. A large-load tariff changes the real cost of a data center lease in ways that do not show up in a headline power price quote. Minimum 10-year commitments and exit fees convert what used to look like a flexible operating cost into something closer to a long-term capital commitment, and security deposits covering grid upgrades add real cash requirements before a single server rack goes live. States and utilities without this structure yet are the ones now facing the most pressure to adopt it, since unprotected ratepayers are the ones showing up at public utility commission hearings.
What this means for site selection and vendor contracts
If your organization is negotiating colocation space, a build-to-suit lease, or even a large cloud reserved-capacity agreement in a PJM state or any other grid region under similar strain, the power cost structure your provider is operating under deserves the same scrutiny as their SLA. Ask directly whether your provider's facility falls under a large-load tariff, what minimum commitment and exit fee terms they accepted with the utility, and whether those obligations get passed through to you as a tenant through higher colocation rates or contract minimums.
The broader lesson extends past Pennsylvania. Electricity cost and availability, not just chip supply or floor space, is becoming a first-order site selection variable, and the states and utilities that move fastest to formalize large-load tariffs will become the more predictable places to build. CTOs planning multi-year infrastructure commitments should treat a region's regulatory posture on data center tariffs as seriously as its tax incentives, because the tariff structure determines whether your power costs stay fixed for a decade or fluctuate with the next capacity auction.
