The reform published by the regulator in March takes effect on 1 January 2026, and it changes how commercial and industrial sites are billed for the grid, not just for the energy they draw. Until now, most C&I rooftops in our service territory paid a demand charge set on a single monthly peak — one bad fifteen-minute window in July could define the bill. From next year that charge splits into a coincident-peak component tied to the utility’s own system peaks and a lower non-coincident component. In plain terms, the grid stops penalising you for being busy and starts pricing when you are busy.
We have modelled the new tariff against 42 of our operating sites, from a 220 kWp warehouse roof in Hamlin to a 3.1 MWp cold-storage array outside Glenmere. The pattern is consistent. A flat solar-only roof shaves energy volume but barely touches the coincident-peak charge, because our regional peaks now land between 18:00 and 20:00 in winter — after the array has stopped producing. On several sites the demand-charge line item falls by less than 8 percent even where self-consumption tops 60 percent.
Where storage rewrites the bill
The math changes the moment you add batteries. On the Glenmere site we paired the existing 3.1 MWp array with a 1.2 MW / 2.4 MWh battery and dispatched it purely against the four forecast coincident-peak intervals. Annual demand charges dropped from roughly 118,000 to 41,000 in the model — a 65 percent cut — while the energy arbitrage was almost incidental. The storage was not there to move kilowatt-hours; it was there to be invisible during four hours a month when the grid is most expensive.
Who wins, then, is straightforward. Sites with steady daytime loads and no storage see modest gains. Sites that can shift or shave their evening coincidence — through batteries, thermal mass, or simply rescheduling a shift — capture most of the upside. Who pays is the operator who assumes a rooftop array alone still clears the demand charge the way it did in 2023. That assumption is now roughly a decade out of date.
There is a timing wrinkle worth flagging. The coincident-peak windows are published day-ahead, so dispatch has to be forecast-driven, not rule-of-thumb. We have retrofitted our monitoring stack on twelve sites to pull the utility’s day-ahead signal automatically; manual dispatch left about a fifth of the savings on the table in our December pilot.
Our read is simple. If you are sizing a C&I rooftop for 2026 and beyond, model the demand charge first and the energy yield second, and treat storage as part of the base case rather than an upgrade. We are happy to run your last twelve months of interval data against the new tariff before you commit to a design.
