Ask a facility manager what their biggest power problem is and, if it's a warehouse, a cold-storage operation, or anything with heavy machinery, the answer is often the same: the demand charge. It's the line on the commercial bill that surprises people, and it's frequently doing more damage than the energy charge everyone focuses on.
What a demand charge is
Energy charges bill you for total kilowatt-hours used over the month. Demand charges bill you for your single highest fifteen-minute spike of power draw — your peak demand — regardless of how briefly it lasted. Fire up every compressor at 2 p.m. on the hottest day of the summer and, for many tariffs, that one interval sets a charge you pay all month.
The result is a bill where two facilities using the same total energy can pay wildly different amounts, purely because one has a spikier load. Flatten the spikes and you attack a cost that has nothing to do with using less energy overall.
Why solar alone doesn't fully fix it
A PV array reliably cuts energy charges — it's generating during the day when you're consuming. But it doesn't reliably cut demand charges, because your peak might land on a cloudy afternoon, or after the sun has dropped, or in a fifteen-minute window when a passing cloud cut production exactly when a compressor kicked on. Solar reduces your average draw; it doesn't guarantee to shave the one peak that sets the charge.
Where storage changes the math
Pair the array with a battery and you can target the demand charge directly. The system watches your load, and when a peak starts to build, it discharges the battery to hold your metered draw below the demand threshold — peak shaving. The grid never sees the spike, so the demand charge never gets set by it.
On the Valley Logistics Hub project, that pairing cut demand charges by 84% — a 1.2 MW array for the energy, a 500 kWh LFP battery sized specifically to the facility's demand curve for the peaks. The battery wasn't sized to run the building; it was sized to erase the fifteen-minute intervals the utility was pricing hardest.
The key is that the storage has to be sized against your actual fifteen-minute demand data, not a round number. That's the modeling we do before we ever quote a commercial project — because on the commercial side, the demand charge is usually where the return lives.
