Home Battery Economics under the Capacity Tariff
Twelve real Flemish bills, twelve months of meter data, one conclusion: the capacity tariff rewards a battery handsomely in some homes and does almost nothing in others. The difference is not the battery. It is the shape of your evening.
Start with the mechanism, because most articles stop at the slogan. Since 2023, every Flemish household with a digital meter pays for grid capacity by the kilowatt of monthly peak — the highest average quarter-hour draw in each month, billed at €53.39 per kW per year in 2026. Not your total consumption. Your worst fifteen minutes. A family that runs the heat pump, the oven and the EV charger at 18:30 pays for that coincidence twelve times a year.
We took twelve Flemish bills from real households — terraced and detached, with and without EVs — and modelled what a 10 kWh battery with peak-shaving enabled actually did to that peak. The spread was the story. Homes with an EV and electric cooking: average peak fell from 4.1 kW to 2.6 kW, worth about €80 a year. Homes with gas cooking and no EV: 2.4 to 2.1 kW, worth €16. The same battery, the same software setting, a 5x difference in grid-tariff savings.
The counter-intuitive finding nobody puts in the brochure: peak-shaving is not where the money is anyway. Even in the best case, the capacity-tariff saving is a side dish. The main course is arbitrage — buying at €0.09 in a negative-price hour and using it at €0.34 in the evening. The households in our set on dynamic contracts captured three to four times more value from shifting than from shaving. If your installer sold you the battery on the capacity tariff alone, the quote deserves a second look.
Second counter-intuitive finding: the tariff punishes ignorance, not behaviour. One household in our set had beautiful equipment and a peak that never moved — because the peak-shaving threshold was left at the factory default of 2.5 kW, above their actual peak. The feature existed; nobody turned the dial. We now ask for the settings screenshot in every quote review.
Third: the monthly structure forgives. Your peak is billed per month, so one bad evening in January does not haunt you all year. This is the detail that makes DIY peak management viable — you do not need perfection, you need your three worst evenings under control.
For the Netherlands, none of this applies yet — salderen’s end makes shifting the whole game there. For Germany, §14a makes the controllable-load conversation the equivalent one. But the Flemish lesson generalises: a battery is a strategy with hardware attached, and the strategy is set in software by whoever commissions the system.
Three checks before you act. First, pull your quarter-hour data from Mijn Fluvius and find your real evening peak — not the number on the quote. Second, if you buy a battery, get the peak-shaving threshold in writing on the commissioning sheet. Third, re-measure after the first winter month: if the peak did not move, the setting, not the battery, is at fault.
Method: twelve households, Fluvius quarter-hour exports, twelve months, modelled dispatch against 2026 tariffs. Figures rounded; household mix is not a statistical sample. The tariff rate is the published VREG figure for 2026.