Demand Charge

A demand charge is a component of a commercial or industrial electricity bill based not on how much energy you consume, but on your single highest power draw — your peak kilowatt (kW) demand — during the billing period, usually measured as the highest average over a 15- or 30-minute interval. It is billed separately from the consumption charge, which covers total kilowatt-hours (kWh) used. The distinction matters because the two measure different things: kWh is how much energy you took; kW is how big a pipe the utility had to keep ready for you. Demand charges exist to recover the cost of that pipe — the generation, transmission, and transformer capacity sized to your worst moment, not your average one. Most residential customers never see one; for larger loads it can be a substantial line item, and for an unprepared mining operation it can be the difference between a profitable site and a puzzling loss.

Why miners feel it acutely

Bitcoin mining is an unusually unforgiving load for demand-charge billing. Because ASICs run at or near full power around the clock, a miner's peak demand is essentially its baseline — there is no quiet period to pull the monthly peak down, and a single 15-minute interval at full fleet power sets the charge for the whole month. Depending on the tariff, the demand line can effectively add meaningful cents per kWh on top of the base energy rate. The classic newcomer mistake is quoting a site's economics from the energy rate alone: a facility that pencils out at the advertised per-kWh price can be underwater once the demand charge is applied. Worse, restart behavior can bite — a fleet that hard-restarts after an outage and slams to full power simultaneously can set a fresh peak in its first interval back online, which is one reason staggered start-up is standard practice in well-run operations.

Reading the tariff first

Everything about managing demand charges depends on how your specific utility defines and measures the peak. Key questions: What is the measurement interval? Is the charge based on the monthly peak or a rolling ratchet, where one bad summer interval sets a minimum demand billed for the next twelve months? Is demand billed only during defined peak windows, or around the clock? Does the tariff distinguish facility demand from billing demand? A ratchet clause, in particular, turns one careless afternoon into a year-long expense — reading the tariff carefully is not bureaucracy, it is the first engineering step.

Managing it

Operators attack demand charges from three directions. First, peak shaving: deliberately curtailing or stepping down the fleet during the utility's peak-setting windows, which mining does better than any other industrial load because hashrate can throttle in seconds — see curtailable load. Firmware-level underclocking gives a graduated version of the same lever: dropping the fleet to an efficient low-power mode during peak windows caps demand while keeping most of the hashrate alive. Second, storage: an on-site battery energy storage system that discharges during would-be peaks caps the demand the meter ever sees. Third, rate strategy: demand charges interact with how energy prices vary by time of day — see time-of-use rate — and sometimes the right answer is a different tariff class or a negotiated interruptible rate that trades curtailment commitments for lower charges.

The craftsman's summary: the utility bills you for your worst fifteen minutes. Know when those minutes are measured, and make sure your fleet never has its worst fifteen minutes then. For anyone modeling a new site, put the demand line into the spreadsheet on day one — it is the difference between quoting your power cost honestly and discovering it on the first invoice.